Comparative Study of Deep Learning Techniques for Breakdown Prediction in Pump Sensor Systems

In the current fourth industrial revolution, data plays a crucial role in providing insights about engineering systems. This paper studies artificial intelligence techniques for predicting the occurrence of anomalies, which are breakdowns in pump sensors. Among artificial intelligence techniques, deep learning methods are in high demand due to their ability to perform automated feature engineering. The techniques studied in this work belong to this class. They include the Long short-term memory (LSTM), Temporal Convolutional Network (TCN) and the Multi-Layer Perceptron(MLP).