Automated fault diagnosis algorithms based on vibration sensor recordings play an important role in determining the state of health of the machines. Data-driven approaches demand a large amount of labelled data to build reliable models. The performance of such lab-trained models degrades when deployed in practical use cases in the presence of distinct distribution target domain datasets. In this work, we present a novel deep transfer learning strategy that fine-tunes the trainable parameters of the lower (convolutional) layers with respect to the changing target domain datasets and transfers the parameters of the deeper (dense) layers from the source domain for efficient domain generalisation and fault classification. The performance of this strategy is evaluated by considering two different target domain datasets and studying the sensitivity of fine-tuning individual layers in the networks using time-frequency representations of the vibration signals (scalograms) as inputs. We observe that the proposed transfer learning strategy yields near-perfect accuracy, even for use cases where low-precision sensors are used for data collection and unlabelled run-to-failure data with a limited number of training samples.
Despite the diverse number of machine learning algorithms reported in the literature for machine fault detection, their implementation is mainly confined to laboratory-scale demonstrations. The complexity and black-box nature of machine learning models, the processing cost involved in appropriate feature extraction, limited access to labeled data, and varying operating conditions are some of the key reasons that curtail their implementation in practical applications. Furthermore, most such models serve as decision support tools, aiding domain experts in root cause analysis, and are not truly autonomous by themselves. To address these challenges, we present a lightweight autoencoder-based unsupervised learning framework to accurately identify machine faults against the changing operating conditions in a real-world scenario. The fault detection strategy is further strengthened by a model agnostic Shapley Additive exPlanations (SHAP)-based method (kernel SHAP) for identifying the most prominent features contributing to fault detection inference, the findings of which are then explored for identifying trends and correlations among prominent features and various types of faults. The framework is validated using two widely used and publicly available datasets for machine condition monitoring, as well as a large industrial dataset comprising 18 machines installed at 3 factories in India, monitored for several months.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.