Spark-based Manufacturing IoT System Design

With the rapid development and application of information technology in traditional manufacturing,the demand for reasonable utilization of data resources is increasing.Discrete manufacturing equipment,as a source of data resources,plays an important role in the development direction of discrete manufacturing.Therefore,it is of great significance to study the Io T system based on discrete manufacturing equipment to realize the use and integration of data.This article focuses on the digital factory of Wuyang Textile Machinery Co.,Ltd.as the application scenario,and studies the machine tools and machining centers of FANUC-0i / 30 i and HEIDENHAIN i TNC530 system.First of all,according to the needs of the discrete manufacturing equipment Io T system for real-time data preview and query operations,a data acquisition scheme based on process data is proposed.By encapsulating the Focas system package and the DNCRemo Tools software package,the common data format parameter transmission method is used to solve the problems caused by the discrepancies of the equipment manufacturer models in the discrete manufacturing equipment.It solves the problem of uneven work in the data collection process and effectively improves the efficiency of data collection and data processing.Then,the storage technology for process data is discussed,and a storage strategy based on process data is proposed.By adopting HDFS distributed files as the storage of process data,using Spark Streaming components to complete data operations,using Hash Lru Cache and querying data structures as caches,the average response time of query operations is between0.5 and 2 seconds,which effectively solves problem that query operations in high QPS takes a long time.Finally,the related technologies of machine tool processing equipment fault prediction are analyzed,and a matrix completion model based on matrix decomposition and a machine condition prediction model based on time series characteristics are established.The matrix completion model was solved using rank estimation and ALS algorithm,and a graph of the loss function and the number of iterations was obtained through experiments.The accuracy rate of the model in the test set reached 90%.The equipment state prediction model mainly uses the LSTM network to obtain the accuracy-iteration generation curve through parameter optimization,and the accuracy rate on the test set reaches more than 90%.The research results of this paper provide a certain reference for the process data acquisition and optimization of operation status prediction of discrete manufacturing equipment.