Prediction of Remaining Useful Life based on Gramian Angular Field and Recurrent Neural Network

Predicting the Remaining Useful Life (RUL) using time-series data obtained from sensors can significantly enhance the reliability and safety of equipment. This is crucial for the health management tasks of in-service equipment. To fully extract the temporal and spatial features from sensor data, this paper proposes a RUL prediction method that integrates Gramian Angular Summation Fields (GASF), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN). Specifically, the sensor data is first converted into image data with temporal correlation through GASF. Subsequently, CNN is utilized for feature extraction from the image data. Based on this, an RNN model is established to predict the RUL by leveraging spatiotemporal characteristics. The superiority of the proposed RUL prediction method is finally validated using a turbofan engine dataset.