Embedded Files
Smart Machine & AI Tech. 2
Smart Machine & AI Tech. 2
Smart Assessment of Remaining Useful life (RUL)
Smart Assessment of Remaining Useful life (RUL)
Bearings are the most commonly used components in rotating machines and the ability to diagnose their faults and predict their remaining useful life (RUL) is critical for system maintenance. This paper proposes a smart system combined with a regression model to predict the RUL of bearings. The method converts the azimuth signal through low-pass filtering (LPF) and a chaotic mapping system, and uses Euclidean feature values (EFVs) to extract features in order to construct useful health indicators (HIs). In fault detection, the iterative cumulative moving average (ICMA) is used to smooth the HIs, and the Euclidean norm is used to find the time-to-start prediction (TSP). In terms of prediction, this paper uses a self-selective regression model to select the most suitable regression model to predict the RUL of the bearing. The dataset provided by the Center for Intelligent Maintenance Systems (IMS) is applied for performance evaluation; in comparison with previous research, better prediction results can be achieved by applying the proposed smart assessment system. The proposed system is also applied to the PRONOSTIA (also called FEMTO-ST) bearing dataset in this paper, demonstrating that acceptable prediction performance can be obtained.
To build up a smart system for data status prognosis via applying Chaotic Mapping Networks (CMN) combing with Convolution Neural Networks (CNN) is the mission, where the open database as well as the data of robot will be employed to further verify the performance of the developed system. The flowchart of the research plan in the first year is given in Fig. 3-2, it is clear that there are two main components in this flowchart, the first part should be Chaotic Mapping Networks (CMN), and the second component is the deep neural networks for computer vision, called Convolution Neural Networks (CNN). The optional output is decided to be time series data for CNN to further learn the image features.
Some Experiment Results
This tech. proposes a smart system integrated with a regression prediction system to predict the Remaining Useful Life (RUL) of ball bearings. The proposed method yields promising results in RUL prediction, and four key conclusions are drawn: (1) Smoothing the curve of Health Indicator (HI) through Low-Pass Filtering (LPF) and Improved Cumulative Moving Average (ICMA) enhances the accuracy of RUL prediction. (2) The Euclidean norm is an effective tracking value in capturing the degradation process of bearings. (3) Our method achieves an average Corrected R-squared (CRA) value above 90% in predicting Integrated Monitoring System (IMS) data. (4) The proposed method also yields satisfactory results when applied to FEMTO-ST experimental data. (5) We have incorporated the use of deep learning tools, including the Fractional Order Score-based Time Series Prediction (FROST) neural network, in the extension of our research. In addition, we aim to explore more appropriate TSP decisions and accuracy improvement techniques, while also applying a chaotic mapping strategy to develop key health indicators.
Related Research Achievements:
Shih-Yu Li, Ho-An Li, Lap-Mou Tam, Chin-Sheng Chen*, “A Smart System for an Assessment of the Remaining Useful Life of Ball Bearings by Applying Chaos-Based Health Indicators and a Self-Selective Regression Model”, Sensors, vol. 23, no. 3, pp. 1267-1284, 2023. (SCI, IF= 3.847, Rank: 12/136=8.82 %, Q1)
Page updated
Report abuse