Optimal control of high pressure gas networks by two different methods

Author: A. J. Osiadacz

The aim of the paper is to present two different methods implemented for optimal control of high pressure gas networks.

The problem is of large scale and is highly non-linear in both objective function and constraints.

The first method is based on the Lagrangian dual formulation.

The second method uses hierachical systems theory.

Discrete state equation for the case in which output pressure are treated as elements of the control vector has been formulated.

Elements of the state vector are non-outlet node pressures and flows through units. In both cases algorithms for optimal control over periods up to a day have been developed.

Both methods assume that the overall objective is to minimise the total cost of operating the compressors over the entire control period.

This objective may be expressed as a time integral; the objective function used is a discrete approximation to this integral, a non-linear function.

Some other published methods are briefly described also.

Results of investigations are included.