Long Term Deflection in Concrete Beams

UAA Civil Engineering

CE 433 - Reinforced Concrete Design
by Dr. Bart Quimby, P.E.
Spring 2002

Long Term vs. Immediate Deflections

Immediate deflections are those deflections that are the immediate response to applied loads.
There is virtually no delay between the application of the load and the deflection that results
from the load application.

Long term deflections are the result of either dry shrinkage or viscous flow (creep) under
sustained loads. ACI 318 mixes the two effects together. Both are time dependent, with the
creep component being the more dominate of the two terms. There are much more complex
methods for determining the effects of resulting from creep and shrinkage than presented in ACI
318. See the text book discussion on these topics to get a better feel for what is happening.

Shrinkage is significant issue, particularly in members that are unsymmetrical reinforced such as
most RC beams. Without the reinforcing, the member would simply get shorter. Adding
reinforcing steel that bonds to the concrete restrains shrinkage. With more restraint on one face
than the other shrinkage occurs differently on each face and results in curvature that adds to that
caused by bending. Adding more steel on the compression side tends to reduce the problem.

Creep is viscous flow of the material under stress. The tension side gets longer and the
compression side gets shorter. Similar to shrinkage, the presence of reinforcing steel (which
does not measurably creep under load) restrains the deformation due to creep. The shortening of
the compression side and lengthening of the tension side results in curvature that greatly
magnifies deflections over time. Creep effects are most noticeable in the short term. The rate of
creep decreases as with time. Figure R9.5.2.5 shows the typical creep vs time curve used in
computing long term deflections by ACI 318.

Sustained Loads

In determining creep deformations, the term "Sustained Loads" comes into play. Sustained loads
are what drives the creep deformations. Two quantities are important here. The first is the
MAGNITUDE of the sustained load and second is the DURATION of the sustained loads. The
determination of these two quantities is left up to the engineer.

The discussion in the textbook considers only the dead load as a sustained load and that it's
duration is more than five years. This is a simplistic approach. It is not uncommon for a portion
of the live load to be sustained as well. For example, in an office occupancy, the desks, book
shelves, file cabinets, copiers, etc. are all part of the live load, but they are sustained over a long
period of time. Part of the live load is accounted for people coming in and out of the space and
for temporary office equipment loadings. In this case, 50% of the design live load may be

Long Term Deflections T.B. Quimby 1

sustained for over 5 years, another 25% may be around for a cumulative total of say 2 years over
the life of the structure and the remaining 25% being truly transient or rarely occurring. The
creep effects of these portions on the load should be included in the long term deflection
calculations as well.

Other occupancies may have different estimates of sustained loading. There are no guides for
determining the magnitudes or durations for any given design situation. These are left to the
judgment of the engineer.

ACI 318 Multiplier

ACI 318 9.5.2.5 presents a multiplier, , that should be used to estimate the effects of creep and
shrinkage. The multiplier is a function of the creep curve and the amount of steel in the
compression zone. The creep curve brings in the time effect. The reinforcement ratio accounts
for presence of compression zone steel.

Note that the addition of more compression zone steel has a significant effect in reducing the
multiplier. When long term deflections are an issue (i.e. when they exceed the allowable), it may
be a good idea to add more steel in the compression zone for that reason alone. The addition of
compression zone steel does little to add to the flexural strength and is often omitted in the
flexural capacity calculation (as has been stated several times before in the course).

The long term deflection component is found by applying the multiplier to the immediate
deflection due to the sustained load. The total deflection is the sum of the long term deflection
component and the immediate deflections.

An example is in order. We will continue the problem started in the notes for immediate
deflection calculations (see those notes for the details)

For this problem, we will assume that 50% of the live load is sustained for five years or more.
An additional 25% of the live load is sustained for twelve months. We will need to compute two
multipliers.

For both cases the compression zone reinforcement ratio is zero since this beam has no
compression steel. This means that equals (zeta) in both cases. Lambda will equal 2.0 for
the five year loads and 1.4 for the twelve month loads.

The immediate deflections will also need to be computed for the various load stages. The
spreadsheet results are shown in Table 1:

Long Term Deflections T.B. Quimby 2

 Table 1
Immediate Deflections

load Ma Mcr/Ma Ie defl

Stage
(plf) (ft-k) (in^4) (in)
DL Only 800 250 0.615 52108 0.599
DL + .5 LL 1600 500 0.307 47375 1.317
DL + .75 LL 2000 625 0.246 47045 1.658
DL + LL 2400 750 0.205 46899 1.996
LL only 1.397

The graph in Figure 1 illustrates the deflection at various stages of loading. The five year and
one year sustained loads/deflections are shown. The sustained deflections are multiplied by their
respective values.

Figure 1
Load vs Immediate Deflection

To check ACI 318 9.5.2.6 requirements "that part of the total deflection occurring after
attachment of nonstructural elements" must be computed. This has several parts in our case.

The total immediate live load only deflection. From the table above, this value is 1.996
in.

Long Term Deflections T.B. Quimby 3

 The long term deflection due to the five year loads. This found by multiplying the DL +
.5LL deflection by the for the five year plus loading. This equals 2.0*1.317 in = 2.635
in.
The long term deflection due to the one year loads. Part of this deflection is already
accounted for in the previous term, so here the additional creep is determined by
multiplying the sustained load deformation not already accounted for by the for twelve
months. The calculation is 1.4*(1.658-1.317) = 0.477 in.

The total long term deflection is the sum of the second two components: 2.635 + 0.477 = 3.112
in.

That part of the total deflection occurring after attachment of nonstructural elements is the sum
of all three components. The result is 4.509 in.

For the beam in question, the limit on this deflection is (50')/240 = 2.50 in. This is much less
than the computed 4.51". Something must be done to limit the long term deflections. The most
likely way would be to add compression reinforcement. This will help in two ways. First it will
increase Ie by increasing the Icr. Secondly, it will reduce the long term effect multiplier lambda.

Lets try adding four #8 bars in the compression zone and see what happens. ' becomes 4(.79
in2)/(18 in)/32.625 in) = 0.00538. This reduces the five year from 2.0 to 1.58 and the one year
from 1.4 to 1.10.

Computing the new Icr requires the derivation of a new set of formulas for the doubly reinforced
situation.

Figure 2
Doubly Reinforced Section

Figure 2 shows the various parts of the problem. Note the (n-1) term in the computation for Cs.
This term is there so that the area of the bars is not double counted. The computation of C
already includes the areas of the actual bars.

Long Term Deflections T.B. Quimby 4

As before, the equilibrium equation is : 0 = T - Cs - C

Note that fs = fc*(d-c)/c (as previously shown) and fs' = fc*(c-d')/c. Making these substitutions
into the equilibrium equation results in:

0 = n*As*fc*(d-c)/c - (n-1)*As'*fc*(c-d')/c - fc*c*b/2

0 = n*As*(d-c) - (n-1)*As'*(c-d') - c2*b/2

0 = (n*As*d + (n-1)*As'*d') - (n*As + (n-1)*As')*c - (b/2)*c2

Solve this quadratic equation for c, the location of the neutral axis. In this case, c = 12.91 in.

The cracked moment of inertia is then found about this axis using principles of statics.

Icr = [bc3/3] + [(d-c)2*n*As] + [(d'-c)2*(n-1)*As]

Again we ignored the moment of interias of the bars about their own centriodal axis since it is
not significant. The resulting cracked moment of inertia for this problem is 57809 in4. The
recomputed the immediate deflections are given in Table 2.

Table 2
Immediate Deflections in Doubly Reinforced Beam

load Ma Mcr/Ma Ie defl

Stage
(plf) (ft-k) (in^4) (in)
DL Only 800 250 0.615 60638 0.515
DL + .5 LL 1600 500 0.307 58163 1.073
DL + .75 LL 2000 625 0.246 57990 1.345
DL + LL 2400 750 0.205 57914 1.617
LL only 1.102

The live load only deflection is less than before and still well within the limiting value.

Long Term Deflections T.B. Quimby 5

The recomputed long term deflections are given in Table 3.

Table 3
Long Term Deflections in Doubly Reinforced Beam

Stage duration Increment Long Term

DL Only 5+ years 1.58 0.515 0.811
DL + .5 LL 5+ years 1.58 0.558 0.880
DL + .75
1 year 1.10 0.272 0.300
LL
zero
DL + LL 0 0.271 0.000
years
Total Long Term Deflection 1.992

That part of the total deflection occurring after attachment of nonstructural elements is the sum
of the LL only deflection and the total long term deflection. The result is 3.093 in. which is still
greater than the permitted 2.50 inches. Further iterations show that seven #8 bars will get the
resulting deflection below the 2.50 inch limit.

Summary

There are several things that should be noted:

The long term deflections can be (and usually are) greater than the immediate deflections
Adding compression steel can have a dramatic effect on long term deflection control even
though it does not add significant flexural strength. It is still acceptable to ignore the
presence of compression steel when doing flexural strength computations.
Always remember that superposition does NOT work with RC deflection computations.
You MUST consider load stages. Differential deflections are found by finding the
differences in load stages. This is different than what you can do in other materials. The
reason for this behavior is addition of cracking to the analysis.
Remember that these calculations can be avoided by adhering to the span/depth ratios in
ACI 318 table 9.5a. The problem with doing this is that the resulting members will
probably have excessive depth.

Long Term Deflections T.B. Quimby 6