An official website of the United States government
Failure prognostics is the process of predicting the remaining useful life (RUL) of machine components, which is vital for the predictive maintenance of industrial machinery. This paper presents a new deep learning approach for failure prognostics of rolling element bearings based on a Long Short-Term Memory (LSTM) predictor trained simultaneously within a Generative Adversarial Network (GAN) architecture. The LSTM predictor takes the current and past observations of a well-defined health index as an input, uses those to forecast the future degradation trajectory, and then derives the RUL. Our proposed approach has three unique features: (1) Defining the bearing failure threshold by adopting an International Organization of Standardization (ISO) standard, making the approach industry-relevant; (2) Employing a GAN-based data augmentation technique to improve the accuracy and robustness of RUL prediction in cases where the deep learning model has access to only a small amount of training data; (3) Integrating the training process of the LSTM predictor within the GAN architecture. A joint training approach is utilized to ensure that the LSTM predictor model learns both the original and artificially generated data to capture the degradation trajectories. We utilize a publicly available accelerated run-to-failure dataset of rolling element bearings to assess the performance of the proposed approach. Results of a five-fold cross-validation study show that the integration of the LSTM predictor with GAN helps to decrease the average RUL prediction error by 29% over a simple LSTM model without GAN implementation.
more »
« less
An Adversarial Learning Approach for Machine Prognostic Health Management
Achieving accurate remaining useful life (RUL) prediction for prognostic and health management (PHM) depends upon sufficient prior degradation apprehension of critical components within the system. However, such prior knowledge is not always readily available in practice. We alleviate this shortcoming by proposing a novel data-driven framework that is capable of providing accurate RUL prediction without the need for any prior failure threshold knowledge. Correlative and monotonic metrics are utilized to identify critical features throughout the degradation progress. Subsequently, we append one-hot health state indicators to extracted degrading features, which are utilized together as adversarial training data for a Long Short-Term Memory (LSTM) network-based model. Finally, we utilize a fully connected layer to project the LSTM outputs into the parameters of a Gaussian mixture model (GMM) in conjunction with a categorical distribution, from which the long-term degradation progress is sampled. We verify the performance of the proposed framework using aeroengine health data simulated by Modular Aero-Propulsion System Simulation (MAPSS), and the results demonstrate that significant performance improvement can be achieved for long-term degradation progress and RUL prediction tasks.
more »
« less
- Award ID(s):
- 1809164
- PAR ID:
- 10105853
- Date Published:
- Journal Name:
- 2019 International Conference on High Performance Big Data and Intelligent Systems (HPBD&IS)
- Page Range / eLocation ID:
- 163 to 168
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Wang, Dong (Ed.)Lithium-ion batteries have been extensively used to power portable electronics, electric vehicles, and unmanned aerial vehicles over the past decade. Aging decreases the capacity of Lithium-ion batteries. Therefore, accurate remaining useful life (RUL) prediction is critical to the reliability, safety, and efficiency of the Lithium-ion battery-powered systems. However, battery aging is a complex electrochemical process affected by internal aging mechanisms and operating conditions (e.g., cycle time, environmental temperature, and loading condition). In this paper, a physics-informed machine learning method is proposed to model the degradation trend and predict the RUL of Lithium-ion batteries while accounting for battery health and operating conditions. The proposed physics-informed long short-term memory (PI-LSTM) model combines a physics-based calendar and cycle aging (CCA) model with an LSTM layer. The CCA model measures the aging effect of Lithium-ion batteries by combining five operating stress factor models. The PI-LSTM uses an LSTM layer to learn the relationship between the degradation trend determined by the CCA model and the online monitoring data of different cycles (i.e., voltage, current, and cell temperature). After the degradation pattern of a battery is estimated by the PI-LSTM model, another LSTM model is then used to predict the future degradation and remaining useful life (RUL) of the battery by learning the degradation trend estimated by the PI-LSTM model. Monitoring data of eleven Lithium-ion batteries under different operating conditions was used to demonstrate the proposed method. Experimental results have shown that the proposed method can accurately model the degradation behavior as well as predict the RUL of Lithium-ion batteries under different operating conditions.more » « less
-
Digital twin is a vital enabling technology for smart manufacturing in the era of Industry 4.0. Digital twin effectively replicates its physical asset enabling easy visualization, smart decision-making and cognitive capability in the system. In this paper, a framework of dynamic data driven digital twin for complex engineering products was proposed. To illustrate the proposed framework, an example of health management on aircraft engines was studied. This framework models the digital twin by extracting information from the various sensors and Industry Internet of Things (IIoT) monitoring the remaining useful life (RUL) of an engine in both cyber and physical domains. Then, with sensor measurements selected from linear degradation models, a long short-term memory (LSTM) neural network is proposed to dynamically update the digital twin, which can estimate the most up-to-date RUL of the physical aircraft engine. Through comparison with other machine learning algorithms, including similarity based linear regression and feed forward neural network, on RUL modelling, this LSTM based dynamical data driven digital twin provides a promising tool to accurately replicate the health status of aircraft engines. This digital twin based RUL technique can also be extended for health management and remote operation of manufacturing systems.more » « less
-
Electric vehicles (EVs) are spreading rapidly in the market due to their better responsiveness and environmental friendliness. An accurate diagnosis of EV battery status from operational data is necessary to ensure reliability, minimize maintenance costs, and improve sustainability. This paper presents a deep learning approach based on the long short-term memory network (LSTM) to estimate the state of health (SOH) and degradation of lithium-ion batteries for electric vehicles without prior knowledge of the complex degradation mechanisms. Our results are demonstrated on the open-source NASA Randomized Battery Usage Dataset with batteries aging under changing operating conditions. The randomized discharge data can better represent practical battery usage. The study provides additional end-of-use suggestions, including continued use, remanufacturing/repurposing, recycling, and disposal; for battery management dependent on the predicted battery status. The suggested replacement point is proposed to avoid a sharp degradation phase of the battery to prevent a significant loss of active material on the electrodes. This facilitates the remanufacturing/repurposing process for the replaced battery, thereby extending the battery's life for secondary use at a lower cost. The prediction model provides a tool for customers and the battery second use industry to handle their EV battery properly to get the best economy and system reliability compromise.more » « less
-
null (Ed.)Abstract: Modeling student learning processes is highly complex since it is influenced by many factors such as motivation and learning habits. The high volume of features and tools provided by computer-based learning environments confounds the task of tracking student knowledge even further. Deep Learning models such as Long-Short Term Memory (LSTMs) and classic Markovian models such as Bayesian Knowledge Tracing (BKT) have been successfully applied for student modeling. However, much of this prior work is designed to handle sequences of events with discrete timesteps, rather than considering the continuous aspect of time. Given that time elapsed between successive elements in a student’s trajectory can vary from seconds to days, we applied a Timeaware LSTM (T-LSTM) to model the dynamics of student knowledge state in continuous time. We investigate the effectiveness of T-LSTM on two domains with very different characteristics. One involves an open-ended programming environment where students can self-pace their progress and T-LSTM is compared against LSTM, Recent Temporal Pattern Mining, and the classic Logistic Regression (LR) on the early prediction of student success; the other involves a classic tutor-driven intelligent tutoring system where the tutor scaffolds the student learning step by step and T-LSTM is compared with LSTM, LR, and BKT on the early prediction of student learning gains. Our results show that TLSTM significantly outperforms the other methods on the self-paced, open-ended programming environment; while on the tutor-driven ITS, it ties with LSTM and outperforms both LR and BKT. In other words, while time-irregularity exists in both datasets, T-LSTM works significantly better than other student models when the pace is driven by students. On the other hand, when such irregularity results from the tutor, T-LSTM was not superior to other models but its performance was not hurt either.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/HPBDIS.2019.8735480
