Abstract
This paper presents a prognostic method for RUL (remaining useful life) prediction based on EMD (empirical mode decomposition)-ESN (echo state network). The combination method adopts EMD to decompose the multisensor time series into a bunch of IMFs (intrinsic mode functions), which are then predicted by ESNs, and the outputs of each ESN are summarized to obtain the final prediction value. The EMD can decompose the original data into simpler portions and during the decomposition process, much noise is filtered out and the subsequent prediction is much easier. The ESN is a relatively new type of RNN (recurrent neural network), which substitutes the hidden layers with a reservoir remaining unchanged during the training phase. The characteristic makes the training time of ESN is much shorter than traditional RNN. The proposed method is applied to the turbofan engine datasets and is compared with LSTM (Long Short-Term Memory) and ESN. Extensive experimental results show that the prediction performance and efficiency are much improved by the proposed method.
Highlights
Prognostics is an engineering discipline with regard to predictive diagnostics, including calculation of the remaining useful life based on the observed system condition [1]
The RUL prediction methods are typically classified as physics-model based approaches, data-driven approaches, and hybrid approaches, and data-driven methods can be divided into statistical model based approaches and AI approaches further [6, 7]
Data processing methods including GBDTfeature selection and Kalman filtering are adopted. e proposed method is verified through a case study on turbofan engines of aircraft. e relevant data of turbofan engines is collected from multiple sensors under variable operating conditions. e proposed method is compared with ESN and LSTM [19]. e results demonstrate the superiority of our proposed method. e contributions of this paper are as follows: (1) We introduce the idea of signal decomposition to multisensor data processing for RUL prediction. e adoption of EMD has a good effect on noise disposal, which provides a reference solution for reducing the noise of the RUL prediction problem
Summary
Prognostics is an engineering discipline with regard to predictive diagnostics, including calculation of the remaining useful life based on the observed system condition [1]. Approaches based on the physical model [8, 9] generate an explicit mathematical model of the degradation processes of machinery. Compared with physics-model based approaches, data-driven approaches are easier to be implemented. E methods can depict the uncertainty of the degradation process and its influence on RUL prediction [6]. They are not good at modeling for nonlinear systems, and AI approaches such as ANNs [12, 13], NF systems [14], and SVM [15] can deal with the issue. Hybrid approaches combine physics-model based approaches and data-driven approaches to take advantage of different approaches
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
