Research On Power Optimization Control Of Variable-speed Variable-pitch Wind Turbine

The twelfth five-year plan (2011-2015) of the national energy science andtechnology was noted that “the variable pitch and variable speed technology aremainly adopted in wind power generation”. Basing on the short-term prediction modelof wind speed and wind power, variable-speed variable-pitch DFIG is regarded as theresearch object in the paper. Then with the goal of the maximum wind power captureand the smooth output power and taking generator speed and pitch angel as controlvariables, the corresponding multi-objective optimization models are established. Theminimization control criterion and variable universe algorithm are proposed to solvethe optimal solution to realize variable speed variable pitch wind turbine poweroptimization control.The power optimization control of1.5MW DFIG is realized as the followingresearch aspects.1、Mathematical models of variable speed and variable pitch wind turbine areestablished, a rotating coordinate system of DFIG is got with the analysis ofcoordinate transformation theory, and then the simulation models are built in theMATLAB/Simulink platform, this makes preparation for wind and power forecastand power optimization control.2、EMD-RBFNN short-term power prediction model is proposed. The windspeed data is decomposed into a series of intrinsic mode function (IMF) componentswith similar time-frequency characteristics and stationary by using EDM to achievethe stationary of the wind speed data. The IMF components are predicted by RBFNNbased on the time-frequency characteristics of different IMF components. Theorthogonal least squares (OLS) is adopted to minimize the error rate. The eachprediction results of IMF-RBFNN are restructured to obtain the last prediction result. Finally, prediction value is got by power conversion.3、Short-term wind speed forecasting based on rough set theory and PCA-Elmanneural network is established. The wind speed prediction model using Elman neuralnetwork (ElmanNN) is introduced, the principle component analysis (PCA) is used toextract the feature of wind speed data, which optimizes the inputs of ElmanNN.Furthermore, excitation function and the structures of network are improved to searchfor the optimum solution of function convergence rate and prediction accuracy. Tosolve large error and prediction accuracy fluctuations of the ElmanNN model at thepeak value of wind speed, the rough set theory is proposed to compensate and correctthe predicted values to further improve forecasted results. Finally, prediction value isgot by power conversion.4、Based on rough set theory and PCA-ElmanNN model, with the goal of themaximum wind power capture and the smooth output power and taking generatorspeed and pitch angel as control variables, the corresponding multi-objectiveoptimization models are established. The minimization control criterion and variableuniverse algorithm are set out to solve the optimal solution to realize variable speedvariable pitch wind turbine power optimization control.At last, the simulation results of1.5MW DFIG are showed that the output poweris increased and the turbulence of power in low frequency is reduced compared withthe traditional power control strategy and it is realized variable speed variable pitchwind turbine power optimization control.