Research On Short-term Wind Power Forecasting And Economic Optimization Algorithm Of The Microgrid

Wind conversion technology is one of the rapidly growing energy technologies in the world,however,wind power appeals to intense intermittent and uncertainty along with the frequent surroundings fluctuation,which causes the enormous difficulty for wind pow-er usage.While the short-term wind power forecasting approaches bring sufficient data in the power system to guarantee its stable operation and reasonable dispatch,thereby it is meaningful to acquire more precise data on wind power forecasting.Meanwhile,the microgrid improves the renewable energy penetration rate and provides reliable energy to the non-electrical area.The microgrid always combines different generator types,thus it becomes an important issue to achieve the microgrid operation optimal status by consider-ing different tariffs of different units.Aiming to solve problems aforementioned,this paper researches the short-term wind power forecasting and microgrid economic optimization algorithms and the contributions are as follows:(1)A particle swarm optimization dynamic grey model(PDGM)is proposed to solve the short-term wind power forecasting by using historical wind power data.This model uses the particle swarm optimization(PSO)algorithm to make the background value changeable,and exponent parameters are optimized by iteration searching and linearizing processing.Furthermore,the residual model is introduced to compensate for the transient surroundings.Furthermore,this paper also proposes a hybrid forecasting approach based on PDGM,wavelet transform and Lyapunov exponent prediction method,which decomposes the power curve to low-and high-frequency by wavelet transform and then they are predicted by PDGM and largest Lyapunov exponent prediction,respectively.The final forecasting value is obtained from the combination of these prediction values.The proposed model and approach are tested by the real wind power data and simulations validate that they improve the forecasting accuracy effectively.(2)A hybrid method using multi-source data is proposed for the short-term wind power forecasting,which comprises the support vector machine(SVM)and grey model.Firstly,the support vector machine is used to forecast the wind speed and direction in terms of different weight calculated by grey correlation analysis through the historical weather data.Then,the wind power is predicted by the grey model with the combination of numerical weather prediction and wind vectors predicted by SVM.PSO algorithm is introduced to continuously optimize its background value to achieve more accurate prediction curves.Finally,the Fourier residual sequence is employed to provide a deterrent effect of uncertainties brought from surroundings.Thereafter,the proposed approach adopts the data collected in an actual wind farm.Simulation results with four seasons validate that the proposed approach performs more effective than other methods in forecasting the short-term wind power.(3)An immune particle swarm optimization(IMPSQ)algorithm is proposed to solve the problem of the optimal economic dispatch of the microgrid with the couple of multi-constraints,discontinuous and multi-generators.The algorithm introduces the artificial immune mechanism into PSO algorithm,and the particles with small affinity in the coordinate plane are subjected to cross and mutation,which guaran-tees the particles distributed uniformly in the solution space and the convergence of the algorithm,thus ensuring the global searching ability and the robustness of the algorithm.In order to test the effectiveness of the algorithm,a microgrid model without the uncertainty of renewable energy is established in this section and the proposed algorithm is applied in solving the island and grid-connected operation modes of this model.With the comparison of other three algorithms,simulations show that the proposed algorithm effectively solve the problem of economic opti-mization of the microgrid.(4)A hybrid particle swarm optimization(HPSO)algorithm is proposed to solve the economic optimization of the microgrid under the uncertain of the renewable ener-gy,which is based on the stochastic weight trade-off particle swarm optimization algorithm,the immune mechanism is introduced to uniformly distribute particles in the coordinate plane.The nonlinear strategy is integrated to balance the searching ability and searching accuracy.Furthermore,the sub-gradient method is presented to accelerate the convergence rate.Each controllable power resource is deployed dynamically.The electricity market mechanism is employed to the microgrid sys-tem to dispatch units with more effective and to realize the economic benefits.The simulation results of the island and grid-connected operation modes show that the proposed algorithm outperforms for resolving the problem.In summary,this paper discusses two problems and divides into four main points:(?)PDGM and a hybrid forecasting approach are proposed to forecast the wind pow-er in short-term using the simple historical wind power data,respectively;(?)A hybrid forecasting approach is proposed for short-term wind power forecasting using multi-source data;(?)IMPSO algorithm is proposed to solve the problem of microgrid economic opti-mization;(?)HPSO algorithm is proposed to solve the microgrid economic optimization considering the uncertain of the renewable energy.