Research On Thermal Characteristics Of Thermal Users Based On Big Data Analysis

Due to the development of computer technology,most thermal power companies have various network systems that operate smoothly and play a key role in operation and management.However,due to the long-term operation of the information system,there are also some problems in the internal structure of thermal power enterprises.How to reasonably retain the historical statistics of various operation and management systems such as industry,marketing and information processing,and how to scientifically study and reasonably divide the accumulated massive social and historical statistics,so as to better provide services for enterprise operation and management has become a thorny problem faced by thermal power enterprises.In order to solve the above problems,combined with big data analysis technology,this paper analyzes the heat consumption characteristics of heat users according to the whole process of data mining of thermal system.Taking a large number of historical data of a thermal power plant as an example,different prediction models are used to predict the steam demand in the next 24 hours,and the prediction effects of each model are compared and analyzed,so as to provide a reliable basis for steam output and effectively reduce energy waste,It provides an important guarantee for the stable operation and production of urban heat supply network.The main research work and conclusions are as follows:1.According to the actual historical statistical data accumulated by thermal power enterprises,integrate the required production data,and process and analyze various data according to the big data processing process.The data set mainly involves the historical time series data of the total daily steam consumption of each heat user and the total steam sold by the thermal power station,as well as various weather parameter information.In the preparation of big data analysis,statistical analysis,normalization analysis,descriptive statistical analysis and correlation analysis of missing and abnormal values of data are also carried out,which provides an accurate,comprehensive and unified big data analysis set for the construction of steam load forecasting model of each heat user.2.Based on the selected evaluation indexes,the total steam sales of the thermal power plant in the next 24 hours and the steam required by two heat users are predicted through the traditional BP neural network algorithm,and the heat consumption habits and steam load trend of heat users are analyzed.At the same time,according to the characteristics of BP neural algorithm and prediction performance evaluation index,its advantages and disadvantages are analyzed,and the corresponding optimization algorithm is proposed according to the shortcomings of BP neural network.3.Aiming at the shortcomings of BP neural network,the model is optimized by using wavelet theory,and the wavelet neural network prediction algorithm is constructed to predict and analyze the heat consumption trend of the thermal power plant again.In addition,this paper also uses different optimization methods to construct mea-bp neural network model and PSO-BP neural network model to predict the total amount of steam required in each thermal system every day.The prediction effects of the above three optimization models are compared and analyzed.4.Through the analysis of the prediction results of various models,it can be seen that the optimized algorithm model has obvious advantages in both prediction accuracy and convergence speed;The prediction accuracy of mea-bp model is closer to the real thermal characteristics of users.