Abstract:Deep learning is gaining attention in data-driven remaining useful life (RUL) prediction of machinery because of its powerful representation learning ability. With the help of deep learning techniques, the machine degradation information can be mined more sufficiently and some promising RUL prediction results have been achieved in several studies recently. These deep learning-based prognostics approaches, however, have the following weaknesses: 1) Their prediction performance largely depends on the hand-crafted feature design. 2) The correlations of different sensor data are not explicitly considered in representation learning. To overcome the above weaknesses, a new deep prognostics network named deep separable convolutional network (DSCN) is proposed in this paper for RUL prediction of machinery. In the proposed DSCN, the monitoring data acquired by different sensors are directly used as the inputs of the prognostics network. Then, a separable convolutional building block with a residual connection is built based on separable convolutions and squeeze and excitation operations. Through stacking multiple separable convolutional building blocks, the high-level representations are automatically learned from the input data. Finally, the RUL is estimated by feeding the learned representations into the fully-connected output layer. The proposed DSCN is validated using the vibration data from accelerated degradation tests of rolling element bearings and the public degradation simulation data of turbine engines, respectively. The experimental results show that the proposed DSCN is able to provide accurate RUL prediction results based on the raw multi-sensor data and is superior to some existing data-driven prognostics approaches. (C) 2019 Elsevier Ltd. All rights reserved.

Thermal Stress Deformation Prediction for Rotary Air-Preheater Rotor Using Deep Learning Approach.

Deep-Learning Strategy Based on Convolutional Neural Network for Wall Heat Flux Prediction

Aero-Engine Remaining Useful Life Prediction Via Physics-Informed Self-Attention Encoder

Soft sensor based on DBN-IPSO-SVR approach for rotor thermal deformation prediction of rotary air-preheater

Thermal deformation prediction based on the temperature distribution of the rotor in rotary air-preheater

Deep Learning-Based Approach for Heat Transfer Efficiency Prediction with Deep Feature Extraction.

Exhaust Temperature Prediction for Gas Turbine Performance Estimation by Using Deep Learning

Fast Reconstruction Method of the Stress Field for the Steam Turbine Rotor Based on Deep Fully Convolutional Network

Real-time power prediction approach for turbine using deep learning techniques

Deep learning method for fast prediction of film cooling performance

Numerical Analysis of the Coupling Between Heat Transfer and Deformation in Rotary Air Preheater

A prediction method of temperature distribution and thermal stress for the throttle turbine rotor and its application

Field Reconstruction and Off-Design Performance Prediction of Turbomachinery in Energy Systems Based on Deep Learning Techniques

Predictive Maintenance in Rotating Machinery Using Deep Learning Techniques

A thermal deformation prediction method for grinding machine’ spindle

Deep Separable Convolutional Network for Remaining Useful Life Prediction of Machinery

An attention-based architecture for predicting thermal stress on turbine blade film hole surface using partial temperature data

Fatigue life prediction of a supercritical steam turbine rotor based on neural networks

Knowledge enhanced ensemble method for remaining useful life prediction under variable working conditions

Remaining Useful Life Prediction Method for High Temperature Blades of Gas Turbines Based on 3D Reconstruction and Machine Learning Techniques

Machine learning-assisted effective thermal management of rotor-stator systems