Study On The Wind Power Prediction Based On MFF-IWMC And Its Application In Risk Assessment

As the installed capacity of wind power increases,wind power irregularity and volatility on the power system caused by the adverse impact gradually appear.Therefore,carry out wind power prediction and turbine operation risk assessment is one of the necessary technologies to guarantee system stability and reduce the adverse impact of wind power on the power system.Currently,the machine learning algorithm is used in the wind power prediction and risk assessment model.Specifically as follows: first,due to the redundancy of power data,unjustified weights occur when weights are determined by applying rough set theory.Second,due to the limitations of the model and the fact that the long time power series do not have a clear regularity,therefore,the direct using of meteorological factors and wind power as the input data set of the model is still insufficient.Third,in the risk assessment model of wind turbine based on output power prediction,predictor model will have the unreasonable phenomenon of repetition of predicted state probability,and the uncertainty and time dependence of risk degree sequence is ignored by the risk assessment model.First,in the traditional rough set weighted Markov chain prediction model,a conditional entropy-based rough set-WMC method for wind power prediction is proposed,which address the problem that the redundancy of power data leads to defects in determining the weights of traditional rough sets.Finally,an arithmetic validation is performed,and the results show that the unreasonable weights is avoided by the proposed method in this chapter avoids and the root mean square error of wind power prediction error is reduced from 16.54% to 15.96% compared with the traditional model.Second,the irregularity of the long time power series will be neglected by the direct input of meteorological factors and power,a multiple joint probability is proposed firstly.The release probability matrix in HMM is improved by multiple joint probabilities,which can solve the problem that meteorological factors is ignored by MC and the irregularity of power series direct input of meteorological factors and power is ignored by the direct input of meteorological factors and power.Second,the problem of ignoring time series correlation in traditional models is solved by applying the weighted forecasting method of determining weights in Chapter 3,and a multivariate wind power forecasting model based on MFF-IWHMM is established.Finally,in the arithmetic validation,the probability of duplication as well as zero is avoided in the improved release matrix.the phenomenon of unreasonable weight determination is also avoided by the proposed method of determining the weights using Chapter 3,and the final RMSE of wind power prediction is reduced from15.96% to 11.56% compared with the conventional model.Thirdly,aiming at the lack of application above model,a wind turbine operation risk assessment method based on wind power prediction model is proposed.In this method,an sensitivity-layer IWHMM wind power prediction and its risk assessment are proposed,which aim at the problem that the above HMM model is unreasonable in selection of final prediction state,and its own uncertainty and time series correlation is ignored in the risk degree series.First,in the prediction model of sensitivity-layer IWHMM proposed in this paper,in which the trend of wind power in the wind speed interval is extracted by using sensitivity,the unreasonable phenomenon of possible probability repetition of the state distribution in the first two chapters is solved.Further,a wind power uncertainty risk assessment model is established.The problem of uncertainty of risk series and time series characteristics in traditional risk assessment can be solved by model,which considers the probability of risk occurrence in the risk degree series and the time information of the series is mined by HMM.Finally,in the arithmetic validation,RMSE by 5.49% and the error rate of risk assessment by 4.03% is reduced by the proposed model which is compared with the traditional risk degree assessment method.By adopting proposed wind power prediction model and risk assessment model of wind turbines,the accuracy of wind power prediction can be improved by the wind power forecasting model of the input data,model and the improvement and optimization of the output in the part of risk assessment,and risk assessment can be made more reasonable and higher accuracy...