Research And Application Of Abnormal Energy Consumption Detection Of Aluminum Extrusion Machine Based On Data Driven

Aluminum profile in the background of China's emerging market rapid development,especially in recent years the development is very rapid.Aluminum profile extruder is an important equipment in the production process of aluminum profile,its quality will directly affect the output and quality of products.In the actu al production,because of the large process force and concentrated load of extrusion equipment,it is easy to be affected by a variety of factors,resulting in abnormal energy consumption,which has an unpredictable impact on the production and product quality of enterprises,and has brought great economic losses to aluminum profile enterprises.In the investigation of the actual production environment of extrusion press,the sensor equipment due to the current and voltage instability,abnormal communication and other factors will lead to the loss of data at some collection points,affecting the accuracy of anomaly detection.In addition,the data features in the detection of abnormal energy consumption of extrusion press are not comprehensive and appropriate,resulting in low detection accuracy.Therefore,we need to do more in-depth research on missing data interpolation and abnormal energy consumption detection methods.In this paper,the aluminum extrusion press as the research object,based on the energy consumption data in the extrusion production process and the energy consumption data in the extrusion production process,the problem of missing data and abnormal detection of energy consumption in the extrusion production process are studied.The specific research contents of this paper are as follows:(1)In this paper,the extrusion process and energy consumption characteristics of the extrusion process are analyzed comprehensively.Firstly,the production system of aluminum profile extrusion machine is introduced,and the working principle and extrusion process of the extrusion machine are deeply analyzed.The energy consumption change and energy flow analysis in the extrusion process are further studied,and the periodicity of energy consumption in the extrusion process is revealed;The influencing factors of energy consumption in extrusion process were analyzed;Finally,the structure diagram of data interpolation and anomaly detection system of aluminum extrusion press is described,which provides a theoretical basis for the next research.(2)The current research methods can't effectively capture the nonlinear correlation between the energy consumption data attributes of extrusion press,and it is difficult to adapt to various forms of data missing mode interpolation,which leads to the problems of low interpolation accuracy and low efficiency.In view of the idea of deep learning,an automatic encoder model for missing data interpolation and denoising(MIDAE)based on DAE is proposed,and two kinds of MV interpolation are designed Methods,namely,MIDAE-Sequential and MIDAE-Batch,are used to better adapt to data with various missing patterns.The experimental results show that the proposed method not only can better apply to a variety of missing data patterns,but also has a high accuracy.(3)According to the missing data interpolation method in Chapter 3,a complete data set is obtained to avoid the influence of missing data on the accuracy of anomaly detection.Aiming at the problem that the data features in the detection of abnormal energy consumption of industrial extrusion press are not comprehensive and appropriate,which leads to low detection accuracy,an abnormal detection method based on GMM-LDA clustering feature learning algorithm and PSO-SVM model was proposed(GMM-SVM).First,the GMM-LDA clustering feature learning algorithm is used to obtain the optimal features of normal or abnormal data,and then combined with PSO algorithm to optimize SVM model to obtain the optimal model.The experimental results indicate that the anomaly detection method presented in this paper is more accurate than other methods.(4)Based on the above research,according to the existing energy management system of an aluminum profile enterprise,this paper designs and develops the ex trusion press data interpolation and abnormal energy consumption detection system with the method proposed in this paper as the theoretical support;through the realization of data interpolation and abnormal energy consumption function,it further illustra tes the feasibility of the method proposed in this paper in aluminum profile enterprises.