Multivariate Time Series Classification of Sensor Data from an Industrial Drying Hopper: A Deep Learning Approach

In recent years, the advancement of industry 4.0 and smart manufacturing has made a large number of industrial process data attainable with the use of sensors installed in the machineries. This thesis proposes an experimental predictive maintenance framework for an industrial drying hopper so that it can detect any unusual event in the hopper which reduces the risk of erroneous fault diagnosis in the manufacturing shop floor. The experimental framework uses Deep Learning (DL) algorithms in order to classify Multivariate Time Series (MTS) data into two categories- failure or unusual events and regular events, thus formulating the problem as binary classification. As classification is a supervised learning technique, any DL algorithm needs labeled data for classification. Moreover, raw data extracted from the sensors contain missing values. Therefore, necessary preprocessing is performed to make it usable for DL algorithms and the dataset is self-labeled after defining two categories precisely. To tackle the imbalanced data issue, data balancing techniques like Ensemble Learning with undersampling and Synthetic Minority Oversampling Technique (SMOTE) are used. Moreover, along with DL algorithms like Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM), Machine Learning (ML) algorithms like Support Vector Machine (SVM), K Nearest Neighbor (KNN), etc. have also been used to perform a comparative analysis on the result obtained from these algorithms. The result shows that CNN is arguably the best algorithm for classifying this dataset into two categories and outperforms other traditional approaches as well as deep learning algorithms