The Operation Scheduling Of Electrical Power System Based On Particle Swarm Optimization

Unit commitment problem is very important in the operation scheduling of electrical power systems and it is a minimization problem which determines the on-off state of each generating unit and the generating capacity of a switched-on unit so as to ensure least generating cost with the fulfillment of multiple constraints. Mathematically, the unit commitment problem is a high-dimensional, non-convex, discrete, nonlinear combinatorial optimization problem, an NP-hard problem and it is quite difficult to obtain the optimal solution. So most of the research of this problem aims at getting a near-optimal solution. The Particle Swarm Optimization (PSO) algorithm is an intelligent optimization algorithm imitating flock behavior and the algorithm performs well for both continuous and discrete optimization problems. PSO is fit for unit commitment. Traditional unit commitment problems are all backgrounded by thermal power generation, which has the disadvantages of high fuel cost, severe environmental pollution and the like. Wind energy is a kind of novel clean energy which is environmentally friendly and low in cost, with wide development prospects. The development of wind turbine generators technique has further expands the application of wind power generation. In this thesis, we proposed an improved PSO to solve the traditional unit commitment problem and some unit commitment problems related with wind power generation, including about three aspects.Firstly, this thesis solves the unit commitment problem with a scheduling cycle of24hours by an improved PSO algorithm. During the improvement on the algorithm, the quantum evolutionary algorithm is used to update binary discrete variables, and asynchronous learning factors with linear-decrease time varying weight to update velocity of particles. In addition, a new repair strategy is utilized to guarantee slight damages to particles repaired. To verify the application performance of the proposed algorithm, the thesis obtains upper bound and the lower bound of the unit commitment problem using the Lagrange relaxation method. Through computational experiments on groups of examples, it is proved that the proposed improved PSO algorithm could obtain a satisfactory feasible solution, thereby proving the feasibility and effectiveness of the algorithm.Secondly, this thesis introduces wind power generation technique into traditional unit commitment problems and finds the solution of a new problem with the previously proposed improved PSO algorithm. The final simulation results are also satisfactory compared with results from the Lagrange relaxation method again.However, wind energy also has the defects of easy interruption, non-stability and so on. In order to overcome these disadvantages, finally, this thesis introduces the battery energy storage system into a wind generating set. The battery energy storage system stores surplus wind energy obtained when wind energy is sufficient for use when wind energy is insufficient by the battery’s charging and discharging so as to improve the utilization efficiency of wind energy significantly. A PSO algorithm with additional particles is applied to the solution to a unit commitment problem involving both wind turbine generators and battery energy storage system, the final simulation results are also compared with that by the Lagrange relaxation method, so as to prove the feasibility and the effectiveness of the proposed algorithm.