Recent advancements in technologies such as in Big data and Internet of Things have made Predictive Maintenance (PdM) a key strategy to reduce unnecessary maintenance costs and improve product quality in the manufacturing sector. The premise of this paper is to implement and explore some of the most promising machine learning models for PdM, a Gradient Boosting (GB) model, and a Support Vector Machine (SVM) model. An innovative methodology for model training is proposed that aims to improve model performance while also allowing for continuous training. Furthermore, it is proposed an automatic hyperparameter tunning approach for the GB and SVM models. A synthetic dataset that reflects industrial machine data was used to validate the proposed methodology. The implemented models can achieve up to 0.92 recall and 94.55% accuracy, highlighting the effectiveness of the proposed methodology.
CITATION STYLE
Mota, B., Faria, P., & Ramos, C. (2023). Predictive Maintenance for Maintenance-Effective Manufacturing Using Machine Learning Approaches. In Lecture Notes in Networks and Systems (Vol. 531 LNNS, pp. 13–22). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-18050-7_2
Mendeley helps you to discover research relevant for your work.