Integral Sliding Mode Current Controller And Hybrid Particle Swarm Optimization For Doubly-fed Wind Power Generator

Doubly fed induction generator (DFIG) with the rotor excitation has been one of the main wind turbine types for the variable speed constant frequency wind power system. It has advantages of excellent speed adjustment, flexible power control ability as well as fractionally rated converter due to the independent control of the excitation current frequency, amplitude and phase angle. Effective excitation current control plays a key role for the DFIG-based wind power system to achieve advantaged operation performance. The researches on the DFIG excitation current control and optimization design are very important to the development of wind power technology and industry, as the DFIG operation characteristics are the prerequisite for the effective excitation control.Focusing on the problem that the DFIG excitation current control tends to be disturbed by system uncertainties, mathematical models of the DFIG excitation current control in no-load stage and power generation stage are built. The sliding mode control strategy is proved to be matched with the uncertainties of the system, such as operating switch. A novel integral sliding mode controller with exponential reaching law for DFIG current control is proposed. Then, the excitation control for DFIG-based wind power system in no-load stage is studied by digital simulation. The simulation is conducted in the cases of ideal grid voltage, fluctuant grid voltage, perturbation of the generator parameters and operating stage switch. The results show that the system under integral sliding mode control accomplishes switches from no-load to generation stages without additional changes in controller, and it is relatively robust to grid voltage disturbance and generator parameter perturbations.The particle swarm optimization (PSO) algorithm has some problems such as slow convergence and high probability of being trapped in local optima when it is applied to the DFIG optimization design and other nonlinear multimodal optimization problem. In this paper, a hybrid particle swarm optimization (HPSO) algorithm, in which a fitness-guided individual fuzzy inertia weight and a diversity-guided adaptive mutation are introduced, is proposed to solve the problems. Different inertia weights should be fuzzily assigned to different particles in the same generation, and the mutation probability and factor are adaptively calculated by the population diversity. The optimization results of benchmark function show that the proposed HPSO algorithm achieves an optimized trade off between the convergence and population diversity, with proper dynamic balance between global and local searching ability. In addition, it also makes improvements in terms of quick convergence, high precision and other commendable optimizing performances such as the absence of premature convergence. Those characteristics make the proposed HPSO algorithm applicable to the nonlinear optimization problems.In addition, an HPSO algorithm integrated with Pareto optimality is proposed for multiobjective optimization design of the DFIG in this paper. The electromagnetic design features are studied, the parameters matching rules between the generator and its excitation converter are discussed, the effects of the harmonics caused by non-sinusoidal excitation currents are analyzed, and an optimization model of the DFIG design is built. Based on the model, the effective material cost, efficiency and flat efficiency curve are separately selected as the optimization goal for the DFIG optimization design by the HPSO algorithm. The optimization results of a DFIG design example show that the proposed HPSO algorithm behaves quick convergence and high precision as it is applied to the benchmark functions optimization. The work of this paper is useful to achieve the overall optimization of the economic and technical performances indices for DFIGs.