Optimal filter design with progressive genetic algorithm for local damage detection in rolling bearings

• Published in: Mechanical systems and signal processing Vol. 102; pp. 102 - 116
• Main Authors: Wodecki, Jacek, Michalak, Anna, Zimroz, Radoslaw
• Format: Journal Article
• Language: English
• Published: Berlin Elsevier Ltd 01.03.2018; Elsevier BV
• Subjects: Bearings; Computer simulation; Damage detection; Decision making; Design; Diagnostic systems; Digital filter; Filter design (mathematics); Filters; Genetic algorithm; Genetic algorithms; Local damage; Machinery; Roller bearings; Rolling; Underground mining; Vibration; Digital filter; Vibration; Local damage; Genetic algorithm
• ISSN: 0888-3270; 1096-1216
• DOI: 10.1016/j.ymssp.2017.09.008

•Prefiltration is used in order to obtain signal with better diagnostic parameters.•Progressive genetic algorithm is proposed for design of dedicated optimal filters.•Method is tested against real and simulated data, both for good and damaged case.•Original concept of termination procedure is proposed.•Diagnostic decision making feature has been proposed for final result determinance. Harsh industrial conditions present in underground mining cause a lot of difficulties for local damage detection in heavy-duty machinery. For vibration signals one of the most intuitive approaches of obtaining signal with expected properties, such as clearly visible informative features, is prefiltration with appropriately prepared filter. Design of such filter is very broad field of research on its own. In this paper authors propose a novel approach to dedicated optimal filter design using progressive genetic algorithm. Presented method is fully data-driven and requires no prior knowledge of the signal. It has been tested against a set of real and simulated data. Effectiveness of operation has been proven for both healthy and damaged case. Termination criterion for evolution process was developed, and diagnostic decision making feature has been proposed for final result determinance.