This paper proposes a condition monitoring method based on Dynamic Related ReliefF-SFA (DRRSFA), which solves the problem of huge variable dimensions and strong autocorrelation in process monitoring. First, the samples are mapped into a new space, and the slow projection components (SPCs) of the samples are calculated. The SPCs not only considers the autocorrelation between the variables, but also characterizes the inherent properties of the whole system. Second, the feature selection algorithm ReliefF is used to select the more weighted features, which is called the principal components(PCs) to achieve dimensionality reduction. Then the corresponding statistics and control limits are calculated based on the obtained PCs. Finally, the process monitoring using the proposed algorithm is performed by testing a numerical example and the actual production process data, and the results show the effectiveness of the proposed method.

With the rapid evolution of artificial intelligence technologies in power systems, data-driven time-series forecasting has become instrumental in enhancing the stability and reliability of power systems, allowing operators to anticipate demand fluctuations and optimize energy distribution. Despite the notable progress made by current methods, they are still hindered by two major limitations: most existing models are relatively small in architecture, failing to fully leverage the potential of large-scale models, and they are based on fixed nonlinear mapping functions that cannot adequately capture complex patterns, leading to information loss. To this end, an LLM-Empowered Kolmogorov–Arnold frequency learning (LKFL) is proposed for time series forecasting in power systems, which consists of LLM-based prompt representation learning, KAN-based frequency representation learning, and entropy-oriented cross-modal fusion. Specifically, LKFL first transforms multivariable time-series data into text prompts and leverages a pre-trained LLM to extract semantic-rich prompt representations. It then applies Fast Fourier Transform to convert the time-series data into the frequency domain and employs Kolmogorov–Arnold networks (KAN) to capture multi-scale periodic structures and complex frequency characteristics. Finally, LKFL integrates the prompt and frequency representations through an entropy-oriented cross-modal fusion strategy, which minimizes the semantic gap between different modalities and ensures full integration of complementary information. This comprehensive approach enables LKFL to achieve superior forecasting performance in power systems. Extensive evaluations on five benchmarks verify that LKFL sets a new standard for time-series forecasting in power systems compared with baseline methods.

The power systems have generated a large amount of multi-view data with the advancement of electrical equipments. Graph-based multi-view methods boost the performance across diverse power systems, via aggregating inter-view and intra-view topological structures of data, which gains greatly attention. However, they focus on pairwise relationships between data to mine potential patterns, which is difficult to capture complex high-order correlations between data. To this end, a novel multi-view hypergraph adaptive fusion approach (MHAF) is proposed for representation learning in various systems, which contains the multi-view hypergraph modelling module and the adaptive correlation aggregation module. The former constructs hypergraphs from different views to capture different high-order relationship among social systems. The later dynamically weights and fuses different views via employing an attention aggregation mechanism, which creates a more expressive fused representation. Two components work seamlessly together to capture the complex patterns of multi-view data, enhancing the system's ability to understand user behavior and provide more personalized guidance. Numerous experiments on LastFM and FilmTrust demonstrate the superiority and advancement of MHAF over baselines.