Approach for fault prognosis using recurrent neural network
- PDF / 1,577,995 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 67 Downloads / 216 Views
Approach for fault prognosis using recurrent neural network Qianhui Wu1 · Keqin Ding2 · Biqing Huang1 Received: 2 January 2018 / Accepted: 4 June 2018 © Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract In general, fault prognosis research usually leads to the research of remaining useful life prediction and performance prediction (prediction of target feature), which can be regarded as a sequence learning problem. Considering the significant success achieved by the recurrent neural network in sequence learning problems such as precise timing, speech recognition, and so on, this paper proposes a novel approach for fault prognosis with the degradation sequence of equipment based on the recurrent neural network. Long short-term memory (LSTM) network is utilized due to its capability of learning long-term dependencies, which takes the concatenated feature and operation state indicator of the equipment as the input. Note that the indicator is a one-hot vector, and based on it, the remaining useful life can be estimated without any pre-defined threshold. The outputs of the LSTM networks are connected to a fully-connected layer to map the hidden state into the parameters of a Gaussian mixture model and a categorical distribution so that the predicted output sequence can be sampled from them. The performance of the proposed method is verified by the health monitoring data of aircraft turbofan engines. The result shows that the proposed approach is able to achieve significant performance whether in one-step prediction task, in long-term prediction task, or in remaining useful life prediction task. Keywords Fault prognosis · Recurrent neural network · Long short-term memory · Turbofan engine · Remaining useful life
Introduction With the development of science and technology, the scale of mechanical equipment and industrial system is gradually expanding. Large scale and complicated structure make equipments and systems more prone to go failure, which brings lots of maintenance costs, reduces production efficiency, and leads to safety accidents and casualties. Making optimal maintenance strategy becomes more and more crucial. In general, maintenance strategy can be categorized into three main types, namely corrective maintenance, preventive maintenance and predictive maintenance (Kothamasu et al.
B
Biqing Huang [email protected] Qianhui Wu [email protected] Keqin Ding [email protected]
1
Department of Automation, Tsinghua University, Beijing 100084, China
2
China Special Equipment Inspection and Research Institute, Beijing 100084, China
2006; Peng et al. 2010; Soh et al. 2012). Corrective maintenance is a fault-based strategy. Maintenance will only be performed after equipment failure occurs. Preventive maintenance is a time-based strategy where the equipment is maintained at a certain time interval, that is, regular maintenance. Predictive maintenance is condition-based. This strategy arranges maintenance activities in advance by monitoring the state of the target components or
Data Loading...
