Research On Power Retailer's Purchasing And Sales Strategy Based On Customer Classification By Big Data

In the initial stage of the new round of power system reform,China has released the electricity distribution and sales business to social capital in an orderly manner,and promoted the sales business into an effective competition pattern gradually.As the most active part of this reform and the most closely related part to power consumers,the electricity sales market has become the most critical part of the field.With the gradual opening up of the market,consumer market power has gradually increased,and the purchasing pattern has begun to diversify.As a key player in the electricity sales market,the retailer's performance is not only related to its own interest,but also directly affects the reform effect of the electricity sales market.At the current stage,there is almost only load contract trading between power sales retailers and consumers,but it cannot be ignored that the market construction of value-added services has always been the long-term goal of this round of power-sales reform.How to integrate customer demand and develop diversified business for increasing market share is the key to survival and profitability retailers in the fierce market competition.Therefore,this paper aims to rely on the application of artificial intelligence technology in the power system with big data,and study how to guide the retailers to comprehensively and reasonably evaluate their customers with multi-value factor system,how to decide real-time electricity price contracts which takes demand response and risk assessments into account,and how to accurately formulate corresponding value-added service strategies based on the customer classification.This paper summarizes the relevant experience in the development process of foreign power market and the study achievements of current electricity purchasing and sales strategy for retailers domestic and abroad.Combined with the application and development of big data and artificial intelligence in the field of power system,a load forecasting model based on deep learning and clustering algorithm is proposed.The application of fuzzy mean value in recurrent neural network is proposed.Furthermore,by studying the current status of power user value assessment,combined with the novel consumer characteristics under the new power reform in China,this paper integrates a novel electricity consumer value evaluation model with consumer's consumption level,load level,credit level,customer viscosity level,demand response level and environmental-friendly level.In this part,the multi-value grading model is established,and the optimized support vector machine classification model is used to realize a reasonable classification of power consumers.And then,some typical power-using portraits are drawn including the levels above,which provide a more operational classification standard and pattern for retailers to quality their customer resources.In addition,considering that after the new reform released,power consumers begin having rights to choose retailers according to their own energy demand and the sales price.So the retailers will finally consider multi-type contracts with their consumers including the fixed price,time-of-use price and real-time price.Therefore,this study establishes a bi-level optimization model for electricity retailers that take demand response and customer satisfaction into account,and obtains an optimal real-time price with different users participating in demand response projects.Finally,in the aspect of value-added services,a differentiated service model based on customer classification is proposed,which prompts users to change energy use,achieve energy substitution,improve environmental protection level,and provides a direction for retailers to expand their emerging business.The above research can improve power retailer's service awareness and the service quality in the situation where the electricity sales market is gradually opening up and developing fast.The customer classification method with consumer portrait and the price decision method with add-value services will strive retailers for greater profit and better development space.