Copper Tube Production Process Energy Consumption Prediction Model Based On GA-BP Network Research And Application

Copper tube production process energy is on-demand supply, if energy supply shortage or excess supply, will cause disruption and energy waste, influence the economic benefits of enterprises, accurately grasp the energy demand, secure energy supply and demand balance is an urgent problem need to solve for copper tube production enterprise. In addition, the presence of copper tube production process energy consumption is abnormal and difficult to find, often resulting in waste of energy.And to solve these problems is simply to achieve accurate prediction of energy consumption, particularly in the process, for complex machining parameters that affect energy consumption, accurate prediction of the corresponding energy consumption, by comparison with actual energy consumption can be effectively detected energy consumption is abnormal. Therefore, the energy demand forecast for the energy and abnormal events detected two problems characteristic data processing and predictive modeling, energy consumption estimation model is proposed based on GA-BP network, security of energy supply and demand balance, reducing energy waste and improve energy efficiency. Contents of this paper as follows:1. According to the copper tube production process, systematic analysis of the production process energy consumption structure, energy distribution and various types of energy-consuming step, reveals the smelting, milling,mill pull, the factors influencing the energy consumption of the stretching of the plate, forming the copper tube production process energy consumption influencing factors set.2. For energy demand forecasting problem, one kind of energy demand forecasting model is proposed based on GA-BP network. To combination historical sequence data and energy consumption factors data as input data, the prediction output includes historical consumption data sequence timing information, power consumption information and energy information factors, and genetic algorithm optimization BP neural network model, by which optimization model to improve prediction accuracy (average relative error is 1.53%).3. For the energy consumption anomaly detection problem, one kind of energy consumption anomaly detection model is proposed based on GA-BP network. Factors affecting energy consumption correlation analysis, the factors associated with a greater degree as the main factors energy consumption, and the use of principal component analysis of the raw data for dimension reduction, combined with multiple linear regression confidence interval set according to prediction errors the method will be extended forecast consumption value to a range, realized anomaly detection.Based on the above studies, the establishment of energy demand and energy forecast anomaly detection model library and using Java language to develop energy demand and energy forecast anomaly detection module, integrated into the energy management system by the module for application verification in the enterprise.