Economic short term scheduling for a run-of-river-plant chain by combined LP and genetic optimization

A simulation and optimization tool for daily operation planning of a cascaded hydro electric power plant system is proposed in the paper. To maximize the value of (weighted) energy generation subject to imposed constraints the optimal water release schedule is refined in a functional hierarchy down from a several days scope to a four hours horizon. A two-step-concept is chosen: coarse pre-optimization by linear programming, followed by a detailed optimizing search procedure, for which a self-adjusting genetic algorithm was designed. An inherent and essential part of this approach is the developed (fully) nonlinear, dynamic simulation model. Test case results and obtained performance improvements versus current scheduling practice are reported