PSO Approaches For Optimal Scheduling Of Cascaded Hydropower And Hydro-thermal Power Stations

The optimizations of cascaded hydropower stations and hydrothermal power systems with cascaded reservoirs have always been the forefront and hotspot of academic and engineering research. Especially in recent years, China is gradually accelerating the development process of hydropower. The large-scale cascaded hydropower stations are developed continuously and put into operation gradually. The joint optimal scheduling for cascaded hydropower stations will be the vested operation mode of the cascaded stations in future. The joint scheduling for cascaded stations must consider the intricate relationships among hydropower stations and departments of water utilization, and the uncertainty hydrological runoff and system load. The joint scheduling for hydrothermal power systems must consider the load dispatch among hydro and thermal plants additionally. All these make the scheduling issues be high-dimensional, strong constrained and nonlinear optimization problems. The solution of the optimization problems will be an obstacle of the joint optimal scheduling of cascaded hydropower stations. Finding a novel algorithm has been one of the most important issues due to the rapid development of cascaded hydropower stations and the obstacle faced. In this thesis, the optimal scheduling of cascaded hydropower stations and hydrothermal power systems with cascaded reservoirs are deeply studied by adopting systems engineering theory, optimization methods, computer science, and techniques of swam intelligence. Some conclusions with theoretical and practical value are obtained. The main innovation achievements are as follows:(1) A mathematical model which maximizes the total benefits of cascaded hydropower stations is established. Multi-guide particle swarm optimizer (MGPSO), which introduces a multi-guide set into the selection of flying leaders, is proposed to solve the optimization problem. This approach is verified through an application to the scheduling of the cascaded hydropower stations on some river in southwest China. Simulations show the algorithm's strong exploration ability in the later evolution stage. After being validated, MGPSO is applied to the comparisons among the different on-grid prices of cascaded hydropower stations in Lancang River Basin. (2) A complex scheduling model for hydrothermal power systems is established with consideration of water balance of hydro reservoirs, system load balance, and physical limitations of generators and reservoirs. A modified PSO is developed to solve this optimization problem. In the proposed algorithm, a mutation operation acting on the selection of particles' personal best, and a migration operation which regenerates newly diverse population are introduced. The simulations show the modified PSO has fast convergence and the optimal solution obtained is more competitive.(3) A small population-based PSO (SPPSO) approach is proposed to solve the complex constrained hydrothermal scheduling problem. In this approach, small population (3-5 particles) instead of large population including hundred of particles is adopted in order to avoid more CPU time consumed and lower convergence speed. The operations of mutation, migration and DE-acceleration are incorporated into the SPPSO to enhance its competitive performance. In addition, a special repair procedure instead of penalty function approaches is also employed to handle the complex constraints of hydrothermal scheduling problem. Three hydrothermal test systems from literature are considered to test the proposed algorithm. The results have been reported through the comparisons with those obtained by other population-based evolutionary techniques. It is found that the proposed SPPSO can obtain more competitive solutions within faster convergence and shorter time consumed.(4) A mathematical model minimizing the total power production cost including both fuel and start-up cost is present for the short-term unit commitment of hydrothermal power systems. Aiming at the solution of this optimization problem, a small population-based hybrid particle swarm optimization (SPHPSO) approach is proposed basing on the improvement and expansion of the SPPSO. The hybrid algorithm employs the binary particle swarm optimization and the SPPSO to handle the variables of discrete unit status and continuous power, respectively. The SPHPSO is used to solve the hydrothermal unit commitment problem after it is validated in thermal unit commitment problem.(5) Aiming at multi-objective hydrothermal scheduling problems, a mathematical model with consideration of fuel cost, emissions of atmospheric pollutants, transmission losses, energy storage, peaking benefits and ecological impact of cascaded hydropower plants is built. In order to solve the short-term environmental economic hydrothermal scheduling problem, one of the famous multi-objective hydrothermal scheduling problems, two solution approaches are proposed. Multi-objective is coverted into a single objective through weight factors and penalty value factors firstly, and then the SPPSO is adopted to solve the converted optimization problem in one of the solution approaches. Small population-based multi-objective PSO basing on Pareto-domination theory (SPMPSO), which is extended from the above SPPSO, is presented aiming at optimizing the different objective simultaneously in the other solution approach. After it is verified in the famous hydrothermal test systems, the SPMPSO is applied to the multi-objective hydrogenation scheduling of the cascaded hydropower stations among Lancang River Basin.