Industry Equipment Anomaly Detection Based On The Analysis Of Temporal Data Mining

Due to the complex, large-scale, automation of modern industrial process, makes the reliability of each link is more demanding, if a link to appear problem, so the whole system is likely to coll apse, therefore, fault diagnosis for industrial processes is becoming more and more important. And based on the data driven method is the most widely used a method of fault diagnosis field, which makes use of the data collected continuously detect the chan ge of system operation and fault information, do not need mathematical model of complex and accurate prior knowledge. Most industrial processes with nonlinear characteristics, using the conventional PCA method to fault detection and diagnosis of its accura cy is low, so this paper mainly aimed at the nonlinear process of fault diagnosis method based on data driven was studied, and its application in TE process and based on the cement kiln calcination stage. The main research content is as follows:First, this paper introduces the research status and development of fault diagnosis, and introduces the TE process and system of rotary kiln cement production equipment, a detailed analysis of the characteristics of their systems and data.Secondly, in view of the cement production line of actual operation of the design software of a production line for long time operation observation, collection and storage of the whole production line to produce a large number of real-time data, these data include the need analysis of the rotary kiln system fault information.Third, this paper introduces the data preprocessing to deal with the problem, the establishment of the actual industrial process monitoring model and monitoring effect is good or bad has a vital role. To princi pal component analysis and kernel principal component analysis at the same time elaborated the theory of knowledge, and to the above mentioned TE process as an object, compares the principal component analysis and kernel principal component analysis method of fault detection results, verify the kernel principal component analysis for nonlinear process fault detection is more effective and accurate.Finally, the kernel principal component analysis in fault diagnosis and identification of the effectiveness of the system in rotary kiln using kernel principal component analysis method to get good effect, show the effectiveness of it. By using matlab and VS joint programming technology, design and implement the operation monitoring system of rotary kiln.