Dynamic Economic Dispatch Of Power System Considering Uncertainty Of Wind Power Forecast Error

In recent years,environmental problems caused by traditional fossil energy were becoming increasingly prominent,wind power as the representative of clean renewable energy has been rapid development.At the same time,due to its inherent randomness,volatility and intermittency,wind power brings new challenges to power system operation and dispatching.Therefore,the study of wind power forecasting method and power system dynamic economic dispatch considering the uncertainty of wind power have become urgent problems to be solved.A short-term wind power forecasting model based on gated recurrent unit network is studied.Firstly,the correlation coefficients between meteorological factors and wind power time series were calculated,meteorological factors with significant correlation with wind power were selected as input variables.Then the time series of multi-input variables were modeled dynamically by gated recurrent unit-recurrent neural network(GRU-RNN),parameters of GRU-RNN were trained by back propagation through time(BPTT)algorithm and adaptive moment estimation(Adam)algorithm.Model forecasting effect is verified by measured data of an actual wind farm.A segment fitting method based on Gaussian mixture model(GMM)to fit wind power forecast error is studied.Based on the historical data of an actual wind farm,the difference of forecast error probability distribution of wind power forecasted value in different intervals was analyzed.Wind power forecasted value was divided into several intervals.Then GMM was proposed to fit the probability distribution of forecast error in each interval respectively,the parameters of GMM were estimated by expectation maximization(EM)algorithm.GMM is compared with other probability distribution models;fitting results show that GMM has better flexibility and fitting precision.A dynamic economic dispatch model considering uncertainty of wind power and load forecast errors is established.Spinning reserve chance constraints based on the random variables of wind power and load forecast errors were included in the model.Wind power and load forecast errors were equivalent to a new random variable;the probability density function of the equivalent random variable was derived from the probability distribution model of wind power and load forecast errors.The stochastic optimization model with chance constraints was transformed into an equivalent deterministic model.In view of the disadvantages of standard bat algorithm(BA),such as easy to get into the local optimal solution and slow convergence rate in the later iteration,a hybrid bat algorithm(HBA)combining quantum behavior and chaotic map was proposed to solve the formulation.Simulation results verify the effectiveness of the proposed model and algorithm.