Adjustable Potential Analysis Model Of Industrial Large Users Based On Machine Learning

Industrial load is a key part of the new load management system.It is of great significance to analyze the adjustable potential of industrial large users to alleviate the contradiction between supply and demand of the power grid and improve the load management ability of the new power system.Large industrial users have huge adjustable potential,but the existing research has not comprehensively considered the effects of regulation costs and price incentives to carry out a time-segment analysis of the bidirectional adjustable potential of large industrial users,which makes it difficult for large industrial users to participate in the demand response.In order to solve the above problems,this paper firstly modeled and analyzed the main production equipment of large industrial users,proposed the adjustment cost parameters corresponding to large industrial users,established the sliding window input model,and made input preparation for the analysis and calculation of adjustment potential.Secondly,the reconstruction method of high adjustable potential data set of large industrial users is proposed,and the selection model of adjustment potential indicators based on Pearson Mutual information is proposed.The various types of adjustment potential indicators are selected.Based on the selected adjustment potential indicators,the Fully Convolutional Networks(FCN)established the data reconstruction model of industrial large users,reconstructed the data set of industrial large users,and obtained the high adjustable potential data set of industrial large users;Finally,a calculation method for the adjustable potential of large industrial users is proposed,and the improved Temporal Convolutional Network(TCN)is established.Based on the data set of high adjustable potential of large industrial users,the adjustable potential range of large industrial users is obtained by time segment and direction.The research content of this paper is helpful to promote the construction of adjustable resource pool,improve the establishment of new load management system,promote the unified management,regulation and service of load resources,further promote the digital transformation of load management,serve the construction of new power system,ensure the safe operation of power system,promote the consumption of renewable energy and improve the level of social energy efficiency.Precise control of support load and regular demand-side management help achieve the goal of "carbon peak and carbon neutrality".