Investigations on generalized Hjorth's parameters for machine performance degradation assessment

Created by W.Langdon from gp-bibliography.bib Revision:1.8051

@Article{YAN:2022:MSSP,

• author = "Tongtong Yan and Dong Wang2 and Tangbin Xia and Zhike Peng and Lifeng Xi",
• title = "Investigations on generalized {Hjorth}'s parameters for machine performance degradation assessment",
• journal = "Mechanical Systems and Signal Processing",
• year = "2022",
• volume = "168",
• pages = "108720",
• keywords = "genetic algorithms, genetic programming, Machine performance degradation assessment (MPDA), Health indicators, Hjorth parameters, Composite fitness function",
• ISSN = "0888-3270",
• URL = "https://www.sciencedirect.com/science/article/pii/S0888327021010396",
• DOI = "doi:10.1016/j.ymssp.2021.108720",
• abstract = "The strategies of machine performance degradation assessment (MPDA) are to detect the time of incipient fault initiation as early as possible and subsequently track machine deterioration evolution by extracting an effective health indicator so that timely maintenance strategies can be scheduled to avoid catastrophic accidents. Three Hjorth's parameters including Activity, Mobility, and Complexity have been studied to extract informative features from ectroencephalography signals, while their theoretical investigations for MPDA are not fully explored. In this paper, theoretical and experimental investigations on Hjorth's parameters for MPDA are conducted, which verifies that Activity is more suitable to MPDA than Mobility and Complexity. A new theorem of Hjorth's parameters for MPDA is accordingly proposed to show that Hjorth_1, Hjorth_2, Hjorth_3, and Activity have a same characteristic with complexity measures rather than sparsity measures for MPDA. After the characteristics of Hjorth's parameters for MPDA are thoroughly studied, generalized Hjorth's parameters are constructed and designed based on Genetic programming (GP). Firstly, Hjorth's parameters are reformulated as a unified mathematical framework. Aiming at solving the limitations of Hjorth's parameters, a new composite fitness function of GP is specifically designed to tailor for MPDA. Subsequently, generalized Hjorth's parameters are constructed by integrating the unified framework of Hjorth's parameters with GP. Three bearing and gear run-to-failure datasets are used to validate the effectiveness of the proposed generalized Hjorth's parameters. Results show that the newly constructed Hjorth's parameters are superior to the original Hjorth's parameters and some state of the art health indicators for MPDA",

}

Genetic Programming entries for Tongtong Yan Dong Wang2 Tangbin Xia Zhike Peng Lifeng Xi

Citations