I am currently analyzing several systems which are all connected to a common flare header. As recommended by other members before in this forum, I combined all the models into one to attain a more accurate result. The combined models resulted in more than 1500 nodes. But for the sake of reporting, I rerun the calculation of each system separately for I am planning to report it individually. By inputting the displacement in the connecting node between the system branch to flare header obtained from calculation result of combined model, I thought that I that the result of separated model would be no difference as compared to combined model. But the result tells otherwise. The loading are far different (usually higher in separated model) and some lines connecting to header are overstressed when displacement is considered which did not occur in a combined model. What is wrong my assumption? Did I miss something here? Please help…..

Dear,

If the inputs are correct, your combined model is more accurate than separate ones.

When you are considering only the six DOF displacements & SIFs at branch point for seperate flare analysis, the analysis ignores the effect of six DOF force-moment generated from this analysis on the main header analysis. If you go to-and-fro between these separate analyses with displacements from main header & force-moment from branches, these errors will reduce.

regards,

sam

Thanks Mr. Sam. I have tried your suggestion in one model and it appears to be effective. The stress have gone down from 140% to just 75% when moment and forces are applied. But I am still having trouble with the loading. The difference between operating and sustained loading is unrealistic. From 3 tons in sustained condition it goes up to 20 tons in operating condition, in contrast to combined model values for operating condition which slightly decreased.

I am insisting to separate this models for better reporting. My senior manager and even the client would not be very happy reviewing one report generated from more than 1500 nodes even for the fact that you have provided a more accurate result. At the same time, I am thinking that separating this models might save additional work in the future. For instance, if client insisted additional load combination for one system will mean I have to rerun the calculation of whole system and prepare the report all over again whereas; this can be isolated in separated model. Please share some thoughts for this one..

You can consider a portion of main header piping with branch system bound by nearest anchors/ enough restraints in three directions. By this way you can have separate, but, a bit more accurate run. Decoupling means loss of accuracy - judgement decides where to stop.

regards,

sam

Why not run the whole asnalysis as one model, then report it out in node ranges that correspond to your individual models? Now you have the best of both worlds.

pipedrifter

"I am insisting to separate this models for better reporting. My senior manager and even the client would not be very happy reviewing one report generated from more than 1500 nodes even for the fact that you have provided a more accurate result."

sam
"Decoupling means loss of accuracy - judgement decides where to stop."

pipedrifter, So you THINK the senior manager/client would be OK to compromise the accuracy of the analysis in lieu of Aesthetically pleasing report?? Have you has a word with your manager/client regarding the acceptability of your (lesser accurate report) approach?

Sam has it nailed perfectly. While CraigB has given a good balanced solution without compromising the accuracy of the analysis.

And in terms of future rerun's I dont see any problem as such if you have proper documentation in the first place.

Good Luck.

If you make stress ISOs for your reports calculate the full model. Copy and paste the model to an other directory, change model name using Windows explorer. Open each of the models and delete what you do not want to report and run Stress ISO module. Using suitable model names you can use the model name as Stress ISO description.

If you use CAESAR II Save As it removes calc results! Set annotations before copying the model. To get full system drawing on one sheet you are better off by using Stress ISO system that is included with CadWorx.

End result is readable reports.

so analysis takes a back seat to reports???? "I am insisting to separate this models for better reporting." bad Priorities....