The research on noninvasive incipient fault diagnosis of power converters is very critical to avoid strenuous periodic check-ups and costly interruptions. Thermal cycling is one of the main ...techniques to accelerate the package-related failure progress. In this paper, first, a custom designed accelerated aging platform that can expose multiple discrete power MOSFETs to thermal stress simultaneously is introduced. Based on the collected experimental data, the variation of the on-state resistance is identified as the failure precursor, and an exponential degradation model that fits successfully with the experimental data are developed. The remaining useful lifetime (RUL) of degraded power MOSFETs is estimated through classical least-squares algorithm run on experimental data filtered by Kalman Filter which deals with the measurement noise and model uncertainties. The essential advantage of the proposed method is that it does not require junction temperature information. The RUL estimation with limited field data is demonstrated on a number of experimental results.
•Recurrent Neural Network for domain adaptation of remaining useful life predictions.•Domains are composed of data with different fault modes and operating conditions.•Only sensor information from ...the target domain is required.•Experiments investigate different combinations of source and target data.•Improved prognostics results in comparison to non-adapted methods on aero-engine data.
In Prognostics and Health Management (PHM) sufficient prior observed degradation data is usually critical for Remaining Useful Lifetime (RUL) prediction. Most previous data-driven methods assume that training (source) and testing (target) condition monitoring data have similar distributions. However, due to different operating conditions, fault modes and noise, distribution and feature shift exist across different domains. This shift reduces the performance of predictive models when no target observed run-to-failure data is available. To address this issue, this paper proposes a new data-driven approach for domain adaptation in prognostics using Long Short-Term Neural Networks (LSTM). We use a Domain Adversarial Neural Network (DANN) approach to adapt remaining useful life estimates to a target domain containing only sensor information. We analyse our approach using the NASA Commercial Modular Aero-Propulsion System Simulation (C-MAPPS). The results show that the proposed method can provide more reliable RUL predictions than models trained only on source data for varying operating conditions and fault modes.
Prognostic and health management of lithium batteries is a multi-faceted approach that can provide crucial indexes for guaranteeing the reliability and safety of the energy storage system. Herein, a ...novel multi-time-scale framework is proposed that focuses on short-term battery state of health estimation and long-term remaining useful lifetime prediction. The proposed method extracts four significant features through in-depth analysis of partial incremental capacity and Gaussian process regression with nonlinear regression is applied to forecasting battery health conditions. First, the advanced signal filter methods are employed to smooth initial incremental capacity curves. After that, the significant feature variables are extracted from different degrees such as intercept, slope and peak by linear fitting the partial incremental capacity curves. Second, the significant feature variables feed to Gaussian process regression to establish a short-term battery degradation model using kernel-modified Gaussian process regression. Third, an autoregressive long-term battery prediction model is established by combining the offline short-term battery model with nonlinear regression. The predictive capability, robustness and effectiveness of proposed methods are verified using four datasets with different cycling test conditions and health levels. The results show that the proposed method can give accurate battery health conditions forecasting.
•Multi-timescale framework is established for forecasting battery SOH and RUL.•The health features are extracted from partial IC curves from different dimensions.•A nonlinear regression RUL model is developed by using the GPR-based model.•Four batteries are used to verify and evaluate the proposed method.
A significant number of wind turbines will reach the end of their planned service life in the near future. A decision on lifetime extension is complex and experiences to date are limited. This review ...presents the current state-of-the-art for lifetime extension of onshore wind turbines in Germany, Spain, Denmark, and the UK. Information was gathered through a literature review and 24 guideline-based interviews with key market players. Technical, economic and legal aspects are discussed. Results indicate that end-of-life solutions will develop a significant market over the next five years. The application of updated load simulation and inspections for technical lifetime extension assessment differs between countries. A major concern is the uncertainty about future electricity spot market prices, which determine if lifetime extension is economically feasible.
The research cores of RUL prediction with machine learning algorithms are analyzed to find future improvement directions. And using RUL prediction results to extend battery lifetime is also explored.
...Display omitted
Lithium-ion batteries are the most widely used energy storage devices, for which the accurate prediction of the remaining useful life (RUL) is crucial to their reliable operation and accident prevention. This work thoroughly investigates the developmental trend of RUL prediction with machine learning (ML) algorithms based on the objective screening and statistics of related papers over the past decade to analyze the research core and find future improvement directions. The possibility of extending lithium-ion battery lifetime using RUL prediction results is also explored in this paper. The ten most used ML algorithms for RUL prediction are first identified in 380 relevant papers. Then the general flow of RUL prediction and an in-depth introduction to the four most used signal pre-processing techniques in RUL prediction are presented. The research core of common ML algorithms is given first time in a uniform format in chronological order. The algorithms are also compared from aspects of accuracy and characteristics comprehensively, and the novel and general improvement directions or opportunities including improvement in early prediction, local regeneration modeling, physical information fusion, generalized transfer learning, and hardware implementation are further outlooked. Finally, the methods of battery lifetime extension are summarized, and the feasibility of using RUL as an indicator for extending battery lifetime is outlooked. Battery lifetime can be extended by optimizing the charging profile serval times according to the accurate RUL prediction results online in the future. This paper aims to give inspiration to the future improvement of ML algorithms in battery RUL prediction and lifetime extension strategy.
Predictive maintenance (PdM) plays an important role in industrial manufacturing. One of the most fundamental ideas underlying many PdM solutions is to estimate Remaining Useful Life (RUL) of ...machines. Recently, advanced deep learning models like convolutional neural network (CNN) and long short-term memory (LSTM) have been widely used for RUL prediction. However, these models also have certain limitations because of the difficulty in dealing with long-term dependencies in time series data. In this study, we propose a novel model based on transformer networks to overcome this difficulty. Rather than using the full structure of a transformer model, we exploit only the encoder combined with a linear layer. The Bayesian Optimization algorithm is applied to find optimal hyperparameters for the encoder. Experiments on widely used turbofan engine datasets show that our proposed method significantly outperforms the state-of-the-art RUL prediction methods by up to 25% in terms of predicting remaining usable life. We also provide a solution for the problem of preserving the privacy and security of data in smart manufacturing by designing a Federated Learning-based architecture for RUL using the proposed transformer-based model.
•A novel transformer-based model with impressive efficiency for RUL prediction.•A Bayesian Optimization to find optimal hyperparameters of the Transformer encoder.•A light-weight system to run on top of weak edge hardware.•A federated learning framework to ensure privacy of data.
Remaining useful lifetime prediction and extension of Si power devices have been studied extensively. Silicon carbide (SiC) power devices have been developed and commercialized. Specifically, SiC ...mosfet s have been utilized for the next generation high-voltage, high-power converters with smaller size and higher efficiency, covering various mainstream applications, including photovoltaic systems, electric vehicles, solid-state transformers, and more electric ships and airplanes. However, the SiC-based devices have different failure modes and mechanisms compared with Si counterparts. Therefore, a comprehensive review is critical to develop accurate lifetime prediction and extension strategies for SiC power converter systems. The SiC power device component-level failure modes and mechanisms are first investigated. Different accelerated lifetime tests and component-level lifetime models are then compared. Power converter system-level offline lifetime modeling techniques and software tools are further summarized. Besides, the SiC power converter condition monitoring strategies and health indicators are surveyed. The online measurement challenges are also studied. Furthermore, the system-level lifetime extension strategies are reviewed. By integrating device physics, statistical modeling, reliability engineering, and mechanical engineering with power electronics, this article is intended to provide a comprehensive overview, address existing challenges, and unfold new research opportunities regarding the SiC power converter real-time lifetime prediction and extension.
•Two artificial neural network models are established to estimate State-of-Health and remaining-useful-lifetime synchronously.•The advanced filter methods are applied to smooth the original IC ...curves.•The indicators of health status are extracted from partial incremental capacity curves.•Correlation analysis method is proposed to extract input data of artificial neural network models.
The state of health (SOH) and remaining useful lifetime (RUL) estimation are important parameters for battery health forecasting as they reflect the health condition of battery and provide a basis for battery replacement. This study proposes a novel on-line synthesis method based on the fusion of partial incremental capacity and artificial neural network (ANN) to estimate SOH and RUL under constant current discharge. Firstly, the advanced filter methods are applied to smooth the initial incremental capacity curves. Then the strong correlation feature values are extracted from the partial incremental curves by using correlation analysis methods. Finally, two ANN models aiming at estimating SOH and RUL are established to estimate the SOH and RUL simultaneously. The training and verification results indicate that the proposed method has highly reliability and accuracy for SOH and RUL estimation.
Display omitted
•The EMD-based drift increment extraction method is proposed.•The drift increment is predicted based on the LSTM network.•A difference approximation method for the drift function derivative is ...proposed.•The drift-increment-based representation of RUL’s PDF is derived.•Advantages in the improvement of the prediction accuracy are justified.
In this paper, a novel remaining useful lifetime (RUL) prediction method that fuses stochastic degradation modeling and machine learning is proposed to improve the fitness of the model and quantify the uncertainty of the prediction results. First, a stochastic degradation model based on the Wiener process is built, and the drift increment is extracted using empirical mode decomposition (EMD). Second, a long short-term memory (LSTM) network is trained to learn the equipment degradation rule and predict the drift increment. The diffusion coefficient of the degradation model is then estimated according to the maximum likelihood principle. The final step is to derive the analytical expression for the probability distribution of remaining useful lifetime (RUL) based on the concept of first hitting time and the difference principle. The lithium battery degradation test confirmed the efficacy of the proposed method, achieving a life cycle average prediction accuracy of up to 97.45%.