Spatial Load Situation Awareness And Regulation Based On IForest-VMD And Neural Networks

At present,under the strategic goal of "double carbon",the energy structure of our country is being optimized and adjusted,and the proportion of traditional fossil energy such as oil and natural gas is gradually declining.Instead,the proportion of electric energy consumption is increasing continuously.The electric energy substitution on the load side is advancing continuously,especially taking the construction of electric vehicle charging station as a typical example.Not only the total amount of electricity demand is increasing,but also the demand distribution is becoming more and more complex.In this context,improving the accuracy of space load situation awareness results can avoid blind construction or unreasonable planning in the distribution network.At the same time,how to achieve effective regulation of load to ensure the balance of power supply and demand,these are all issues that must be considered.First of all,this paper introduces the research background of space load situation awareness,summarizes the research results at home and abroad.The existing space load forecasting methods are simply introduced and classified according to different standards;The influence factors of power load are introduced and analyzed;The regularity of spatial load from the point of view of time and space are analyzed;The power geographic information system needed in this paper are introduced and established.Secondly,a spatial load situation awareness method based on isolation forest,variational mode decomposition,multi-layer perceptron and gated recurrent unit（i Forest-VMD-MLP-GRU）is proposed.In the part of situation perception,the isolation forest algorithm is used to identify the outliers of the measured data of type I cell load under the given spatial resolution in the electric power geographic information system,and the Lagrangian interpolation method is used to modify it,so as to determine the reasonable type I cell load data.In the part of situation comprehension,the type I cell load data after the situation perception is decomposed by variational mode decomposition method.The components of different central frequencies are obtained,and the trend components and low frequency components are determined according to their energy values.In the part of situation forecast,the trend component is predicted by the multilayer perceptron,and the low frequency component is predicted by the gated recurrent unit,and then the predicted results of the two components are reconstructed.The results of class I cell load situational awareness in target year are obtained.The gridding technique is used to transform it into the result based on class II cell.The effectiveness of the above method is verified by an example.Finally,a load regulation strategy based on demand response is proposed.According to the relevant data,summarize and analyze the existing enterprise energy efficiency evaluation index,and select the appropriate relevant index to construct the enterprise energy efficiency evaluation index system.The score index of each enterprise is calculated by AHP method and entropy weight-TOPSIS method respectively,and then the comprehensive score is calculated.On this basis,all enterprises are divided into different energy efficiency grades according to their scores,and then sign different contracts with power companies or choose different electricity price packages,so as to guide the production and load of enterprises to change.Considering the participation of various types of users to the demand response,and on the basis of the situation awareness results of the IVMG method,the optimized situation awareness results under the demand response are calculated.An example is analyzed to verify the effectiveness of the method.