Intelligent Data Analytics for Maintenance in Industry 4.0

Industry 4.0 or smart manufacturing/production is the new era of industry. It involves IoT, cloud computing and intelligent data analytics. A wide range of data is produced by production and assembly manufacturing lines that can be used for performance improvement by using Data analytics. Traditionally, several mathematical models are the base of most of the automated machines however, a small mistake in the formula can result in a huge material and human cost. Today these errors are found and fixed by a human. The machinery maintenance is manual, and calendar-based which can be performed more efficiently by data-driven machine health monitoring. Common and unwanted events can be diagnosed with data analytics. Therefore, to address these problems a decision-making system based on contextual information and data analytics has been developed. The problem with data driven approach is that the data obtained from the industry is imbalanced. Imbalanced data means it contains more samples in one class than the rest of the classes. For this reason, SMOTE (Synthetic Minority Over-Sampling Technique) was used for making the classes even. Missing data samples were removed as it does not increase efficiency. For increased efficiency of the model it is necessary to extract useful features from the data set. This Feature extraction was performed with the help of expert knowledge. The specific business case which was addressed in this work is the real-time prediction of shims Dimensions in power transfer Units (PTU) using Data analytics. The model trained with imbalanced data was biased in predicting samples as good. However, after using synthetic data the performance increased remarkably.