Optimal Design and Cost/Benefit Analysis of Hydroelectric Power Systems by Genetic Algorithms

Local optimization of individual hydroelectric power plant parameters such as tunnel diameters or dam height is a standard feature in the design of hydropower stations. Combining such local optimization problems into a global optimization process for a set or sequence of hydroelectric plants with traditional optimization techniques is an interesting and challenging problem considering the potential of powerful modern computer hardware and software. It is, however, often a formidable task and a problematic one, due to the often excessive computational burden and the associated complexity of the problem. Furthermore, global optimization is often difficult due to factors such as the dynamic behavior and nonlinearity of the objective function as well as multiple local optima. Therefore attempts have been made to solve this problem for each hydro plant by efficient and modern methods such as those based on evolutionary computational techniques. In this paper this problem is developed by addressing simultaneously a sequence of hydroprojects rather than a single project, and performing the global optimization simultaneously on all projects using both a proposed one-shot approach and iterative techniques, using Lagrange multipliers as economic signals between stages in the expansion process, while the method of Genetic Algorithms (GA) is used at each stage in the parameter design and optimization of each hydroelectric project.Finally a case study is presented and discussed based on data from the Icelandic electrical power system for optimization of individual projects. Therefore, these general tools and specific case studies should be important for the design and operation of hydro-dominated power systems in general. The stage wise computations were carried out using the HYDRA software developed at the University of Iceland.