Multivariate time series is widely derived from industrial facilities, such as power plants, manufacturing machines, spacecraft, digital devices etc., and anomaly detection and location is of great importance to industrial preventive maintenance. However, anomalies in multivariate time series always result from their unusual change of temporal or correlative features, and it is challenging to capture these complex characteristics. Besides, achieving consistent anomaly detection and location performance is also a tricky issue. In this article, a novel anomaly detection and location framework that combines generative adversarial networks and autoencoder is proposed to capture time dependent and correlation features of multivariate time series with on need of anomalous sequences. Firstly, multi-time scale correlation computation is utilized to encode multivariate time series into multiple cross-correlation graphs, which can be fed into the proposed deep architecture for extracting more distinguishable features. On this basis, a robust cost function with multiple loss issues is designed, and reconstruction matrix deviation from original space of encoder-encoder structure is utilized to detect and locate abnormal time series, ensuring the consistency of detection and location tasks and the framework reliability. Extensive experiments on five industrial datasets are conducted to indicate our model is a generic and excellent framework for anomaly detection and location of multivariate time series.

Accurate remaining useful life prognosis of bearings in wind turbines can effectively help to schedule maintenance strategy and reduce operational costs at wind farms. Unscented particle filter is good at state tracking in nonlinear problem. A robust model-based approach based on improved unscented particle filter is presented to deal with bearing life prognosis in wind turbines, which involves: (1) The mean of sigma points after unscented Kalman transform is regarded as the particles in particle filter to guarantee the particles aggregation; (2) Several past measurements are utilized to estimate the likelihood function of current step; (3) Uniform distribution is adopted for resampling particles to make them diversity. The presented remaining useful life prognosis approach depends more on the measurement, rather than the initial parameters of degradation model, which makes it practicable for the on-site wind turbines. Three life-cycle bearings from wind turbine high-speed shafts demonstrate the effectiveness of the proposed approach.

Accurate and reliable state-of-health (SOH) prediction becomes increasingly vital to ensure the safe and reliable operation of lithium-ion batteries (LIBs). The existing data-driven methods for LIBs’ SOH prediction are developed with an ideal database, i.e., a huge run to failure data with the consistent distribution of training and testing sets. However, due to the individual quality differences and complex operating conditions, data distribution shifts among different batteries may be obvious, and the entire life-cycle samples are difficult to collect in real world. Therefore, there exists distribution discrepancy between source and target domains. Besides, temporal distribution shift may also exist with incomplete target domain, called time covariate shift (TCS). Thus, the model trained with source and incomplete target domains will cause prediction bias in unseen target data. To address such issues, a novel transfer learning approach using multiple feature alignment transformer (MFA-Transformer) model is conducted for the SOH prediction of LIBs. Firstly, a multi-layer feature alignment is performed via encoder-decoder structure of transformer framework, and Multi-Kernel Maximum Mean Discrepancy (MK-MMD) is adopted to tackle data distribution discrepancy. Then, a new loss item based on Weibull distribution is utilized to enhance the data alignment effect. Moreover, a shift compensation strategy using Shape-based Distance (SBD) estimation is designed to dynamically eliminate the prediction bias resulting from TCS. Finally, experiments on two public LIBs datasets validate the effectiveness of the proposed method, which can offer a promising solution for industrial prognostic without entire life-cycle data.

Series capacitor compensation is used to improve the utilization of existing power systems. Subsynchronous resonance (SSR) can be caused by series compensated lines, which would lead to turbogenerator shaft breakdown. A novel approach was presented in this paper to analyze the characteristics and influence factors of SSR in a series compensated transmission system. The system model of SSR, including the various modules of the electromechanical network, was established, and the eigenvalue results under 70% series compensation level were analyzed by eigenvalue analysis method for the the institute of electrical and electronics engineers (IEEE) first benchmark model (FBM). Compared with the results of the power systems computer-aided design (PSCAD) modeling and simulation, the effectiveness of eigenvalue analysis method was proven. After that, the eigenvalue analysis method was used to study, in detail, the effects of system series compensation levels, synchronous generator parameters, speed governing system parameters, and excitation system parameters on SSR characteristics. The research results show that the series compensation level has the greatest influence on the torsional mode damping of the system. The parameters of generator reactance, speed governing system, and excitation system have some effect on the torsional mode damping. The parameters of excitation system significantly affect the low-frequency oscillation damping.