Online Degradation Assessment of Shore Bridge Hoisting Gearbox Based on Improved Symbolic Sequence Entropy and Logistic Regression Model

Abstract

Aimed at the problem of degradation assessment of shore bridge hoisting gearbox, an online evaluation method of degradation state based on improved symbol sequence entropy (O_ISSE) and logistic regression is proposed. First, a threshold factor is introduced to retain the “coarse-grained” information of the signal change direction and amplitude, reduce the “sensitivity” of original algorithm to the impact component, and propose an improved symbol sequence entropy (ISSE).Then, the sliding window Weibull fitting method is used to effectively filter out the influence of fluctuations in the ISSE characteristic sequence to form O_ISSE. Finally, a logistic regression model is trained and established, and the H value of health factor of the unknown sample are calculated to realize its status recognition online.The example analysis of the life data of the hoisting gearbox of a dock in Shanghai is conducted. The results show that the ISSE and logistic regression model can describe the complexity of signal, track and identify performance degradation status accurately.

Keywords

• shore bridge
• degradation feature extraction
• symbol sequence entropy
• weibull distribution
• gearbox