Research On Application Of Data Analytics Based Kalman Filtering Algorithm To Energy Diagnosis In Iron And Steel Industry

Iron and steel enterprises are energy-intensive industries,with the increasing of energy exhaustion,the research on iron and steel energy has a very important theoretical and practical significance in energy-saving and emission reduction.Due to the complexity of the iron and steel production process,it is very difficult to establish the mechanism model.And,the high temperature and high heat production environment have caused a lot of noise in the measured value.These problems have brought difficulties to the energy management in steel production.In this thesis,the energy diagnosis problem is extracted from real-world energy management of iron and steel enterprises.The purpose of energy diagnosis is to analyze the energy consumption of each process and each product based on the statistical energy data,and to identify the abnormal situation of energy consumption,so as to provide effective data support for operators analyze energy consumption and optimize the entire energy system to achieve energy-saving purpose.At present,the research on the energy diagnosis in iron and steel enterprises mostly stays at the stage of qualitative analysis.In this thesis,based on the real-world production data of iron and steel enterprises,energy consumption model is established and corresponding evaluation criteria for energy diagnosis is proposed.The main research works are as follows:(1)The problem of diagnosing energy medium consumption is proposed based on the energy system in iron and steel enterprises.Then,by studying the steel production process and analyzing the measured data collected from practical energy management system,it is found that it is difficult to establish the mechanism model of the energy consumption process in steel production,and that the real-time measurement data contains a lot of noise.Based on the characteristics of these problems,a data-driven energy consumption relationship model is established.(2)Based on the analysis of the characteristics of various statistical models,the LSSVM model is chosen as the data modeling method.The nonlinear fitting ability of LSSVM model can express the consumption relationship between steel energy medium,and its control of noise can weaken the error caused by measurement noise to a certain extent.Then,an intelligent optimization strategy is designed to optimize the parameters of the steel medium diagnosis problem.Finally,a large number of data experiments are used to verify the effectiveness of the model in the iron and steel energy diagnosis.(3)Combining the energy consumption model based on LSSVM with the traditional Kalman filter algorithm,a Kalman filter model based on LSSVM is established.Based on the selected standard data,the diagnostic algorithm uses the statistical results of standard data filtering to formulate the evaluation criteria.By comparing the difference between the real measurement data and the evaluation criteria,the degree of anomaly of the current system is given.Finally,the algorithm is applied to the diagnosis of the data of six typical processes of iron and steel enterprises,and the diagnostic effect of the model is analyzed and verified.(4)A power information management system is developed,and using the data collected by the system and data analytic technique,the economic situation and future economic trend of the local area in recent years is analyzed.In this process,the diagnosis algorithm proposed in this thesis is applied to the data analysis to verify the effectiveness of the algorithm.