Minimum time step specification during transient analysis in TLNET (Pipelinestudio)

[BENSMIL]

Hi,

I have been modelling a pumping station ( Booster + Main pumps) and pipeline (68 km 10" long) for analysing the transient conditions like pump trips and shutdown valve closures using TLNET software (Pipelinestudio). I have observed that the peak surge pressure during any case is dependent on the minimum time step which is specified to the program. Lower minimum time step gives higher surge pressures especially at the pump suction. I have used time steps of the order of 0.01 sec, 0.1 sec and 1 sec. I would like to know if any of you have come across a simillar situation while doing simillar studies.

What could be the correct time step to use in such cases?

Regards,
Benoy

 

[BigInch]

The smallest time step always results in the highest accuracy.
You should try to reduce the time step until the differences between the results can be considered acceptable margins of error, ie, they don't make any difference in the pipe and fitting selections, or affect any other critical parameters, ie. relief tank size, etc.

Independent events are seldomly independent.

 

[BENSMIL]

I have reduced the time step to the least possible on my machine i.e 0.001 sec. However the peak surge pressure keeps going up. The event is the main pump trip initiated at 10 sec. The trend is attached for various time steps.

 

[Latexman]

I would expect those results IF the valve closing time was decreased. It doesn't make sense to me with decreasing step time. Are you sure of the variable that is being changed?

Good luck,
Latexman

Need help writing a question or understanding a reply? forum1529

 

[MortenA]

The way i remember it, TLNET is a finete difference solve scheme. The min. time step parameter is only used when the solver need to reduce time step in order to obtain the solution. The solver will then extend the time step when the solution is more stable.

I think the manual give some hints. There is a relationship between the minimum timestep to obtain a stable solution and the length of a pipe segment plus the wavespeed of the fluid. I cant remember details.

In general for liquid systems, if your are investigating water hammer phenomena the the minimum timestep should be less than half of the length of the distance from the source to the valve that you close(i think - its been a while and i dont use the software anymore). TLNET does not consider column separation as far as i remember so downstream the valve is not that important as far as i remember.

Best regards

Morten