An Adaptive Process Monitoring Method Based on an Improved ADALINE Neural Network

In the manufacturing of parts, the machining process is a critical factor for ensuring product quality. Process condition monitoring is essential for this process. Traditional monitoring approaches that usually use static thresholds are inefficient and lag behind process monitoring, which limits the regulation capability of process monitoring. In this paper, a new adaptive process monitoring method based on process state information is proposed. More precisely, two critical issues are addressed. First, adaptive control limits improve the dynamic monitoring ability. Second, designing monitoring strategies for quality trends improves the monitoring efficiency. In this paper, a framework for an adaptive process monitoring method is designed. First, the modulus maxima is calculated to improve the preprocessing method of wavelet threshold denoising. Second, an adaptive monitoring strategy is designed for different quality trends to improve the monitoring efficiency. Third, the real-time adaptive update of the control limits is realized for machining condition monitoring based on the artificial adaptive linear element (ADALINE) neural network. Finally, experiments are conducted to validate the feasibility of the proposed approach, and the results show that the proposed approach has good applicability as needed in practice.