Research On Short-Term Electric Load Forecasting Based On VMD-HSBSO-ELM

High-precision short-term load forecasting is a prerequisite to ensure the safe,reliable,high-quality,and economic operation of the power system.It is also of great significance to the optimal combination of units,economic dispatch and optimal power flow of the power system.Due to the many random factors that affect power load,and its own strong nonlinearity and non-stationarity,traditional short-term load forecasting methods have been difficult to meet the demand for load forecasting accuracy of contemporary power systems due to the limitations of their own algorithms.Therefore,it is of great significance to study modern high-precision short-term load forecasting methods.In view of this,this thesis studies a short-term power load forecasting model that combines the improved Extreme Learning Machine(ELM)algorithm and the Variational Mode Decomposition(VMD)technology.Firstly,this thesis builds four short-term load forecasting models: traditional ELM,BP neural network,Long ShortTerm Memory(LSTM)neural network and Support Vector Machine(SVM),and compares and analyzes them through simulation experiment.The results verify the effectiveness of the ELM algorithm.Secondly,in view of the fact that the input weights and hidden layer biases of the traditional ELM algorithm are randomly given,the prediction accuracy is difficult to meet the requirements.A Brain Storm Optimization algorithm with a harmonious search mechanism(HSBSO)is used to improve it.And a short-term load forecasting model based on HSBSO-ELM is proposed,and through simulation experiments compared it with the ELM forecasting model improved by other optimization algorithms,verifying that the HSBSO algorithm is more advanced;Finally,in view of the strong non-stationary and non-linear characteristics of the original power load,this thesis applies VMD technology to short-term load forecasting.First,the VMD technology is used to decompose the original power load sequence into multiple sub-sequences,and then combined with the HSBSO-ELM prediction model to predict each sub-sequence separately,and finally the prediction results of each subsequence are reconstructed to obtain the final predicted value.The simulation experiment results can prove that the short-term load forecasting model based on VMD-HSBSO-ELM researched in this thesis has extremely high forecasting accuracy and is feasible in practice.