Intelligent early-warning quality control for electromagnets based on long short-term memory network with Bayesian optimisation and exponentially weighted moving average chart

The quality of electromagnet products is crucial for the smooth operation of mechanical equipment. However, the manufacturing of electromagnet products generates vast amounts of data, which means that real-time monitoring of product quality and ensuring its adherence to standards are challenging. Consequently, traditional quality-control processes for electromagnet product manufacturing lack real-time feedback on quality. To solve this problem, this paper proposes a novel and intelligent early-warning quality control method for an electromagnet manufacturing process. First, a framework for intelligent early-warning quality control is introduced. Second, an intelligent early-warning quality control algorithm is developed. This algorithm integrates a bidirectional long short-term memory network based on a Bayesian optimisation algorithm (BOA-BiLSTM) and an exponentially weighted moving average (EWMA) control chart method. Third, the implementation process of the early-warning quality control method is analysed using a round-tube electromagnet applied in an electromagnetic trip device. Finally, the practical feasibility of the early-warning quality control method is verified using a dataset of electromagnet product quality. The results demonstrate the effectiveness and accuracy of the BOA-BiLSTM-EWMA method for early-warning quality control in an electromagnet manufacturing process.