A condition monitoring approach for machining process based on control chart pattern recognition with dynamically-sized observation windows. (April 2020)

• Record Type: Journal Article
• Authors: Lu, Zhiyuan; Wang, Meiqing; Dai, Wei
• Is Part Of: Computers & industrial engineering. Volume 142(2020)
• Publication Date: 2020-04
• Subjects: Abnormal machining condition monitoring -- Control chart pattern recognition -- Dynamically-sized observation windows -- Online data; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2020.106360
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: A CCPR model with dynamic size observation windows for online data is developed. The sensitive OWSs related to the abnormal machining conditions are selected. The monitoring model does not rely on cutting parameters but online data patterns. The proposed method has the excellent generalization ability to monitor abnormity. Abstract: In the manufacturing of metallic parts, the machining process is a critical factor for ensuring product quality. Machining condition monitoring is essential for the intelligent process. Existing machining condition monitoring approaches usually detect abnormal conditions for a fixed machining procedure, which is unrealistic and impractical for real practical applications. In this paper, a novel generalized machining condition monitoring approach based on control chart pattern recognition (CCPR) with dynamically-sized observation windows for online data is proposed. More precisely, two critical issues are addressed. First, the development of a CCPR model that handles patterns with stochastic sample size. Second, a procedure for selecting the window size for detecting abnormal machining conditions. An information fusion framework is implemented to assist the machining conditions monitoring by combining data from multiple sensors and multiple sized observation windows. Experiments are conducted to validate the feasibility of the proposed approach for two machining processes with the different cutting parameters. The results demonstrate the applicability of the proposed approach for conducting condition monitoring for machining process under different machining environments as needed in practice.