Study On Big-data Analysis Platform Technology For Manufacturing

As the big data technology becomes one of the standard technologies of the industry 4.0,manufacturing enterprises enter the era of intelligent manufacturing,which lead to increasing high demand of production and administration based on data.In the manufacturing,the IOT-based data acquisition and control module has been widely applied,which forms a very large data source.It is difficult for conventional database technology to capture,store,manage,and analyze the large-scale data set tasks.The big data technology provides enterprises with efficient data analysis capability.It is able to improve the ability of management and decision-making for quality control,craft improving and service upgrade.To solve the problem that manufacturing enterprise is unavailable to get valuable information from history database through the traditional data analysis method,this paper puts forward a solution of big data analysis method and analyze some cases.First of all,the infrastructure of the platform is introduced in this paper.Then,core function modules of big data platform is designed and implemented in detail,such as data acquisition module,data analysis module,platform visualization module and so on.Especially,taking KNN as an example,a distributed parallel implementation is given to analysis module aiming at operation features of large data platform.Aiming at the specific business requirements of a manufacturing enterprise,the design and implementation of data preprocessing,feature extraction,clustering model and algorithm is given.The general process of enterprise business data analysis case of tightening craft is described in detail.In the process,an improved clustering analysis algorithm for large-scale unbalanced data is proposed,which is named Second Time K-means(STKmeans).The results of enterprise data analysis show that the clustering effect is much better than the traditional clustering method.In addition,the practicability of the platform is verified.