The Study Of The Internal Economic Operation Of Hydropower Plant Based On The Operating Region Partitioning

The internal economic operation of a hydropower plant is making the start-stop sequence, optimizing the number, the commitment and the loads dispatch of hydroelectric generating units through computation under constraints for the best economic benifits. The internal economic operation of hydropower plant involves the economic dispatch (ED) and the unit commitment (UC). This dissertation builds mathematical models for the two problems based on the Genetic Algorithm with emphasis on the units operation stability.In order to achieve the stable and efficient operation, the ED model is built based on the accomplishment of our research group, that is, to partition the operating region into the stable region, the transient region and the restricted region. The output regions of a unit under certain heads were separated by the restricted region into discrete feasible regions. The economic dispatch results must avoid the restricted regions, and the chance of the operation in the transient region is reduced as much as possible by adding the penalty function. Thus, the operation stability of the units is ensured with the consideration of the economic operation and the continuity of the output. To solve the problem of the discrete feasible regions, some operations are introduced to the improved Genetic Algorithm for the discrete feasible constraints, including adaptive crossover operator, adaptive mutation operator, local search operation, migeration operator between sub-populations and simulated annealing algorithm. The numerical experiments show that all these operations improve the local search ability and the optimization speed. During building the UC model, the start-stop costs are changed into water consumption to take account, in order to prevent the units from frequent start-stop.The internal economic operation of hydropower plant is performed for the the Shuikou Hydropower Plant. Firstly, the combined characteristic curve is extended and fitted by artificial neural network. The fitting precision is higher than that of the least squares fitting. For Shuikou Hydropower Plant, the proposed models and the opetimal algorithm are applied to the optimization on some day. The comparison between the optimization and the actual operation shows the effectiveness and feasibility of the proposed models. The water consumption decreases remarkably, the units are prevented from running in the restricted region, and the operating time in the transient region is reduced obviously. Therefore, the internal economic operation for the Shuikou Hydropower Plant is meaningful to the economic benefit and the operation stability. Finally, economic dispatch results for various loads under the heads of 40m, 49m and 60m are presented for reference.