Linear friction weld process monitoring of fixture cassette deformations using empirical mode decomposition. (October 2015)

• Record Type: Journal Article
• Authors: Bakker, O.J.; Gibson, C.; Wilson, P.; Lohse, N.; Popov, A.A.
• Is Part Of: Mechanical systems and signal processing. Volume 62/63(2015)
• Page Start: 395
• Page End: 414
• Publication Date: 2015-10
• Subjects: Linear friction welding -- Process condition monitoring -- Hilbert–Huang transform -- Time-frequency analysis -- Empirical mode decomposition -- Non-stationary signal; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2015.02.005
• Languages: English
• ISSNs: 0888-3270

Abstract

Abstract: Due to its inherent advantages, linear friction welding is a solid-state joining process of increasing importance to the aerospace, automotive, medical and power generation equipment industries. Tangential oscillations and forge stroke during the burn-off phase of the joining process introduce essential dynamic forces, which can also be detrimental to the welding process. Since burn-off is a critical phase in the manufacturing stage, process monitoring is fundamental for quality and stability control purposes. This study aims to improve workholding stability through the analysis of fixture cassette deformations. Methods and procedures for process monitoring are developed and implemented in a fail-or-pass assessment system for fixture cassette deformations during the burn-off phase. Additionally, the de-noised signals are compared to results from previous production runs. The observed deformations as a consequence of the forces acting on the fixture cassette are measured directly during the welding process. Data on the linear friction-welding machine are acquired and de-noised using empirical mode decomposition, before the burn-off phase is extracted. This approach enables a direct, objective comparison of the signal features with trends from previous successful welds. The capacity of the whole process monitoring system is validated and demonstrated through the analysis of a large number of signals obtained from welding experiments. Abstract : Highlights: A cassette force monitoring system for a linear friction welding machine is proposed. A significant part of the process monitoring is the analysis of the trends in forces. For best possible accuracy, data is de-noised before other data processing happens. For robust data processing, the trend-finding methodology uses IMFs as filter bank. An automated adaptive process IMF selection strategy for filterbank is established.