In currently used steady-state simulation programs, setpoints should be given for all devices. After a simulation is done, it has to be checked if no constraints are violated. For instance : a discharge pressure for a compressor is given and as a result the calculated compressor flow is to large. A new run has to be done with the compressor flow cant rolled on its maximum flow. (As a result somewhere else it may go wrong etcetera.) In the new approach constraints can be given for all devices. After performing a simulation the simulator will automatically find all the setpoints so that no constraints are violated. Moreover if no solution is possible a list of possible causes will be generated. This new approach is used for the planning, where planning is based on reliability, and a large number of simulations has to be done.


In this paper we describe an algorithm which is implemented in a software package for performing steady-state network calculations. The package is used by the (long term) technical planning of Gasunie. Traditionally the planning activities where focused on the transport capacity at design conditions and some other worst case scenarios. This involved steady state calculations for a very limited number of transport situations. Lately the planning became more involved with questions about reliability. This implied that calculations had to be done for various combinations of supply and demand, but also the possibility of failing equipment had to be included. The central quest ions in these calculations are: Is it possible to satisfy all demand with the equipment and the supply available? And if not what is the best we can do? The approach until now was: guess a combination of setpoints and control modes, perform a steady state calculation, inspect the results for violations of the constraints, and repeat until a satisfying solution is found. This approach can be very time consuming. It is an iteration process with manual labour in it. The new algorithm finds a best solution within the given constraints automatically. Instead of checking combinations one can now give attention to the real planning problems, such as identifying the bottlenecks in the transmission system and finding solutions for them. In the operational planning answers have to be found for questions like: Is it possible to shut down a certain piece of equipment for maintenance given the expected load on the system? Or: Is it possible to accommodate a request for additional transport capacity? Also this type of questions require a search for the best solution with the available resources and is therefore a field where the new model could be used. However in Gasunie this is not yet the case. The new algorithm is based on Linear Programming (LP). This technique gives us also the possibility to find answers to a total different question: If no feasible transport solution is found, what are the causes for this failure?

This content is only available via PDF.
You can access this article if you purchase or spend a download.