For a standard system anlysis with wind/ earthquake loading, the following load cases are run in CAESAR:
L1: W + P1 + T1 (OPE)
L2: W + P1 + T1 + WIND (OPE+WIND)
L3: W + P1 (SUS)
L4: L1 - L3 (EXP)
L5: L2 - L1 (OCC,WIND EFFECT)
L6: L5 + L3 (SUS + OCC)
Cases 3, 4 and 6 are the B31.3 stress cases. Which case should be used in checking nozzle deflections/ loadings etc.? Is it case 2?
Why doesn't B31.3 recognise wind loading in computing expansion stresses? i.e why isn't case 2 used along with case 3 to determine a 'worst case' expansion stress range (L2-L3), rather than adding the wind effect to the sustained case.

Hi Ben,

This “pipe stress analysis” is an interesting thing. The “pipe stress” part of it is really to a large degree prescribed by the B31 Codes. We look at the secondary stresses (thermal “expansion stress range” and “displacement stress range”) in an entirely different way that we look at the primary stresses “sustained stresses” or “additive stresses” of weight and pressure. CAESAR II is designed to address the Code requirements for stress analysis, and it sets up the load cases accordingly).

In considering the thermal expansion stress range (secondary stresses), we are looking at the “complete range of temperature” from the minimum temperature to the maximum temperature. For example, if the piping system is erected on a day when the temperature is 70 degrees (assume that there is no cold spring and the system has “zero” thermal stress at 70 degrees), and the maximum design temperature is 500 degrees, there will be a 530 degree difference in temperature. However, when that system is shut down on a cold winter day, it may see a temperature of, for example, 10 degrees. In this case the piping sees a temperature change of minus 60 degrees. So the maximum range of temperature change (delta T) would be 530 + 60 = 590 degrees (this is one complete thermal cycle, we have Code rules for considering partial thermal cycles too). In this case, we allow CAESAR II to assume the “stress-less” temperature is 70 degrees and the total delta T is 590 degrees (linearly elastic behavior to loading). We set up our model with an analysis T1 temperature of 590 degrees. Also for stress calculation, the Code disallows considering the effect of “cold spring” (but who uses THAT any more?).

Also, the Code says that we must use the “cold” modulus of elasticity (Young's Modulus) for STRESS analysis. So, we use the Young's modulus for the pipe material at 70 degrees. When CAESAR II automatically sets up the load cases, you will get the proper cases for Code stress analysis. It is also common practice (if we have some “surplus” stress margin) to assume that at the ends of the piping system (a vessel nozzle for example) we have an “anchor”. That is to say six degrees of fixity. For static analysis this would be a conservative assumption (it would result in calculated stresses greater than the actual stresses). When the secondary stresses are shown to be within the Code allowable stress range, we turn our attention to the sustained loading. The calculated “expansion” (or “displacement”) STRESS RANGE is compared to the B31 “allowable STRESS RANGE”, Sa. Where Sa = f X (1.25 X Sc + 0.25 X Sc).

The sustained loadings are to a large degree (if we have used a wall thickness that is appropriate for the design pressure) a function of how well the piping system is supported. If we are successful in our design, we will have a CAESAR II Code stress report that says that we have satisfied the requirements for stress analysis (and/or “flexibility analysis”).

However, when we look at the loadings on the terminal equipment nozzles (turbines, pumps, pressure vessels, et. al.) we cannot use the model as it was developed for stress analysis. We are interested in calculating the forces and moments applied by (transferred by) the piping to the nozzles at the operating temperature and also at the coldest “out-of-service” temperature (of course this logic will have to be “adjusted” when you consider cryogenic systems). So, there are two analyses needed to do this. First, the model is set up to consider the temperature change from 70 degrees to operating temperature (lets assume 500 degrees, although the operating temperature may be less) and so the delta T for this case is 530 degrees (not the 590 degrees from the “thermal expansion case”. Also, it is OK to use the “at temperature” Young's modulus (more flexible material) and to include the effects of “cold spring” AND nozzle rotation (depending upon how much conservatism the piping engineer wants to include or exclude). Then, it would be prudent to look at the "winter day” loadings on the nozzles – delta T would be 70 to 10 degrees (minus 60 degrees), the cold Young's modulus would be used and the effects of “cold spring” AND nozzle rotation might be included. Note that we are looking at the TWO temperature excursions separately – this is because the resulting calculated nozzle stresses will have opposite signs. JUST REMEMBER THAT ANY THERMAL PIPE STRESS RANGES THAT ARE CALCULATED IN THESE ANALYSES SHOULD BE IGNORED.

OK, that sets the stage for looking at your questions.

Why doesn't B31.3 recognise wind loads in computing expansion stresses? It is because wind loadings are considered primary loadings and result in primary (non-self limiting) stresses (they are not diminished by the system deflecting). These primary stresses are considered in concert with the “longitudinal stresses due to weight and pressure”. The Secondary (or “self limiting”) stresses are considered as a “range” of stresses as discussed above. As an aside, at this time we consider “occasional seismic loadings” to be primary loadings but we are rethinking that now due to data resulting from full size, shake table tests that clearly produced ratcheting failures. Stay tuned to you B31 Code book.

What would be the load case to be considered for evaluation of the nozzle loadings? None of the above. Remember, when you are looking at the nozzles in the “hot” (i.e., operating) condition, you need only the delta T associated with the temperature excursion from the thermal “zero stress” temperature, to the OPERATING temperature. It is assumed that the “temperature at erection time” will result in the smallest loadings on the nozzles (note that this is only true if there is NO intentional (or unintentional) pre-stress (i.e., “cold spring”) in the system). Then, you should look at nozzle loads given the case of the temperature excursion from the “stress-less” temperature (i.e., 70 degrees) to the coldest temperature (and if you are unfortunate, the effect of “cold spring” would be included in this analysis) with the cold Young's modulus. And of course, operating temperatures that are less than “as erected” temperatures require adjustments.

I hope this will be of some value to you.

Best regards, John.

John,
What you have said about primary/secondary stresses and stress range cases is correct, and to consider the hot and cold operating cases separately for nozzle loading, but you have not answered whether or not wind or other loads such as siesmic should also be included in nozzle load evaluation. My answer is that all possible operating cases should be included in nozzle load evaluation, combined with and without other loads such as wind and seismic, ie. piping at ambient, at cold temp and at high temperature for completeness. The equipment vendor may have higher allowable loads for such upset cases as seismic.

John,

I'm not sure about your approach to include the restraint loads
of the minimum temperature case into the evaluation of e.g. nozzle loads.
Imagine a pump discharge line shut-down on a cold winter day
with an support next to the pump resting on a steel structure; The steel
will have the same temperature-change as the pipe and the vertical loads
on the nozzle are expected to be close to the sustained results. Furthermore,
a shut down condition with a pump out of service might not be the basis
for the allowable nozzle loads indicate by the pump supplier.

Just an idea…

Veit

Thank you Mike and Veit. Your's are valuable observations

I agree to a large degree with Mike and Veit. As Mike observed, it is very important when evaluating either the pipe stress or the nozzle stress (due to loadings applied through the piping) to be sure to compare the calculated stresses to the appropriate allowable stresses.

The allowable stresses will vary depending upon the nature (and/or frequency) of the loading (and in some cases, the situation which produced the loading). It is important that the calculated thermal expansion stress ranges be compared to the Code allowable stress range. And it is important that the calculated (Primary) sustained stresses be compared to the ALLOWABLE STRESS AT TEMPERATURE from Appendix A of the B31 Code. The minimum temperature allowable stress may be greater than the maximum (design) temperature allowable stress.

Mike also hints at another interesting topic for discussion - what are to be considered the "on-design" loadings (the loadings that were included as creditable loadings than must be considered in the design) and the "off-design" loadings (loads that were NOT considered in the design because they are considered improbable in the design life of the system). We would all agree that the 25 year maximum wind (with gusts) is something that we must include as a normal (creditable) design load. But, what about the 100-year maximum wind? Or, what about flooding? What about seismic events? We can have “on-design” occasional loadings (e.g., seismic loads, heavy wind loads, relief valves lifting, etc.) and the B31 Code gives us allowable stresses for these. But for the “off-design” loads (e.g., a truck collides with a system, a mine subsides causing foundations to sink, a 100 year hurricane hits, sabotage, etc.) what would be the appropriate maximum allowable stress for these? Thankfully, these are evaluated “after the fact” and we do not have to set up CAESAR II loading combinations for them.

Wind loadings on exposed piping that is not housed within a building are normal (creditable), and these loadings should be included in the design basis and considered in the sustained load case. Furthermore, I think that if Ben's L5 case includes the 25 year wind it should be combined with the sustained loadings of pressure and weight as a normal sustained load combination that should be compared to 1.0 X Sh. However if it includes the 100 year wind, then it should be considered an occasional load combination and the stress for comparison should be 1.33 X Sh. Seismic events are occasional in many areas of the world (in Texas we can count the number in our lifetime on one hand) but perhaps they are weekly events (for some lesser magnitudes) in other areas of the world - if you can expect several seismic events within the design life of the system, perhaps this is another "on-design" occasional load case.

I think that the evaluation of the mechanical integrity of nozzles should include (as Mike points out) all the creditable, "on-design" loadings - maximum temperature and minimum temperature (with cold-spring included) and say, 25-year winds AND IT SHOULD INCLUDE “ON-DESIGN”, OCCASIONAL LOADS. As Mike points out, the equipment manufacturer might have (as does there B31 Codes) a greater maximum allowable stress (or loading) to compare these stresses (or loadings) to if they occur very infrequently. There is a point though, where there must be an operating procedure in place that requires inspection of vessel nozzles (and flange bolts, and....) in the event of a sustained wind greater than ....., or a fire, or a flood, or a seismic event greater than magnitude 4 - make your own list. It is not practical to design the nozzles or the piping for all possible "off-design" loading. The design basis document must unambiguously define (envelope) the creditable design life loadings. Whenever these loading are exceeded by an “off-design” event, lets go have a look. Then it becomes a matter of having a procedure in place for evaluating by analysis the effect of “off-design” loadings after they happen. The procedure would address “what to do if” and the degree of NDE required if the calculations indicate that the pipe or nozzles might have exceeded yield.

Veit, I think that we should evaluate the minimum temperature loading on the nozzles (even considering the effect of the contracting restraint) but we might compare the stresses that occur in the nozzles to a rather high allowable. Maybe (with the concurrence of the manufacturer) nearly up to yield. The limits on loadings on rotating equipment must include consideration of misalignment due to casing deflection because the bearings may bind if the casing is deflected just a little while the shaft is rotating – when it is out of service, wiping out bearings is not an issue. These high loadings will be cyclic (maybe annual) and so the S-N curves for the material at issue will have to be considered in how close to yield we can go for a reasonable number of cycles in the nozzle life. Certainly, if we will have the loading even infrequently, we should put these nozzles on the NDE list for periodic inspection and bolt replacement.

Defining the load cases that are to be considered and combined by CAESAR II has a certain amount of “art” mixed in with the logic. You have to play “what-if” a little to decide what combinations are possible and how they must be fit into the Code stress “categories” of primary and secondary stresses. Also, what is “on-design” and what is “off-design”. CAESAR II can address the majority of likely load combinations that might be needed based upon your input. But only you will know if more is needed.

Anybody have any further thoughts? Thanks to everybody (so far) for their contributions to the discussion. How "bout the rest of y'all?

Best regards, John.

This has not been my experience. A wind load is pretty much always treated as occasional (OCC in Caesar). We've talked to structural engineers about wind loading, since ASCE 7-95 (now 7-98) is the basis for the wind pressures we use. B31.3 considered the multiplier for increased allowables based on duration and frequency. I've sought input from structural engineers if there is any time element involved in the increase in allowables they use for structural steel under wind loading.

The answer basically is no. Wind is considered a temporary loading, much like a hydrotest, and allows for the use of increased stress values, essentially reducing the factor of safety, to allow these situations to be considered without unnecessarily increasing the cost of the design.

In looking at the ASCE criteria, the terminology (which I don't have in front of me, I'm going on memory) suggests the calculated wind speeds are of a transient nature.

As such, I would have a hard time including a wind load in a stress review that did not consider the increased allowables for a temporary loading. To include wind in a simple sustained case review seems excessively conservative.

Now that would depend on the assumed wind speed, wouldn't it? For instance, on a current offshore project (spar and semi-submersible), we are using 75 mph wind as a Maximum Operating Condition. This is the maximum wind speed with which the platform will continue to operate. We also have Extreme Conditions (Process Shut-in) with 158 mph winds. And finally, Transit conditions with 66 mph wind. All in addition to the normal SUS and EXP load cases. Combine these with the rockin' 'n' rollin' of the waves (different for each condition), and you've got a real tedious situation.

The Maximum operating winds are considered in the primary stress check. Extreme and Transit conditions are assumed to be OCC. These are checked, of course, in all 4 directions per "operating" condition.

hi all

Casear II does not give allowable for (OPE) cases for B31.3. explain

Pleasance read the B31 code which you are doing the work under.


Ben my comments on your questions are:
Q1
"Which case should be used in checking nozzle deflections/ loadings etc.? Is it case 2?"

A1
It depends upon the code requirements of the equipment or the equipment manufacturer or requirement and or what the contract requires. For instance ASME Section VIII Div1 provides little guidance whereas Section VIII Div2 provides detailed guidance. The bottom line is you have to satisfy the code stresses with your load cases and may have to add load cases for the equipment load review.


Q2
"Why doesn't B31.3 recognize wind loading in computing expansion stresses?"

A2
The wind loading we are discussing here is the uniformly applied wind pressure. The piping system must be able to resist this load without breaking or collapsing. To do so the system basically should remain below yield, however due to the transient nature of these loads it is traditional to allow an increase to 1.33 Sh or yield at temp. Note these uniform wind loads are added to pressure and weight effects both which are also "collapsing loads".

The expansion stresses are for fatigue type loads which generally are displacements. Note WAM or wind anchor motion is a displacement stress and should be reviewed as a fatigue type load if the displacements are significant.

Pleasance -

Check out this thread which explains the code philosophy.

Richard, I would love to see how many load cases you have to set up to consider all your different wind cases, waves etc etc. Does CAESAR II have enough lines in the load case editor to allow you to build all these?? I thought defining all our load cases was bad enough when we use 5 temp and pressure cases.

I believe you get 99 thats all you have....

And yes you may run out of room and need to run a dupe file to get all things covered. On the other hand maybe your being too thorough a lot of people only run one temp case and ignore all that messy and expensive code stuff...

Gee I wonder why theres no code allowable stress for my operating case?