A method for editing multi-axis load spectrums based on the wavelet transforms

• Published in: Measurement : journal of the International Measurement Confederation Vol. 162; p. 107903
• Main Authors: Shangguan, Wen-Bin, Zheng, Guo-Feng, Rakheja, Subhash, Yin, Zhihong
• Format: Journal Article
• Language: English
• Published: London Elsevier Ltd 01.10.2020; Elsevier Science Ltd
• Subjects: Accelerated tests; Algorithms; Down-sampling; Durability; Durability tests; Editing; Envelope damage identification method; Materials durability; Mechanical components; Mechanical properties; Multi-axis load spectrum editing; Sampling; Spectrum analysis; Suspension systems; Time domain analysis; Valleys; Wavelet transform; Wavelet transforms; Down-sampling; Multi-axis load spectrum editing; Durability tests; Wavelet transform; Envelope damage identification method
• ISSN: 0263-2241; 1873-412X
• DOI: 10.1016/j.measurement.2020.107903

•A wavelet based spectral editing method is proposed for multi-axis load spectra.•A modified down-sampling algorithm is proposed for pre-processing of load spectrums.•The edited load spectrum revealed time duration reduction of around 70%. A wavelet transform-based spectrum editing method is proposed for deriving effective load spectrums for accelerated durability testing of mechanical components. A modified down-sampling algorithm is also proposed for pre-processing of a measured load spectrum for preserving peaks and valleys in the original spectrum by reassigning the neglected peak/valley samples. The effectiveness of down-sampling and wavelet transform based spectrum editing methods is demonstrated considering multi-axis load spectrum acquired for the handling bushing of an automotive suspension control arm. The results showed with identical pseudo damage values, while the conventional method revealed relative error of 0.4% and 3.3% for down sampling ratios of 2 and 4, respectively. The edited load spectrum obtained from the proposed wavelet transform-based method revealed time duration reduction ratio of about 68%, which was considerably lower than 75% obtained using time-domain editing method. The results also show very good agreements between the proposed spectrum and time-domain editing methods, while the proposed method is considered to be more efficient for accelerated durability tests.