Research On Unit Commitment Based On Improved Discrete Particle Swarm Optimization

The unit commitment (UC) optimization is a practical application of value engineering. It can bring significant annual financial savings and social benefit, and causes the electricity workers and researchers'widespread interest. Optimization if unit is a typical non-linear optimization problem, and involves many non-linear, discrete, stochastic, uncertainties and other factors. In the view of mathematics, it is a NP hard combinatorial optimization problem with many constraints and it is difficult to find the optimal solution in theory. So far, there is still not a practical algorithm which can not only consider all the constraints but also achieve an ideal computing speed and accuracy. How to improve the accuracy and speed of solving the unit commitment problem is still of great significance to the optimal operating of power system. The simulated evolutionary optimization solution is best suited to solve the highly non-linear, discrete and combinatorial optimization problems that traditional optimization methods can't solve.As a result of the search mechanism is simple, easy to program, PSO has been widely used in solving practical problems. In order to improve the convergent speed and accuracy of traditional DPSO, this paper proposes a novel DP SO. Simulation results show that proposed method can get better optimization effect, which validates the effectiveness of the method.Secondly, propose the basic principles and mathematical models of UC in detail. This paper proposes a new discrete particle swarm optimization combining of Comprehensive learning particle strategy (SCLPSO) to solve unit commitment (UC) problem in power system. The UC problem is decomposed into two embedded optimization sub-problems:one the unit on/off status schedule problem with integer variables that could be solved by SCLPSO and the other load economic dispatch problem with continuous variables that could be solved by the primal—dual interior point algorithm. The constraint-handling method is based on repair algorithms not based on penalty functions to raise solution quality. The simulation results show that SCLPSO can get better optimization effect, which validates the effectiveness of the method.Thirdly, research on the conventional UC problem and the optimization method. According to the characteristics of unit commitment(UC) and the security constraints, this paper proposes an improved comprehensive learning strategy based on novel particle swarm optimization. This method can solve security constrained unit commitment well. In addition, this paper proposes a way to produce high-quality particle swarm. This makes the initial particle in the feasible region and that improves the quality of the solution. The feasibility and effectiveness of the proposed method are demonstrated for two test systems with IEEE30 and IEEE118, and the computational results are compared with the custom Benders decomposition and the commercial software CPLEX. The wind power has become one of the most important ways to cope with the worldwide increasing environmental and energy problems. The unit commitment problem with wind power integrated is not a traditional certain problem because of the uncertainty of the wind power and the solution which is economic and reliable is difficult to attain by traditional ways. Considering the present wind power prediction level and the prediction error, the occasional wind curtailment model of the unit commitment with wind power is well established. In order to achieve the higher system flexibility and economy, the energy storage system is also introduced into UC problem with wind power, and the introduction of energy storage system not only can reduce the operation cost, but also can alleviate the impact of wind power volatility on the UC problem.