Condition Monitoring of Slow-speed Gear Wear using a Transmission Error-based Approach with Automated Feature Selection

Abstract

Gear flank changes caused by wear do not only affect the dynamic behavior of gear systems, but they can also compromise the load-carrying capacity of gear teeth up to critical failure. To help avoid unintended consequences like downtime or safety risks, a condition monitoring system needs to be able to estimate the current wear during operation based on available sensor measurements. While many condition monitoring approaches in research rely on vibrational analysis with manual feature engineering, gearboxes running at slow speed do not reveal much excitation information for this purpose. We therefore introduce an approach for slow-speed gear wear monitoring that is based on the dynamic gear transmission error and that contains an automated feature selection process. For this purpose, we extract a large set of features from the preprocessed transmission error samples. Applying combined filter and embedded feature selection methods enables us to automatically identify and remove features with low relevance. The selection process consists of filtering features with no statistical dependence on the target wear value, removing redundant features with a correlation analysis and a recursive feature elimination process with cross-validation based on a random forest regressor. The remaining relevant set of features is the basis for model training and subsequent wear estimation. For this, the present research employed two independent ensemble models, random forest regression and gradient boosted regression trees. To train and test the proposed approach, we conducted slow-speed gear experiments with developing gear wear on a single-stage spur gear test rig setup. The results of both models show good gear wear estimation performance compared to the actual wear mass loss, even for small quantities. Hence, the proposed transmission error-based approach with automated feature selection is able to quantify the degree of slow-speed wear and offers a possible way for condition monitoring and fault diagnosis.

Analytics-at-scale of Sensor Data for Digital Monitoring in Nuclear Plants (3^rd Annual Report)

“…Hybrid and ensemble methods integrate filters and wrappers alike, thereby benefiting from their complementary approaches [10], [14]. Three of the most common feature selection techniques-apart from the ones used in this report-are mutual information (MI) [15], recursive feature elimination [16], and analysis of variance (ANOVA) tests. MI uses entropy as a means of determining the amount of information gained by each input feature.…”

“…[15] used MI as the basis for a minimally redundant, maximally relevant feature selection method for multi-class support vector machine classification of railcar conditions. Recursive feature elimination recursively trains a model, calculates a cross-validation score, and then removes the least important feature, as determined via the internal feature ranking [16]. Feature importance ranking is common to methods such as random forest (RF).…”

“…The cross-validation score is used to determine at what point enough features have been selected to adequately describe the system. To estimate the wear on a gear transmission system, [16] used recursive feature elimination to reduce a set of 5,650 features down to the top 15. ANOVA is a statistical method that uses hypothesis testing to determine whether a given result or feature is significant [17].…”

Development of Short-Term Forecasting Models Using Plant Asset Data and Feature Selection

“…Other common feature selection techniques-apart from the ones used in this paper-include mutual information (MI) (Shahidi, Maraini, & Hopkins, 2020), recursive feature elimination (Sendlbeck, Fimpel, Siewerin, Otto, & Stahl, 2021), and analysis of variance (ANOVA) tests (Bechhoefer, Schlanbusch, & Waag, 2016). MI uses entropy as a means of determining the amount of information" gained by each input feature.…”

“…MI has been used as the basis for a minimally redundant, maximally relevant feature selection method for multi-class support vector machine classification of railcar conditions (Shahidi et al, 2020). Recursive feature elimination recursively trains a model, calculates a cross-validation score, and then removes the least important feature, as determined via the internal feature ranking (Sendlbeck et al, 2021). Feature importance ranking is common to methods such as random forest (RF).…”