Design And Implementation Of Secondary Development Interface Based On Big Data Platform

CNC system is widely used in machinery,electronics,automatic control,terminals,robots and other fields,and it will produce a large amount of electronically controlled data internally during the process of production.These data are an important part of industrial big data.NC big data potentially reflects the status of CNC machine tools such as running status,machining program information,internal current status,workpiece processing quality,and efficiency.We can determine the status of CNC machine tools at the data level through data mining.Reasonable use of the data laws can help us improve the productivity of CNC machine tools and the quality of product.In view of the low value density of industrial big data,we first need to screen data value density of complete industrial big data,extract that part of high value density data,and conduct data mining and research on this basis.On the needs of the development and application of big data generated in CNC system,we design the "high-value" data extraction module by referring to the ETL model,and a data extraction method based on distributed parallel computing framework MapReduce is implemented.The streaming data access method was implemented based on the publish-subscribe mechanism of Redis from studying in the existing data acquisition and storage system of the industrial big data.The data calculation module is based on Spark's proprietary data type RDD.We realized the preprocessing of the NC big data by reasonably using of the RDD,and carried out feature value extraction.At the same time,based on SparkMLlib to achieve common machine learning algorithms and provides algorithm support for value mining of the big data from CNC.According to actual development requirements,data service package based on WebService technology is implemented.The encapsulated WebService interface is published to the Tomcat server to achieve the purpose of supporting cross-language and cross-platform calls for data services.For the requirements of the offline development environment,we built an offline development and debugging platform by using VMware virtualization technology and the integration of the basic services required for related application development.Based on the above research,the functional testing of the method mentioned in this article was conducted,and the data extraction method performance was also tested.The results show the effectiveness of the proposed method,which can provide tools for the development of machine tool big data based on these methods,and facilitate developers to develop big data in the field of digital control.