Search

The deep learning (DL)-based method for predicting remaining useful life (RUL) has gained lots of attention in the industrial equipment sector. Due to the complexity of modern industrial equipment and the necessity of monitoring multivariate time-series data to obtain comprehensive health information, DL models with spatio-temporal feature extraction have been developed to achieve accurate RUL prediction results. However, in the three typical approaches for spatio-temporal feature extraction (parallel, sequential, and nested), each module is independent and separate, making them ineffective in fusing spatio-temporal information at the same time. Most existing approaches use separated modules to extract spatial and temporal features, where convolution, graph, and recurrent neural networks (RNNs) are often applied. To overcome these limitations, this article introduces a unified paradigm for spatio-temporal feature fusion. By extending the message-passing graph neural network (GNN), the spatio-temporal propagation (STP) model is constructed. Using GNN as a single structure, the model can simultaneously match information propagation at both spatial and temporal scales. STP allows a flexible and intuitive way to construct RUL prediction models. An implementation of STP-GNN with an attention mechanism is given and discussed, and a case study on RUL prediction of turbofan engines is reported. Experimental results verify the effectiveness of STP-GNN, and highlight its interpretability in the process of RUL prediction.

Accurate prediction of the remaining useful life (RUL) of a general-purpose heat source radioisotope thermoelectric generator (GPHS-RTG) is critical for mission scheduling and power allocation in spacecraft. The key challenges in this task are long-term predictability and uncertainty quantification due to the limited data and variations in load conditions. However, little reported literature can be found addressing both issues. In this study, a component-level mechanism-based nonlinear additive Wiener process framework was proposed to improve the effectiveness of RUL prediction. Generally, a novel component-level mechanism empirical function developed via energy transfer path and kinetics was adopted to regulate the degradation trend, and a modified nonlinear additive Wiener process achieved uncertainty quantification and probabilistic prediction of RUL. In practice, the empirical function was regarded as a nonlinear drift of Wiener process, while the drift coefficients were randomized to represent component heterogeneity. This study derived parameter estimation by expectation-maximization (EM) algorithm and provided a probability density function (pdf) under the first hitting time (FHT) of RUL. Experimental results on NASA’s GPHS-RTG dataset demonstrate that the proposed empirical function outperforms existing counterparts and long short-term memory (LSTM) networks, exhibiting strong long-term predictive capability. The proposed RUL prediction framework exhibits higher accuracy than traditional one- and two-stage nonlinear Wiener processes, which is helpful for task planning. Moreover, the proposed empirical function enables the identification of component-level power losses, helping the detailed power analysis.

Fault detection (FD) algorithms based on supervisory control and data acquisition (SCADA) data have been widely used in the operation and maintenance (O&M) of wind turbines (WTs). However, the performance of FD models will degrade due to the time-varying operating conditions (TVOC). Incremental learning (IL) methods can be employed to update the models online to adapt to TVOC, but the error accumulation caused by false negatives will lead to a continuous decrease in the FD rate (FDR). A novel IL method considering false negatives is proposed to improve the performance of WT FD models. First, a data buffer is built to cache some new normal data used in model updating. Second, a processing strategy for false negatives is proposed to block some high-risk data from being added to the buffer, thereby weakening the error accumulation. Third, when the buffer is full, update the FD model using the data in it. A real-world SCADA dataset with gearbox faults and four different FD algorithms with various model updating strategies are used in the experiments. The results demonstrate that the proposed method can lower the false alarm rate (FAR) of all four FD algorithms. After processing false negatives, for example, for the FD model based on Gaussian kernel regression, its FDR increased by about 11%, and for the FD model based on multivariate state estimation technique, its FDR increased by about 14%. The results of hyperparameter experiments and early stopping experiments show that the proposed method has good potential for practical applications.

Remaining useful life (RUL) prediction based on machine learning assumes that there are enough representative data for training models. However, it is impossible to have so many representative data considering security, economy factors, and so on. Thus, an incremental learning based RUL prediction approach is proposed to address this problem. First, a novel sequence input vector is constructed from the limited condition monitoring data, and it is proved that the input subspace have orthogonal properties, which is a necessary assumption to ensure the existence of a projector. Second, a projector is constructed to find a weight configuration for avoiding catastrophic forgetting. Finally, an integrated gate recurrent unit model is constructed to map the relationship between condition monitoring data and RUL. A benchmark-bearing case study, whose results indicate that the approach can update fundamental model with the acquisition of new degradation cases, demonstrates the effectiveness.