Professional Documents
Culture Documents
Hydraulic Components and Systems
Hydraulic Components and Systems
and Systems
by
Fluid power or oil hydraulics is an essential area of knowledge for anyone who has a
technical interest in moving machinery. In for example farm tractors and implements,
industrial trucks, earth-moving equipment, cars, we find applications of brute force with
very precise control through hydraulic systems. Your girl friend can park your big two-
ton automobile with only a slight effort at the steering wheel...the touch of a handle lifts
a huge amount of load in the bucket of a loader, because the hidden giant hydraulics is
there.
The purpose of this note is to put the tools of fluid power theory and practice into to the
hands of the reader. The note has been prepared as an aid to basic training in hydraulic
system design. In very simple language and with minimum use of mathematics, you are
told the why’s and how’s of hydraulics. The note explains the fundamental principles of
pressure and flow, describes the operation of the basic hydraulic components, tells how
these components are combined to do their many jobs, and explores the fundamental
considerations of hydraulic equipment design and use.
The note is organised in four major sections, Fundamentals (Chapter 1), Hydraulic
Components (Chapter 2 and Chapter 3), Actuator Control (Chapter 4), System Design
(Chapter 5) and System Analysis (Chapter 6). The fundamentals are just what the name
implies, the basics of hydraulics. Note that this section also includes calculating
pressure losses. The section on Hydraulic Components will give an understanding of the
most important components in the system. The sections on actuator control describe
approaches to speed and power of hydraulic actuators. In System Design a basic
approach for arriving at a hydraulic system capable of performing a specific task is
outlined. Finally, in the last section an introduction is given to both steady state and
dynamic modeling of hydraulic systems.
i
Contents
CHAPTER 1 HYDRAULIC PRINCIPLES
1.1 Hydrodynamics vs. Hydrostatics
1.2 Pressure and Flow
Pressure provides the push flow makes it go
1.3 Hydraulic System Components
1.4 Hydraulic Losses
Flow in lines Losses in fittings The Orifice equation Leakage flow
1.5 Summary
ii
Choose hydraulic fluid Select line dimensions Select control elements
5.3 Determine overall efficiency of system Tank and cooler sizing Filtering
Dynamic Design Considerations
5.4 Servo systems Servo valves Selecting servo valves
Load Sensing Systems
5.5 Pressure compensation Fixed displacement pump Variable displacement pump
System efficiencies
iii
1
Hydraulic Principles
Design
1.3 Hydraulic System Components……………………………… 10
Today, there are many thousands of pressure-operated machines and they are so distinct
from earlier devices we must divide hydraulics into two sciences – hydrodynamics and
hydrostatics. Hydrodynamics can be called the science of moving liquids under
pressure. A water wheel or turbine (Figure 1.1) represents a hydrodynamic device.
Energy is transmitted by the impact of a moving fluid against vanes. We are using the
kinetic energy that the liquid contains.
1:1
Figure 1.2 Force excerted by use of pressure
Most of the hydraulic machines in use today operate hydrostatically – that is through
pressure. This note is limited to the pressure hydraulics branch, and will be using both
the terms “hydraulic” and “fluid power” as is costumary in the industry.
In studying the basic principles of hydraulics, we will be concerned with forces, energy
transfer, work and power. We will relate these to two fundamental conditions or
phenomena that we encounter in a hydraulic system. They are pressure and flow.
Pressure and flow, of course, must be inter-related in considering work, energy and
power. On the other hand each has its own particular job to do.
Because these jobs are often confused, try to keep them distinct as we consider
separately...and then together...the phenomena of pressure and flow.
What is pressure?
To the engineer, pressure is a term used to define how much force is exerted against a
specific area. The technical definition of pressure, in fact, is force per unit area.
1:2
Pressure in a confined liquid
If you have tried to force a stopper into a bottle that was completely full of water, you
have experienced the near-incompressibility of a liquid. Each time you tried to push the
stopper in, it sprang back immediately when you let it go.
When a confined liquid is compressed, there is a pressure build-up. The pressure is
transmitted equally throughout the liquid container. This behaviour of a fluid is what
makes it possible to transmit a push through pipes, around corners, up and down, and so
on. In hydraulic systems, we use a liquid, because its near-in-incompressibility makes
the action instantaneous, so long as the system is full of liquid.
( abs )
patm 1 atmosphere 760 mmHg 101,350 Pa (1.1)
Absolute pressure is a scale with its zero point at the complete absence of pressure, or a
perfect vacuum. Gauge pressure ignores atmospheric pressure and is always measured
relative to atmospheric pressure.
p( abs ) patm
( abs )
p( gauge) (1.2)
490.5 N
4 2
106 Pa 10 bar (1.3)
4.9 10 m
1:3
Now we have examined the phenomenon of pressure and how it is measured, we can go
to see how it behaves in a hydraulic circuit
Pascal’s law
Pascal’s Law tells us that:
Pascal might have used the hydraulic lever to prove his law. He found that a small
weight on a small piston will balance a larger weight on a larger piston...provided that
the piston areas are in proportion to the weights.
Thus in Figure 1.4, a 20 kg weight on a 5 cm2 area piston balances a 1000 kg weight on
a 250 cm2 piston. (ignoring the weights of the pistons themselves)
If the small piston is the pressure source, the pressure would be the weight divided by
the piston area.
20 kg 9.81 m / s 2
Pressure = 392400 Pa 3.92 bar (1.4)
5 104 m2
The resulting force on the large piston is equal to this pressure multiplied by the piston
area. Thus;
Force 392400 Pa 250 104 m2 9810 N (1.5)
meaning that the weight on the big piston is 1000 kg. We have multiplied force 50 times
in this example; in other words obtained leverage of 50 to 1.
p F / A or F p A (1.6)
When using this relationship always use the SI units. If any other units are given,
convert them to these units before solving the problem.
1:4
F (force) --- Newton (N); p (pressure) --- Pascal (Pa); A (area) --- square meters (m2)
Series operation is not common, but here it is used to illustrate that pressures add up in
series. Anything that creates a back pressure on the device that moves the load, adds to
the load, and increases the pressure requirement of the system.
1:5
Figure 1.6 Pressure in parallel connection. The medium loaded piston is moving.
What is flow?
Flow is much easier to visualise than pressure. In our kitchen sink, for instance, we have
atmospheric pressure. The city water works has built up a pressure in our pipes. When
we open the tap, the pressure difference forces the water out. Thus, movement of the
water is caused by a difference in pressure at two points.
In a hydraulic system, flow is usually produced by the action of a hydraulic pump; a
device used to continuously push on the hydraulic fluid.
Flow rate is the measure of how much volume of the liquid passes a point in a given
time. It is usually measured in litre-per-minute (l/min). Flow rate is important in that it
determines the speed at which the load moves, and therefore is important to the
consideration of power.
Suppose we are pumping a constant flow rate through the conduit shown in Figure 1.7.
If A1 and A2 are cross-sectional areas, it is obvious that a constant flow rate Q (l/min)
will result in a lower velocity, when the diameter increases or a higher velocity when
1:6
the diameter decreases. In fact, the velocity of oil in a hydraulic line is inversely
proportional to the cross-sectional area (or to the diameter squared). Thus
The speed of the cylinder, then must be proportional to flow and inversely proportional
to the piston area (or the diameter squared).
1:7
as an orifice. Accounting for the pressure drops around a hydraulic circuit is a key
factor in circuit analysis. Anytime a pressure drop occurs and no mechanical work is
delivered, fluid energy is converted to heat energy. Efficiency is increased when these
pressure drops are minimised.
Energy is the capacity to do work, and is expressed in the same units as work. We are
familiar with several forms of energy. The load of 100 kg just mentioned, when it is
raised, has potential energy. It is capable of doing work when it is lowered. A body in
motion has kinetic energy, capable of doing work. Coal contains heat energy; a battery
electrical energy; a steam boiler pressure energy.
In Figure 1.11, we have just slightly upset the balance so that the small piston is forced
down and pushes the large piston up.
1:8
Figure 1.11 Energy transfer in a hydraulic lever
For simplicity’s sake we ignore the effects of friction. At the small piston, we have
moved 20 kg downward for 50 mm. In doing so, we gave up 20 9.81 0.05 = 9.81 Nm
of potential energy. We changed the potential energy to pressure energy. The
displacement flow moved the big weight up 1 mm. Thus, the 1000 kg weight received
an increase of 9.81 Nm of potential energy. At each piston, then, we did 9.81 Nm of
work. Thus, energy was transferred without loss from the 20 kg weight to the 1000 kg
weight.
Bernouilli’s principle
Bernouilli’s principle tells us that the sums of pressure, potential and kinetic energy at
various points in a system must be constant, if flow is constant. The potential energy
term is normally negligible in hydraulic systems.
When a fluid flows through areas of different diameters (figure 1.13), there must be
corresponding changes in velocity.
Figure 1.12 Bernoulli’s principle Figure 1.13 Effect of velocity and friction
At the left, the section is large so velocity is low, and pressure high. In the middle
velocity must be increased because the area is smaller. Again at the right, the area
increases to the original size and velocity again decreases. Bernouilli’s principle states
the pressure in the middle section must be less than elsewhere because velocity is
greater. An increase in velocity means an increase in kinetic energy. Since energy
cannot be created, kinetic energy can only increase if the static pressure decreases. In
the outer sections kinetic energy is converted back to pressure. If there is no frictional
loss, the pressure in the outer sections will be equal. In Figure 1.13 is shown the
combined effects of friction and velocity changes. Pressure drops from maximum at C
to atmospheric pressure at B. At D velocity is increased, so the pressure decreases. At E,
the pressure increases while most of the kinetic energy is given up to pressure energy.
1:9
Again at F, pressure drops as velocity increases. The difference in pressure at D and F is
due to friction losses.
Power
Power is the rate of doing work or the rate of energy transfer. To visualize power, think
about climbing a flight of stairs. If you walk up, it’s relatively easy. But if you run up,
you are liable to get to the top out of breath. You did the same amount of work either
way...but when you ran up, you did it at a faster rate, which required more power.
The standard unit of power is Watt...named after James Watt that related the ability of
his steam engine to the pulling power of a horse. By experiments with weights, pulleys
and horses, Watt decided that a horse could comfortably do 736 Nm/s (1 Nm/s = 1 W),
hour after hour. This value has since been designated as one horsepower (hp).
Power, though, is force multiplied by distance divided by time:
F (Force) D (Distance)
P (Power) (1.9)
T (Time)
The horsepower used in a hydraulic system can be computed if we know the flow rate
and the pressure. In Figure 1.14 we have a flow rate Q into the piston chamber giving
the piston a velocity V. Due to the outer force F a pressure p builds up in the piston
chamber as well. While force is equal to pressure multiplied area and flow is equal to
velocity multiplied area, we can write the equation that relate/convert the mechanical
power to hydraulic power.
We can see from the relationship in Equation (1.10) that an increase in either pressure
or flow will increase the power. Also, if pressure or flow decreases, power decreases.
In the following chapters, you will be studying the components that go into hydraulic
circuits, and how they are put together to do their jobs. From the beginning of Chapter 2
and in the end of this chapter we will talk about components and their interrelationships,
so we will give them some basic attention now.
1:10
We have already mentioned that a pump is required to push the fluid. We have implied,
too, that a cylinder is the output of the system. Besides a pump and cylinder, we also
require valves to control the fluid flow; a reservoir to store the fluid and supply it to the
pump; connecting lines; and various hydraulic accessories.
We will look at two very basic and simple systems now, and see how the basic
components are classified.
In the bottom view, the pump is stoked downward. The inlet check valve is closed by
pressure and the outlet valve opens. Another “slug” of liquid is pumped under the large
piston to raise it.
1:11
To lower the load, we open a third valve, a needle valve, which opens the area under the
large piston to the reservoir. The load then pushes the piston down and forces the liquid
into the reservoir.
Motor-reversing system
In Figure 1.16, we have an entirely different kind of system. Here, a power-driven pump
operates a reversible rotary motor. A reversing valve directs fluid to either side of the
motor and back to the reservoir. A relief valve protects the system against excess
pressure, and can bypass pump output to the reservoir if pressure rises to high.
Pump classifications
The pump in Figure 1.15 is called a reciprocating pump. Most pumps are of the rotary
type as in Figure 1.16, and are driven by engines or electric motors. Rotary pumps can
be constant displacement; meaning they deliver the same amount of fluid every stroke,
revolution or cycle. Flow rate varies in proportion to drive speed - or by using variable
displacement pumps, which can have their delivery rates changed by external controls
while drive speed remains constant.
Actuator classifications
The actuator is the system’s output component. It converts pressure energy to
mechanical energy. A cylinder is a linear actuator. Its outputs are force and straight line
motion. A motor is a rotary actuator. Its outputs are torque and rotating motion.
1:12
The large piston actuator in Figure 1.14 is a single-action cylinder. This means it is
operated hydraulically in one direction only and returned by other means...in this case
by gravity. A double-acting cylinder operates hydraulically in both directions.
The motor in Figure 1.15 is a reversible motor. Other motors are uni-directional or non-
reversible. They can only rotate in one direction.
Valve classifications
In Chapter 3 we will study three classes of valves. They are:
- Directional control valves tell the oil where to go by opening and closing passages.
The check valves in Figure 1.14 are directional valves. They are called one-way valves,
because they permit only one flow path. The reversing valve in Figure 1.15 is a four-
way directional valve, because it has four flow paths.
- Flow control valves regulate flow to control the speed of an actuator. The needle
valve in Figure 1.14 has a flow control function. It restricts flow so that the load can’t
come down too fast.
Classification of lines
The lines which connect the components of our systems are classified according to their
functions. The principal kinds of lines are:
Pressure lines
Drain lines
Working lines : Suction lines Non working lines :
Return lines Pilot lines
- Working lines are lines which carry the mainstream of fluid in the system; that is, the
fluid involved in the energy transfer. Starting at the reservoir, we have a suction line
which carries the fluid to the pump inlet (Figure 1.15). From the pump to the actuator, is
the pressure line, which carries the same fluid under pressure to do the work. After the
pressure energy in the fluid is given up at the actuator, the exhaust fluid is re-routed to
the reservoir through the exhaust or return line.
- Non-working lines are auxiliary lines which do not carry the main stream of flow. A
drain line is used to carry leakage oil or exhaust pilot fluid back to the reservoir. A pilot
line carries fluid that is used to control the operation of a component.
1:13
Circuit diagrams
We have used some very simple schematic diagrams in this chapter to illustrate
hydraulics principles and the operation of components. However, outside of
instructional materials we almost exclusively use a standard type of shorthand (i.e.
symbols to DIN ISO 1219). Symbols for hydraulic systems are for functional
interpretation. Symbols are neither dimensioned nor specified for any particular
position. We will return to these diagrams in more detail in Chapter 2 and 3.
In Figure 1.17 is shown the graphical diagram for the reversible motor circuit.
When a fluid flows through a line, the layer of fluid particles next to the wall have zero
velocity. The velocity profile a distance away from the wall develops because of
viscosity. The more viscous the fluid, the greater the change in velocity with the
distance from the wall.
Viscosity
The most important of the physical properties of hydraulic fluids is the viscosity. It is a
measure of the resistance of the fluid towards laminar (shearing) motion, and is
normally specified to lie within a certain interval for hydraulic components in order to
obtain the expected performance and lifetime. The definition of viscosities is related to
the shearing stress that appear between adjacent layers, when forced to move relative
(laminarly) to each other. For a Newtonian fluid this shearing stress is defined as:
1:14
dx
xy (1.11)
dy
where
xy is the shearing stress in the fluid, [N/m2]
is the dynamic viscosity, [Ns/m2]
x is the velocity of the fluid, [m/s]
y is a coordinate perpendicular to the fluid velocity, [m]
In Figure 1.18 the variables associated with the definition of the dynamic viscosity are
shown.
The usual units for the dynamic viscosity are P for Poise or cP for centipoise. Their
relations to the SI-units are as follows: 1 P 100 cP 0.1 Ns / m2 .
For practical purposes, however, the dynamic viscosity is seldom used, as compared to
the kinematic viscosity that is defined as follows:
(1.12)
where
υ is the kinematic viscosity, [m2/s]
is the dynamic viscosity, [Ns/m2]
is the density, [kg/m3]
The usual unit used for υ is centistoke, cSt, and it relates to the SI units as follows:
m2 mm 2
1 cSt 10 6 1 .
s s
A low viscosity corresponds to a "thin" fluid and a high viscosity corresponds to a
"thick" fluid.
Reynolds number
A key issue in fluid power circuits are forces due to fluid inertia and forces due to
viscosity. In general, flow dominated by viscosity forces is said to be laminar, and
inertia dominated flow is said to be turbulent. In Figure 1.19 is shown the velocity
profile of the two kinds of flow. Laminar flow is characterised by an orderly, smooth,
parallel line motion of the fluid. Inertia dominated flow (turbulent) is characterised by
irregular, eddylike paths of the fluid.
1:15
Figure 1.19 Laminar and turbulent flow profile
In fluid power it is only inertia and viscous forces that matters. Experience has shown
that it is either the inertia forces or the viscous forces that dominate, giving two types of
flow regimes. Osborn Reynolds performed a series of experiments in 1833 to define the
transition between laminar and turbulent flow. He defined a useful quantity which
describes the relative significance of these two forces in a given flow situation. The
dimensionless ratio of inertia forces to viscous force is called Reynolds number and
defined by
U dh
Re (1.13)
where is fluid mass density, υ is the kinematic viscosity, U is the average velocity of
flow, and d h is a characteristic length of the flow path. In our case d h is taken to be the
hydraulic diameter which is defined as:
4 flow area 4 A
dh (1.14)
flow perimeter Ch
For a hydraulic hose or pipe it is convenient in many cases to use the following formula
for Reynolds number
Vline Dline
Re (1.15)
where Vline is the fluid velocity in the line and Dline is the inner diameter of the line.
We normally use Re = 2300 as transition number for hydraulic hoses and pipes.
Darcy’s equation
Friction is the main cause of loss of fluid energy as the fluid flows through a line.
Because of friction, some fluid energy is converted to heat energy and exchanged into
the surrounding atmosphere.
Loss along a line can be calculated as a pressure loss p , directly using Darcy’s
equation.
1:16
L V2
p line (1.16)
Dline 2
Laminar flow
For laminar flow, the friction factor can be given as
64 / Re (1.17)
Substitution of Equation 1.17 into Darcy’s equation gives the Hagen-Poiseuille equation
128 L Q
p (1.18)
Dline
4
Turbulent flow
When the flow is turbulent, the friction factor is a function of Reynolds number and the
relative roughness of the line, but for smooth lines and Reynolds numbers less than
100.000 the equation given by Blasius can be used to calculate the friction factor.
1
0.3164 (1.19)
Re0.25
Tests have shown that pressure losses in fittings are proportional to the square of the
velocity of the fluid
p 2
Vline (1.20)
2
where is a friction factor that must be determined by tests. Some common factors are
The flow restrictions or orifices are a basic means for the control of fluid power. An
orifice is a sudden restriction of short length in a flow passage and may have a fixed or
variable area.
Equation 1.21 is Bernoulli’s equation with negligible gravity forces.
1:17
p u2
= constant (1.21)
2
As an important case where Equation 1.21 is used consider flow through an orifice
(Figure 1.20).
Since most orifice flow occurs at high Reynolds numbers, this region is of great
importance. Experience has justified the use of Bernoulli’s equation in the region
between point 1 and 2. The point along the jet where the area becomes a minimum is
called the vena contracta. The ratio between the area at vena contracta A2 and the
orifice area A0 defines the so called contraction coefficient Cc .
Cc A2 / A0 (1.22)
After the fluid has passed the vena contracta there is turbulence and mixing of the jet
with the fluid in the downstream region. The kinetic energy is converted into heat. Since
the internal energy is not recovered the pressures p2 and p3 are approximately equal.
Now it is possible to use Bernoulli’s equation 1.21 to calculate the relation between the
upstream velocity u1 to the velocity u0 in vena contracta. Therefore
2
u22 u12 ( p1 p2 ) (1.23)
A1 u1 A2 u2 A3 u3 (1.24)
1/2
A 2 2
u2 1 2 ( p1 p2 )
A1 (1.25)
1:18
In the real world there will always be some viscous friction (and deviation from ideal
potential flow), and therefore an empirical factor Cv is introduced to account for this
discrepancy. The factor Cv is typically around 0.98. Since Q A2u2 the flow rate at
vena contracta becomes,
Cv A2 2
Q ( p1 p2 ) (1.26)
1 ( A2 / A1 )2
CvCc
Cd (1.27)
1 Cc2 ( A0 / A1 )2
Now, combining Equation 1.22, 1.26, and 1.27 the orifice equation (in Danish
blændeformlen) can be written
2
Q Cd A0 ( p1 p2 ) (1.28)
At low temperatures, low orifice pressure drop, and/or small orifice openings, the
Reynolds number may become sufficiently low to permit laminar flow. Although the
analysis leading to Equation 1.28 is not valid at low Reynolds numbers, it is often used
anyway by letting the discharge coefficient be a function of Reynolds number. For Re <
10 experimental results show that the discharge coefficient is directly proportional to
the square root of Reynolds number; that is Cd Re . A typical plot of such a result
is shown in Figure 1.21.
Figure 1.21 Plot of a discharge coefficient versus Reynolds number for an orifice
1:19
So the orifice equation states that the flow through an orifice is proportional to the
square root of the pressure drop across the orifice
Solving for the pressure drop p , we obtain the orifice equation with the flow rate Q as
the independent variable; a form more useful for our analysis of fluid power circuits.
1
p keq Q 2 ; keq (1.29)
2 C A02 2
d
This relationship shows there is a quadratic relation between p and Q. Most valves
form some type of orifice in their internal geometry, and thus have a pressure vs. flow
characteristic following Equation 1.29. Knowing the flow through the valve, the
designer can estimate the pressure drop. The technical data sheets supplied by almost all
valve manufacturers have a pressure vs. flow curve, from which it is possible to
determine keq.
Many hydraulic components have precision made parts with very fine clearances
between internal moving elements, but still, due to high pressure, they make leakage
pathways.
Dryden (1956) presents the following expression for leakage flow as a function of
pressure drop across a rectangular leakage pathway.
w h3
Q p (1.30)
12 L
where w is the width of the rectangular opening and h the height of the rectangular
opening (Figure 1.22)
1:20
1.5 Summary
Now that we have studied basic principles and have an idea of how hydraulics work,
lets close this chapter with some important characteristics of hydraulic systems and a
short summary.
1. High power density (high power output per unit mass of system)
2. Operation may commence from rest under full load
3. Smooth adjustment of speed, torque, force are easily achieved
4. Simple protection against overloading (relief valve opens to protect system)
Accounting for the pressure drop in a hydraulic system is a very important factor in
circuit analysis and design. Pressure drops occurring without delivering mechanical
work, means that hydraulic energy is converted to heat.
Pressure drops occurring in different types of valves, is in most hydraulic circuit the
major source of losses, and dominates over the losses in the lines and fittings.
----- oo 0 oo -----
1:21
2
Characteristics of
Pumps and Motors
Definitions
Hydraulic pumps and motors are used to convert mechanical energy into hydraulic
energy and vice versa. It is transmission of power by fluid using positive displacement
pumps and motors. A positive displacement pump is one in which each revolution of the
pump shaft is associated with a fixed quantity of fluid delivered and similarly a positive
displacement motor is one in which each revolution of the motor shaft is associated with
a fixed quantity of fluid accepted.
In non-positive pumps and motors, such as turbine and centrifugal pumps the flow is
continuous from inlet to outlet and results from energy being directly imparted from the
fluid stream. These machines are basically low pressure with high volume output.
Another way to classify positive and non-positive displacement machines is whether the
inlet is sealed from the outlet.
2:1
If the inlet and outlet are connected hydraulically so that the fluid can
recirculate in the pump/motor when pressure builds up, the pump/motor
is non-positive displacement
It the inlet is sealed from the outlet, the pump will deliver fluid anytime
the inlet is kept supplied and the pump is driven. Such a pump is
classified as positive delivery or positive displacement, and requires a
relief valve to protect it from pressure overloads.
While non-positive pumps and motors are quite ineffective and not suited for control
purposes only positive displacement pumps are used in hydraulic systems.
The pump is driven by a prime mover which is usually an electric motor or a petrol or
diesel engine. The energy input from the prime mover to the pump is converted into
high-pressure energy in the fluid which is transmitted through pipes and in turn is
converted into rotational energy by a motor or translational energy by a cylinder.
Hydraulic pumps are classified by their design – gear, vane and piston. In all cases a
moving element in a fixed container displaces fluid from an inlet to an outlet port. The
pressure at the inlet port is due to the head of fluid from the reservoir and at the outlet
port by the resistance imposed by the work load on the hydraulic motor/cylinder. Plus
the resistance due to friction in pipes, valves, etc.
Gear-type pumps
These employ the principle of a pair of meshing gears, contained in a housing fitted
with suction and discharge ports as shown in Figure 2.1
Stator
Idler
Suction Discharge
Rotor
One gear (the rotor) is driven by the prime mover and as it rotates and drives the idler
gear. The fluid is displaced between the gears and the inner surface of the stator. The
fluid is thus carried around from the inlet port to the outlet port, the main seal between
2:2
the two being provided by the meshing gears. Some fluid is trapped between the gear
teeth and the design of the pump allows for this fluid to be transferred back to the
suction port.
There must obviously be clearance between the stationary casing and the tips of the
moving gear. As pressure builds up on the outlet side of the pump the pressure
difference across the gears results in leakage of fluid. This leakage is commonly called
“slip” and is a factor present to at greater or lesser extent in all positive displacement
pumps. The volumetric efficiency is reduced as the leakage increases and the
mechanical (or mechanical-hydraulic) efficiency is reduced because of the power
wasted in producing the slip. The performance (efficiency, losses, etc.) of pumps and
motors is dealt with in detail in Section 2.2 and 2.3.
In a gear pump slip occurs not only across the gear teeth but also between the gears and
the side faces. To reduce this as far as possible these side plates are hydraulically loaded
to maintain close contact onto the sides of the gears.
Vane-type pumps
The construction of a vane pump is shown in Figure 2.2. Here the moving member
consists of a rotor in which close fitting vanes are carried radially in slots and their tips
bear against the stationary housing. The vanes are free to move in and out of the slots.
As the rotor turns, the vanes are thrust against the stator to form an effective seal. Side
plates are used to keep the oil confined to an area the width of the rotor and vanes. The
rotor centre is eccentric to the stator centre so that the chambers between adjacent vanes
expand when connected to the suction port and compress when connected to the
discharge port. The pump in Figure 2.2 is equipped with a double stroke which
increases the displacement and helps balance the rotor. The connection to the suction
and discharge ports are via kidney-shaped connections in the side plates.
Stator
Rotor
Suction Discharge
Vanes
2:3
Vane pumps can discharge a variable volume by changing the degree of eccentricity
between the rotor and stator, see Figure 2.3. In that case, a ring with adjustable
eccentricity relative to the rotor is added to the pump. If the rotor is dead centre with the
ring there is no pumping action. At maximum eccentricity, the greatest volume of fluid
will be pumped. The degree of eccentricity is adjusted by the use of suitable controls
outside the casing.
The slip in a vane pump is in general less than that in a conventional gear pump and a
higher volumetric and mechanical efficiency is maintained because the vane wear is
compensated for by the outward movement of the vanes.
Stator
Rotor
Displacement
adjustment
Suction Discharge
Ring
Piston-type pump
In the positive displacement piston pump the rotor is a cylinder block and the moving
members are pistons sliding in the cylinder bores. There are no seals on the pistons
which are close-fitting in the bores. The pumping action is obtained by reciprocating the
pistons relative to their bores and feeding the fluid to and from the cylinders by inlet
and discharge valves.
Piston-type pumps are commonly used for applications that require high pressures and
accurate control of the discharge volume. There are many different designs, but
generally all designs are based on the radial piston-type or the axial piston-type.
Axial piston pumps present the widest variety in design. The axial piston pump shown
in Figure 2.4, contains a cylinder block shaped rotor with pistons that are equally spaced
around the cylinder block axis. The cylinder bores are parallel to the axis. The pistons
reciprocate parallel to the rotor centre line. The drive shaft rotates the rotor that slides
against the stationary end plate. The stationary end plate mates with the surface of the
cylinder barrel to prevent leakage of fluid. The pistons are connected to the drive shaft
via ball and socket joints. As the cylinder barrel block rotates in contact with the
stationary valve plate, the cylinders are alternately ported to the suction (expanding
chamber) and discharge (compressing chamber) connection via the kidney shaped
2:4
connections so that expanding chambers are conn. The bent axis design shown in Figure
2.4 has variable displacement because the end plate can be rotateed in the stator. When
the rotor center line is aligned with the drive shaft center line there is no pumping
action.
Stator
A Displacement
adjustment
Stationary
Rotor
Drive shaft end plate
AA
Pistons
A
Discharge
Suction
Figure 2.4 Bent axis axial piston pump with variable displacement.
As an alternative to the bent axis design we have the swash plate design, shown in
Figure 2.5. Here, the rotor and drive shaft are always aligned whereas the end plate =
swash plate is rotated relative to theese. The angle between the cylinder block and the
swash plate causes the pistons to reciprocate generating a pumping action similar to the
one described for the bent axis design.
Stationary
Stator A
Swash plate end plate
AA
Piston
Rotor
Drive shaft
Piston
A Discharge
Displacement
adjustment Suction
Figure 2.5 Swash plate axial piston pump with variable displacement.
The swash plate angle can relatively easily be adjusted during operation facilitating a
wide range of pump control options that will be described in more details in Chapter 4.
Adjusting the angle increases or decreases the piston stroke to increase or decrease the
volume output. An important feature is the ability to reverse the direction of the fluid.
This can be done by tilting the swash plate in the opposite direction.
2:5
2.1.2 Motor classifications
The purpose of this chapter is to give an overview of the governing equations for
hydraulic pumps and actuators as components in a hydraulic system. Hence, the chapter
does not go into much detail about the different types of pumps and actuators and when
to choose one or the other.
The main parameter for a hydraulic pump is the stroke displacement, D, defined as the
displacement pr. revolution of the pump. It relates to the pump flow as follows:
QtP D n (2.1)
where
QtP theoretical pump flow, [volume/time]
D stroke displacement, [volume/revolution]
n rotational speed of the pump, [revolutions/time]
n (2.2)
2
where
n rotational speed, [revolutions/time]
angular velocity, [radians/time]
2:6
Defining the unit displacement, D , as the displacement pr. radian leads to:
D 2 D (2.3)
QtP D (2.4)
where
QtP theoretical pump flow, [volume/time]
D stroke displacement, [volume/revolution]
D unit displacement, [volume/radian]
angular velocity, [radians/time]
Some pumps have adjustable displacements. In that case the displacements may be
determined from:
The power supplied to the fluid by the pump may be determined based on the general
expression for hydrokinetic power:
P Q p (2.6)
where
P power, [power]
Q flow, [volume/time]
p pressure, [pressure]
Accordingly, the theoretical power put into the hydraulic system, e.g. the fluid, by the
pump may be determined as:
The pressure and flow variables of Equation (2.7) are shown in Figure 2.6.
2:7
pP QtP
pP M
pS
Figure 2.6 Diagram illustration of motor driven pump. The pressure and flow variables
associated with the pump are shown.
pP pP pS (2.8)
The theoretical mechanical power delivered to the pump from some power source is
given as:
Pt ,PS P M tP (2.9)
where
Pt ,PS P theoretical power supplied to the pump by the power supply, [power]
M tP theoretical input torque to the pump from the power supply, [torque]
angular velocity, [radians/time]
Without any losses, the power delivered to the pump by the power supply equals the
power supplied by the pump to the fluid, i.e., combining Equation (2.7) and (2.9) gives:
Inserting Equation (2.4) and (2.3) into Equation (2.10) leads to:
2.2.2 Efficiencies
The expressions developed for the pump flow and pump torque in the previous section
correspond to an ideal pump without power losses of any kind. In reality the pump will
produce less flow than the theoretical value and the pump will require more torque
than the theoretical value. Hence, only a part of the power supplied to the pump will end
up as hydrokinetic power, whereas the power loss will heat up the pump and its
surroundings including the hydraulic fluid.
The fact that the pump delivers less flow than theoretically expected is expressed by
means of a volumetric efficiency:
QP
vP (2.12)
QtP
where
vP volumetric efficiency of the pump
QP (actual) flow of the pump, [volume/time]
2:8
QtP theoretical pump flow, [volume/time]
The volumetric loss is mainly due to leakage in the form of laminar clearance flow from
the high pressure chambers to the low pressure chambers within the pump. This leakage
flow is mainly laminar (although some models also include turbulent leakage) and
therefore proportional to the pressure rise and inverse proportional to the viscosity:
pP
QlP KlP (2.13)
where
QlP leakage flow within the pump, [volume/time]
K lP leakage constant for the pump, [volume]
pP pressure rise across the pump, [pressure]
dynamic viscosity, pressure time]
The volumetric constant tends to vary with the displacement for the same type of pump
(larger pumps will have larger dimensions including the leakage clearances).
Combining Equations (2.1), (2.12) and (2.13) gives:
It is seen that the volumetric efficiency will depend on pressure rise, rotational speed
and the viscosity (mainly temperature), i.e., change with the working conditions. The
variation of the volumetric efficiency may be viewed graphically for 2 different
situations, see Figure 2.7. The curves are based on Equation (2.14).
vP vP
1.0 1.0
p P
n
Figure 2.7 The volumetric efficiency for a hydraulic pump as function of the pressure rise
across the pump and the rotational speed of the pump, respectively.
The volumetric efficiency decreases linearly with increasing pressure rise when
rotational speed and viscosity is held constant. When varying the rotational speed for
fixed viscosity and pressure rise it is seen that the volumetric efficiency goes to zero and
2:9
below. At zero volumetric efficiency the pump is just capable of producing enough
theoretical flow to make up for its own internal leakage.
It should be noted that in case of insufficient suction pressure, the volumetric efficiency
will decrease with increasing pump speed because there is not enough time for the
suction chambers to be properly filled with fluid.
The fact that the pump requires more input torque than theoretically expected is
expressed by means of a hydro-mechanical efficiency:
M tP
hmP (2.15)
MP
where
hmP hydro-mechanical efficiency
M tP theoretical input torque to the pump, [torque]
MP (actual) input torque to the pump, [torque]
1. Mechanical friction due to mechanical contact between parts of the pump moving
relative to each other. This loss is proportional to the pressure rise.
2. Viscous (laminar) friction due to shearing of fluid films between parts of the pump
moving relative to each other. This loss is proportional to the speed of the moving
parts and the viscosity.
3. Hydrokinetic (turbulent) friction due to turbulent pump flow around restrictions,
bends, etc. within the pump. This loss is proportional to the square of the flow.
4. Static friction due mainly to sealings. This loss is constant.
The different hydro-mechanical losses may be expressed as either pressure drops, or
more conveniently, as extra input torques to the pump:
M vP KvP n (2.17)
M hP KhP n 2 (2.18)
M sP cst. (2.19)
where
M mP input torque required to overcome mechanical friction, [torque]
M vP input torque required to overcome viscous friction, [torque]
M hP input torque required to overcome turbulent friction, [torque]
M sP input torque required to overcome static friction, [torque]
K mP , K P , K hP pump dependant constants
pP pressure rise across the pump, [pressure]
n rotational speed, [revolutions/time]
dynamic viscosity, pressure time]
Introducing Equation (2.11) in the above and rearranging leads to an expression that
shows the dependency of the hydro-mechanical efficiency on pressure, speed and
viscosity.
1.0
hmP
1 (2.21)
C0
pP
C1 C2 n C3 n 2
where
C0..3 pump dependent constants
In Figure 2.8 the variation of the hydro-mechanical efficiency for constant pump speed
and constant pressure rise, respectively, may be viewed graphically. The curves are
based on Equation (2.21).
As seen in Figure 2.8 the hydro-mechanical efficiency goes to zero as the pressure rise
goes to zero for fixed pump speed. This corresponds to the pump meeting no resistance
(no demand to pressurize the fluid), but still demanding some input torque from the
power supply in order to overcome the viscous, turbulent and constant friction losses.
PP F
P (2.22)
PPS P
where
P efficiency of the pump.
PPF (actual) power delivered by the pump to the fluid, [power]
PPS P (actual) power delivered by the power supply to the pump, [power]
hmP hmP
1.0 1.0
n
p P
Figure 2.8 The hydro-mechanical efficiency for a hydraulic pump as function of the
pressure rise across the pump and the rotational speed of the pump, respectively
2:11
The power delivered by the pump to the fluid is found by combining Equation (2.7) and
(2.12)
Similarly, the power delivered by the power supply to the pump is found by combining
Equation (2.9) and (2.15)
M tP D pP
PPS P M P (2.24)
hmP hmP
Inserting Equation (2.23) and (2.24) into Equation (2.22) yields the following
expression for the pump efficiency
1 D pP 1
MP D pP (2.27)
hmP 2 hmP
The dependency of the total efficiency on the pressure rise across the pump, the speed
of revolution and the viscosity is complex, especially as the volumetric efficiency is
best at low pressure and high speed, whereas the hydro-mechanical efficiency is
best at high pressure and low speed.
The main purpose of the hydraulic motor is to transform hydraulic power into
mechanical power by means of a rotating output shaft. Hydraulic motors are exclusively
made as positive displacement motors, where the driven volumes are separated in
pressurized and relieved (depressurised) chambers. By means of a flow to the motor,
typically provided by some hydraulic pump, these chambers will vary in such a way that
the output shaft of the motor is rotated. Depending on the load on the shaft from the
driven mechanical system the hydraulic motor will demand a certain pressure drop, i.e.,
a certain pressure level at its intake, in order to move. The hydraulic motor may be
thought of as an inverted hydraulic pump.
2:12
2.3.1 Basic equations
As for the hydraulic pump, the main parameter for a hydraulic motor is the stroke
displacement, D, defined as the displacement pr. revolution of the motor shaft. It relates
to the motor flow demand as follows:
QtM D n D (2.28)
where
QtM theoretical motor flow demand, [volume/time]
D stroke displacement, [volume/revolution]
n rotational speed of the motor, [revolutions/time]
D unit displacement, [volume/radian]
angular velocity of the motor, [radians/time]
Notice that the flow through the motor is referred to as a flow demand, i.e., the flow
required in order for the motor to meet a certain shaft speed.
Some motors have adjustable displacements. In that case the displacements may be
determined from:
The displacement control parameter must be kept above some minimum vale, min , as
the motor shaft otherwise will rotate with infinite speed for any flow.
The theoretical power put into the motor from the hydraulic system, e.g. the fluid, may
be determined as:
The pressure and flow variables of Equation (2.30) are shown in Figure 2.9.
2:13
QtM
pi
ML
n
pM
po M tM
Figure 2.9 Diagram illustration of a hydraulic motor driving a load. The pressure and flow
variables as well as the torques associated with the motor are shown. The motor is subjected to a
positive load.
The theoretical mechanical power supplied by the motor to the load is given as
Pt ,M L M tM (2.31)
where
Pt ,M L theoretical power supplied by the motor to the load, [power]
M tM theoretical output torque from the motor, [torque]
angular velocity, [radians/time]
Without any losses, the power delivered to the motor by the fluid equals the power
supplied by the motor to the load, i.e., combining Equation (2.30) and (2.31) gives
In steady state conditions (constant shaft speed), the torque delivered to the output shaft
is opposite in direction but in size to the torque load on the shaft.
M tM M L (2.34)
In Figure 2.9 the load torque, M L , is opposite to the direction of the speed, n .
Whenever this is the case we have a positive load, and the motor is said to be motoring.
As an example, this corresponds to the motor driving a wire drum that is hoisting a load.
For a positive load the pressure drop is, by definition:
pM pi po (2.35)
2:14
Alternatively, we have a negative load whenever the load torque, M L , is in the same
direction as the speed, n , see Figure 2.10.
QtM
pi
M tM
n
pM
po ML
Figure 2.10 Diagram illustration of a hydraulic motor driving a load. The pressure and flow
variables as well as the torques associated with the motor are shown. The motor is subjected to a
negative load.
As an example, this corresponds to the motor driving a wire drum that is lowering a
load. In this case the return pressure must be the highest, and the motor is being driven
as a pump by the external load. The governing equations become:
The pressure drop for a negative load is, by definition, given as:
pM po pi (2.37)
From (2.36) it is seen that the power supplied to the motor by the fluid will be a
negative number which underlines the fact that the motor is driven as a pump and
therefore supplies power to the fluid. Similarly, we get a negative theoretical
mechanical power supplied by the motor to the load
Pt ,M L M tM (2.38)
where
Pt ,M L theoretical power supplied by the motor to the load, [power]
M tM theoretical output torque from the motor, [torque]
angular velocity, [radians/time]
2:15
Eqs. (2.32) and (2.33) are the same for both positive and negative loads although it
should be remembered that the sign of pM in (2.33) changes with the type of loading
according to (2.35) and (2.37).
2.3.2 Efficiencies
The expressions developed for the motor flow demand and motor torque in the previous
section correspond to an ideal motor without power losses of any kind. In reality the
motor will require more flow than the theoretical value and the motor will deliver less
torque than the theoretical value. Hence, only a part of the hydrokinetic power supplied
to the pump will end up as mechanical power, whereas the power loss will heat up the
motor and its surroundings including the hydraulic fluid.
The influence of the efficiencies depend on whether the motor is subjected to a positive
load or a negative load. First, we will investigate the influence of the efficiencies on a
motor subjected to positive load and later this will be supplemented with the similar
equations relating th negative load.
The fact that a motor subjected to a positive load requires more flow in order to reach a
certain shaft speed, n, than theoretically expected is expressed by means of a
volumetric efficiency:
QtM
vM (2.39)
QM
where
vM volumetric efficiency of the motor
QtM theoretical flow demand of the motor, [volume/time]
QM (actual) flow demand of the motor, [volume/time]
As for pumps the volumetric loss is mainly due to leakage in the form of laminar
clearance flow from the high pressure chambers to the low pressure chambers within the
motor. This way some of the flow will pass through the motor without helping to rotate
the motor shaft. The clearance flow is proportional to the pressure drop and inverse
proportional to the viscosity
pM
QlM KlM (2.40)
where
QlM leakage flow across the motor, [volume/time]
KlM leakage constant for the motor, [volume]
pM pressure drop across the motor, [pressure]
dynamic viscosity, pressure time]
The volumetric constant tends to vary with the displacement for the same type of motor
(larger motors will have larger dimensions including the leakage clearances).
Combining Equation (2.28), (2.39), and (2.40) gives
2:16
QtM 1.0 1.0
vM
QtM QlM 1.0 QlM 1.0 KlM pM (2.41)
QtM Dn
It is seen that the volumetric efficiency will depend on pressure drop, motor speed and
the viscosity (mainly temperature) under working conditions. The variation of the
volumetric efficiency may be viewed graphically for 2 different situations, see Figure
2.11. The curves are based on Equation (2.41).
M M
1.0 1.0
n const . const .
p P const . const .
n
p P
Figure 2.11 The volumetric efficiency for a hydraulic motor as function of the pressure drop
across the motor and the rotational speed of the motor shaft, respectively.
MM
hmM (2.42)
M tM
where
hmM hydro-mechanical efficiency of the motor
M tM theoretical output torque from the motor, [torque]
MM (actual) output torque from the motor, [torque]
M tM M mM M M M hM M sM
hmM (2.43)
M tM
where
hmM hydro-mechanical efficiency of the motor
2:17
M tM theoretical output torque from the motor, [torque]
M mM loss in output torque due to mechanical friction, [torque]
M M loss in output torque due to viscous friction, [torque]
M hM loss in output torque due to turbulent friction, [torque]
M sM loss in output torque due to static friction, [torque]
1
hmM D0
pM
D1 D2 n D3 n 2 (2.44)
where
D0..3 motor dependant constants
pM pressure drop across the motor, [pressure]
n motor speed of revolution, [revolutions/time]
dynamic viscosity, pressure time]
In Figure 2.12 the variation of the hydro-mechanical efficiency for constant motor speed
and constant pressure drop, respectively, may be viewed graphically. The curves are
based on Equation (2.44).
When the motor runs with a constant speed, the hydro-mechanical efficiency will be
zero, when the pressure drop has a value where it is just capable of maintaining
equilibrium with the internal torque losses. This leaves nothing to rotate any load and
the motor is on the border line of acting like a hydraulic pump driven by the load.
hmM hmM
1.0 1.0
p M
n
Figure 2.12 The hydro-mechanical efficiency for a hydraulic motor as function of the pressure
drop across the motor and the rotational speed of the motor, respectively.
In the following equations we look at the influence of the efficiencies on the motor
equations. The total efficicency for a motor subjected to a positive load is:
PM L
M (2.45)
PF M
2:18
where
M total efficiency of the motor.
PM L (actual) power delivered by the motor to the load, [power]
PF M (actual) power delivered by the fluid to the motor, [power]
The power delivered by the fluid to the motor is found by combining Equation (2.39)
and (2.28)
Q pM D pM
PF M QM pM tM (2.46)
vM vM
Similarly, the power delivered by the motor to the load is found by combining Equation
(2.42) and (2.33)
Inserting Equation (2.46) and (2.47) into Equation (2.45) yields the following
expression for the motor efficiency:
Further, expressions for the flow demand and motor torque can be readily computed as:
1 1
QM Dn D (2.49)
vM vM
D ( pi po )
M M hmM hmM D ( pi po ) M L (2.50)
2
MM ML (2.51)
QM
vM (2.52)
QtM
M tM
hmM (2.53)
MM
2:19
PM F PF M n D pM
M vM vM hmM
PLM PM L 1 D pM (2.54)
hmM 2
where
M total efficiency of the motor.
PM F (actual) power delivered by the motor to the fluid, [power]
PLM (actual) power delivered by the load to the motor, [power]
The flow demand and motor torque for the negative load become similar to those for the
pump (see Eqs. 2.26 and 2.27)
As in the case with hydraulic pumps the dependency of the total efficiency on the
pressure drop across the motor, the motor speed and the viscosity is complex, with the
volumetric efficiency best at low pressure and high speed and the hydro-mechanical
efficiency best at high pressure and low speed.
The main purpose of the hydraulic cylinder is to transform hydraulic power into
mechanical power by means of a translating piston rod. In general, a piston attached to
the piston rod, uses the cylinder housing to seal of 2 pressure chambers. Each chamber
is connected to the remaining hydraulic system. When hydraulic flow is led to either
one of the pressure chambers, the piston and piston rod moves in the corresponding
direction.
A Ar (1 ) A Aa A
H H
H Ld 2 H Ld
2:20
Depending on the load on the piston rod from the driven mechanical system the
hydraulic cylinder will demand a certain pressure drop across its piston, in order to
move.
The above parameters are all shown in Figure 2.13. The area ratio is the ratio between
the annulus area (piston area - piston rod area) and the piston area.
Aa A Ar
(2.57)
A A
0 1
In order to set up equations for the flow demand of the cylinder and the relationship
between load, cylinder force and pressures it is necessary to discuss the possible modes
of operation of a cylinder. As in the case of the motor, a cylinder may be subjected to
both positive and negative loads. Further, since a cylinder, in general, is assymetric it is
also necessary to distinguish between the direction of motion that is divided into out-
stroke and in-stroke. This gives a total number of four combinations that are shown in
Figure 2.15.
2:21
Out-stroke and positive load Out-stroke and negative load
FtC FL FtC FL
p1 p1
p2 p2
vC vC
Qi Qo Qi Qo
FtC FL FtC FL
p1 p1
p2 p2
vC vC
Qo Qi Qo Qi
Out-stroke
Qi vC A (2.58)
Qo A vC (2.59)
FtC p1 A p2 A FL (positive load) (2.60)
FtC p2 A p1 A FL (negative load) (2.61)
In-stroke
Qi vC A (2.62)
Qo A vC (2.63)
FtC p2 A p1 A FL (positive load) (2.64)
FtC p1 A p2 A FL (negative load) (2.65)
In Eqs. (2.60), (2.61), (2.64) and (2.65) definitions of the theoretical cylinder force are
introduced that ensure a positive values. Also, steady state conditions are assumed
which means:
FtC FL (2.66)
2.4.2 Efficiencies
The equations developed in the previous section are all based on ideal conditions
without any power loss in the hydraulic cylinder. The hydraulic cylinder has, just like
the hydraulic pump and hydraulic motor, both a volumetric and a hydro-mechanical
efficiency.
The volumetric efficiency means that the cylinder will require more flow than
theoretically expected, in order to meet a certain velocity. The volumetric loss is caused
2:22
by leakage across the sealings between the piston and the cylinder housing and the
sealings between the piston rod and the housing. These losses are, however, so small for
a typical cylinder that they may be ignored. This means that the volumetric efficiency
approximately is 100% (or 1.0).
The hydro-mechanical efficiency, however, cannot be disregarded. It is almost
exclusively caused by mechanical friction between the piston seals and the cylinder
housing and the piston rod and the cylinder housing seals. There exist several
definitions but the most common for a cylinder subjected to a positive load is:
FC
hmC (2.67)
FtC
where
hmC hydro-mechanical efficiency of the cylinder
FC (actual) force applied by the cylinder on the piston, [force]
FtC theoretical force applied by the cylinder on the piston, [force]
Equation (2.67) can be used to derive a general expression for the friction force
FC FtC FmC
hmC FmC (1 hmC ) FtC (2.68)
FtC FtC
where
hmC hydro-mechanical efficiency of the cylinder
FC (actual) force applied by the cylinder on the piston, [force]
FmC mechanical friction force, [force]
In these notes Equation (2.,68) is used as the correlation between the theoretical
cylinder force (Eqs. (2.60), (2.61), (2.64) and (2.65)) and the friction force.
Disregarding the volumetric loss the total efficiency of the cylinder corresponds to the
hydro-mechanical efficiency, i.e.
C hmC (2.69)
where
C total efficiency of the cylinder
hmC hydro-mechanical efficiency of the cylinder.
The four modes of operation when taking into account the mechanical friction are
shown in Figure 2.16. The figure corresponds to Figure 2.15 except that the friction
force has been added.
2:23
Out-stroke and positive load Out-stroke and negative load
Figure 2.16 Cylinders shown in all four modes of operation including the friction force.
The equations relating flows and velocities are the same, i.e., Eqs. (2.58), (2.59), (2.62)
and (2.63) also applies for a non-ideal cylinder. The equations relating the pressures in
the cylinder chambers and the external load change:
Out-stroke
FL
FC FtC FmC FL FtC p1 A p2 A (positive load) (2.71)
hmC
FL
FC FtC FmC FL FtC p2 A p1 A (negative load) (2.72)
2 hmC
In-stroke
FL
FC FtC FmC FL FtC p2 A p1 A (positive load) (2.73)
hmC
FL
FC FtC FmC FL FtC p1 A p2 A (negative load) (2.74)
2 hmC
FC FL (2.75)
----- oo 0 oo -----
2:24
3
Characteristics of
Valve Operation
3.1 Introduction
Valves are used in our hydraulic systems to control the operation of the actuators. Very
often, in fact, we find the valves referred to as the “control”, particular where a number
of them are built into a single assembly.
The valves assert their authority in the circuit by regulating pressure; by creating special
pressure conditions; by deciding how much oil will flow in portions of the circuit; and
by telling the oil where to go.
We group hydraulic valves into three general categories: directional controls, pressure
controls, and flow controls. Some valves, however, have multiple functions that fall into
more than one of these categories.
Valves are rated by their size, pressure capabilities and pressure drop vs. flow. They are
usually named for their functions, but may be named according to their design as well.
This chapter gives an overview of the most common valve types and describe their
characteristics.
In general fluid passing through orifices will try to close the orifices. For fixed orifices
this is not a problem, but for orifices with variable discharge areas, the two bodies
generating the orifice geometry will be pulled together, normally affecting the expected
3:1
behaviour of the valve or increasing the actuation demand. In Figure 3.1, two typical
variable orifices are shown.
Q
p2
Q p2
x
p1
Q Q
Di D D
p1
Q
Control volume x
Control volume
Figure 3.1 Typical variable orifices in hydraulic systems. To the left a spool-drill orifice and to
the right a poppet-seat orifice.
To the left, a spool moves within a drilled valve house generating a band-shaped orifice
with discharge area: A D x . To the right, a poppet is moved relative to a valve
house seat, generating a conical band-shaped orifice with discharge area:
A D x sin / 2 .
For both cases the basic momentum conservation is utilized. It states that the sum of
external forces on a system corresponds to the time derivative of the linear momentum
of the system:
d m v
F (3.1)
dt
where
F sum of external forces on a system, [force]
m v sum of linear momentum of the system, [massvelocity]
Employing Equation (3.1) on a control volume containing steady state flow it may be
reformulated to:
F m
v (3.2)
Equation (3.2) states that the sum of external forces on a control volume corresponds to
the sum of mass flow times velocity leaving the control volume.
Defining a control volume, V, as a ring with inner diameter Di and outer diameter D, see
Figure 3.1, it is seen that flow enters perpendicular and leaves through the band-shaped
orifice at some angle, , relative to the spool centre axis.
The required resulting force on the ring of fluid is determined from:
Q Q 2 cos
Fx ,fl m
vx Q cos 2 cos Q p1 p 2 (3.3)
CC A CD A
where
3:2
Fx,fl required axial force on the control volume from the spool, [force]
m mass flow, [kg/time]
vx exit flow velocity in axial direction, [length/time]
mass density, [mass/volume]
Q flow, [volume/time]
CC contraction coefficient
A discharge area of the band-shaped orifice, [area]
angle between the spool axis and the orifice flow direction
CD discharge coefficient
p1-p2 pressure drop across the orifice, [pressure]
From Figure 3.1 it is seen that the control ring volume is acted upon by an axial force to
the right. Only the spool can supply this force, hence, the ring acts upon the spool with
an equal and opposite force, trying to move the spool to the left, i.e., closing it. The
same result would be obtained if the direction of flow was reversed, because the sum of
the mass flow times the reversed velocity would now be entering the control volume,
still requiring a force to the right to be supplied to the control volume. An analytical
solution to a problem of similar nature suggests = 69 but in general the angle may
vary from 0 - 90, depending on the opening, the radial clearance and any variation in
spool and valve house geometry relative to a perfect sharp edge.
Basically, the same expression as Equation (3.3) is obtained when, applying Equation
(3.2) to the poppet valve in Figure 3.1 to the right. Defining a rotational control volume
with an outer diameter, D, and embracing the poppet as shown, the required resulting
force on the cylinder of fluid is determined from:
Q 2 cos
Q 2 2 cos Q p p (3.4)
Fx ,fl m
vx Q cos 1 2
CC A 2 CD A 2
As in the case with the spool, only the poppet can act upon the control volume in axial
direction, hence, the poppet is pulled towards its seat.
For the poppet it seems that the jet angle is more easily obtained as simply half the
poppet angle. However, the jet angle may vary substantially, especially at small
openings, and specifically, the jet will have a tendency to jump from one configuration
to another at a certain combination of flow and opening.
Directional control valves belong to the group of valves controlling flow direction.
Their purpose is to direct pump flow to an actuator as well as allow return flow from the
same actuator to the reservoir. They are classified according to the number of service
ports and number of possible configurations (positions). Hence, the directional control
valves shown in Figure 3.2 are referred to as a 2/2-way, a 3/2-way and a 4/3-way
directional valve.
3:3
2/2 way 3/2 way 4/3 way
A A A B
a b
a b a 0 b
P P T P T
By convention, the designation of the service ports are almost always referred to as P
(pump connection), T (tank connection), A and B (actuator connections). Similarly, the
designation of the position of the valve is referred to as a-b (2 positions) and a-0-b (3
positions). The port connections shown for the 4/3-way valve in Figure 3.1 are typical
for the 3 positions a, 0, and b, respectively. There exist, however, several different port
connections for both 2-, 3-, and 4-working ports, see Figure 3.3, and they may be
combined to yield directional control valves with almost any thinkable functionality.
Figure 3.3 A few of the building blocks for directional control valves. Also a 4/3-way valve
with specification of intermediate positions is shown.
By distinguishing between the port connections while shifting from one position to
another the variety is even further multiplied. An intermediate position is displayed by
means of hidden lines, see Figure 3.3.
A B
A B
P T A
T P T
P
A B
A B
P T
T A
A
P T
T P T P
A B
A B P T
T A
P T
T P T
Figure 3.4 Schematic drawings of a spool directional control valve and a ball directional
control valve in the different working positions.
3:4
Another way of classifying the directional control valve is by the type of moving body
generating the different flow paths. The type of body is either a ball/poppet or a spool,
see Figure 3.4.
The ball/poppet type valve is leak proof but unsuitable for high power flow, because the
actuation is along the same line as the main flow. Contrary to this, the main flow is
perpendicular to the actuation, for spool type valves. However, in order for the spool to
move, a certain clearance between spool and valve housing is required causing leakage.
Directional control valves may be actuated in a great variety of ways. Normally the
moving body is kept in a neutral position by a centering spring. Hence, the actuator
must overcome the spring force in order to activate the valve. The types of actuation
vary from manual, electric, pneumatic, hydraulic and electro-hydraulic. In Figure 3.5
the standardized symbols used for different types of actuation are shown alone and
applied to a 4/3-way directional control valve.
Figure 3.5 Above is shown the standardized symbols for 3 different types of manual actuation
(roller, pedal and general) together with electrical, hydraulic and pneumatic actuation.
A B
A0 ,P A A0 ,BT
T P T
A B
A0 ,AT A0 ,PB
T P T
Figure 3.6 Spool of directional control valve shown in a-position and b-position. The band
shaped discharge areas are shown as double-hatched rectangles.
3:5
To distinguish proportional/servo directional control valves from directional control
valves an extra set of lines is added to their symbol as shown in Figure 3.5, bottom.
The governing equations for a directional control valve concerns the flow from/to the
service ports as well as the pressure drop(s) across the valve.
In Figure 3.6 the spool is shown in the a- and b-position. In general, the flow will be
turbulent after passing through the discharge areas, A0,PA and A0,BT or A0,PB and
A0,AT . Thus, the orifice equation may be used to describe the pressure drop:
2
Q CD A p p Q2 (3.5)
2C2 A 2
D
where
Q flow through the orifice, [volume/time],
CD discharge coefficient,
A discharge area, [area]
mass density, [mass/volume]
p pressure drop across the orifice, [pressure]
For the type of orifices normally encountered in directional control valves the discharge
coefficient will lie around 0.55-0.65.
For an ordinary directional control valve with everything closed in neutral position the
governing equations for flow and pressure drop become:
(3.6)
Q P QA Q B QT 0
0
where
K coefficient that mainly depends on geometry,
[volume/(time·pressure½)]
p pressure, [pressure]
Q flow, [volume/time]
The different constants in Equation (3.6) vary from valve to valve, as they almost
exclusively depend on the geometry of the orifices, i.e., spool geometry, spool stroke as
well as house geometry. A typical chart for a directional control valve is a Q-p curve
for each flow passages, see Figure 3.7. The valve in the figure is symmetrical in the sense,
that the P A port and the P B port are identical, and the A T port and the B T port
are identical.
3:6
Q [l/min]
90
80
70
60 AT & BT
50
PA & PB
40
30
20
10
0
0 5 10 15 20
p [bar]
For a proportional/servo directional control valve the discharge area is not a constant
but varies with spool travel, hence, the governing equations will include the spool
travel, x, as a variable:
0 Q P QA Q B QT 0 (3.7)
A P B ( x )
b Q P B K P B pP pB
Q P Q B Q PB A 0 , P B
A AT ( x )
Q A Q T Q AT Q AT K AT pA pT
A 0,AT
where
K coefficient that mainly depends on geometry,
[volume/(time·pressure½)]
A spool travel dependant discharge area, [area]
A0 maximum value for a discharge area (full spool travel), [area]
x spool travel, [length]
p pressure, [pressure]
Q flow, [volume/time]
3:7
Often, the spool is machined to obtain a tailor-made flow dependency on the spool
travel, i.e., linear, progressive or regressive. A typical chart for a proportional
directional valve is a Q-x curve for each flow passage and for different pressure drops,
see Figure 3.8.
Q [l/min]
80
PA & PB
70
60
50 p = 100 bar
p = 80 bar
40 p = 60 bar
p = 40 bar
30
p = 20 bar
20
10
0
x
0 0.2 0.4 x0 0.6 0.8 1
Figure 3.8 Computed discharge flow from a proportional directional control valve as a function
of the relative spool travel. The value x0 represents the full spool travel (stroke).
In Figure 3.8 there is no flow until the relative spool travel is 0.2, indicating that
directional control valves normally have a certain dead band, in order to keep leakage
down.
A major problem for the actuation of directional control valves are the flow forces. As
mentioned in section 3.2, the flow force will always try to close any orifice through
which fluid is flowing. In the case of spool valves this means that the actuation has to
overcome an extra force beside the force from the usual centering spring.
pA pB
K f ,P A Q P A p P p A K f , B T Q B T p B p T
pP pT
QPA QBT
Figure 3.9 The flow forces acting on a 4/3-way spool in the a-position.
3:8
Employing Equation (3.3) the flow force acting on a spool of a 4/3-way directional
control valve can be set up:
where
Ff flow force on the spool, [force]
Kf geometry and flow dependant coefficient,
[(force·time)/(volume·pressure½)]
p pressure, [pressure]
Q flow, [volume/time]
Notice how the flow forces from both throttlings work in the same direction, against
actuation. Information on the flow force coefficients, Kf, is not given as part of standard
catalogue information unlike the discharge flow coefficients.
Check valve
Check valves belong to the group of valves controlling flow direction. They act as
rectifiers in a hydraulic system, allowing (almost) free flow in one direction and
preventing flow in the opposite direction. When the pressure in the inlet port, port A,
reaches a certain value, crack pressure, the ball is lifted from its seat, compressing the
spring and allowing flow to port B. (see Figure 3.10)
A
B
2r Q
A B
pA pB
Figure 3.10 The standard symbol and a schematic drawing of a check valve.
3:9
The governing equations for a check valve are based on the orifice equation and static
equilibrium:
2
Q C D A( x ) p A p B (3.9)
p A p B r 2 k sp x x ic (3.10)
k sp x ic
p cr pB (3.11)
r2
where
Q flow across the valve, [volume/time]
CD discharge coefficient
A discharge area, [area]
mass density, [mass/volume]
pA inlet pressure, [pressure]
pB outlet pressure, [pressure]
r inlet radius, [length]
ksp spring stiffness, [force/length]
x ball/poppet travel, [length]
xic initial compression of the spring, [length]
pcr crack pressure required to open the valve, [pressure]
Equation (3.9) and Equation (3.10) constitutes 2 equations with 3 variables Q, x and pA -
pB. Hence, knowing the flow through the valve, the pressure drop and the position of the
ball/poppet may be determined. In the above static equilibrium, Equation (3.10), the
flow force is not included, as it may be neglected for check valves because of the small
pressure drops. The variation of the discharge area depends on the configuration: ball or
poppet, see Figure 3.11.
rb
p
x
x
2r 2r
An expression for the discharge areas is: A(x) 2 r x sin x 2 sin 2 cos .
3:10
However, as long as the ball/poppet travel, x, is small compared to the inlet radius, r,
then the discharge areas may approximately be determined as:
p
Poppet (3.13)
2
r2 r2 x
Ball arctan b (3.14)
r
where
r seat radius, [length]
x ball/poppet travel, [length]
discharge area projection angle
p poppet angle
rb ball radius
Equation (3.12) only holds for x smaller than some maximum value x0 where the poppet
is blocked by the valve house. After that the check valve acts a simple restrictor (orifice
with constant discharge area).
For a simplified analysis the ball/poppet travel is disregarded, and the pressure
drop is set equal to the crack pressure, independent of the flow.
p [bar]
5
No spring
0
0 10 20 30 40 50 60
Q [l/min]
Figure 3.12 Computed Q-p curves for a ball check valve with different springs, i.e., crack
pressures. Notice how the curves join, as the ball reaches its full travel.
3:11
A pilot operated check valve allows flow in the opposite direction when activated by a
pilot piston, see Figure 3.13. This feature is very useful when lowering loads that were
held by the check valve functionality.
A B
X
B
pX Q
2 rX 2r
pB
pA
A x
Figure 3.13 Standard symbol and schematic drawing of a pilot operated check valve. The valve
is piloted open.
The governing equations for the pilot operated check valve correspond to those of the
check valve except for the static equilibrium:
p x p A rx2 p B p A r 2 k sp x x ic (3.15)
r 2 k sp x x ic
p cr ,x p B p A pA (3.16)
rx2 rx2
where
pcr,x required crack pressure in the pilot line, [pressure]
rx radius of the pilot piston
pB inlet pressure, [pressure]
pA outlet pressure, [pressure]
r inlet radius, [length]
ksp spring stiffness, [force/length]
x ball/poppet travel, [length]
xic initial compression of the spring, [length]
Normally, the pilot piston will open the check valve to its maximum value, and it will
work as a simple restrictor.
A pressure control valve may have the job of limiting or otherwise regulating pressure
or creating a particular pressure condition required for control.
3:12
All pure pressure control valves operate in a condition approaching hydraulic balance.
Usually the balance is very simple: Pressure is effective on one side or end of a ball,
poppet or spool – and is opposed by a spring. In operation, the valve takes a position
where the hydraulic pressure exactly balances the spring force.
The function of any relief valve is to protect the hydraulic system from excessive
pressure if the pressure increases above a predetermined maximum. A relief valve is an
automatic relieving device that is actuated by the static pressure upstream of the valve.
Relief valves are designed to return the hydraulic fluid directly to the reservoir. A relief
valve is normally closed until the system pressure approaches a reset value, called the
cracking pressure. As system pressure continues to increase, the amount of flow through
a properly sized relief valve will increase until the entire pump output passes through
the valve. When system pressure decreases, the valve closes smoothly and quietly.
The pressure relief valve acts like a spring loaded check valve, however, the spring load
and the poppet design are very different, as the relief valve typically is supposed to
throttle pressure drops up to several 100 bars.
B
pB
A B
p
x
pA
Q A
2r
Figure 3.14 Standard symbol and schematic drawing of pressure relief/safety valve. The valve
is open (blowing), i.e., the inlet pressure, pA, is above the crack pressure.
When the pressure in port A reaches a certain value, crack pressure, the poppet is lifted
from the seat, compressing the spring and allowing flow to port B. As long as flow is
send across the relief valve the pressure in port A will stay at, approximately, the crack
pressure. The crack pressure can be varied because the initial compression of the spring
is adjustable. A pressure safety valve is designed in exactly the same way as a pressure
relief valve. They differ in functionality, as the safety valve is only expected to bleed of
flow, when the system pressure accidentally moves above a preset value.
In general, both relief and safety valves are made with either a spring loaded ball or a
spring loaded poppet.
3:13
The governing equations for a pressure relief valve corresponds rather closely to those
of a spring loaded check valve, see Equations (3.9)...(3.11). The only difference in the
mathematical expressions is the inclusion of flow forces in the static equilibrium for the
ball/poppet. There is a distinct difference, however, in the higher crack pressure levels,
and the stiffer springs.
2
Q C D A( x ) p A p B (3.17)
p A p B r 2 K f Q p A p B k sp x x ic (3.18)
k sp x ic
p cr pB (3.19)
r2
where
Q flow across the valve, [volume/time]
CD discharge coefficient
A discharge area, [area]
mass density, [mass/volume]
pA inlet pressure, [pressure]
pB outlet pressure, [pressure]
r seat radius, [length]
Kf geometry and flow dependant coefficient,
[(force·time)/(volume·pressure½)]
ksp spring stiffness, [force/length]
x ball/poppet travel, [length]
xic initial compression of the spring, [length]
pcr crack pressure required to open the valve, [pressure]
Equations (3.17) and Equation (3,.18) constitutes 2 equations with 3 variables Q, x and
pA-pB. Hence, knowing the flow through the valve, the pressure drop and the position of
the ball/poppet may be determined. The variation of the discharge area with the
ball/poppet travel is described in Equation (3.12). According to Equation (3.2) the flow
coefficient can be expressed as:
p
K f 2 cos (3.20)
2
where
Kf geometry and flow dependant coefficient,
[(force·time)/(volume·pressure½)]
mass density, [mass/volume]
p poppet angle
There are a number of complex phenomena related to the flow across the ball/poppet
that reduces the validity of Equation (3.4). It does, however, in most cases give a
reasonable estimate of the actual flow forces.
3:14
p [bar]
250
Crack pressure
200 bar
200
No flow force
150
Crack pressure
100 bar
100
No flow force
Crack pressure
50 bar
50
No flow force
0
0 20 40 60 80 100
Q [l/min]
Figure 3.15 Computed p-Q curves for 3 relief valves with different crack pressure settings.
The influence of the flow force is illustrated by not including it in the thin line curves.
In Figure 3.15, p-Q curves are shown for a pressure relief valve. As may be seen the
presence of a flow force increases the slope of the p-Q curve, especially at high
pressure drops. It corresponds approximately to having an extra spring. This flow
induced force is a hydraulic reaction force acting as a result of accelerating fluid
through an orifice. The pressure relief valve has 2 inherent problems: The slope of the
p-Q curve is unwanted from a static functionality point of view. However, if the slope
is decreased (weaker spring, flow force compensation), the result is often instability. A
solution to this basic antagonism is the pilot operated relief valve.
xX
2 rX pX
A B
pB
B
x
QX
pA
A Q QX
2 r
Figure 3.16 Standard symbol and schematic drawing of pilot operated pressure relief valve.
The valve is piloted open, i.e., the inlet pressure, pA, is above the pilot crack pressure, pcr,X.
3:15
The pilot operated relief valve works as follows, see Figure 3.16: The main poppet is
subjected to the same pressure on both sides, i.e., the light spring closes the main poppet
effectively. The crack pressure is set by adjusting the spring load on the pilot poppet.
When the pilot flow (in this case also the inlet flow) reaches the set value, a flow across
the pilot poppet and subsequently across the fixed orifice, O, is introduced causing a
pressure drop from pA to pX. The pressure drop is flow dependent, and as the flow
across the pilot poppet is increased the pressure drop across O, will finally become
larger than the crack pressure of the light main spring. The main poppet is lifted from its
seat, and only flow forces prevent the valve, see Figure 3.17, from having a perfect flat
p-Q curve, without any of the stability problems mentioned earlier.
250
p [bar] Crack pressure
200 bar
200
No flow force
150
Crack pressure
100 bar
100 No flow force
Crack pressure
50 bar
50 No flow force
0
0 20 40 60 80
Q+QX [l/min]
Figure 3.17 Computed p-Q curves for 3 pilot operated relief valves with different pilot crack
pressure setting. The influence of the flow force is illustrated by the thin line curves
The curves in Figure 3.17 are based on the following governing equations for a pilot
operated pressure relief valve:
2
Q C D A( x ) p A p B (3.21)
p A p X r 2 K f Q p A p B k sp x x ic (3.22)
2
Q X C D0 A 0 p A p X (3.23)
3:16
2
Q X C DX A X ( x X ) p X p B (3.24
k sp x ic
p cr pX (3.26)
r2
k sp,X x ic,X
p cr ,X pB (3.27)
rX2
where
Q flow across the main poppet, [volume/time]
CD discharge coefficient for the main poppet orifice
A discharge area of the main poppet orifice, [area]
mass density, [mass/volume]
pA inlet pressure, [pressure]
pB outlet pressure, [pressure]
pX pilot pressure, [pressure]
r seat radius of the main poppet, [length]
Kf geometry and flow dependant coefficient,
[(force·time)/(volume·pressure½)] concerning the flow forces
around the main poppet orifice
ksp spring stiffness of the main poppet spring, [force/length]
x main poppet travel, [length]
xic initial compression of the main poppet spring, [length]
CD0 discharge coefficient for the fixed orifice
A0 discharge area for the fixed orifice
QX pilot flow, [volume/time]
CDX discharge coefficient for the pilot poppet orifice
AX discharge area of the pilot poppet orifice, [area]
rX pilot poppet seat radius, [length]
ksp,X pilot spring, [force/length]
xX pilot poppet travel, [length]
xic,X initial compression of the pilot poppet spring, [length]
pcr crack pressure required to lift the main poppet, [pressure]
pcr,X crack pressure required to lift the pilot poppet, [pressure]
The pilot flow is, normally, only a few l/min, and the crack pressure of the main poppet
is around 5 bar. As for the directly operated pressure relief valve the crack pressure can
be varied by adjusting the initial compression of the pilot spring.
For a simplified analysis of both types of pressure relief valves, the inlet pressure is
constant and equal to the crack pressure as soon as there is any flow across the
valve.
3:17
3.4.2 Pressure reducing valve
A pressure reducing valve is used to limit pressure level from the normal operating
pressure of the primary hydraulic system to the required pressure of a secondary
hydraulic circuit
The purpose of pressure reducing valves is to maintain a desired pressure downstream
of the valve, independently of (but lower than) the upstream pressure. In Figure 3.15 a
schematic drawing is shown together with the standard symbol for a pressure reducing
valve with bypass.
If the pressure in the regulated port, pB, goes up, the spool moves to the right against the
adjustable spring and the reservoir pressure, pT. This tends to close the connection
between ports A and B, thereby increasing the throttling and decreasing p B. If pressure
peaks appear in the regulated line the spool will travel even further to the right and first
close the connection between ports A and B. Next, after passing a certain dead band, xD,
the connection between B and T is opened. As the pressure in port T is quite low, this is
in fact a pressure relief function. It is build into almost all pressure reduction valves.
A B
x
Q A
pA
2 r
pB pT
B T
x xD
Figure 3.18 Standard symbol and schematic drawing of pressure reducing valve. The valve is
throttling, i.e., the pressure in port A is reduced to the desired pressure level (set by adjusting
initial compression of spring) in port B.
2
Q C D A( x ) p A p B (3.28)
p B p T r 2 K f Q p A p B k sp x ic x (3.29)
3:18
k sp x ic
p cl pT (3.30)
r2
For x x D :
2
Q T C D,T A T x p B p T (3.31)
p B p T r 2 K f ,T QT p B p T k sp x ic x (3.32)
k sp x ic x D
p cr ,T pT (3.33)
r2
where
Q flow from A to B, [volume/time]
CD discharge coefficient for the A to B orifice
A discharge area of the A to B orifice, [area]
mass density, [mass/volume]
pA upstream pressure, [pressure]
pB downstream pressure, [pressure]
pT reservoir pressure, [pressure]
r radius of the spool, [length]
Kf flow force coefficient, [(force·time)/(volume·pressure½)]
concerning the flow forces around the A to B orifice
ksp spring stiffness of the spring, [force/length]
xic initial compression of the spring, [length]
x spool travel, [length]
pcl downstream pressure where the A to B connection is closed,
[pressure]
QT flow from B to T, [volume/time]
CT discharge coefficient for the B to T orifice
AT discharge area for the B to T orifice
Kf,T geometry and flow dependant coefficient,
[(force·time)/(volume·pressure½)] concerning the flow forces
around the B to T orifice
xD dead band, [length]
pcr,T crack pressure required downstream to open the B to T
connection, [pressure]
The p-Q curve for the pressure relief function (B to T) corresponds to those discussed
in Section 3.4.1. For the pressure reducing part (x > 0), some typical curves are shown
in Figure 3.19, where the reservoir pressure, PT , has been set to nil.
The influence of the upstream pressure is complex, as increased pressure will tend to
push more flow through the A to B orifice, however this increase the valve closing flow
force and downstream pressure. In general, catalogue material will not clearly indicate
at what upstream pressure curves, like the ones shown in Figure 3.19, are determined.
The closing pressure can be varied because the initial compression of the spring is
adjustable.
3:19
Just like pressure relief valves, pressure reducing valves may be pilot operated gaining
some of the same advantages, i.e., the slope of p-Q curve is reduced without loss of
stability.
p B [bar]
100
90
80 p A 300 bar
p A 100 bar
70
60
50
40
30
20
10
0
0 20 40 60 80 100
Q [l/min]
Figure 3.19 Computed p B-Q curves for a pressure reducing valve at different inlet pressures.
The desired downstream pressure, pcl , is set to 80 bar.
Flow control valves provide volume control in hydraulic circuits. Flow is controlled by
either throttling or diverting the flow. Throttling the flow involves decreasing the size of
an opening until all of the flow cannot pass through the orifice. Bypassing the flow
involves routing part of the flow around the circuit so that the actuator device receives
only the portion of flow needed to perform its task.
Hydraulic circuits that use flow control devices are called metered circuits. If an
actuator has the inlet flow controlled, the circuit is a “meter-in” circuit. If an actuator
has the outlet flow controlled, the circuit is a “meter-out” circuit. Flow control circuits
can either be non-compensated or compensated circuits.
Non-compensated flow controls are simple valves that meter flow by throttling. The
amount of flow that passes through the valve is determined by the position of the valve.
As the valve is closed, flow decreases. One of the most common non-compensated
valves is the adjustable needle valve, considered in section 3.5.1. Compensated flow
control valves are considered in the following sections.
Its purpose is to act as a simple orifice, generating a pressure drop large enough, for
some pressure relief valve further upstream to continuously bypass flow to tank. In that
3:20
case the restricting valve will work in parallel with the pressure relief valve controlling
the speed of the downstream actuator. The speed control will, however, be load
dependant.
B
pB
A B
pA
Q A
Figure 3.20 Standard symbol and schematic drawing of a restrictor valve with adjustable
orifice.
Often, the discharge area of a restrictor valve can be varied, and in Figure 3.21 the Q-p
curve for a restrictor valve with different settings of the discharge area is shown.
p [bar]
400
A = 8 mm2
350
300
250
200 A = 12 mm2
150
100
A = 16 mm2
50
0
0 20 40 60 80 100
Q [l/min]
Figure 3.21 Computed Q-p curve for a restrictor valve with different settings of the
adjustable discharge area.
The curves simply reflect the basic orifice equation, which is the only governing
equation of the restrictor valve:
2
Q CD Α p A p B (3.34)
where
Q flow, [volume/time]
CD discharge coefficient
A restrictor area, [area]
pA upstream pressure, [pressure]
pB downstream pressure, [pressure]
3:21
3.5.2 2 way flow control valve
The 2-way flow control valve belongs to the group of flow rate controlling valves. Its
purpose is to provide a constant flow independent of downstream pressure, i.e., actuator
load. This is also referred to as a pressure compensated flow valve.
The pressure independent flow is obtained by means of a differential pressure
controller, see Figure 3.19, positioned before a fixed orifice. The differential pressure
controller is, essentially, a spool subjected to the pressure before and after the fixed
orifice, on each end. The smaller pressure (after the orifice) works together with a
spring against the higher pressure (before the orifice). This way a pressure drop,
approximately equal to the initial compression of the spring, is maintained across the
fixed orifice, yielding an approximately constant flow. If the intermediate pressure, p I ,
increases, the flow will go up, but simultaneously the spool will tend to close, thereby
reducing the flow. If the opposite happens, i.e., the downstream pressure, pB, increases,
the flow will go down, but simultaneously the spool will tend to open, thereby
increasing the flow.
A B
pB
pI
2r
pA
A
x
Q
Figure 3.22 Standard symbol and schematic drawing of a 2-way flow control valve. The valve
is shown working as it regulates the pressure drop across the fixed orifice by throttling the flow
before the main (fixed) orifice.
The governing equations for the 2-way flow control valve are:
2
Q C D Ax p A p I (3.35)
2
Q C Df A f p I p B (3.36)
3:22
k sp x ic x p I p B r 2 K f Q p A p I 0 (3.37)
where
Q flow, [volume/time]
CD discharge coefficient of the variable orifice
A discharge area of the variable orifice, [area]
mass density, [mass/volume]
pA upstream pressure, [pressure]
pI intermediate pressure, [pressure]
CDf discharge coefficient of the fixed orifice
Af discharge area of the fixed orifice, [area]
pB downstream pressure, [pressure]
ksp spring stiffness, [force/length]
xic initial compression of the spring, [length]
x spool travel, [length]
r spool radius, [length]
Kf flow force coefficient, [forcetime/volumepressure½]
In Figure 3.23 some typical curves Q-pA for a 2-way flow control valve are shown for
different pB values.
Q [l/min]
40
35
30
25
20
10
0
0 50 100 150 200 250
p [bar]
Figure 3.23 Some Q-p curves for a 2-way flow control valve set to deliver 40 l/min. The
curves differ in load/downstream pressure, pB.
Despite minor variations due to spring stiffness and flow forces, constant flow (and
thereby constant actuator speed) is obtained. Both the area of the so-called fixed orifice
as well as the initial compression of the spring can, in general, be adjusted.
It should be noted that the 2-way flow control valve only works when it is subjected to a
prescribed input pressure, pA, and not a prescribed flow, Q. Hence, it should work
together with a pressure relief valve, continuously bleeding of flow and keeping system
pressure approximately constant.
3:23
3.5.3 3 way flow control valve
The 3-way flow control valve belongs to the group of flow rate controlling valves. Its
purpose is to provide a constant flow independent of downstream pressure, i.e., actuator
load. Just like the 2-way flow control valve is also referred to as a pressure compensated
flow valve.
The pressure independent flow is obtained by means of a differential pressure
controller, see Figure 3.24. The differential pressure controller is, essentially, a spool
subjected to the pressure before and after the fixed orifice, on each end. The smaller
pressure (after the orifice) works together with a spring against the higher pressure
(before the orifice). The controller measures the flow via the pressure drop across the
fixed orifice. If there is too much flow the pressure drop will increase and the spool will
move to the right increasing the opening of a passage from the inlet to the reservoir.
Contrary, if there is too little flow the pressure drop will decrease moving the spool in
the opposite direction.
A B
pB
2r
pA
pT
A T
Qin x
Figure 3.24 Standard symbol and schematic drawing of 3-way flow control valve. The valve is
shown working, as it regulates the pressure drop across the fixed orifice bypassing excess flow
to tank.
The governing equations for the 3-way flow control valve are:
2
Q C Df A f p A p B (3.38)
3:24
2
Q T C D Ax p A p T (3.39)
p A p B r 2 k sp x ic x K f Q T pA pT 0 (3.40)
Q in Q Q T (3.41)
where
Q flow used by the remaining system (actuators), [volume/time]
CDf discharge coefficient of the fixed orifice
Af discharge area of the fixed orifice, [area]
mass density, [mass/volume]
pA upstream pressure, [pressure]
pB downstream pressure, [pressure]
CD discharge coefficient of the variable orifice
A discharge area of the variable orifice, [area]
pT reservoir pressure, [pressure]
r spool radius, [length]
ksp spring stiffness, [force/length]
xic initial compression of the spring, [length]
x spool travel, [length]
Kf flow force coefficient, [forcetime/volumepressure½]
Qin total flow sent into the valve, [volume/time]
A couple of Q-Qin curves for a 3-way flow control valve are shown in Figure 3.25 for 2
different values of pB. The reservoir pressure, PT, has been set to zero.
Q [l/min]
45 p B 0 bar
p B 200 bar
40
35
30
25
20
15
10
0
0 20 40Q
in l / min 60 80 100
Figure 3.25 Computed Q-Qin curves for a 3-way flow control valve. The curves differ in
load/downstream pressure, pB
3:25
Despite minor variations due to spring stiffness and flow forces, constant flow (and
thereby constant actuator speed) is obtained. Both the area of the so-called fixed orifice
as well as the initial compression of the spring can, in general, be adjusted.
If compared to the 2-way flow control valve, it is clear that the bypassed flow to port T
(reservoir) is not used to drive any actuator. The 2-way flow control valve utilizes all
flow that it receives. However, the 2-way flow controller also throttles the working fluid
twice, whereas the 3-way flow controller only throttles the working flow once keeping
down the temperature of the fluid.
Another basic difference is the fact that the 3-way flow controller only works when
subjected to a prescribed input flow Qin = Q + QT and not if subjected to a prescribed
pressure, pA. Hence, it should work together with a pump delivering constant or
controlled flow.
----- oo 0 oo -----
3:26
4
Actuator Control
4.1 Introduction............................................................................... 1
4.5 Accumulators............................................................................ 21
Types Governing equations
4.1 Introduction
In general, the flow, the pressure or the power, i.e., pressure times flow, that is directed
to each individual actuator may be controlled in a number of ways, depending on the
type of system. In the following some classic approaches to speed, pressure and power
control of hydraulic actuators is described.
Load
4:1
A safety valve with crack pressure a certain percentage above the highest load pressure
is necessary to avoid pressure overload. There are no power losses associated with this
system, i.e., all the power put into the pump is directed to the actuator. Hence, the
system is simple = inexpensive and efficient. There are, however, some major
drawbacks. Direct control is only applicable to systems with one actuator working at the
time and always at the same speed. Also the rotational speed of the primary mover (the
motor driving the pump) must be insensitive to moment variations in order for the speed
control to be properly pressure compensated/load independent. Finally, using AC-
motors with rotational speeds fixed at certain discrete values it may be a bit of a puzzle
to find a pump/actuator combination that gives the desired actuator speed.
In this type of systems the speed control is accomplished by the delivered pump flow
rate. In the case where the output actuator is a cylinder or a fixed displacement motor,
variation of the speed is only possibly by variation of the flow, thus we must use a
variable displacement pump. The higher cost, in comparison to fixed displacement
pumps, can to some degree be accepted due to a higher operating efficiencies
With a variable displacement pump, the pump pressure may be controlled by adjusting
the pump displacement. However, it requires a downstream load that increases with
increasing flow and decreases with decreasing flow.
In Figure 4.2 a diagram corresponding to direct pump pressure control is shown
including a fixed orifice, QL, that represents the flow dependant load.
p P
Load
orifice
PC
QP
D- D+ Qmax
Direct pressure control works as follows: If pump pressure, pP, increases, the piston of
the positioning cylinder, PC, will move to the left. This reduces the pump displacement
and, subsequently, the pump flow. As the pump flow is reduced, the pump pressure
(pressure drop across load orifice) will decrease. Contrary, a decrease in pump pressure
will lead to an increase in pump displacement and, subsequently, an adjusting pump
4:2
pressure increase. The spring of the positioning cylinder is compressed corresponding to
the desired pump pressure level.
The simple direct control has no loss associated with it but is not well suited for normal
hydraulic applications because the spring of the positioning cylinder needs to be very
stiff as it is acting directly against the system pressure. This will lead to a noticeable
slope in a pP-QP curve, see Figure 4.2 to the right, because the stiff spring is
compressed corresponding to the travel required to change the pump displacement.
A much more usual pressure control is the pilot operated pump pressure control, see
Figure 4.3.
Load
V2
orifice
p P
O2
QL
O1
QC
QP
pC
pP PC
QP
Qmax
O0
D- D+
Figure 4.3 Pilot operated pump pressure control applied to a load orifice
It works as follows: When pump pressure increases, the spool of the 3/2-way
proportional directional valve, V2, will move up. This will widen the orifice, O1, that
connects the pump pressure, pP, and the control pressure, pC, and narrow the orifice, O2,
which connects the control pressure to the tank reservoir. These variations in orifice
areas will increase the control pressure and move the piston of the control cylinder to
the left, reducing pump displacement, pump flow and pump pressure. Contrary, a
decrease in pump pressure will move the control spool down narrowing O1 and
widening O2. This will lead to a decrease in control pressure and, subsequently, an
increase in pump displacement, pump flow and pump pressure. The spring of the
proportional directional control valve is compressed to give the desired pump pressure
level. The spring of the positioning cylinder is rather weak, set to work against the
control pressure that is reduced by means of the fixed bleed off orifice O0 .
For the pilot operated pump pressure control, the control pressure can be reduced
significantly; hence a much weaker spring can be employed in the positioning cylinder.
In comparison, the stiff spring acting on the control valve is only compressed a small
fraction, hence the pump pressure is almost constant in a pP-QL curve, see Figure 4.3 to
the right.
4:3
The efficiency of the pilot operated pressure control is:
PL Q L p L (Q P Q C ) p P Q
1.0 C (4.1)
PP Q P p P Q P p P QP
where
efficiency of the pump pressure control
PL power delivered to the load, [power]
PP power delivered to the pump, [power]
QL flow across the load orifice, [volume/time]
pL pressure drop across the load orifice = pump pressure, [pressure]
pP pressure rise across the pump = pump pressure, [pressure]
QP pump flow, [volume/time]
QC control flow, [volume/time]
As indicated by Equation 4.1 the control flow should be minimized, hence the discharge
areas of the different orifices, O0..2, are kept as small as possible. This also means that
the flow through the orifices will tend to be of a more laminar nature. It also means that
the control operation may become temperature (viscosity) dependant. Together with the
unwanted transition from turbulent to laminar flow also the sensitivity to contamination
sets a lower bound on the size of the orifices and thereby the control flow. Typical
values for control flow are 2-4 l/min.
LS
V2
O2
QL QC
O1
QP
pC
pP PC
O0
D- D+
Figure 4.4 Pilot operated pump pressure control based on load sensing signal from actuator(s).
Clearly pump pressure control is not well suited for more than one actuator unless extra
valve control is added. Also, it is obvious that if the pump pressure control is applied to
a load = downstream pressure that is insensitive to variations in flow then the pump
4:4
displacement will either be zero (load pressure above set pressure) or maximum (load
pressure below set pressure).
A popular version of pump pressure control is so-called load sensing (LS) where the
highest pressure level in the actuators of the system is used as reference pressure on the
control spool rather than the tank reservoir pressure. The corresponding diagram is
shown in Figure 4.4.
With a variable displacement pump the pump flow may be controlled by controlling the
pressure drop across a fixed orifice. This can be achieved by means of direct control or
pilot operated control. Due to the same conditions as described for pump pressure
control, see section 4.3, this is, however, almost exclusively carried out as pilot operated
control and only this configuration will be gone through here. The set up may look as
shown in Figure 4.5.
V2 QP
OC
O2
QL
Qset
O1
QP QC
pC
pP
PC
p L
O0
D- D+
It works as follows: If the load flow, QL, increases, the pressure drop across the control
orifice, OC, increases. This causes the control spool of the 3/2-way proportional
directional control valve, V2, to move up widening the orifice connecting the pump
4:5
pressure and control pressure, O1, while narrowing the orifice connecting the control
pressure to tank, O2. As a result the control pressure will increase thereby reducing the
pump displacement and, thus, readjust the load flow. Contrary if the load flow decrease
the control spool will move down, reducing the control pressure. This will cause an
increase in pump displacement and thereby a readjustment of the load flow.
For the pilot operated pump flow control, the control pressure can be reduced
significantly as compared to the system pressure, hence a relatively weak spring can be
employed in the positioning cylinder. As a result the load flow is almost constant in a
QL-pL curve, see Figure 4.5 to the right. The slope of the curve is caused by increased
throttling from pump pressure to control pressure until, finally, the spool generated
orifices loose the control ability (the connection between pump pressure and control
pressure almost closed off).
The efficiency of the pilot operated flow control is:
PL Q L p L p Q
1.0 C 1.0 C (4.2)
PP Q P p P p P QP
where
efficiency of the pump flow control
PL power delivered to the load, [power]
PP power delivered to the pump, [power]
QL flow across the load orifice, [volume/time]
pL pressure drop across the load, [pressure]
pP pressure rise across the pump = pump pressure, [pressure]
pC pressure drop across the control orifice, [pressure]
QP pump flow, [volume/time]
QC control flow, [volume/time]
Similar to pump pressure control the control flow should be kept at a minimum. For
pump flow control, however, also the pressure drop across the control orifice reduces
the efficiency, hence the pressure drop across it should be kept at a minimum without
compromising functionality. The lower limit is determined by the minimum spring
stiffness of the control spool required in order to avoid instability. Typical values for
pressure drops across control orifice are 10-20 bar.
In this strategy, using a variable displacement pump, the pump power, i.e., the power
delivered by the pump, is kept constant. There are 2 different ways of obtaining this; an
approximate approach and an exact approach.
The approximate approach corresponds in principle to pressure control in its build up,
see Figure 4.6.
The difference is the extra spring on the control spool. This rather stiff spring is not
activated until the control spool has travelled a certain distance. The effect on the pP-
Qload curve is an increase in slope as shown in Figure 4.6 to the right.
In the curve is also shown the exact hyperbola corresponding to constant pump power.
By inserting an extra spring and thereby forcing a jump in slope, this hyperbola is
approximately met. In some cases up to three springs are used.
4:6
V2
p P
O2
QL Constant power
O1
QC
QP
pC Actual characteristic
pP
PC
QP
Qmax
O0
D- D+
Figure 4.6 Pilot operated pump power control by means of extra spring(s).
Exact power control is achieved if the set up in Figure 4.7 is used. In this case the
desired increase in control spool spring stiffness is obtained by means of gearing. As the
pump is de-stroked the measuring piston, MP, is moved towards the pivot point of the
lever arm, hence, requiring more pump pressure to keep the control spool open. In other
words, as pump flow goes down the pump pressure increases in the same ratio,
maintaining constant pump power.
O2 V2 p P
QL
O1
QC
QP
pC
MP
pP PC
QP
Qmax
O0
D- D+
Figure 4.7 Pilot operated pump power control by means of a measuring piston.
4:7
The pP-Qload curve which corresponds to the exact hyperbola is shown in Figure 4.7 to
the right. The curve is limited by the maximum flow of the pump whereas a separate
pressure overload is required.
The expression for the efficiency of a pump power control corresponds to that of a
pump pressure control, see Equation 4.1. The difference is that the power control is
much better suited for applications with a wide range in actuator speed and load. The
cost is a more complicated system.
Meter-in speed control is obtained by inserting a 2-way flow control valve upstream
relative to the actuator, see Figure 4.8.
Load
pP
QL
QP QR
A constant displacement pump can be used. A pressure relief valve that continuously
keeps the pump pressure at a certain level is required. It by-passes the pump flow above
the set value of the 2-way flow control valve. This way the pressure at the inlet of the 2-
way flow control valve is held constant as required in order for it to work properly.
4:8
The efficiency of the meter in set up may be determined from:
Q L p L Q p L
1.0 R (4.3)
Q P p P QP p R
where
efficiency of the meter in speed control
Qset set flow value of the 2-way flow control valve, [volume/time]
pL load dependant pressure drop across the actuator, [pressure]
QP pump flow, [volume/time]
pP pressure rise across the pump, [pressure]
pR crack pressure of the pressure relief valve = pump pressure,
[pressure]
QR flow across the pressure relief valve, [volume/time]
Obviously, the crack pressure setting of the pressure relief valve should be set as close
as possible to the highest load pressure. Similarly the constant pump flow should be
chosen as close as possible to the actual flow demand, controlled by the 2-way flow
control valve. Hence, the overall efficiency will be low for a system with great variation
in load pressures. Meter-in is well suited for more actuators, however, a 2-way flow
control valve is required for each actuator.
Meter out speed control is obtained by inserting a 2-way flow control valve downstream
relative to the actuator, see Figure 4.9.
Load
pP
QL
QP QR
4:9
A constant displacement pump can be used. A pressure relief valve that continuously
keeps the pump pressure at a certain level is required. It bypasses the pump flow above
the set value of the 2-way flow control valve. Considering a motor as actuator or a
hydraulic cylinder with area ratio = 1.0, the efficiency of the meter out is:
Q L p L Q p L
1.0 R (4.4)
Q P p P QP p R
where
efficiency of the meter out speed control
Qset set flow value of the 2-way flow control valve, [volume/time]
pL load dependant pressure drop across the actuator, [pressure]
QP pump flow, [volume/time]
pP pressure rise across the pump, [pressure]
pR crack pressure of the pressure relief valve = pump pressure,
[pressure]
QR flow across the pressure relief valve, [volume/time]
The same considerations with respect to efficiency and suitability described in the
previous section about meter-in hold for meter-out.
A comparison between meter-out speed control and meter-in speed control yields:
Temperature of fluid in actuator low, because not throttled after leaving pump.
Pressure level on actuator always maximum.
Back pressure on hydraulic cylinders with low area ratio very high.
Negative (runaway) loads may be controlled.
By-pass speed control is obtained by inserting a flow control valve parallel to the
actuator. In Figure 4.10 this is shown with a 2-way flow control valve.
Load
QL
QP
Qset
pP
Figure 4.10 By-pass speed control by means of a 2 way flow control valve.
4:10
A constant displacement pump can be used. A pressure safety valve is used as guard
against pressure overload. It does not work continuously but only when pressure peaks
appear. This means that the pump pressure will never go above the actual load pressure.
The excess pump flow is bypassed by means of 2-way flow control valve. The
efficiency of by pass speed control is:
Q L p L Q P Q set p L Q
1.0 set (4.5)
Q P p P Q P p P QP
where
efficiency of the by pass speed control
QL flow into the actuator, [volume/time]
pL load dependant pressure drop across the actuator, [pressure]
QP pump flow, [volume/time]
pP pressure rise across the pump = load pressure, [pressure]
Qset set flow value of the 2-way flow control valve, [volume/time]
Potentially, by pass speed control has a better efficiency than meter in and meter out,
because the pump pressure never rises above the necessary pressure. Hence, the only
loss associated with by pass speed control is the flow through the 2-way flow control
valve. A comparison between by pass speed control and meter in speed control yields:
Load
QL Qset
QT
QP
pP
Figure 4.11 By pass speed control by means of a 3-way flow control valve.
4:11
Alternatively, by pass speed control may be obtained by directly employing a 3-way
flow control valve, see Figure 4.11.
Here the excess flow is bypassed to tank directly by the valve. The only difference from
the by pass speed control by means of a 2-way flow control valve is that the load
sensitivity of the primary mover has been eliminated at the cost of some minor pressure
drop, e.g., 5 bar, across the 3-way flow control valve. The efficiency of this type of by
pass speed control becomes:
A negative load is defined as a load that tries to move an actuator in the same direction
as the flow. Negative loads always represent a cavitation threat to the hydraulic system.
They typically appear when a load has to be lowered, see Figure 4.12 to the left.
F F
A A
pP pP
pL
pL
A A
Figure 4.12 Two systems holding a load and trying to lower it.
4:12
To the left a system without any load holding capability. To the right a system with
perfect load holding capability but without proper functionality.
For the simple system shown above, the only pressure build up below the cylinder
piston comes from pushing flow through the tank connection of the directional control
valve as well as other restrictions (piping, filter, coolers, etc.) in the flow path before it
reaches the tank reservoir. A load holding pressure, pL, that will maintain equilibrium
with the load and the pump pressure, pP, is necessary. However, if the flow required to
build up this pressure is larger than what the pump is capable of delivering (including
the flow gearing of the differential cylinder in Figure 4.12) then the pressure line of the
pump will cavitate, i.e., in order to obtain equilibrium, pP will try to become negative.
This is impossible and the load runs away.
At the same time, any static load always represents a load drop threat. This may be
caused by pipe/hose bursting and in less critical cases (load dropping slowly) by leakage
from the pressurized regions to the tank reservoir. The latter is especially a problem
when using spool based directional control valves (which is the typical case), as they
cannot be made leakage proof.
These 2 problems: Load drop and runaway loads, may be dealt with in several ways.
Basically, the load drop due to pipe/hose bursting is dealt with by mounting a seat valve
directly on the actuator, see Figure 4.12 to the right. This gives a leak proof load holding
capability, however, it is necessary to lift the poppet/ball from the seat when the load is
supposed to be lowered. The opening pressure may be picked up in 3 different ways.
F F
A A
pP pP
pL
pL
A A
Arv
p p
Figure 4.13 To the left is shown part of a system using a pilot operated check valve to lower a
load. To the right is shown the same system but with a restrictor-check valve added.
4:13
The pilot operated check valve, described in section 3.3, uses pP as opening pressure,
see Figure 4.13 to the left. By increasing the pilot area relative to the seat area, see
Figure 3.13, a relatively low pumping pressure (low power consumption) can pilot the
check valve open when the load is lowered. The check valve has a relatively weak
spring so that it does not offer too much resistance when the load is raised. Ignoring its
contribution, the necessary pump pressure to pilot open the valve may be determined
from:
pL 1 F 1 F
pP p P p P (4.7)
P p A p A
where
pP pump pressure required to pilot open the valve, [pressure]
pL pressure on the piston side of the cylinder, [pressure]
p ratio of the pilot area to the seat area, (rx/r)2 , see Figure 3.13
F load, [force]
A piston area of the cylinder, [area]
cylinder area ratio
The weak spring does, however, also mean that the pilot operated check valve cannot
work as a metering (pressure build up) unit to control a runaway load. It would be much
too unstable as minor pressure fluctuations would cause consistently closing and
opening of the valve. Hence, a pilot operated check valve only solves the load holding
problem. Therefore a pilot operated check valve will normally be mounted in series with
a restrictor valve, see Figure 4.13 to the right. The restrictor valve is adjusted to offer
the flow restriction necessary to build up a sufficient load holding pressure, p L. At the
same time it will have a dampening effect on piston speed and pressure fluctuations in
the system, and hence improve stability. If the pressure drop across the pilot operated
check valve is disregarded and it is assumed that it is piloted fully open, i.e., the pilot
piston is resting against a mechanical stop, then the governing equations for the system
shown in Figure 4.13 are:
F
pL pP (4.8)
A
QP 2
C D A rv pL (4.9)
where
pL pressure on the piston side of the cylinder, [pressure]
F load, [force]
A piston area of the cylinder, [area]
cylinder area ratio
pP pump pressure required to pilot open the valve, [pressure]
QP pump flow, [volume/time]
CD discharge coefficient of the restrictor valve
Arv discharge area of the restrictor valve, [area]
mass density of the fluid, [mass/volume]
4:14
Combining Equation (4.8) and (4.9) yields an expression for the correct setting of the
discharge area of the restrictor valve for a given pump flow, a desired pump pressure
and a maximum load, Fmax:
QP
A rv
2 Fmax (4.10)
CD pP
A
The so-called counter balance valve, see Figure 4.14 to the left, which principally
corresponds to the pressure relief valve discussed in section 3.3 uses pL as opening
pressure.
F F
A A
pP pP
pL
pL
A A
pcr pcr
p
Figure 4.14 To the left is shown part of a system lowering a load by means of a counterbalance
valve. To the right is shown a system lowering a load by means of an over centre valve. Both
systems have a check valve in parallel with the load holding valve.
The crack pressure is set a certain percentage above what is required to maintain
equilibrium with the load. This way both the load holding problem as well as the load
runaway problem is taken care of. A separate check valve that allow by pass of the flow
when lifting the load is necessary.
Neither the pilot operated check valve nor the counter balance valve are well suited for
varying loads because they are set to handle the maximum load, thereby causing very
high pumping pressures when lowering smaller loads. For systems with greatly varying
loads a so-called over centre valve may be used. It is a combination of the pilot operated
check valve and the counter balance valve as it uses both pP and pL as opening pressure,
see Figure 4.14 to the right. The pump pressure will typically act on an area 3-10 times
larger than the area acted upon by the load holding pressure. The ratio between the area
4:15
acted upon by the pump pressure and the area acted upon by the load holding pressure is
referred to as the pilot area ratio. Ignoring the pressure in the return line the governing
equations for the system shown in Figure 4.14 to the right are:
F
pL pP (4.11)
A
pcr p L p p P (4.12)
where
pL pressure on the piston side of the cylinder, [pressure]
F load, [force]
A piston area of the cylinder, [area]
cylinder area ratio
pP pump pressure required to pilot open the valve, [pressure]
pcr crack pressure of the over centre valve, [pressure]
p pilot area ratio of the over centre valve
The crack pressure is defined as the load holding pressure required to open the over
centre valve without the aid of any pump pressure. The crack pressure is adjusted by
adjusting the initial compression of the spring. Combining Equation (4.11) and (4.12)
yields an expression for the correct setting of the crack pressure for a given pump flow,
a desired pump pressure and a maximum load, Fmax:
Fmax
p cr p p P (4.13)
A
In general, the effect of using both pL and pP as opening pressures tends to stabilize the
system. However, too high pilot area ratios will cause instabilities even for over centre
valves. The effect of changes in the load is strongly reduced, see Figure 4.15.
pP
Fmax / A
2.4
1.2
0.6
Over center valve
0
0 0.2 0.4 0.6 0.8 1
F / Fmax
Figure 4.15 The pump pressure dependency on the load for the different types of load holding
valves
4:16
4.3.5 Braking
Figure 4.16 Simple hydraulic system shown lowering a load and breaking.
When the valve goes to neutral the load will, due to inertia, continue its motion. This
will lead to a compression of the fluid in branch A and a decompression of the fluid in
branch B. Hence pL increases and pP decreases. For large inertia loads and systems with
insufficient damping this may easily cause severe problems, both with respect to
overloading in branch A due to pressure peaks and cavitation in branch B. These
problems are typically handled by inserting a shock valve and a suction valve, see
Figure 4.17.
The shock is simply a pressure safety valve, see section 3.3, dimensioned to a rather
small flow. It is set a certain percentage above the maximum expected static load
pressure. The suction valve is a check valve, see section 3.2, with a very weak spring, so
that the back pressure can crack it open and refill the branch in danger of cavitating. If
the back pressure/tank reservoir pressure is not high enough, an extra check valve may
be inserted in the return line with a somewhat stiffer spring, ensuring sufficient suction
pressure.
4:17
Figure 4.17 Simple system shown breaking a load. The system is equipped with a shock valve,
a suction valve and a check valve in the return line to maintain a certain back pressure.
4:18
The closed circuit is used for applications with frequent reversals of the motor speed.
For such systems it is preferrable to change the amount as well as direction of flow
smoothly via the pump avoiding an expensive and complicated valve control system.
Closed circuit hydrostatic transmissions are widely used for propelling small and
medium sized skid-steered vehicles employed in agriculture or on building sites where
the demand on maneouvreability is so high that the relative poor efficiency of a
hydrostatic transmission can be neglected. Open circuit hydrostatic transmissions
requires a directional control valve in order to reverse the speed of the motor. It is
primarily used in applications with a high demands on torque to volume ratio on the
actuator.
In practice, any hydrostatic transmission requires a number of extra valves to ensure
proper functionality during negative loads and braking, see also sections 4.3.4 and 4.3.5.
For a closed hydrostatic transmission used for vehicle propelling the system shown in
Figure 4.19 can be considered as a minimum.
7
8 8
1 5 3
6
2 7 4
In any pump or motor there will be internal leakage from the high pressure side to the
housing of the component. In order to reduce the pressure level and thereby the
requirements to the sealings there must be a drain connection to a reservoir. For a closed
circuit this must be added as a separate line and both the pump (1) and the motor (3) has
each their drain connection to tank, (2) and (4), respectively. Depending on the physical
layout these drain connections may be joined before reaching the reservoir. Because of
the drainage the closed circuit needs continuous refilling. Therefore, a small circulation
pump (5) is introduced that continuously pumps fluid from a reservoir into the closed
system. It is mechanically connected to the pump and maintains a certain minimum
pressure (for example 15 bar) by means of the pressure relief valve (6). Whenever the
pressure level in either branch of the closed circuit falls below the minimum pressure
the branch is refilled via the check valves (7). A major concern for closed circuits is the
lack of cooling because the main pump does not pull fluid from a reservoir, i.e., the
same fluid is constantly recirculated. This problem is also greatly reduced by the small
circulation pump, since it continuously refills the closed circuit with fluid that has been
cooled down in the reservoir. When the motor is accelerated or braked by highly
dynamic external loading one branch will cavitate and the other will experience high
pressure peaks. To avoid this, two shock valves (8) are added. Together with the check
valves (7) they correspond to the suction-shock valve arrangement of the cylinder
actuator in Figure 4.17.
4:19
As mentioned, the overall efficiency is the major drawback of a hydrostatic
transmission. In Figure 4.20 a closed hydrostatic transmission where both pump and
motor has variable displacement is shown.
p1
Q Q
n1 M 1 n2 M 2
D2
D1
p2
Figure 4.20 Closed hydrostatic transmissions with variable displacement pump and motor.
D1 1 D1, max
(4.14)
D2 2 D2, max
2 D2, max
i (4.16)
1 D1, max
n
n2 1 (4.17)
i
From Equation (4.16) it is clear that the displacement of a pump may change sign, i.e.,
1 1 1 with 1 0 simply causing the motor to stand still no matter the pump
speed. However, a motor with variable displacement will normally will have some
safety measure that ensures: 2, min 2 1 , where 2, min is some positive value, since
2 0 would give infinite motor speed for any finite flow.
The theoretical torque is:
D1 p 1 D1, max ( p1 p2 )
M1 D
2 2
M 2 2 2, max M 1 i M 1 (4.18)
D p 2 D2, max ( p1 p2 ) 1 D1, max
M2 2
2 2
It is not a practical problem that i for 1 0 since the output torque is determined
by the external load. Hence 1 0 simply implies that theoretically no pump torque is
needed to drive the motor when the pump displacement is zero.
4:20
Disregarding the losses the input power and output power balances:
n1
P2 2 n2 M 2 2 i M 1 2 n1 M 1
i
(4.19)
P 2 P1
Q v1 n1 D1
n
1 n2 v1 v 2 1 (4.20)
Q n2 D2 i
v 2
1 D1 p
M1
hm1 2 M i M
2 hm1 hm2 1 (4.21)
D2 p
M 2 hm2
2
P2 2 n2 M 2 2 v1 v 2 n1 hm1 hm2 M 1
(4.22)
P2 v1 v 2 hm1 hm2 P1
In a real system other losses would include the pressure drops in the lines of the two
branches as well as the continuous pressure supply to the circulation pump.
4.5 Accumulators
4:21
Gas
Gas Bladder
Piston
Fluid
Fluid
The two types of accumulators are illustrated in Figure 4.21. In general, the piston
accumulator is used whenever high pressure and high safety are required. The
drawbacks as compared to the bladder accumulator are the necessary sealing between
the piston and the cylinder as well as the low response time due to the inertia of the
piston.
Because the main purpose of an accumulator is to be loaded with high pressure and,
simultaneously, releases its energy within short time intervals it is important to have it
connected to the remaining hydraulic system in a safe way. This is typically done by
means of a valve block that as a minimum has a system isolator valve, a manually
actuated unloading valve and a pressure relief valve, see Figure 4.22.
Pressure relief
Unloading valve Isolator
valve valve
Figure 4.22 Accumulator shown built into a hydraulic system. In this system the functionality
of the accumulator is to maintain the supply pressure as constant as possible.
The isolator valve protects the pump from accumulator back flow, the unloading valve
allows for the accumulator to be de-energized after shutdown and the pressure relief
valve protects the accumulator.
4:22
4.5.2 Governing equations
The total volume of the accumulator is fixed. It is the sum of the gas and fluid volume:
Va Vg V f cst (4.23)
When the accumulator is in its initial state (preloaded) the gas takes up all the volume,
see Figure 4.23.
Vg ,min
V0 Va Vg ,max
pmax
p0 pmin
V f ,max
V f ,min
Vf 0
Figure 4.23 Accumulator shown in three states: initial=preloaded, loaded with minimum
working pressure and loaded with maximum working pressure.
The accumulator is loaded by pressurizing the fluid whereby the gas volume is reduced.
The accumulator is typically unloaded by connecting it to a part of the hydraulic system
that has a lower pressure. The pressure of the gas follows that of the fluid (force
equilibrium across the separator) during loading and unloading and the gas can be
considered as undergoing a polytropic proces:
The polytropic exponent varies with the proces. If no experience or measurements are
available approximation must be made regarding the proces. One way is to choose the
proces that yields the most conservative design, however, it is preferrably to get more
precise data from the supplier. Nitrogen gas is by far the most used gas and the
polytropic exponent for Nitrogen varies strongly both with temperature, pressure and
duration of proces. It is also different depending on whether the accumulator is being
loaded or unloaded. If no data is available it is suggested to use the following rough
approximation for the polytropic exponent corresponding to an adiabatic and an
isothermal proces, respectively:
1.4 t p 10 s
n (4.25)
1.0 t p 300 s
4:23
In general, it should be avoided (especially for bladder accumulators) that the pressure
falls below the preload pressure, p0 . As a rule of thumb the preload pressure should be
90% of the minimum working pressure:
When an accumulator is used for the earlier mentioned purposes it is normally used in
the following way: initially, it is loaded isothermally with the minimum pressure. Next,
it is loaded to maximum pressure and at a certain time it is unloaded towards an
actuator. Alternatively, it is loaded isothermally with an intermediate pressure and then
passive against a prescribed inlet flow. In both cases it is normally given how much
volume of fluid that the accumulator should accumulate/discharge. Also, the acceptable
limits on the working pressure would typically be specified as pmax and pmin . The
maximum pressure, pmax , is a design parameter that often is set by means of a pressure
relief valve or simply by a maximum acceptable pressure level. The desired variation in
volume, V , is normally given by the requirements on the actuator motion. Knowing
these values it is possible to compute the maximum and minimum gas volume, see
Figure 4.23:
1
p nL
(4.27)
pmax Vgn,Lmin pmin Vgn,Lmax Vg , max max Vg , min
pmin
1
pmax n L
V Vg , max Vg , min 1 Vg , min
pmin
(4.28)
1 1
Vg , min V Vg , max V
1
1
pmax n L pmin n L
p 1 1 p
min
max
Since, the initialization should be considered isothermal the total volume of the
accumulator may be determined as:
pmin
p0 V0 pmin Vg , max V0 Va Vg , max (4.29)
p0
Hence, the data necessary to specify an accumulator, namely the total volume and
corresponding preload pressure may be determined from Equations (4.26) and (4.29),
respectively.
---- oo 0 oo ----
4:24
5
Synthesis of
Hydraulic Systems
5.1 Introduction............................................................................... 1
5.1 Introduction
The designer rarely needs to design the individual components. His task is to conceive a
strategy for solving the problem – of cause appropriate to the particular application, then
to represent this strategy on a circuit diagram, and at last to select the components for
his system from a wide range of commercial stock.
This chapter will outline a basic approach for arriving at a hydraulic system capable of
performing a specific task – giving a stepwise approach to synthesis of hydraulic
systems. Hence, the chapter will address how to combine and size some of the basic
components of a hydraulic system: Pumps, valves, actuators, lines, filters, coolers and
tank reservoir.
When designing a hydraulic system the following steps should be addressed, see also:
"Design and Steady-state Analysis of Hydraulic Control Systems" by Jacek S. Stecki et
al., ISBN 83-86219-94-7:
1. Define required operating cycle for entire system.
2. Define the required operating cycle for each sub function.
3. Decide on system concept:
4. Setup a hydraulic diagram with the necessary components.
5. Decide on pressure level.
5:1
6. Select and size the actuators.
7. Select and size the primary mover and pump(s).
8. Select the type of hydraulic fluid.
9. Select size of hydraulic lines.
10. Select control elements.
11. Determine overall efficiency of system.
12. Select and size tank reservoir and/or cooling elements.
13. Select suitable filtering.
14. Do a system analysis and determine reliability.
15. Prescribe procedures for system assembly, monitoring, operation and component
replacing.
16. Estimate costs.
5:2
power requirements of each actuator. Hence, for a hydraulically actuated system it
should be possible to set up the required force and speed for each cylinder and the
required torque and angular velocity for each motor, see Figure 5.1..2.
Velocity Force
Steps 1 and 2 should return curves as the ones shown in the above figures.
Steps 3 and 4 correspond to finding a concept for the hydraulic actuation. Basically, this
means determining how the fluid power should be generated and directed to the
different actuators. At this stage it should be determined:
Number of pumps
For each pump how should it be driven and controlled
For each actuator how should it be controlled
In hydraulics it is customary to use fewer pumps than actuators (often only one pump).
This is due to costs because the pump is one of the most expensive components in any
hydraulic system. More pumps can be justified if for instance two simple inexpensive
pumps can replace one complicated and expensive one. Otherwise, adding pumps are in
general done with a view to improve the efficiency.
When there are more actuators than pumps it is, in general, necessary to insert valves to
direct and control the fluid power to each actuator. The basic conceptual building blocks
are the types of pump and actuator control described in Chapter 4 together with the
5:3
basic valves described in Chapter 3. As an example, a pressure controlled pump may be
used in conjunction with flow controlled actuators if the speed of the mechanical system
should be controlled. Alternatively, if the force supplied by the actuator is the main
concern then pressure controlling valves should be utilized.
The proper combination of pumps, valves and control methods cannot be chosen
without doing some sizing, i.e., the correct topology of the hydraulic system must be
determined iteratively and should include computation of the dimensions of the main
components. An important tool during these iterative steps is the standardized approach
to depicting hydraulic systems as so-called CETOP-diagrams. These diagrams are of a
conceptual nature and are well suited both during the design phase but also as final
documentation. However, a CETOP-diagram alone is not enough to properly describe
the operation of a hydraulic system and it should always be accompanied by some kind
of written description of the intended behavior of the system during operation.
The pressure level is typically chosen to lie within 150-250 bar. If pressure levels within
this interval results in very small flow demands, e.g., less than 3-5 l/min or very large
flow demands, e.g., > 1000 l/min then pressure level should probably be reduced or
increased, respectively. Alternatively, the entire system concept might need to be
reevaluated.
2 M max
D (5.1)
p
where
D displacement of the motor, [volume/revolution]
M max maximum load torque on the motor, [moment]
p chosen pressure level, [pressure]
The greater than sign allows for an estimated hydro-mechanical efficiency of the motor
and pressure drops within the system to be taken into account at this stage.
If the actuator is a cylinder, the piston area may be determined from:
Fmax
A (5.2)
p
where
A piston area of the cylinder, [area]
Fmax maximum load on the cylinder, [force]
p chosen pressure level, [pressure]
The greater than sign allows for an estimated hydro-mechanical efficiency of the
cylinder, back pressure on the annulus area as well as pressure drops within the system
to be taken into account at this stage.
5:4
5.2.5 Primary mover and pump sizing
The necessary pump flow is determined based on the situation during the system
operating cycle, where the total flow demand is at its maximum:
motors cylinders
Qmax Di ni Ai vi (5.3)
i i max
where
Qmax maximum required theoretical flow, [volume/time]
Di displacement of the i'th motor, [volume/revolution]
ni required rotational speed of the i'th motor, [revolution/time]
Ai piston area (or by reversed flow the annulus area) of the i'th
cylinder, [area]
vi required speed of the i'th cylinder, [speed]
The power demand to the primary mover may, relatively simple, be determined from the
maximum required theoretical flow and the pressure level:
The greater than sign allows for an estiamte of the total efficiency of the pump and the
volumetric efficiency of the actuators to be taken into account at this stage.
It should bet noted that in mobile hydraulics the primary mover is not a free choice,
since it is the engine of the vehicle. Otherwise, the primary mover is chosen among
stationary combustion engines or electric (AC and DC) motors.
Knowing the rotational speed of the primary mover, the pump displacement may be
determined:
Qmax
D (5.5)
n
where
D pump displacement, [volume/revolution]
Qmax theoretical maximum flow demand, [volume/time]
n rotational speed of the primary mover, [revolutions/time]
Here the greatet than sign accounts for an estimated volumetric efficiencies of pump and
actuators. The rotational speed of the primary mover, n, will depend on the torque
required by the pump. However, this torque is a function of the pump displacement.
pD
M (5.6)
2
where
M output torque from the primary mover, [moment]
5:5
p chosen pressure level, [pressure]
D pump displacement, [volume/revolution]
The greater than sign covers an estimated hydro-mechanical efficiency of the pump at
this stage. Clearly Equation (5.5) and (5.6) depend on each other. Hence, if the
dependency between torque and rotational speed is relatively complex for the primary
mover, then the pump displacement needs to be determined iteratively.
Hydraulic fluids are gone through in detail elsewhere. Its main purpose is to transfer the
hydraulic power, lubricate the moving parts and protect against corrosion. The dominant
type of hydraulic fluid is mineral oil. They are replaced by water-based fluids in
applications with fire hazards, and by environmental friendly fluids (biologically
degradable) when special attention to the surroundings must be addressed.
The mineral oils are categorized according to DIN 51524 into: H, HL, HLP, HV and
HLPD. Except for type H that is rarely used the others are supplied with additives that
improve
corrosion protection, HL, HLP, HV and HLPD
oxidation protection, HL, HLP, HV and HLPD
wear reduction, HLP, HV and HLPD
viscosity independency on temperature, HV and HLPD
ability to self-clean and resistance to dissolve water, HLPD
The main parameter of a hydraulic fluid is the viscosity and its dependency on
temperature. It is necessary that the viscosity of the fluid at start and during operation
stay within the acceptable limits of the components of the system. Hence, it is necessary
to know the temperature of the surroundings as well as the overall efficiency of the
system to get an exact estimate of the possible viscosity range of the fluid. This should
be compared with the acceptable viscosity range for each component of the system.
However, these computations can only be performed when the entire system sizing has
been carried out. As an initial approximation it may be expected that the operating
temperature may lie in the range 40-60 C.
The hydraulics lines consist of pipes and hoses. First of all, the pipe and hose
dimensions must be chosen so that their allowable pressure is above the pressure level
of the system. Also, the dimensions should be chosen so a suitable fluid velocity is
obtained. The following values are suggested:
Suction lines (from tank reservoir to pump) : 0.5 – 2.0 m/s
Delivery line (from pump to actuators): 3.0 – 10 m/s
Return line (from actuators to tank reservoir): 1.0 – 3.0 m/s
Especially, in the suction line is it essential to avoid any pressure drops so that the
suction pressure of the pump may be adequate for it to work optimally. Both in the
delivery as well as the return line there is a trade off between strength (smaller
diameters needs less thickness to withstand same pressure) and low velocity (larger
diameters).
5:6
The pressure drop in any hydraulic line depends on several parameters. Firstly, the type
of flow, laminar or turbulent, has to be established. This is done based on the
dimensionless Reynolds number:
4v A
Re (5.7)
O
where
v mean fluid velocity, [length/time]
A cross sectional area of the flow path, [area]
kinematic viscosity, [area/time]
O circumference of the flow area, [length]
Hence, for a circular line (pipes and hoses), the Reynolds number is:
4 Q
Re (5.8)
d
where
Q flow in the pipe, [volume/time]
kinematic viscosity, [area/time]
d inner diameter of the line, [length]
At low Reynolds numbers the flow is laminar, whereas it becomes turbulent rather
abruptly around a critical value of approximately 2300.
The pressure loss due to flow in a line is given by:
L v2
p (5.9)
2d
where
is a dimensionless resistance number
L is the length of the line, [length]
is the mass density, [mass/volume]
v is the mean fluid velocity, [length/time]
d is the inner diameter of the line, [length]
The resistance number can be found when the Reynolds number and the type of flow
have been established. For laminar flow it is:
64
(5.10)
Re
For turbulent flow it is:
0.3164
(5.11)
Re0.25
Equation (5.8) (5.11) is the necessary tools to determine the pressure losses in the
hydraulic lines of the system. Knowing the approximate mechanical system as well as
the position of the actuators, it is possible to reasonably estimate the different line
lengths.
5:7
Another pressure loss source in the hydraulic lines is associated with any bends or
branching of the flow path. As a rough estimate on the pressure drop around a bend the
following formula may be employed:
v2
p (5.12)
2 90
where
dimensionless pressure drop coefficient number
mass density, [mass/volume]
v mean fluid velocity, [length/time]
d inner diameter of the line, [length]
angle of the bend, see Figure 5.1, [degrees]
The pressure drop coefficient depends on the ratio between the bend radius and the inner
diameter, see Figure 5.3, and may roughly be determined
0.3
0.6
rb (5.13)
d
where
dimensionless pressure drop coefficient number
rb radius of the bend, [length]
d inner diameter of the line, [length]
rb
Figure 5.3 Illustration of the variables associated with the pressure drop across a bend in a
hydraulic line.
The necessary control elements are determined from the design concept. The valves
may either be mounted on a plate or screwed into a housing as cartridge valves. In
general, it is a good idea to keep the valves close together and to use ports with the same
dimensions. I.e., if the delivery line is, e.g. ½", then the valves should, preferably have
½" connections. It is important to know the demands from the valves to contamination
(filtering) and viscosity (fluid and fluid temperature).
5:8
motors cylinders
2 M i ni Fi vi
P (5.14)
out i i
Pin M PM n PM
where
efficiency of the system at a certain time
Pout power delivered by the system, [power]
Pin power put into the system, [power]
Mi output torque delivered by the i'th motor, [moment]
ni rotational speed of the i'th motor, [revolutions/time]
Fi output force delivered by the i'th cylinder, [force]
vi speed of the i'th cylinder, [length/time]
M PM output torque from the primary mover, [moment]
nPM rotational speed of the primary mover, [revolutions/time]
Normally, the efficiency will vary during the operating cycle of the entire system.
Hence the overall efficiency of the system should be determined as:
tc Pout
tot dt (5.15)
t 0 Pin
where
tot overall efficiency of the system
Pout power delivered by the system, [power]
Pin power put into the system, [power]
tc duration of the operating cycle for the entire system, [time]
The losses associated with the efficiency of the system will always lead to heating of the
fluid. In general, all the power losses in the system are used to generate heat in the fluid.
Assuming that the temperature of the fluid is uniform within the entire system and that
the only heat exchange with the environment is via the surface area of the tank reservoir
then a heat balance may be set up for a steady state situation:
s e (5.16)
where
s heat flux generated by the system, [power]
e heat flux transmitted to the surroundings/environment, [power]
The average heat flux generated by the system may be determined as:
The heat flux transmitted to the environment from the tank reservoir may be determined
as:
5:9
e k c Ar T (5.18)
where
e heat flux transmitted to the surroundings/environment, [power]
kc heat transmission coefficient for the tank, [power/(areadegree)]
Ar area of the cooling surface of the tank reservoir, [area]
T temperature difference between the hydraulic fluid and the
surroundings, [degree]
Typical values for the heat transmission coefficient are: 10..15 W/(m2K) for steel
reservoirs and 6..9 W/(m2K) for cast iron reservoirs. Combining Equation (5.17) and
(5.18) leads to an expression for the necessary surface area of the tank reservoir based
on an acceptable fluid temperature:
s
Ar (5.19)
k c T
As a rule of thumb the tank reservoir should be approximately 3..5 times the pump flow.
In cases where the requirements to the reservoir surface area becomes too large, heat
exchangers are widely used. They are either water or air driven, and their power
consumption capability is easily determined once a tank reservoir has been chosen:
he s kc Ar T (5.20)
where
he required cooling capability of the heat exchanger, [power]
If steady state is not reached for the temperature during an operating cycle then
Equation (5.19) may be considered somewhat conservative and the more general
equation for the development of the fluid temperature as a function of time may be
employed:
kc Ar
t
s s M fl c fl
T( t ) TS T0 TS e (5.21)
k
c r A k c Ar
where
t is the time, [time]
M fl is the mass of the fluid, [mass]
c fl is the specific heat of the fluid, [energy/(massdegree)]
Ts is the temperature of the surroundings, [degree]
T0 is the initial temperature of the fluid, [degree]
5.2.11 Filtering
In order for the hydraulic system to function properly over a longer period proper
filtering is crucial. Filters may, basically, be inserted in suction lines (protect pump),
5:10
delivery lines (protect valves and actuators) and return lines (remove picked up
contamination). Return line filtering is most common as the demands to and hence cost
of a return line filter is smaller than that of a delivery line (high pressure) filter. Filters
in the suction line protect the expensive pump but might cause cavitation if the pressure
drop across it becomes too large. Filters typically come with a by pass check valve that
allow free flow if the filter is blocked.
The efficiency of the filter is expressed by means of a value. It is defined as follows:
nin
x (5.22)
nout
Where
x efficiency w.r.t. to particles of the size x m
nin number of particles of the size x m pr. 100 ml entering the filter
nout number of particles of the size x m pr 100 ml leaving the filter
Hence a filter with 10 75 will pick up 98.7% of 10 m particles trying to pass
through it. In general, the component information will include requirements to fluid
cleanliness and recommendations to the necessary filtering.
Fluid cleanliness is classified according to ISO 4406 and consists of 2 numbers that
define the allowable number of 5 m and 15 m particles, respectively, in a 100 ml
sample.
The design approach described in Section 5.2 is well suited for systems with a
predominantly static behavior. However, many hydraulically actuated systems have a
strongly dynamic nature and this needs to be taken into account. The hydraulic system
influences the dynamic performance of the entire system and, simultaneously, the
dynamic performance of the entire system influences the choice of hydraulic
components. The most important type of hydraulically actuated systems where dynamic
design considerations have to be made are, in general, referred to as servo systems.
A servo system is a tracking control system that receives a reference signal, measures its
own output and controls the output to be in accordance with the reference. For a
hydraulic servo system the output to be controlled is typically a position, speed or force
and, accordingly, the servo system is labelled either a position servo, speed servo or a
force servo. In Figure 5.4 a typical valve controlled hydraulic servo system is shown
with its basic components. A typical hydraulic servo consists of:
Constant pressure source
Directional control valve
Hydraulic lines between valve and actuator
Actuator
Mechanical system - payload
5:11
Output sensor
Servo amplifier
Actuator
Directional
uref K control valve
Servo
amplifier
Constant
pressure supply
Figure 5.4 Hydraulic servo system with linear actuator.
In order to reduce disturbances and non-linearities in the system the supply pressure is
normally held at a constant value by means of accumulators together with either
variable displacement pumps or pressure control valves. This is done at the expense of
efficiency which can become quite small depending on the characteristics of the
working cycles.
The valve is a high-end directional control valve and it is characterized by a high
bandwidth (low response time), internal control loop that positions the spool according
to an electrical input signal and with finely machined tolerances that ensures smooth
transition in performance when the valve is operating around the neutral position.
Directional control valves for servo applications are referred to as either servo valves
(servo applications) or proportional valves (spool position is propotional to input
signal). The distinction between the two types is not well defined but there is a tendency
to refer to valves depending on the type of actuation used to move the spool:
Single-stage - the spool is directly actuated by means of an electrical linear force
motor. The inner control loop is realised via electrical position feedback.
Two-stage - the spool is hydraulically actuated by means of an electrically
actuated pilot stage. The inner control loop is realised via mechanical position
feedback.
Three-stage - the main spool is hydraulically actuated by means of a two stage
valve. The inner control loop of the main spool is realised via electrical position
feedback.
The two- and three-stage control valves have higher bandwidth but also requires extra
filtering (built in), requires a pilot flow, costs more and their response time depends on
the level of the supply pressure. The three-stage control valve is the most expensive and
is only used for high power applications where flow and pressure demands cannot be
met by a two-stage valve. The valves are also, classically, referred to according to the
layout of the spool around neutral position, see Figure 5.5.
5:12
A A A
T P T P T P
This greatly influences the performance of the valve around its neutral position and for
such applications this must be taken into account. Normally, a number of steady state
characteristics are given for a commercial valve that describe how the valve is lapped.
The hydraulic lines between the valve and the actuator are normally reduced in size as
much as possible in order to maintain as high an effective stiffness as possible of the
fluid. This is important in order to have as low a response time as possible for the
hydraulic part of the servo system. Appropriate peak fluid veclocity for such lines may
be as high as 30 m/s.
The actuators are the usual rotary=motor and linear actuators=cylinder gone through in
chapter 2. Normally, the inertia of the actuators are negligible in hydraulically actuated
systems when compared with that of the payload. The use of cylinders pose a number of
challenges that is not encountered when using motors. For high speed applications the
cylinder friction may pose a problem and more expensive cylinders where extra
measures have been taken to reduce both stiction and coulomb friction are used. Also,
the effective stiffness of hydraulic cylinders change significantly with stroke, see
Section 6.3, and this must be considered in the design phase. Also, cylinders with
differential areas should be avoided for applications where the valve operates around its
neutral position because the dynamic characteristics of the system will be changing
abruptly all the time. Because of this, cylinders with identical pressure areas, i.e. 1 ,
are used frequently, see Figure 5.4. When using differential area cylinders different
layouts may be utilized, see Figure 5.6.
5:13
When a hydraulic motor is used as actuator the controlled variables of the different
types of servo systems are: rotation, rotational speed and torque. For such systems the
effective stiffness is almost constant, see Section 6.3.
The mechanical system should be well understood, especially its effective inertia and its
effective stiffness, see Section 6.3. Often the mechanical system to be actuated is quite
stiff as compared to the hydraulic system, however, this is not always the case, and in
such situations the extra flexibility introduced by the mechanical structure must be
included when computing the eigenfrequency of the system. Special care must be paid
to any type of sliding connections where friction may appear since this can greatly
reduce the controllability of the mechanical system. Ideally, the possible implications of
friction should be investigated either via model based sensitivity analysis or, if this is
not sufficient for a reliable design, via experimental work.
The sensor obviously depends on the type of servo and can be in the shape of a
potentiometer, an encoder, a tachometers or an accelerometer for a position or speed
servo. For a force servo it may be sufficient to measure the pressure drop across the
actuator, however, if friction is high it may be necessary to measure the actual force
produced at the piston rod or at the payload.
The servoamplifier receives the measured signal and a command reference signal. The
difference is fed into a controller that generates a command signal to the valve. In some
cases the electronics is integrated in the valve. It is important to note that depending on
the type of servo the valve is operating around neutral (position servo) or around a
metered out position (speed and force servo), see Figure 5.7. Hence, a differential area
cylinder can be used for a force servo because the jumps in dynamic characteristics
around the neutral valve position are avoided.
Figure 5.7 Position (equal area cylinder), speed (motor) and force servo (differential area
cylinder).
When generating motion references for servo systems it is necessary to take into
account the natural eigenfrequency of the hydraulic mechanical system. In Section 6.3 it
is gone through in detail how to compute the eigenfrequency of hydraulic mechanical
5:14
systems. This requires the computation of the spring stiffness of the hydraulic system
and the effective inertia of the mechanical system. It should be noted that if the
mechanical system has a flexibility that cannot be ignored it should be treated as a
spring in series with the spring representing the hydraulic system.
The natural eigenfrequency gives an indication of the limit that the mechanical system
imposes on the desired motion. Consider a 2nd order under-damped system with a
natural eigenfrequency, n , a damping, and a mass, m. The mass is traveling at a
speed, y v0 , and the motion of the mass should be ramped down via the reference
input x , see Figure 5.8. In principle this corresponds to halting a hydraulically
controlled payload by ramping down the input flow.
v0
k
m
b y
x
Figure 5.8 A second order system with a mass traveling at a speed y v0 .
Initially, x y v0 and then the reference velocity is ramped down according to:
t
x v0 1 (5.23)
tR
In (5.23) t is the time and t R is the ramp time during which the mass is decelerated. The
analytical solution to the motion of the mass is:
y C0 C1 t C2 t 2 C0 e t cos( t ) sin( t )
v v
C0 2 0 C1 v0 C2 0 (5.24)
n t R 2 tR
n n 1 2
In Figure 5.9 the variations in the reference motion and the motion of the mass are
shown. The reference velocity becomes zero at t t R . At that instant the position error,
referred to as the overshoot, is:
v0
e xt t R yt t R 1 e t R cos( t R ) sin( t R ) (5.25)
n2 t R
The relative overshoot is the absolute position error relative to the nominal travel:
5:15
e 2
1 e t R cos( t R ) sin( t R ) (5.26)
xt t R n2 t R2
1.2
y*
1
x*
0.8
[ ]
0.6 rad
n 100 0.2
s
0.4
m
v0 1 t R 0.06 s
0.2 s
0
0 3 6 9 12 15
t* t n [ ]
Figure 5.9 Reference motion and actual motion of a 2nd order system subjected to a ramp down
of the velocity. The relative overshoot is shown.
In Figure 5.10 the relative overshoot is plotted as a function of the ramp time and the
damping. The eigenfreguency has no influence on the curves.
100 [%]
100
80
60
40 0.2
20
0.8
0
0 2 4 6 8 10 12 14 16
t R* t R n [ ]
Figure 5.10 Relative overshoot as a function of ramp time and damping.
Since most hydraulic systems have dynamic characteristics that are not far from those of
the system in Figure 5.8 the above results may be utilized for a hydraulic servo system.
From Figure 5.10 it is clear that for a typical hydraulic mechanical system ramp times
should be considered whenever motion is prescribed. Also, it is clear that a useful ramp
time depends on the acceptable overshoot. Hence, if high overshoot is acceptable small
ramp times may be prescribed and vice versa. As a rule of thumb, the ramp time of a
prescribed motion should obey the inequality:
6
tR (5.27)
n
5:16
It is very important to keep in mind that the eigenfrequency of a real system may be
substantially smaller than the one computed for a model of the system. The discrepancy
is mainly because the modeled stiffness typically is higher than the actual stiffness. This
can be taken into account in many ways, for example by using a smaller stiffness in the
computation or by means of experimental work that can reveal the actual
eigenfrequency(ies) of the system to be controlled.
For stiff systems with high natural frequency the limitation on the ramp time might
come from other sources such as:
pressure level required to generate the ramp acceleration
response time of directional control valve
response time of variable displacement pump if present
These lower bounds on the ramp time must be considered from application to
application, however, in general, ramp times should be kept as small as possible because
this reduces the maximum speed and thereby the energy usage and the pump flow. This
can be illustrated by considering a simple task of translating a payload of mass, m 1kg ,
a distance, s 1m , within a given time t 1s . In that case the maximum velocity and
the maximum kinetic energy depend on the ramp time as shown in Figure 5.11.
2 2
m 1kg s 1m t 1s
1.5 1.5
v0
m
s 1 E kin 1 [J]
0.5 0.5
0 0
0 0.1 0.2 0.3 0.4 0.5
tR
[ ]
t
Figure 5.11 Maximum velocity and maximum kinetic energy as function of ramp time.
Considering losses it should be noted that most of the flow losses in valves and lines
also increase with the maximum velocity.
Let us consider the basic operation; to move a payload a certain distance within a certain
time. Let s denote the travel associated with the operation. let t denote the time
available for the operation and let be a dimensionless number, 0.5 , that denotes
the ratio between the ramp time and total time. Initially, some considerations should be
made concerning the shape of the velocity profile. In Figure 5.12 the velocity profile
associated with a linear ramp, i.e., constant acceleration, and a cosine-shaped ramp, i.e.,
a sinus-shaped acceleration, are shown. Clearly, the sinusoid has the advantage that it
increases the degree of continuity at the ramp boundaries. This will, especially for very
stiff systems, reduce impact loads, vibrations and noise. The disadvantage is the
increased demands on the valve control and the higher maximum acceleration (half way
through the ramp).
5:17
1 s
v0 (5.28)
1 t
tR
[ ]
t
1
Sinus
0.8
v 0.6 Constant
[ ]
v0 0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
t
[ ]
t
Figure 5.12 Two different velocity profiles; constant acceleration and sinusoid acceleration.
Both time and velocity are normalized.
t
v0 0 t tR
t R
v( Linear ) v0 t R t t t R
t*
v0 1 t t R t t
t R
v0 t
1 cos 0 t tR (5.30)
2 t R
v( Sinusoid) v0 t R t t t R
v t*
0 1 cos t t R t t
2 t R
t* t t t R
Servo valves are highly specialized components and as such, they are normally chosen
from the catalogue data of experienced manufacturers. In order to choose a servo valve
from a catalogue two values need to be computed:
1. The minimum rated flow, Qr ,min , of the valve.
2. The minimum bandwidth, v ,min , of the valve
To compute the minimum rated flow it is necessary to perform steps 1..5 in the
systematic design approach, see Section 5.2, but only for the degree of freedom
controlled by the servo valve. As an example, the servo system of figure 5.4 is
considered, see Figure 5.13.
5:18
y , y , y
pL p1 p2
p1 p2 F
m
QL QL
pmo pmo
ps
Figure 5.13 Servo system with 4/3-way servo valve supplying fluid power to an equal area
actuator that translates a payload.
In this case the prescribed motion of the piston, y , y and y , and the applied force on
the piston, F , must be determined over the entire operating cycle. Next, the pressure
level, i.e., the supply pressure, p s , should be chosen. The value of the supply pressure
depends on the chosen valve type. In general, as high a pressure level as possible should
be chosen with a view to reduce the size and cost of the pump and other flow
transmitting components. Because the servo valves use throttling across their metering
orifices for functionality it is necessary to take this into account when sizing the
actuator. The pressure drop across the equal area cylinder is introduced, see Figure :
p L pcyl p1 p2 (5.31)
Using a symmetrical valve we have the following correlation between the supply
pressure, the load pressure and the load flow:
2 2 ps pL 1
QL C d w x p mo C d w x Cd w x ps pL (5.32)
2
In Equation (5.32) the area gradient of the metering orifices of the valve is w and the
spool travel is x . The pressure drop across each of the metering orifices is introduced as
pmo 0.5 ( p s p L ) . The fluid power delivered to the cylinder piston is:
1
PF C p L QL p L C d w x ps pL (5.33)
In order to maximize the available power for a given valve opening the following can be
set up:
5:19
PF C
0
pL ps pL
0
p L p L
(5.34)
pL 2
ps pL 0 pL ps
2 ps pL 3
Hence, the load pressure should be set to 67% of the supply pressure in order to
maximize output power.
Knowing the load pressure the size of the actuator can be computed from:
F m ymax
p L A F m y A (5.35)
pL
F m ymax
A 1.3 (5.36)
pL
Q( t ) A y ( t )
F ( t ) m y( t )
pL ( t )
A (5.37)
ps
Q NL ( t ) Q( t )
ps pL ( t )
The maximum value of the no-load flow represents the basic flow requirement to the
valve: in a situation where the valve is fully open and there is no load on the piston the
valve flow should be larger than or equal to the maximum no-load flow:
In catalogues the ability of the valve to transmit flow is given as a rated flow. The rated
flow corresponds to a no-load flow for a supply pressure equal to a certain rated
pressure, pr . For multiple stage servovalves this rated pressure is, classically, equal to
70 bar which corresponds to 35 bar pressure drop across each metering orifices in the
no load situation. So the flow demand may finally be transformed to a value that is
comparable with the rated flow:
5:20
pr
Qr ,min 1.1 Q NL ,max
ps (5.39)
Qr Q r ,min
A safety factor of 1.1 is normally introduced at this point. This concludes the
computation of the minimum required value of the rated flow. It is recommended to
choose a valve that has a rated flow that is as close as possible to the minimum required
value. A large valve will be costly and compromise the accuracy of the total system.
Secondly, the minimum bandwidth of the valve must be determined. Here, experience
dictates that the valve must be faster than the hydraulic-mechanical system, i.e., the
valve should be able to operate at frequencies higher than the lowest eigenfrequency of
the hydraulic-mechanical system, n . A rule of thumb is that the operating frequency
that corresponds to a 90 phase lag for the valve should be three times larger than n :
v ,min 3 n (5.40)
Most servo valves has a frequency characteristic that roughly ressembles that of a
critically damped 2nd order system. In Figure 5.14 this has been illustrated with three
idealised Bode plots of the same valve. It is important to note that the valve has a much
higher bandwidth if only a certain percentage of the total spool travel is activated.
Bode Diagram
0
-2
Magnitude (dB)
-4
-6
-8
-10
90% 25% 5%
-12
0
-45
Phase (deg)
-90
-135
10
0
10
1 v @ 90% 10
2 v @ 5% 3
10
Frequency (rad/sec) v @ 25%
Figure 5.14 Idealised Bode plots of servo valve for three different levels of actuation.
In Figure 5.14 the effective bandwidth of the valve ( 90 phase lag) is referred to as v
with reference to the actuation level. The valve actuation that corresponds to
computation of the minimum rated flow should off course be used to determine whether
the valve is suitable:
5:21
v v ,min (5.41)
Equations (5.39) and (5.41) represents the criteria used in practice when choosing servo
valves.
Fluid power systems for systems that have several actuators with substantial variations
in their power demand may easily encounter very poor efficiencies. In many cases this
problem is handled by introducing a so-called load sensing (LS) system. An LS-system
is characterized by one or more components that are capable of
sensing the current load situation as pressure signals, and
adjusting the power demand to the pump(s) of the system accordingly.
These type of systems are especially popular within mobile hydraulics where large
variations in actuator loading and several actuators in activity simultaneously are
common practice. In the following different LS system configurations will be
introduced.
Actuator Actuator
Figure 5.15 Simplified hydraulic diagram with two actuators controlled by electro-hydraulic
proportional valves.
5:22
The electrical input to the proportional valves can be generated in several ways. If there
is an operator-in-the-loop then the input is normally generated by means of a manually
operated joy stick. If there is no operator, then some kind of closed loop control is
normally employed that requires position or velocity feedback from the actuator, see
Fig. 5.16.
Operator Operator
input Actuator Actuator input
Valve Valve
electronics electronics
Sensor Sensor
Control feedback feedback Control
reference Actuator Actuator reference
Valve Valve
electronics electronics
Figure 5.16 Typical actuation of electro-hydraulic proportional valve with and without an
operator in the loop, respectively.
In any case, some kind of valve electronics (either integrated into the valve or a stand
alone component) is required to manipulate the input signal into a spool position. This is
typically obtained via an integrated feedback control system with a positional sensor
attached to the spool and a power unit capable of driving a servo system that controls
the distribution of the valve actuation hydraulic pressure.
The system shown in Fig. 5.15 will always be difficult to control by an operator in the
loop. The main problem is caused by the multiple actuators that may require fluid power
at different levels and simultaneously. This means that the operator must adjust the
spool position continuously to ensure that heavily loaded actuators receive their
designated share of the total available pump flow. In most cases this is either very time
consuming or actually impossible. Because of this, the pressure compensated valve has
been developed. In Fig. 5.17 the design shown in Fig. 5.15 has been modified so as to
include pressure compensators for each actuator circuit.
The compensator that has been added upstream to each valve will try to maintain a
constant pressure drop across the main spool P-to-A or P-to-B orifices depending on the
current spool position. This means that the only parameter in the orifice equation that
may still vary is the discharge area of the spool. The discharge area always depends on
the spool position, i.e., the valve flow is proportional to the discharge area and thereby
to the spool position and thereby to the input signal. Hence, the name proportional now
refers to the valve flow and not only the spool position.
5:23
Actuator Actuator
Compensators
In the following section different types of systems using pressure compensators are
introduced.
Actuator Actuator
Qact Qrel Q
Qrel
pcr pmax p
5:24
Also, the crack pressure of the pressure relief valve should be set to a value slightly
higher, p , than the maximum operational actuator pressure. This design will be
referred to as Design A.
Next, the fixed displacement system is turned into an LS-system, see Fig. 5.19. The LS
pressure is obtained by means of check valves that compare relevant pressure levels
with the largest value reported back as the LS pressure of the entire circuit. This LS
pressure is compared to that of the other circuit(s) and the highest value qualifies as the
overall LS pressure of the valve.
Actuator Actuator
LS
Load V2
V2
orifice
p P O2
QL
O2 QC
O1 O1
QC
QP
pC pC
pP PC
PC
Qmax
QP pP
pLS
O0
D- D+
O0
D- D+
Figure 5.20 Simplified diagram notation for variable displacement pumps with pilot operated
pressure control and LS pressure control.
5:25
These pumps may be used in a non LS system and an LS system as shown in Fig. 5.21
and Fig. 5.22.
Actuator Actuator
The pressure relief valve now remains closed at all time and the pump only supplies the
required flow.
Actuator Actuator
pLS
5:26
5.5 System efficiencies
In the previous section four different hydraulic circuits were presented that all are
capable of handling systems with several actuators that are manually controlled by an
operator in the loop. A simplified load case is considered. The total duration of the load
case is 60 seconds and in that period both actuators are subjected to high pressure load
(typically positive load), low pressure load (typically negative load) and idle (not
actuated). If the actuators are chosen carefully, then it is reasonable to set the same
pressure demands for both circuits at high and low pressure. The power difference is
normally seen in the pump flow, hence, the volume flow demand of actuator 2 is set to
the double of the volume flow demand of actuator 1. Finally, the different periods of
high pressure, low pressure and idle for the two actuators are offset relative to each
other as would be the case in many applications.
In Table 5.1 the volume flow, the pressure and the corresponding power level used to
investigate the design put forward in the previous sections are displayed for both
actuators in the 60 second window of the load case.
Actuator #1
Volume flow
Interval [s] Pressure [bar] Power [kW]
[l/min]
0-10 50 180 15
10-20 50 180 15
20-30 50 30 2.5
30-40 50 30 2.5
40-50 0 0 0
50-60 0 0 0
Actuator #2
Volume flow
Interval [s] Pressure [bar] Power [kW]
[l/min]
0-10 0 0 0
10-20 100 30 5
20-30 100 30 5
30-40 100 180 30
40-50 100 180 30
50-60 0 0 0
For comparison the accumulated energy consumption at the pump is displayed for each
design together with the accumulated energy demand at the actuators. Please note, that
for Design A and B a flow safety margin of 10 l/min is used, i.e., the pump flow is
always 160 l/min. Also, for all design a pressure safety margin of 20 bar is used, i.e., the
pump pressure must be 20 bar higher than the maximum/current pressure demand.
These number yields the accumulated energy curves shown in Fig. 5.23.
5:27
E pmp [ kJ ]
3500
A
3000
2500
B
2000 C
D
1500
1000 Eact [ kJ ]
500
0
0 10 20 30 40 50 60
t [s]
Figure 5.23 Accumulated energy consumption at the pump for Design A..D as compared to the
accumulated energy demand at the actuators.
As can be seen, Design D has the best efficiency for this load case and, indeed, for
almost any application. However, it is also clear that as long as both actuators are
working, but with different pressure levels, then Design D will also experience energy
losses. For this load case this is relevant for the intervals 10-20 seconds and 30-40
seconds, respectively. The only way to remove these types of losses are by adding more
pumps (ideally, one pump for each actuator), however, for a practical system this will
increase costs substantially.
----- oo 0 oo -----
5:28
6
Analysis of
Hydraulic Systems
6.1 Introduction............................................................................... 1
6.1 Introduction
6:1
equired. In the following subsections methods for steady state and dynamic modeling
and simulation are presented.
This yields the following basic equations for some of the most common components in
hydraulic systems. The pressure node is simply a volume of fluid without significant
pressure variations.
n n
Pressure node Q1 Qi 0 Qi 0
i 1 i 1
Qi
n
nD
Pump Q Q n D Q
D 1000
6:2
pA
Q n D nD
M n Q
1000
Motor Q
D ( p A pB )
D M
pB 2 M 0.0159 D ( p A pB )
F
Q1 0.06 v A1
Q2 Q1 v A1
pB Q2 0.06 v A2
v
Cylinder Q2 v A2
A2
p A pB A2
F p A A1 pB A2 F A 1
pA 10
A1
Q1
Q
2
Orifice pA pB
Q Cd Ad ( p A pB ) Q 0.89 Cd Ad p A p B
Cd Ad
Q
Q0 p A pB Q0 p A pB
Check valve
pA pB Q0 p A pB Q0 p A pB
kd
Q
Vg
p V gn p0 V0n p V gn p0 V0n
Accumulator
Vf Va V f V g cst Va V f V g cst
6:3
that the slope of the p-Q characteristics should be included. In that case the governing
equations simply change into:
Q0 p A p B pcr
p A p B pcr (6.1)
Q p A p B pcr
pressure
where is the slope of the p-Q characteristic of the valve. If the slope for
flow
some reason is not constant but varies significantly this may also be introduced. It is
simply a question of how detailed information is required at the current stage of the
design evaluation. In Table 6.2 only a few valves are shown, however, similar equations
may be set up for any type of valve.
The most important aspect of the governing equations for the valves in Table 6.2 are the
fact that there are at least/typically two modes of operation. Hence, a governing
equation exists for both modes and for each mode there is an inequality that must be
fulfilled in order for the mode to be active. In the following the choice of mode of
operation is referred to as the configuration parameter of the valve. For a pure steady
state analysis the configuration parameter of each valve must be chosen beforehand, i.e.,
it is necessary to make a number of qualified guesses. Only af choosing/guessing the
configuration parameter of each valve can the governing equations be formulated and
solved. After solving the equations each choice of configuration parameter must be
validated by checking wether the corresponding inequality is fulfilled. If not, the
configuration parameter must be changed and the analysis redone. Potentially, this leads
to 2 n possible different system configurations, where n is the number of valves with
two modes of operation. In practice, however, the mode of operation of most valves are
easily recognized for a given situation.
A step-wise approach for steady-state analysis of any hydraulic system can now be set
up:
1. Identify pressure nodes.
2. Identify the components that demarcate each pressure node.
3. Choose/guess configuration parameters for all components with more than one
mode of operation.
4. Set up equations:
a) Flow continuity for each pressure node.
b) Flow continuity for each pump and actuator.
c) Static equilibrium for each actuator.
d) Flow through restrictions (orifices, filters etc.).
e) Equations associated with the choise of configuration parameters.
f) Static equilibrium for spool and poppet valves.
5. Solve equations numerically.
6. Are the computed variables physically meaningful?
Yes: Analysis completed.
No: Have all combinations of configuration parameters been examined?
Yes: Analysis cannot be carried out.
No: Go to 3) and choose/guess configuration parameters differently.
6:4
As an example the hydraulic system shown in Figure 6.1 has been subjected to this type
of analysis. From the figure it is seen that a total number of 9 equations has been
established. In this case, there is a single component with two operation modes, namely
the pressure relief valve, and the guess is that the valve is closed. For a given system
with a known pump speed and a known output torque the 9 equations may be solved to
yield the 9 unknowns: p1..3 Q1..5 nm . Hence, pressure, flow and actuator speed are
the classical output of steady state analysis, however, it is also possible to prescribe the
motor speed and introduce the motor displacement as a variable thereby changing the
equation solving from pure analysis to component sizing. After solving the equations it
is necessary to examine that all pressures are non-negative and that the choice of
configuration parameter is correct. In this case this is simply done by ensuring that p1 is
smaller than pcr .
Q4
Q1 Q2 Q3 0 Mm
Flow continuity, nodes nm
Q3 Q4 0 Dm
Q4 Q5 0
p2 p3
Flow continuity, Q1 n p D p 0
pumps and actuators
Q4 nm Dm 0
Q3 Q5
Actuator equilibrium
Dm p2 p3 C D AD
Mm 0
2
Q2
p1
2
Q3 CD AD p1 p2 0
Q1
Restriction flow
2
Q5 CD AD p3 0
Dp pcr
np
Operation mode / pT 0
Q2 0
Configuration parameter
Figure 6.1. Hydraulic circuit and corresponding steady state equations.
Clearly, the number of equations might seem excessive, and the system of equations
could easily be reduced substantially by substituting the different expressions into each
other.
The set of equations are non-linear because of the orifice equations and therefore have
to be solved numerically, typically by means of Newton-Raphson iteration. Often the
numerical solver will have difficulties solving the equations simply because they are
formulated in SI-units. If this is the case then one should simply reformulate the
problem using FLP-units. Also, the numerical solver may encounter problems with the
orifice equation because the sign of the pressure drop may become negative during
iteration. This may be avoided by using the following formulation:
6:5
2
Q C D AD SIGN( p A p B ) p A pB
1 x0 (6.2)
SIGN( x ) 0 x 0
1 x 0
Also, numerical problems will be encountered if for some reason the correct solution
includes zero flow through an orifice. In that case the orifice equation must be replaced
with a restriction flow equation that reflects the laminar regime, i.e., Q p .
( Q V )
p (6.3)
V
This equation gives the pressure gradient in a given volume of fluid, V. The net-flow
into the volume is Q (positive if flow enters the volume) and the time derivative of the
expansion (displacement flow) is V (positive if the volume is expanding). If this value
is positie at a certain time the pressure is going up and vice versa. The effective stiffness
of the fluid, which greatly depends on temperature, dissolved air, hosings and tubings, is
. In this case we have a 1st-order differential equation and only one initial condition is
required, namely the initial pressure in the volume: p.
As an example consider the volume shown in Figure 6.2 which is bounded by 2
cylinders, 2 motors, 2 pumps and 2 orifices.
6:6
x1 x2 Aa 2
A p1
V p1 n p1
Vm1 nm1
V p
V p2 n p2
Vm2 nm2
Cd 1 Ad 1 Cd 2 Ad 2
p1 p p2 p
The differential equations for the pressure gradient of this volume may be written using
Equation 6.3 with the following values for Q and V :
Q D p 1 n p 1 D p 2 n p 2 D m1 n m1 D m 2 n m 2
2 2
C d 1 Ad 1 p p1 C d 2 Ad 2 p2 p (6.4)
V A p1 x1 Aa 2 x 2
Just like a mechanical system typically contains more than one body a hydraulic system
will typically need to be modelled with several volumes. In a hydraulic system the
different volumes are typically separated by either: Pumps, motors, cylinders or orifices.
Displacement flows in volumes are typically caused by hydraulic cylinders or
accumulators. If we consider the steady state equations for an accumulator then the time
derivative yields a linear equation containing both the pressure and volume gradient:
6:7
In the above equation the polytropic exponent n depends on whether the
compression/decompression of the accumulator is predominantly adiabatic, see also
section 4.5. In general Equation (6.5) is a simplification because n may vary
significantly, hence, care should be taken in the modeling if the performance of the
hydraulic system is very sensitive to the value of n.
Equation (6.5) introduces another state variable, namely the fluid volume of the
accumulator. The initial fluid volume is readily derived from the total accumulator
volume (which is constant), the preload conditions and the current pressure:
1 1
p n p n (6.6)
V f Va 0 V0 V g 0 V0
p p
Hence, if a volume is connected to an accumulator it has two states: the pressure and the
fluid volume of the accumulator and their gradients must be solved simultaneously
using Equations (6.3) and (6.5).
In general, a purely hydraulic system may be solved numerically in the following steps:
1. Identify all volumes in the system and set up the pressure build up equation for each
and identify all accumulators and set up the (Circuit diagrams useful here).
2. Identify all orifices and their dependancy on the motion of mechanical parts in
valves.
3. Determine initial pressure for each volume and initial fluid volume for each
accumulator.
4. Calculate pressure gradients for each volume and volume gradient for each
accumulator.
5. Calculate acceleration of all movable mechanical parts in valves.
6. Update pressure in each volume and volume of each accumulator.
7. Update position and velocity of all movable mechanical parts in valves.
8. If the analysis is not yet concluded then go back to 4.
Typically som mechanical bodies, cylinders or valves will be part of the system. This
means that their position and velocity must be updated simultaneously in order to
update volumes, orifice flows and net-flows into volumes for the next step.
As may readily be observed from Equation (6.7) the position and velocity of the
mechanical system is needed in order to do a dynamic simulation of the hydraulic
system. In fact, any type of dynamic simulation of a hydraulic system requires a
simultaneous dynamic simulation of the actuated mechanical system and, depending on
the level of detail, also of the movable mechanical parts within the hydraulic valves.
Hence, dynamic simulation of hydraulics is closely connected to dynamic simulation of
mechanics.
The mechanical system may, in general, be divided into a number of bodies. In the
planar case, the governing dynamic equations for a body are:
m r F
(6.7)
J M
6:8
In Equation (6.7) m is the mass and J is the mass moment of inertia with respect to the
mass center of the body. Furthermore, r is the coordinates of the mass center of the
body and is the rotation of the body, both measured relative to some reference
coordinate system. F is the sum of all the forces acting on the body, normally
divided into applied forces (gravity, springs, dampers, actuators, friction, wind
resistance, rolling resistance etc.) and reactive forces (connections with other bodies and
ground). M is the sum of the force moment of F with respect to the mass center
and all the moments (force couples) acting on the body.
The dynamic equilibrium equations contain 3 scalar equations corresponding to the 3
degrees of freedom of a body. All of them are differential equations, and in general
needs to be solved numerically. Being 2nd order differential equations 2 initial
conditions is required for each coordinate, namely the initial position and initial
velocity:
r r
Knowing the positions and velocities of the bodies will typically also be necessary in
order to determine the resulting forces and moments, i.e., the right hand side of
Equation (6.7).
Alternatively, the steady state equations of the actuators, see Table 6.2, may be
generalized to take into account the dynamics of the mechanical system. For the motor a
simplified dynamic equation can be set up:
D p M
J eff M M tM M (6.8)
2
In Equation (6.8) the sign conventions are as follows: rotation of the shaft, (and its
time derivatives; and ), is defined as positive in the same direction as the
hydraulically generated torque on the shaft produced by a positive pressure drop across
the motor. The applied moment on the output shaft, M, is positive in the opposite
direction. The introduction of the hydro-mechanical efficiency, ´ hmM , should be done
with care. Normally, the hydromechanical efficiency is measured in a situation where
the motor is motoring, i.e., the direction of the applied moment is opposite to that of the
angular speed. In that case:
D p M
J eff hmM M (6.9)
2
However, if the motor is working as a pump, i.e., the direction of the applied moment is
in the same direction as that of the angular speed (negative load M<0) we get:
1 D p M
J eff M (6.10)
hmM 2
6:9
may be determined from energy considerations. The applied moment may be computed
as follows:
dWext
M (6.11)
d
In Equation (6.11) Wext is the work done by the external forces/moments (gravity,
friction, rolling resistance etc.) on the mechanical system actuated by the motor.
The effective mass moment of inertia may be computed according to:
2 E kin
J eff (6.12)
2
Motor i Jd r
i
M fric
Drum
Gearbox
m
Payload
Figure 6.3. A hydraulic motor is driving a winchdrum via a gearbox. The drum is connected to
a payload and is also subjected to a certain rotational friction.
For an infinitesimal rotation of the motor shaft the work done on the system by the
external loading can be computed and, subsequently, the moment applied to the output
shaft of the motor:
d d
Wext m g dy M fric d d m g r M fric
i i
(6.13)
m g r M fric
M
i
6:10
2 2
1 1 1 1
E kin J d d2 m v 2 J d m ´ r
2 2 2 i 2 i
(6.14)
J mr2
J eff d
i2 i2
Notice, that the mass moment of inertia of the motor rotor and the gearbox have been
neglected. Because of the high torque pr. volume ratio of hydraulic motors neglecting
the inertia of the motor rotor and the gearbox is normally a reasonable assumption.
Next, consider the system shown in Figure 6.4. It is a four-wheel drive vehicle subjected
to a total rolling resistance of FR , and propelled by a hydraulic motor via a geared chain
z2
drive i . Due to symmetry the vehicle is modeled as a planar system.
z1
Vehicle
mv z1
m wh J wh
Motor
rwh FR
z2 2
FR
2
Figure 6.4. A four-wheel drive vehicle propelled up an incline by a hydraulic motor is shown.
Power is transmitted from the motor to the wheels via chain drives.
As in the previous example a infinitesimal rotation of the motor is used to determine the
moment applied to the output shaft of the motor:
d d
Wext ( mv 2 m wh ) g dy FR ds ( mv 2 m wh ) g rwh sin FR rwh
i i
(6.15)
r FR ( mv 2 m wh ) g sin
M wh
i
6:11
2 2
1 2 1 1
E kin 2 J wh wh ( 2 m wh mv ) v 2 J wh ( 2 m wh mv ) ´ rwh
2 2 i 2 i
(6.16)
2
2 J wh ( 2 m wh mv ) rwh
J eff
i2 i2
Having determined the effective mass moment of inertia it is possible to compute the
eigenfrequency of the hydraulic-mechanical system composed of the lines to and from
the motor, the motor and the mechanical system represented by the effective mass
moment of inertia:
k
n
J eff
(6.17)
D2
k
2 D V L
In Equation (6.17) V L is simply the total volume of the fluid lines leading up to the
motor and away from it.
Similar to that of the motor a simplified dynamic equation can be set up for the
cylinder:
In Equation (6.18) the sign conventions are as follows: piston travel, x (and its time
derivatives; x and x ), is defined as positive when the cylinder is extracting. The
applied force on the piston, F, is positive in the opposite direction. As in the case of the
motor the introduction of the hydro-mechanical efficiency, ´ hmC , should be done with
care. Equation (2.47) is used in its general form to get an expression for the friction
force as a function of the hydromechanical efficiency:
FC FtC FmC F
hmC 1.0 mC FmC ( 1 hmC ) FtC (6.19)
FtC FtC FtC
6:12
FtC FtC FmC ( 1 hmC ) FtC ( 1 hmC ) A p1 p 2
(6.21)
meff x FtC FmC F 2 hmC A p1 p 2 F
Note that in Equation (6.21) both FtC A p1 p2 and F have negative values.
If the cylinder is retracting the theoretical cylinder force is defined as
FtC A p2 p1 , see Equation (2.44). If the cylinder is pulling a load we have:
Note that in Equation (6.22) FtC A p2 p1 has a positive value whereas F has a
negative value.
If the cylinder is retracting and is pushed by the load we have:
Note that in Equation (6.23) FtC A p2 p1 has a negative value and F has a
positive value.
The effective mass, meff , and the applied force, F, must be related to the piston of the
cylinder. In general, they are both functions of x and x and may, similar to the values
associated with the motor, be determined from energy considerations. The applied force
may be computed as follows:
dWext
F (6.24)
dx
2 E kin
meff (6.25)
x 2
6:13
m
L AC
m
L
rm
A
x
L AC
0
C
A
nder
Cyli
B
Figure 6.5. A hydraulic cylinder is rotating an arm with a payload at the end.
For a infinitesimal extraction of the cylinder the work done on the system by the
external loading can be computed and, subsequently, the force applied to the cylinder
piston:
In Equation (6.26) it is utilized that the arm and the piston must have the same absolute
velocity in point C.
The effective mass may be computed as:
2
1 1 2
1
E kin m v 2 m Lm m L2m
x
2 2 2 L AC cos
(6.27)
2
Lm
meff m
L AC cos
6:14
Having determined the effective mass it is possible to compute the eigenfrequency of
the hydraulic-mechanical system composed of the lines to and from the cylinder, the
cylinder and the mechanical system represented by the effective mass:
k
n
meff
(6.28)
A2 A2
k
V1 V2
In Equation (6.28) V1 and V2 are simply the total volume of the fluid leading up to the
cylinder piston side and to the cylinder rod side, respectively. This include the volume
in the lines as well as the volume in the cylinder. Hence, in general the volumes are
functions of the piston position.
The actual solving of the coupled set of differential and algebraic equations can be
performed in several ways. Today, a several software packages exist that allow for
relatively fast and easy modeling and simulation of physical systems. The packages may
vary with respect to modeling concepts and solver algorithms, however, the
fundamentals remain the same and in the following the basic architecture of the
numerical simulation of hydraulic-mechanical systems is presented.
As described in the previous chapter the state variables must be initialized. Normally,
that include:
X p Vf x x (6.29)
Here, X is an algebraic vector of state variables consisting of all volume pressures, all
accumulator fluid volumes, all cylinder positions and velocities and all motor angular
positions and angular velocities.
Based on the state variables Equations (6.3), (6.5), (6.8) and (6.18) may be set up to
yield a linear set of equations:
M ( X ) X Y ( X ) X M 1Y (6.30)
Setting up the coefficient matrix M ( X ) and the right hand side Y ( X ) is done by
solving the algebraic equations, i.e., the steady state equations of components where the
dynamic properties may be ignored.
Often, the equations are not coupled very closely meaning that inverting the coefficient
matrix M ( X ) can be avoided or divided into the inversion of smaller sub matrices.
Having computed the time derivative of the state variables the next step is the time
integration. The simplest way of doing this is by means of a so-called forward-Euler:
6:15
t ( new ) t dt
(6.31)
X ( new ) X X dt
As an example let us consider the hydraulic system shown in Figure 6.6 (see also
ate analysis and Figure 6.1) and set up the combined set of differential and algebraic
equations.
p3 V3
p 2 Q3 Q4 p 3 Q4 Q5 p2 V2 Q4
V2 V3
Cd Ad
Flow continuity pumps and actuators
Q3 Q5
Q1 nP DP Q4 nM DM p0 V0 n
2 2 Q1
Q3 Cd Ad p1 p2 Q5 Cd Ad p3
n P DP
One important result from doing dynamic time domain simulation is that the operation
mode is no longer a guess, now it can be evaluated directly from the state variables. In
this case since we have access to p1 we simply compare it with the crack pressure of
the pressure relief valve.
In order to run a simulation obviously some input must be prescribed in time. Typically,
this can be the load on the motor output shaft, the motion of the directional control
valve and the angular speed of the pump.
----- oo 0 oo -----
6:16
7
Control of
Hydraulic Servo Systems
7.1 Introduction............................................................................... 1
7.1 Introduction
Hydraulic servo systems are always controlled by means of some kind of closed loop
control utilizing some feedback signal to control a servo valve or a variable
displacement pump. In the following focus will be on valve controlled circuits where
one or more high response valve, typically referred to as a servo valves, are the control
elements. There exist a wide variety of control methods that are closely related to the
type of task that the hydraulic servo system should undertake and also to the type of
actuator and its connection to the valve and payload. Again, it has been necessary to
narrow the focus to position and velocity control of symmetrical actuators.
Let us consider the hydraulic servo system shown in Figure 7.1. The set of parameters
includes the normalized position of the spool of the servo valve, 1 u 1 . For this
system it is possible to set up the ideal governing equations:
QL A y (7.1)
m y pL A F (7.2)
1 u
Qv Cd Ad 0 u ( ps pL ) (7.3)
u
7:1
In (7.3) it is assumed that the servo valve is symmetric and that the maximum discharge
area is Ad 0 . Also note that both the load pressure, pL , the load flow, QL , and the spool
travel, u , may take on negative as well as positive values.
y y y
A
m F
p1 pL p1 p2 p2
VL1 VL 2
Qv Qv
u
u1 u0 u 1
ps
Figure 7.1 Hydraulic servo system using a symmetrical cylinder to translate the payload.
Taking into account the flexibility of the fluid we get the following differential equation
for the chamber pressures:
p1 ( Qv A y ) (7.4)
VL1 y A
p 2 ( A y Qv ) (7.5)
VL2 ( h y ) A
In (7.5) the stroke of the cylinder, h , is introduced. Assuming that the cylinder is in
midstroke, y 0.5 h , and that the volume of the hoses are identical, VL1 VL2 VL , we
can derive the following expressions for the load pressure gradient:
p L p 1 p 2 ( Qv A y ) ( A y Qv )
VL1 y A VL 2 ( h y ) A
4
p L ( 2 Qv 2 A y ) ( Qv A y )
h V0
VL A (7.6)
2
7:2
In (7.6) the total fluid volume, V0 A h 2 VL , between the valve and the actuator has
been introduced.
In order to further examine the dynamic characteristics of the hydro-mechanical system
in the frequency domain it is necessary to linearize the governing equations and do a
laplace transformation. The linearized governing equations are:
The linearization is carried out for a steady state situation, denoted by superscript ss ,
hence, all the variables must be considered as deviations from this situation.
( ss ) ( ss )
QL A y (7.17)
( ss ) ( ss )
pL A F (7.18)
1 u( ss )
Qv( ss ) Cd Ad 0 u ( ss )
( ps pL( ss ) ) (7.19)
u
( ss )
( ss ) ( ss ) ( ss )
Qv A y QL (7.20)
Special attention must be given to the linearization of the only non-linear equation (7.3)
leading to the linear equation (7.9). The theoretical definition of the two coefficients of
(7.9) are readily derived from the Taylor expansion:
Qv 1 u( ss )
K qu Cd Ad 0 ( ps pL( ss ) ) (7.21)
u ss u
( ss )
The coefficient, K qu , will be referred to as the flow gain. It is the sensitivity of the load
flow relative to the spool travel and with the load pressure held constant. It appears
directly in the total gain of the control of the hydraulic servo systems and is of major
importance in control of hydraulic servo systems in general.
7:3
( ss )
Q Cd Ad 0 u
K qp v
pL ss u
( ss )
( ss ) (7.22)
2 ( ps pL )
( ss )
u
u( ss ) ( ss )
2 ps pL
K qu u( ss ) (7.23)
K pu
K qp u( ss )
This coefficient is the pressure gain, and it reflects the sensitivity of the load pressure
relative to the spool travel with the load flow held constant (locked piston). Although it
does not enter directly into the control equations it is of major importance when
evaluating the ability of the control loop to handle disturbances.
Clearly, the steady state situation strongly influences the value of the three coefficients.
Especially, the so-called 0-position is often chosen as steady state reference because it
yields a conservative controller design. The 0-position is defined as:
pL( 0 ) 0
u( 0 ) 0 (7.24)
Qv( 0 ) QL( 0 ) 0
(0 ) 1
K qu ,t C d Ad 0 ps
(0 ) (7.25)
K qp ,t 0
(0 )
K pu ,t
Note that an extra index, t, has been added to indicate that these are theoretical values.
In fact, the theoretical flow coefficient is in good accordance with values observed in
practice for any steady state situation. On the contrary, the actual values of the flow-
pressure coefficient and the pressure coefficient differ strongly from the theoretical
values when the spool is in neutral or close by, u( ss ) 0 . In the 0-position this is most
evident. An acceptable way of handling the discrepancy, is simply to introduce a
( ss )
minimum steady state spool travel, u umin , that replaces the actual steady state spool
position, u( ss ) , if its numerical value becomes to small:
7:4
u( ss ) ( ss )
u u u
( ss ) ( ss )
ulim u (7.26)
( ss ) ( ss )
u u u
Typical values that reflects the leakage would be u 0.002 0.05 depending on the
wear conditions of the leakage paths between spool and housing. Using (7.26) it is
possible to set up more realistic expressions for the coefficients in the 0-situation:
(0 ) (0 ) 1
K qu K qu ,t Cd Ad 0 ps
( ss )
(0 ) Cd Ad 0 ulim
K qp (7.27)
2 ps
(0 )
(0 ) K qu 2 ps
K pu (0 )
( ss )
K qp ulim
Next, the governing equations (7.7 ... 7.10) are subjected to Laplace transformation:
QL ( s ) A s y( s ) (7.28)
2
m s y( s ) pL ( s ) A F ( s ) (7.29)
Qv ( s ) K qu u( s ) K qp pL ( s ) (7.30)
1
s pL ( s ) ( Qv ( s ) A s y( s )) (7.31)
C
In (7.31) the total capacitance of the hydraulic fluid, C , has been introduced. It is
defined as:
V0
C (7.32)
4
In Figure 7.2 the equations (7.28 ... 7.31) are shown as block diagrams with the spool
position as input and the piston speed as output. The disturbance, F , is neglected in the
simplifications.
7:5
K qp
F
u s y
1 pL 1
K qu A
C s m s
K qu K mh
u s y u s y
A s 2
2 mh s
mC m K qp 1
2
s2 2
s 1 2
mh mh
A A
The transfer function between piston speed and spool position describes a 2nd order
system with the gain K mh , the eigenfrequency mh , and the damping mh . The index
refers to mechanical-hydraulic system. The transfer function and the associated
parameters are given as:
s y K mh
Gmh ( s ) 2
u s 2 mh s (7.33)
1
2
mh mh
K qu
K mh (7.34)
A
A
mh (7.35)
mC
K m
mh qp (7.36)
2 A C
In the time domain the piston speed can be written as a function of time when subjected
to a the step input, u in , to the dimensionless spool travel. The function is:
1 2
y ( t ) K mh
t tg1
t 2 mh
K mh e mh mh sin mh 1 mh (7.37)
uin 2
1 mh mh
The size of the step input must lie in the interval 1 uin 1 . The time it takes for the
piston speed to reach the steady state value the first time is called the crossing time, t cr ,
and the subsequent relative overshoot is referred to as . They can be computed as:
7:6
1 2
1 mh
tg
mh (7.38)
t cr
2
mh 1 mh
mh
1 2 (7.39)
y max uin K mh mh
e
uin K mh
k ln
mh k (7.40)
1 k
2
1 2
tg1 mh
mh (7.41)
mh
2
t cr 1 mh
The gain, K mh , can be taken from the steady state value of the piston speed as:
y ( t )
K mh (7.42)
uin
K qu A K mh (7.43)
A2
C 2 (7.44)
m mh
C
K qp 2 A mh (7.45)
m
The governing equations have very much the same structure for a hydraulic servo
system with a motor rather than a symmetrical cylinder as actuator. If the motor has a
displacement of D and drives an inertia of J then we get the following transfer function
from the valve spool position, u , to the angular speed of the motor, s
s K mh
Gmh ( s )
u s2 2mh s (7.46)
2
mh
1
mh
7:7
K qu
K mh (7.47)
D
D
mh (7.48)
J C
K qp J
mh (7.49)
2 D C
D
D (7.50)
2
The total volume used to compute the capacitance, C , is also different from the cylinder
actuator, V0 D 2 VL
In position control the task of the hydraulic servo system is to follow either a reference
position of the cylinder piston or a reference rotation of the motor output shaft. In that
case, some position feedback from the actuator is required. A typical setup is shown in
Figure 7.3.
yref e uref u v y
1
Gc ( s ) Gv ( s ) Gmh ( s )
s
The Gv block represents the closed loop spool position control inside the valve and the
dynamics of the electrical-mechanical system of the valve. For simplified analysis this
block is normally modeled as a second order system:
1
Gv ( s ) 2
s 2 v s (7.51)
1
v2 v
Similarly, the Gmh block represents the second order system derived in the previous
section, see (7.33) and (7.46):
K mh
Gmh ( s ) 2
s 2 mh s (7.52)
1
2
mh mh
7:8
0
The performance of any hydraulic servo system depends -10 on the bandwidth of the
product of these two transfer functions. As explained-20in Chapter 5, the typical approach
is to select a valve with a deadband three times larger, or, v 3 mh . If this design rule
-30
is observed, the overall dynamics of the valve and -40 hydraulic-mechanical system will
approximately be that of the hydraulic-mechanical system. If the valve has a smaller
(dB)
deadband then the total deadband of the system is markedly
-50
reduced. In Figure 7.4 the
bode plot of Gmh is shown together with the product -60
Gv Gmh for v mh and
v 3 mh . -70
-80
-20
-30 -90
v
-40
v
(deg)
(dB)
-50 -180
-60
v 3 mh v 3 mh
-70
-270
-80 v mh
v mh
-90
-360
0 1 2
-100
01
1
10 10
10 100
10
10 100 (rad/sec)
[ rad / s ]
[ rad / s ]
Figure 7.4 Bode plot of transfer functions of different combinations of servo valve dynamics
-90
and hydraulic-mechanical system dynamics. The parameters of the transfer function are:
K mh 0.3 , mh 10 rad
s
, mh 0.2 and v 0.8 .
(deg)
-180
Clearly, choosing v mh yields a system with distinctly poorer dynamic performance
whereas choosing v 3 mh justifies the simplification:
-270
-360 K mh
10
0 G1
10 vmh 2 10
2 Gv Gmh
s 2 mh s
1
(rad/sec)
2
vmh vmh
(7.53)
Equation (7.53) is valid for frequencies up to and around mh . Frequencies beyond this
value is rarely of interest in hydraulic servo systems. So we simply disregard the valve
dynamics and continue working with an effective eigenfrequency of the valve-
mechanical-hydraulic system, vmh , that is 90% of mh . It must be emphasized again,
that this approach is only valid for v 3 mh .
In Figure 7.3 the control block contains the control law. In this section we will
investigate pure proportional control as well as proportional control compensated with a
lead-lag network. If the control law is simply proportional control, Gc K p , we get the
following open loop transfer function for the position servo:
7:9
1 K p K mh
Gpo K p Gvmh
s s2 2 mh s (7.54)
s 2 1
vmh
vmh
20
We want to choose the gain of the controller, K p , as high as possible in order to obtain
10
0
a good performance with respect to precision and reaction time. The upper limit is
introduced via the stability criterion of Nyquist that simply states:
-10
In order for the closed loop transfer function to be-20stable then the gain of the open loop
transfer function must be less than unity when the -30
phase lag is 180 .
(dB)
In Figure 7.5 the bode plot of a typical open loop -40 transfer function is shown. The
damping is set to mh 0.2 which is a typical value-50 for the, in general, poorly damped
hydraulic-mechanical systems. -60
Apo 20 log10 Gpo [ dB ]
-70
Gpo [deg]
20 -80
-90
10 Gpo 1
0
Gain margin -135
-10
-20
(deg)
(dB)
-30 -180
-40
-50
-225
-60
-70
-80 -270
-901 cr vmh 10 100 101
0
vmh
10 10
1 100
10
2
[ rad / s ]
(rad/sec)
[ rad / s ]
Figure 7.5 Bode plot of position servo open loop transfer function, Gpo . The parameters of the
-135
-180
In order for the system to be stable then the gain must be less than unity for vmh .
-225
Based on this, the maximum value of the controller gain can be derived from (7.54) as:
( s1 j vmh ) 1 102
-270
10
0 Gpo10
(rad/sec)
K p K mh
1
( j vmh )2 2 mh j vmh
j mh 1 (7.55)
2 vmh
vmh
2 mh vmh
Kp
K mh
The result in (7.55) is quite general within servo systems. In practice, a certain safety
factor must be introduced, mostly because mh is difficult to assess exactly without
experimental work. This safety factor is normally expressed in dB (deciBell) based on
the typical presentation of transfer function gain
7:10
Apo ( ) 20 log10 Gpo ( s j ) (7.56)
Let the safety factor be Apo Avmh . In that case the maximum value of the controller
gain can be derived from:
20 log10 Gpo ( s j vmh ) Avmh
Avmh (7.57)
2 mh vmh
Kp 10 20
K mh
2 mh vmh
K p 0.7079 (7.58)
K mh
In Figure 7.5 the cross frequency, cr , is indicated as the frequency at which we have
unity gain. An alternative interpretation of Nyquist means that at this frequency the
phase lag must not yet have reached 180 . The equation for the cross frequency is:
cr K p K mh (7.60)
1
Gpc 3 2
s 2 mh s s (7.61)
1
2
K p K mh vmh K p K mh vmh K p K mh
In Figure 7.6 the bode plot of a closed loop transfer function corresponding to the open
loop transfer function of Figure 7.5 is shown.
The bandwidth of the entire system is normally taken to be the frequency, bw , at
which we have a closed loop transfer gain of 3 dB :
Apc ( ) 20 log10 Gpc ( s j )
(7.62)
Apc ( bw ) 20 log10 Gpc ( s j bw ) 3
For low damping, mh 0.4 it is usually a good approximation to state that:
7:11
20
10
-10
-20
cr K p K mh bw (7.63)
(dB)
-30
-40
This often simplifies analysis of position servos
-50 substantially since the main dynamic
performance parameter, bw , can be estimated as
-60
the product of two known gains.
-10
-90
-20
(dB)
(deg)
-30
-40
-180
-50
-60
-70
-80 -270
10 bw vmh 10 100 101
0
vmh
10 10
1 100
10
2
[ rad / s ]
(rad/sec)
[ rad / s ]
Figure 7.6 Bode plot of position servo closed loop transfer function.
-90
To further investigate the simple proportional controller we examine the steady state
(deg)
y0
For a step input with the size y0 we insert yref in (7.64):
s
s3 2 mh s 2 s
K K
K p K mh 2 p K mh vmh p K mh y0
( ss ) vmh
e 2
s 0 (7.65)
s
3 2 s s s
K K mh K 1
K p K mh 2 p mh vmh p K mh
vmh s 0
7:12
Hence, the is no steady state error. However, if we take a ramp input with the speed v 0
v0
we insert yref in (7.64):
s2
s
3 2 mh s 2 s
K K
K p K mh
2 p K mh vmh p K mh v0 v0
( ss ) vmh
e s (7.66)
s3 2 mh s 2 s s 2 K p K mh
K K 1
K p K mh 2 p K mh vmh p K mh
vmh s 0
Hence, whenever we have a constant velocity input of the reference speed, we will have
a steady state position error.
This position error can be reduced in different ways. A typical approach is to add an I-
term in the controller, thereby changing it into a PI-controller:
1
Gc K p 1 (7.67)
Ti s
s
1 0
LL
Gc K p (7.68)
s
1 LL -20
LL
(dB)
-40
This will add open-loop gain for frequencies below cr which, in turn, will improve the
accuracy (but not the bandwidth) of the position servo, see also Figure 7.7.
-60
A 20 log10 G [ dB ] G [deg]
20 -80
-90
LL-compensated Not compensated
0
-135
Not compensated
-20 LL-compensated
(deg)
(dB)
-180
-40
-225
-60
-270
-80 0 1 2
-901 cr vmh 10 100 1
10 vmh10 10 100
10
(rad/sec)
[ rad / s ] [ rad / s ]
Figure
-135 7.7 Bode plot of position servo open loop transfer function with and without lead-lag
compensation. In this case, the lead-lag parameters are LL 2.25 rad s
(app. 80% of
cr 3.0 rad ) and LL 5 .
(deg)
-180
s
-225
-270
10
0
10
1
10
7:13
2
(rad/sec)
The lead frequency, LL , and the lead-lag frequency ratio, LL , should preferrably be
chosen based on the particular system, however, good results will normally be achieved
by ensuring that:
----- oo 0 oo -----
7:14