Degradation index construction methodology for mechanical transmission based on fusion of multispectral oil data

Abstract

Purpose

The purpose of constructing a degradation index (DI) is to better characterize the degradation degree of mechanical transmission compared with relying solely on spectral oil data, which leads to an accurate estimation of the failure time when the transmission no longer fulfills its function.

Design/methodology/approach

The DI is modeled using a weighted average function with two desirable properties: maximizing the monotonic trend and minimizing the variance of failure threshold between different transmissions. The method includes concentration modification, data selection and data fusion steps that lead to a reasonable mechanical transmission degradation model. The proposed methodology was verified through a case study involving multispectral oil data sampled from several power-shift steering transmissions.

Findings

The results show that the DI outperforms all spectral oil data. Compared with the existing spectral oil data-based degradation modeling approach for mechanical transmissions, the present methodology provides an accurate RUL prediction.

Research limitations/implications

There are several important directions for future research: First, more degradation data (i.e. ferrography) that are tailored to the degradation modeling of mechanical transmission need to be involved. Second, more effective degradation data selection methodologies that are applicable for multiple data types need to be developed. Third, kernel methods that can fuse the nonlinear degradation data need to be investigated.

Originality/value

The novelty of this methodology lies in integrating the multiple degradation data in a unified DI. And the main contribution of this paper is to establish a new direction in degradation modeling and RUL prediction of mechanical transmission.

Citation

Yan, S., Ma, B. and Zheng, C. (2019), "Degradation index construction methodology for mechanical transmission based on fusion of multispectral oil data", Industrial Lubrication and Tribology, Vol. 71 No. 2, pp. 278-283. https://doi.org/10.1108/ILT-04-2018-0154