Production system performance prediction model based on manufacturing big data

Abstract

Existing production systems are short of real-time performance status of production process active perception, resulting in the production abnormal conditions processed lag, leading to the frequency problems of deviations in production tasks execution and planning. To address this problem, in this research, an advanced identification technology is extended to the manufacturing field to acquire the real-time performance data. Based on the sensed real-time manufacturing data, this paper presents a prediction method of production system performance by applying the Dynamic Bayesian Networks (DBN) theory and methods. It aims to achieve the prediction of the performance status of production system and potential anomalies, and to provide the important and abundant prediction information for real-time production control.