The Control System: Error - T
The Control System: Error - T
CHAPTER 8
THE CONTROL SYSTEM
In previous chapters, the dynamic behavior of several basic systems was examined. With this background, we can extend the
discussion to a complete control system and introduce the fundamental concept of feedback. To work with a familiar system, the
treatment will be based on a stirred-tank heater.
Each of these components can be readily identified as a separate physical item in the process. In general,
these four components will constitute most of the control systems that we consider in this text; however, the
reader should realize that more complex control systems exist in which more components are used. For
example, some processes require a cascade control system in which two controllers and two measuring
elements are used. A cascade system is discussed in Chap. 17.
For computational purposes, it is convenient to represent the control system of Fig. 8–1 by means of the block diagram shown in
Fig. 8–2 . Such a diagram makes it much easier to visualize the relationships among the various signals. New terms, which appear
in Fig. 8–2 , are set point and load. The set point is a synonym for the desired value of the controlled variable. The load refers to
a change in any variable that may cause the controlled variable of the process to change. In this example, the inlet temperature Ti
is a load variable. Other possible loads for this system are changes in flow rate and heat loss from the tank. (These loads are not
shown on the diagram.)
If the signal to the comparator were obtained by adding TR and Tm, we would have a positive feedback system, which is
inherently unstable. To see that this is true, again assume that the system is at steady state and that T= Tm = TR . If Ti were to
increase, T and Tm would increase, which would cause the signal from the comparator ( e in Fig. 8–2 ) to increase, with the
result that the heat to the system would increase. However, this action, which is just the opposite of that needed, would cause T
to increase further. It should be clear that this situation would cause T to “run away” and control would not be achieved. For this
reason, positive feedback would never be used intentionally in the system of Fig. 8–2 .
Note: In this analysis, it is assumed that the flow rate of heat q is instantaneously available and independent of the temperature in the tank. In some
stirred-tank heaters, such as a jacketed kettle, q depends on both the temperature of the fluid in the jacket and the temperature of the fluid in the
kettle. In this introductory chapter, systems (electrically heated tank or direct steam-heated tank) are selected for which this complication can be
ignored.
Notice that the reference temperature To cancels in the subtraction. If we introduce the deviation variables
or
where
In Fig. 8–3 b, the input variables (1/ wC) Q (s) and 𝑇𝑖′ (𝑠)are
summed before being operated on by the transfer
function 1 (𝜏𝑠 + 1). Note that in all cases the units in the block
diagram must be consistent. The quantities being combined at a
summing junction must have the same units. The readers should
convince themselves that this is indeed the case
Note that the rearrangement in Eq. (8.13) is merely for mathematical
convenience, and not what actually happens. Physically the two inputs (heat
and inlet temperature) have independent effects on the outlet temperature.
To = reference temperature used to evaluate enthalpy of all streams entering and leaving tank
Hg =specific enthalpy of steam supplied, a constant
Hls = specific enthalpy of condensed steam flowing out at Ts, as part of total stream
The term Hls may be expressed in terms of heat capacity and temperature (assuming no phase change occurs between Ts and To).
In a practical situation for steam, Hg will be about 1000 Btu/lbm. If the temperature of the bath T never deviates from Ts by more
than 10o the error in using the term f (Hg-Hls) instead of f ( Hg-Hl) will be no more than 1 percent. Under these conditions, Eq.
(8.16) represents the system closely, and by comparing Eq. (8.16) with Eq. (8.1), it is clear that
Therefore, q is proportional to the flow of steam f, which may be varied by means of a control valve. It should be emphasized
that the analysis presented here is only approximate. Both f and the deviation in T must be small. The smaller they become, the
more closely Eq. (8.16) represents the actual physical system.
An exact analysis of the problem leads to a differential equation with time-varying coefficients, and the transfer
function approach does not apply. The problem becomes considerably more difficult. A better approximation will
be discussed in Chap. 20, where linearization techniques are used.
Substituting in numerical values for the variables, we obtain the actual transfer function for this stirred-tank heater.
1𝑘𝑔 1000𝑘𝑔
= 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 𝑜𝑓 𝑤𝑎𝑡𝑒𝑟 =
𝐿 𝑚3
The steady-state heat input qs may be found from the steady-state energy balance, Eq. (8.2).
𝒗𝒐 𝑳 𝟏𝒌𝒈
𝒘= = 𝟐𝟎𝟎 ∗ = 𝟐𝟎𝟎𝒌𝒈/𝒎𝒊𝒏
𝝆 𝒎𝒊𝒏 𝑳
Thus,
A plot of the outlet temperature (in deviation variables) is shown in Fig. 8–8 a. The actual outlet temperature is shown in Fig.
8–8 b. Note that for the uncontrolled mixing tank, a step change of 10o C in the inlet temperature ultimately produces a 10o C
change in the outlet temperature. This result is just what we would expect after considering the physics of the situation.
Qualitatively, this response is the same as for part ( a); however, the ultimate temperature change due to the
increased heat input is 3oC.