Gas Pipeline Optimization Using Adaptive Algorithms

Transmission gas pipeline network consume significant amounts of energy. Then, minimizing the energy requirements is a challenging task. Due to the nonlinearity and poor knowledge of the system states, several results, based on the optimal control theory, are obtained only for simple configurations. In this paper an optimization scheme in the face of varying demand is carried out. It is based on the use of a dynamic simulation program as a plant model and the Pareto set technique to sell out useful experiments. Experiments are used for the identification of regression models based on an original class of functions. The nonlinear programming algorithm results. Its connection with regression models permits the definition off-line and for a long time horizon, of the optimal discharge pressure trajectory for all the compressor stations. The use of adaptive algorithms, with high frequency, permits to cancel the effect of unknown disturbances and errors in demand forecasts. In this way, an on-line optimization scheme using data of SCADA system is presented.