Design And Implementation Of Industrial Time Series Big Data Analysis System For Heat Treatment Equipment

Digital heat treatment process is an important part of industrial heat treatment process.With the improvement of hardware equipment in digital workshop and the development and promotion of information technology,a large number of data about heat treatment process or equipment are produced,most of which are time series data.Most of these time series data are closely related to the production process of heat treatment equipment.By analyzing this part of data,we can find the process information contained in it,and use the hidden information to serve the heat treatment production line,so as to speed up the production or reduce the cost of the enterprise.By classifying the data of heat treatment process,the influence of a certain factor on product quality can be judged.However,industrial time series big data has the characteristics of large amount of data,high data dimension and strong correlation of data points within the sequence.This brings great difficulties to the processing and analysis of industrial time series big data.Aiming at the problem that it is difficult to analyze industrial time series big data,in order to achieve the goal of effective classification of time series data in heat treatment,this paper mainly does the following work:(1)The feature representation algorithm of time series data is studied.Combined with the characteristics of heat treatment time series data,the piecewise linear representation algorithm is improved: the time series data is discretized into two time series data,and the key points of the discrete sequence are extracted by exclusion method.Experimental results show that the method can reduce the correlation degree between time series data,expand the range of key points,eliminate the interference of continuous local trend,and ensure the quality of key points extracted.(2)The similarity measurement algorithm of time series data is studied.According to the extracted key points,this paper studies the Euclidean distance measurement algorithm of time series data,and proposes the Euclidean distance measurement algorithm based on the weighted slope of key points.Experiments show that the algorithm gives full play to the value of key points,improves the measurement accuracy of Euclidean distance,and increases the classification accuracy of 1nn for time series data.(3)Based on the above research,combined with the characteristics of heat treatment process and the status of heat treatment data analysis and management system,this paper designs and develops an industrial time series big data analysis system for heat treatment equipment.On the basis of real-time data collection of heat treatment temperature,the system can analyze the heat treatment temperature series offline,judge the influence of temperature series on the quality of heat treatment products,assist quality inspectors to detect the quality of products,and reduce the workload of quality inspectors.At the same time,the system also provides temperature curves about equipment and products,which helps staff to query specific process.