How do analysts manage modeling piping systems containing multiple equipment and/or valve scenarios when temperature, pressure, or displacement (T/P/D) vectors exceed the limit of nine conditions?

For example: four boiler feedwater pumps, one acting as spare, and any one, two, or three of the pumps may operate to meet load. There are 14 T/P/D conditions that can occur. That will take 6 models to analyze all shutdown/startup cases, and switchover combinations.

It is anticipated that there will be multiple piping revisions to the models as attempts are made to get pump nozzle allowables in compliance.

An approach is to start with the base model and make six copies. Then edit all the T/P/D vectors for each element for each model. Finally, edit the load case combinations for each model. Since it is not possible to copy/paste complete T/P/D vectors (in the Element List), this is a very lengthy process.

With a change in the base model then, the choices are two as I see it: 1) Change the base model and follow the steps above, or 2) Change the piping in all six of the models, and update the affected T/P/D vectors in those six models.

Am I missing a simpler, or less time consuming approach?

The choice of revision method 1 or 2 then, would be based on the extent of piping modifications that are necessary.

- Ken

Judgment may reduce your workload. For example, you may find from preliminary runs that the load case with pumps A and D (opposite ends of the string) running is essentially identical to A, B, D and A, C, D. Similarly, you may find that A, C and A, B, C share common results and B, D vs. B, C, D.

But in extreme cases, you may have to perform many runs to cope with all possible load cases. My personal record for load cases is in the low 40's and we have a job in-house where the norm is in the high 50's.

I think you will find it far better to have duplicate models around, so you can save record input files for all of the output. They're easier to check that way.

Base models are the way to go, but where are the displacements coming from in a thermal system?

CraigB & TJN -

CraigB: Yes, it appears that some load cases have similar results. But to determine whether there are similar results, all models and load cases have to be set up and analyzed. If piping changes, or boundary condition changes are made, would those same load cases produce similar results? Maybe, but maybe not. Would have to look at all models and load cases again to know. After awhile, one may be able to gather some feel for the response of the system for certain small changes to the system. Still, I've been wrong too many times to trust myself that I can figure out how a complicated piping system will respond to a change.

TJN: I am favoring the base model approach because it would be hard to cross check six models that all piping changes were made the same in each model.

The displacements come from thermal growth of the pumps.

All: I have not had to model operating equipment groups of more than 3 units. Have others modeled 4 or more operating units (pumps, HXs, Boilers, etc.) in a group regularly? It looks like a doubling of the engineering analysis effort for that one extra unit. It seems that there is a large incentive to design for no more than three units if at all possible to keep engineering/analysis costs within reason.

I have now 3 pump system to calc. 3 cases and even that is excessive. Old design from 60's. They put pumps far a part and connected with flexible pipes so no problems - less time consuming solution.

Jouko -

Yes, keeping pumps far apart is wisdom from the past...putting 3 or 4 pumps in a shoebox is the challenge now for piping designers faced with first cost (minimum squarefoot layout) driven design nearsightedness.

Yes, for n equipment items, there are ((2^n) -1) load cases when it is possible for 1 to n equipment items run at once. So you are right - each additional piece of parallel equipment can potentially double your analysis workload.