Research On Optimal Dispatching Of Distribution Network Considering Uncertainty Of Distributed Power Source Forecast

Wind power and photovoltaic are affected by weather conditions,and their output power has greater randomness and volatility.As a large number of distributed power sources are connected to the distribution network,it has an impact on the safe and stable operation of the distribution network.Therefore,this paper studies wind power photovoltaic forecasting methods,uncertainty analysis of forecasting errors,optimal dispatch of distribution networks considering forecasting uncertainty.Firstly,by analyzing the working principle of the wind-solar unit,construct wind power and photovoltaic power generation models,and describe the operation rules of wind power photovoltaics.Conduct research on the complementarity of wind farms and photovoltaic power plants in the same area,combined with the complementary evaluation index Kendall to analyze the daily average complementarity,monthly complementarity and annual complementarity of distributed wind power and photovoltaic,obtain the optimal time scale of wind power photovoltaic complementary.Secondly,build a wind power and photovoltaic power prediction model based on WOA-LSSVM,this method uses the WOA to optimize the penalty factor and kernel function width of the LSSVM,the wind farm is tested for different models under 4 hours,24 hours,and 72 hours.Carry out prediction tests on different models of photovoltaic power plants on sunny,cloudy and rainy days,combined with the values of the power prediction evaluation indexes RMSE and MAE,the results show that the WOA-LSSVM prediction model is superior to the PSO-LSSVM and LSSVM models.Then the uncertainty analysis method of wind power and photovoltaic power forecast error based on WOA-LSSVM forecast model is established.Applying cloud model knowledge to qualitatively analyze the uncertainty of wind power and photovoltaic forecast errors,obtain the error cloud drop maps of different scales of wind power at 4 hours,24 hours and 72 hours,under different weather conditions of photovoltaic sunny,cloudy and rainy days,using non-parametric kernel density estimation to quantitatively calculate the uncertainty of wind and photovoltaic power forecast errors.As a result,it is found that non-kernel density estimation can more accurately describe the error probability distribution of wind power and photovoltaic power.Calculate the confidence interval under different conditions of wind power and photovoltaic based on non-nuclear density estimation,and solve to the coverage of 97%,95%,90%,85% confidence level.Finally,construct a regional distribution network economic dispatch model considering the uncertainty of wind power and photovoltaic power prediction,based on the prediction error confidence interval,the uncertainty of wind power and photovoltaic forecast is quantified.Therefore,the uncertain factors in the scheduling model are determined.Introduce the uncertainty caused by prediction into the scheduling model in the form of confidence intervals,by calculating the confidence interval under different confidence levels,obtain the 95% and 90% confidence level of the optimal distribution upper and lower limits in the model.And using PSO algorithm to calculate the scheduling cost under different confidence,providing a more scientific and reasonable dispatching scheme for distribution network optimization dispatching.