question about devideing big system in to a few small.

Hello to all.

I have big system for calculation, (10-15 lines). I want to split lines so that my calculations will be simplest. Is it correct if I put displacement (dx,dy,dz,rx,ry,rz) in branch points,flange connection etc. so that i known impact rest of piping.

Thanks for answers.

Usually "no". The reason is that the influence of one part on another part will be lost. You can break the system at anchors without suffering this problem.

Thanks for answer.

I can put anchor but in real condition it will not be anchor. Can i put forces and moments (with oposite direction).

Here is my system and pump suction piping that I want to devide.

http://i56.tinypic.com/20iig6v.jpg

1.First of all don't divide the system.10-15 lines is not a big system.(in my experience i have done some systems with 20-30 lines because of getting actual effects from one line to another line.)
2.At last you want to the divide the system, you have to divide the system up to stop support and next guide support.
3. As per your snapshot you can divide your system from horizontal equipment to pump- take the main line from horizontal equipment and up to stop and guide support. And take the suction lines from pump,these suction lines should connected with header, so take the line up to stop and guide support.

regards
venkatachalam N

I agree with you. The main reason was because i have to check discharge on pumps (one working the other stand-by) for Tope and Tmax.desig. For rest of the system i have different temperatures, so i just want to simplify load cases combinations. Easiest way is to divide system with axial stops and guides (with or without gaps). But what about moments.

You might find it easier to just give node names (e.g. P-301 Suction, P-302 Suction) to the terminal points so you can identify them quickly in your analysis. Then you can extract the data to your report with a minimum of fuss and no possibility of error.

I do not have problem to identify nozzles. I have to verify nozzle loads for two temperature cases, one on Tope (one pump discharge is working and other is standby)For standby portion up to first valve i have to consider temperature 50% of Tope, after temperature is Tambient. Same think is for another pump. After this I have to do the same but with Tdesign. It is strange for me, I think Tdesign is enough for verifying nozzle loads. This is according specification for stress analysis from client. All other lines in the system are with different temperatures then pump discharge lines.

I just wander is it OK if I divide just discharge lines in to one system because it is very sensitive piping. Pumps are inline piston type and allowable are very small.

Rest of other piping in the system are OK regarding flexibility.

Typically, pump nozzles are considered immovable by the piping system (yet they can have their own casing thermal growth). This is usually conservative in light of the pipe strain loads.
If this is how you see your nozzles, then there is no reason to model beyond the nozzle, away from the pipe. This "anchor" breaks the system and no information (forces&moments, displacements&rotations) can pass through.

I have done like you wrote Mr.Dave. Discharge nozzles are OK but for suction nozzles I had to ask for approval manufacturer for little bigger axial allowable load.

Thanks All for very useful answers.

Kemal.

Kemal,

Not always design temperature is the worst case scenario, sometimes when you deal with pumps "one operating, one spare" is more critical than design in terms of loads on pump nozzles.

Regards,