A routine task of every natural gas provider is the short-term (operative) planning of the load distribution in the pipeline network to satisfy the predicted demand over a planning period of 24 to 48 hours. Variable operating costs are dominated by the energy reqired for the gas transport; thus minimizing the cost requires sophisticated control strategies for the technical equipment. Mathematically this leads to highly dimensional mixed-integer optimization problems involving nonlinear partial differential equations.
Goal of the project was the development of suitable nonlinear models in cooperation with the industrial partners Ruhrgas AG and PSI AG, and the development of structure-exploiting optimization algorithms for the case of prescribed combinatorial decisions.
