Search

State-of-health (SOH) is a measure of a battery's capacity in comparison to its rated capacity. Despite numerous data-driven algorithms being developed to estimate battery SOH, they are often ineffective in handling time series data, as they are unable to utilize the most significant portion of a time series while predicting SOH. Furthermore, current data-driven algorithms are often unable to learn a health index, which is a measurement of the battery's health condition, to capture capacity degradation and regeneration. To address these issues, we first present an optimization model to obtain a health index of a battery, which accurately captures the battery's degradation trajectory and improves SOH prediction accuracy. Additionally, we introduce an attention-based deep learning algorithm, where an attention matrix, referring to the significance level of a time series, is developed to enable the predictive model to use the most significant portion of a time series for SOH prediction. Our numerical results demonstrate that the presented algorithm provides an effective health index and can precisely predict the SOH of a battery. [ABSTRACT FROM AUTHOR]

Accurate predictive modeling of traffic flow is critically important as it allows transportation users to make wise decisions to circumvent traffic congestion regions. The advanced development of sensing technology makes big data more affordable and accessible, meaning that data-driven methods have been increasingly adopted for traffic flow prediction. Although numerous data-driven methods have been introduced for traffic flow predictions, existing data-driven methods cannot consider the correlation of the extracted high-dimensional features and cannot use the most relevant part of the traffic flow data to make predictions. To address these issues, this work proposes a decoder convolutional LSTM network, where the convolutional operation is used to consider the correlation of the high-dimensional features, and the LSTM network is used to consider the temporal correlation of traffic flow data. Moreover, the multi-head attention mechanism is introduced to use the most relevant portion of the traffic data to make predictions so that the prediction performance can be improved. A traffic flow dataset collected from the Caltrans Performance Measurement System (PeMS) database is used to demonstrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

An incipient fault refers to the first change point when a system starts to deteriorate. Early detections of incipient faults are crucial to the safety, reliability, and effective predictive maintenance of complex engineering systems. However, it is very difficult to detect incipient faults at the initial stage of the system-degradation processes. To address this issue, an innovative data-fusion method is introduced to detect the incipient faults by integrating data collected from multiple sources instead of a single data source. The data-fusion problem is formulated as a convex optimization problem, aiming at reducing the false-alarm rate and the time to detect incipient faults. We demonstrate the proposed data-fusion method using a data set generated by the degraded aircraft engines. The numerical results have demonstrated that the proposed method can assist in reducing both the false-alarm rate and the time to detect incipient faults. [ABSTRACT FROM AUTHOR]

Transformer, built on the self-attention mechanism, has been demonstrated to be effective in numerous applications. However, in the context of prognostics and health management, the self-attention mechanism in the Transformer is not effective in selecting the most important features that are highly correlated with the remaining useful life (RUL) of a component. To address this issue, we developed a novel conditional variational transformer architecture consisting of four networks: two generative networks and two predictive networks. The first generative network uses the transformer encoder–decoder as well as both condition monitoring data and RUL as input to extract the most important features in one feature space from condition monitoring data. The second generative network uses the transformer encoder and condition monitoring data to extract features in another feature space. The two predictive networks use the extracted features in two different feature spaces to make predictions. A KL-divergence is used to minimize the distance between the two feature spaces learned by the first and second generative networks so that the feature space extracted from the second generative network can approximate the feature space extracted from the first generative network. We demonstrated that the proposed method is effective in predicting the RUL of bearings using two datasets. [ABSTRACT FROM AUTHOR]

Graph Convolutional Networks (GCNs) have recently been used to predict the remaining useful life (RUL) of bearings due to its effectiveness in revealing correlations in condition monitoring data. However, traditional GCNs use a single graph only, either a temporal-correlated graph or a feature-correlated graph without considering both temporal and feature correlations of condition monitoring data. Additionally, traditional GCNs rely heavily on pre-defined graphs to aggregate correlated features. However, the topology of these pre-defined graphs may vary depending on a pre-defined threshold for cosine similarity or covariance which might affect prediction accuracy and robustness. To address these issues, we introduce a spectral graph convolutional operation that can handle both temporal-correlated and feature-correlated graphs, which allows one to consider both the temporal and feature correlations simultaneously. Moreover, we introduce a self-attention mechanism to construct the temporal-correlated and feature-correlated graphs automatically without defining a threshold. Such a mechanism allows the predictive model to learn graphs automatically during training so that the prediction accuracy and robustness can be significantly improved. The proposed method is demonstrated on two bearing datasets, and the experimental results have shown that it outperforms both traditional GCNs and other deep-learning methods in predicting RUL of bearings. [ABSTRACT FROM AUTHOR]

Recent advances in sensing technologies have enabled engineers to collect big data to predict the remaining useful life (RUL) of complex systems. Current modeling techniques for RUL predictions are usually not able to quantify the degradation behavior of a complex system through a health index. Although some studies have been conducted to learn the health index of degradation systems, most of the existing methods are highly dependent on pre-defined assumptions which may not be consistent with the real degradation behaviors. To address this issue, we introduce a time-dependent directed graphical model to characterize the probabilistic relationships among sensor signals, RUL, operational conditions, and health index. Based on the graphical model, a dynamic conditional variational autoencoder is proposed to learn the health index. The experimental results have shown that the proposed method can learn an effective and reliable health index that measures complex system degradation behavior. Moreover, the learned health index improves the accuracy of RUL predictions. [ABSTRACT FROM AUTHOR]