Basic Caesar II / pipe analysis queries - Structural Engineer

[Seng1984]

I'm a Structural Engineer, appointed to design various support brackets for a pipe run analysed using Caesar II, however the Mechanical Engineer responsible for the pipe analysis seems content to press print all, fire off reams of information (most of which is likely irrelevant to me) and is exceedingly slow at responding to clarification requests.

Whilst I'll continue to press for a response I've been attempting to become better informed in order to ask 'better' questions and as such would appreciate some confirmation (or correction) regarding what I've managed to pick up:

Pipe run has been designed in accordance with BS EN 13480.

OPE cases - operating cases at the various defined temperatures and pressures - absolute unfactored values for displacements, reactions and stresses.

SUS cases - conditions immediately after installation, pressurised but not heated - absolute unfactored values for displacements, reactions and stresses.

EXP cases - thermal ranges based on the extremes of temperature applicable - used for pipe analysis and not support design as not absolute values.

Cold pulls (or pushes) - intentional pre-stressing of a pipe to move the absolute reactions etc. to a more reasonable range (e.g. + or - 150kN instead of -300 or +0kN) - not to be considered as part of EXP cases.

Any assistance on the above would be very welcome.

 

[richay]

I'm not really sure what you're asking, perhaps an explanation of the above load case types? If so, here is a brief set of explanations:

OPE - the operating conditions, likely what you want to look at for restraint design. From a production/monetary point of view, this is where you want to be all the time.

SUS - this is a load case setup to evaluate the stresses for the "primary - force based" failure criteria. You should look at the restraint loads here to be thorough, but it is unlikely that this case will govern displacements or loads. This case is required by most Piping Codes for stress compliance.

EXP - this is a load case setup to evaluate the stresses for the "secondary - expansion" failure criteria. This is a "range case", the difference between two other cases. You really shouldn't look at any values except stresses from these cases. If you do look at restraint loads, be aware these are the "change" in loads between the two referenced cases.

COLD SPRING - the pre-loading/pre-stressing of the system to lower nozzle loads on equipment. These cases could affect your restraint design, so to be thorough, take a look at these in your design.

You'll find a wealth of information on (Pipe) Flexibility Analysis in the CAESAR II Discussion Forum at

http://www.coade.com.

Richard Ay
Intergraph CAS

 

[Seng1984]

Richard,

Your response is greatly appreciated and covers the kind of explanation of the various load cases within the output provided that I was hoping for; it also confirms my own research.

 

[ZippyDDoodah]

I'm also a structural engineer. Here are a few important points - pipe stress analyses are almost always nonlinear, so "unfactored" piping loads are unacceptable for structural design since superposition in load combinations is not valid for nonlinear load cases. You must insist that the piping engineer provides you factored load cases per your requirements for structural design (ASD or LRFD factors). "DEAD" case piping loads should include cold pull strain load factored same as DEAD if in fact the piping was cold pulled, since that strain load coincides with ambient temperature empty piping before it becomes operational. Cold pull is not common in my experience.

Operating loads should be factored per your design code within the nonlinear piping stress analysis - Dead, Live, Thermal, cold pull strain + possibly Wind or Seismic.

Having said that, seismic loads in piping stress analysis are unusable trash 99% of the time because piping stress engineers wrongly analyze seismic loads as if the piping is acting in isolation when in fact the piping is on a frame support structure which usually supports heavy equipment in addition to piping. Piping engineers routinely ignore mass and flexibility of the support structure and mass distribution of equipment, cable trays, cladding, etc. in seismic analysis. For example, 300 tons of equipment on your pipe rack? That's all ignored in piping stress seismic analysis, which piping engineers use for piping design! Unjustifiable and insane? Damn right it is. Yet you'll never find a piping engineer who acknowledges the irrational insanity of their seismic design procedures. Show me the forum post where a piping engineer acknowledges this.

Also with wind load, piping engineers ignore the wind load that the structure and equipment "catches", which in physical reality drags the piping along with the structure. For that reason and others, piping stress analysis of wind loads is similarly trash, yet piping engineers use wind load calcs, as well as their crazy seismic load calcs, for piping design code checks and for checks against piping reactions at equipment conditions. Ridiculous? Absolutely.. yet this is typical design practice for piping engineering.

I would like to hear a justification from piping engineers as to how they rationalize their calculation of piping reactions under seismic loads (and wind loads). Truth is, I don't believe it can be justified. Piping engineers - Be honest, own up, and comment on this problem on public forums..

 

[nickelkid]

ZippyDDoodah, as noted input/support from Structural folk is required for an honest piping analysis. I have worked several piping analysis with structural input, often with better results. Unfortunately most projects do not want someone performing a pipe analysis that will speak up, as this requires time.

 

[ZippyDDoodah]

nickelkid, thanks for your response, but I have to ask, what was the extent of the "input/support" from your structural folks? For each seismic and wind load case, did you obtain structural displacements for each of those load cases and then apply them to your piping stress model at each pipe support location as imposed displacements? And then redo that effort for each design change? Because that's exactly what would need to be done in order to have a reliable piping stress analysis.. I doubt there was time for that. What's typically done is the piping engineer wrongly analyzes and designs the piping with the assumption that the piping is supported on a rigid undeflected structure because there's not enough time to do anything else.

In my experience, piping engineers are usually under the impression that support structure flexibility will only help them, an assumption which is often not true. Piping engineers spend a lot of time to consider 1" or 2" of imposed displacement at vessel connections, yet those same piping engineers will ignore 4" of imposed displacement at pipe supports from the frame support structure under seismic or wind load. It's irrational, but unfortunately it's usually how it's done.

Even if you went to the time and trouble to use and apply structural frame imposed displacements for each occasional load case in the piping stress model at each pipe support location, even then, although much better than nothing, those calculated imposed displacements would still be somewhat problematic - since piping loads on the structural model used for seismic, wind, and other load cases would only be an estimate, because piping load distribution could shift under thermal and occasional loads with nonlinear pipe supports. Structural models don't usually account for nonlinear pipe support behavior which is why we typically rely on piping engineers for those loads/reactions.. Except for seismic and wind load cases. And there lies a significant dilemma that's rarely acknowledged.

 

[nickelkid]

ZippyDDoodah, you are right on the money with displacements! The last analysis I did had significant displacements at support locations resulting from structure, especially during occasional loads, and sustained. The seismic cases I have been involved with have not been in areas that it was the limiting factor, but wind. On 2 occasions when I arrived at marginal results analyzing wind loads, they both failed after I obtained displacements for supporting structure. When practical I will estimate and have stiffness validated of associated structure when it appears that it may help or hurt... Typically I like to make a simple run first, prior to adding displacements or stiffness's.

With all due respect, the biggest problem I find with most piping analysis is that the fundamental input is incorrect, but that does fall in line with what you are saying...

NB I worked under a Principal Engineer that was Structural. Fortunately we learned from each other, prior to him, my analyses would be worthy of your claims!

 

[BigInch]

While in theory much or all of what you say can be true, IMO in the vast majority of cases structural displacememts are not usually considered, since it is the relative displacements of structural joints that affect piping and in general piping is very flexible in relation to the structures supporting them. There is little to no important overall action going from structure to pipe in typical cases, so it is usually only necessary to consider loads from pipe to structure, not structure to pipe. I believe that the typical structural analysis is far too conservative to be applied to piping analysis as you suggest anyway, as levels of redundancy and interactions of structures are far more complex than what the typical analysis would indicate and usually the actual structural displacements will be far less than what is seen in theory, even if they ever did reach design magnitudes after all the factors are applied to each load. That's basically why all the simplifications you do not like have worked well over the last 100 years, regardless of the fact that they probably could be made more theoretically correct today. Remember engineering is not always concerned as much with making reality match theory as it is in simplifying a theory to make design practical and cost effective. As long as one is careful to include structural, or trench to pipe displacements where such are important, as at tank foundations with nearby nozzles, etc., most other cases of structure induced deflections will never reach critical stress levels. Most pipe can deflect 1/2 inch in a 20ft span (even considering simple supports) without experiencing any overstress at all, whereas such a deflection in a structural member, even if not overstressed, would cause unsightly cracks in concrete beams, or otherwise make a floor unsuitable for human occupation due to vibratory response alone. It is of little concern to the piping near the top of a vertical vessel or an offshore platform if it moves 4 inches laterally with the platform, if that pipe is 150 ft long and is not even in fact completely fixed against rotation at the base. Then again I've had the opposite case where vessel foundation settlement was so severe I've had to remove the nuts from the anchor bolts of smaller vessels and let the pipe hold it up in the air.

 

[ZippyDDoodah]

"Most pipe can deflect 1/2 inch in a 20ft span (even considering simple supports) without experiencing any overstress at all, whereas such a deflection in a structural member, even if not overstressed, would cause unsightly cracks in concrete beams"

"in general piping is very flexible in relation to the structures supporting them. There is little to no important overall action going from structure to pipe in typical cases"

The second quote seems to be the prevalent opinion among piping engineers in my experience, yet in reality, industry standard lateral deflection limit for pipe rack structures is Height/100. This limit is spelled out in ASCE 7 and Process Industry Practices (PIP). That means a 40ft tall pipe rack is designed to have almost 5" of lateral deflection while still meeting drift requirements, 60ft rack over 7" lateral deflection, etc.

Large diameter steam pipes are often routed in higher levels of the pipe rack where deflections are highest. Air coolers are often supported on top of the rack with medium to large diameter piping. Pipe racks have piping which is often routed like an octopus with pipes branching out at various locations going to different areas and connecting to pumps, vessels and other equipment - which are usually NOT deflecting in phase with the pipe rack.

L.C. Peng's book 'Pipe Stress Engineering' page 205: "In a process or power plant, the sizes of structural members, such as columns and beams, are comparable to, or smaller than, the sizes of many of the pipes. The effectiveness of the structure members as pipe supports can be an issue on the integrity of the piping system. Sometimes it is even hard to say which is supporting which.... Most support structures experience various movements, including thermal expansion, earthquake/wind displacement, and so forth. These movements impose load on the piping and have to be included in the analysis.... It should be noted that the pipe may also move in the opposite direction of the support structure if the structure is flexible"

The imposed displacements at pipe support locations are real, they are often significant in magnitude, and it's a problem that so many piping stress engineers ignore or diminish the importance of these coupled effects. Just because it's not convenient to consider these imposed displacements, does not justify ignoring them or claiming that they don't matter when they do.

 

[BigInch]

There are obviously cases where interaction can be significant. The trick is to know when they do and when they don't. If it isn't obvious, then the inexperienced engineer has no alternative except to do a full analysis. The obvious will become more so as experience is gained.

 

[LenBog]

ZippyDDoodah wrote: The imposed displacements at pipe support locations are real, they are often significant in magnitude, and it's a problem that so many piping stress engineers ignore or diminish the importance of these coupled effects. Just because it's not convenient to consider these imposed displacements, does not justify ignoring them or claiming that they don't matter when they do.

Totally agree with this statement and two solutions of this problem could be discussed:

First Solution. Integration of Piping Stress and Structural analytical models into one single model. I can foresee two formidable obstacles here:
1.1 There is no such engineering analytical program and likely no software-developer would ever come up with such a product (the market niche is too narrow hence sales would be very limited)
1.2 Even if such software ever to be developed - there is no User; i.e. the Engineer who would combine professional expertise for both Piping-Stress and Structural field and work simultaneously for two different Disciplines/Departments of the same EPC firm.

Second Solution. To develop a 3D graphic 'Viewer' (or 'Navigator' if you will) which would be importing both Structural and Piping-stress analytical models (only work-lines and nodes; without 3-D shapes). Aside of these 3D 'sticks & dots' you'll see panes with various information about user-selected entities (weather piping or structural components) - imported from stress or structural output and properly compiled for convenient preview.
For example Piping-Stress Engineer could see displacements of support nodes belonging to a few different piping systems but automatically associated with the single structural beam supporting them (these nodes are 'hovering' above the beam - the shoe height plus the pipe radius is the anticipated 'gap'). Moreover, he can easily derive the rigidity of the virtual 'spring' and introduce this spring at the given support node.
For Structural Engineer such 'Viewer' should be augmented by very different functionality: after cutting off the piping located beyond his 'Box' (the volume of the structure designed) he would quickly preview each level (with associated piping) see that all support-nodes are matching the beams and anchors/guides are where he has anticipated, 'map' stress-combinations to his 'piping' load-cases and export the Structural Model impregnated by all relevant 'Piping Loads' (reactions from the support-nodes of the stress-model would be automatically projected onto the supporting beam and recorded as point forces (or member loads) written at the proper 'mapped' load-cases.

I'd like to hear your opinions - especially from ZippyDDoodah

Cheers Len