A Short-Term Bus Load Forecasting Model Based On Data Mining Technology And CEEMD-ELM

In 2018,China's power industry will also face new challenges and tasks.Continuously enhancing the construction of the power grid,ensuring and improving the safety and reliability of power grids and operating economy are the most important work of the power grid.Precise grid bus load forecasting will directly affect grid safety early warning analysis,grid transmission capacity calculation,operation plan arrangement,power generation planning,safety constrained scheduling,reactive power optimization scheduling,and optimal power flow analysis results.And it also plays an important basic role in congestion management and security check.At present,a large number of scholars have studied the bus load forecasting model,but there are still some problems.Generally,days before the forecast day are usually selected as historical similar days without analyzing the differences between them.Many present prediction algorithms can not solve the problems of randomness,instability and nonlinearity of the bus load,generating a large error in the bus load forecasting results.The preprocessing algorithms used can not preserve the characteristics of the original bus load sequence well and some prediction models only optimize the parameters in the prediction algorithm without any preprocessing of the original bus load sequence at allIn this paper,the characteristics of two typical 220KV bus loads in Hunan Provincial Power Grid were analyzed.Some influencing factor was discovered through data mining(DM)including temperature,humidity.week type and holiday type.In addition,several algorithm of signal decomposition preprocessing were introduced,including empirical mode decomposition(EMD)algorithm,ensemble empirical mode decomposition(EEMD)algorithm,and complete ensemble empirical mode decomposition(CEEMD)algorithm.And a bus load forecasting method based on CEEMD and extreme learning machine(ELM)was proposed in this paper.In order to further optimize the prediction model and improve bus load forecasting accuracy,a short-term bus load forecasting model based on data mining technology and CEEMD-ELM was proposed.Firstly,hierarchical clustering method was used for clustering the historical daily bus load.Secondly,a decision tree based on the clustering results was constructed.Thirdly,according to the properties of the forecast day,such as temperature,humidity,weekday and vacation information,historical daily bus load was obtained to train the forecasting model of complete ensemble empirical mode decomposition and extreme learning machine through the decision tree.Fourthly,non-stationary time series of historical daily bus load were decomposed into a series of stable components by using CEEMD and each component was forecasted based on the ELM.Finally,the forecasts of components were composed to get the final forecast result.Compared with traditional single extreme learning machine and extreme learning machine based on data mining technology,the analysis of two 220KV buses load forecasting experiments in Hunan Province shows that the method has higher accuracy.