Uncertainty aggregation characterization of multi spatial–temporal distributed energy resources (DERs) is crucial for effective decision-making and control in power systems. In this article, we propose a cloud-edge-end collaboration approach to quantify the aggregated uncertainty of power generation from multi spatial–temporal DERs. First, considering the temporal dynamic and electrical topology correlation of DERs, a local uncertainty aggregation model based on a spatial-temporal graph neural network (STGNN) is developed. This model can effectively extract the spatial-temporal characteristics of data. Second, addressing the data silo problem caused by the unwillingness of various stakeholders managing the DERs to share data due to privacy concerns, an uncertainty aggregation model training mechanism based on an adaptive secure federated learning is proposed. This mechanism enables collaborative modeling of uncertainty aggregation models across stakeholders while preserving user privacy. In addition, it improves the quality of local model training by adaptively extracting parameter information from the global model for local model initialization. Moreover, since the probability distribution of the aggregated uncertainty is unknown, this article combines STGNN with the weighted quantile regression model to characterize the aggregated uncertainty without prior assumptions about the distribution, and by assigning differentiated weights to aggregation results under different confidence levels based on their importance, the proposed method can better meet the diverse needs of power grid. Finally, simulations conducted on the IEEE 33-bus system and IEEE 69-bus system validate the effectiveness of the proposed method.

The intermittency of solar irradiance is the main cause of rapid fluctuations in the power output of photovoltaic (PV) systems. These fluctuations hinder the large-scale integration of solar power generation equipment into the grid, which in turn hinders the process of utilizing solar energy resources to reduce carbon emissions. The main way to solve this dilemma is to achieve high-precision forecasting of solar irradiance. Although various methods exist to forecast the variations of solar irradiance, few focus on fully utilizing multimodal data information and fuzzy method to improve the forecasting performance. Therefore, a forecasting method combining multimodal data fusion and fuzzification is proposed to forecast ultra short-term global horizontal irradiance (GHI). First, a modal conversion method is designed to convert temporal modal data to spatial modal data. Then, the fused data are formed by fusing the converted data with normal and under exposure all-sky images. Subsequently, the fuzzy method is used to generate fuzzy GHI data with low nonlinear features. Last, we utilize deep neural networks to learn potential patterns between fused data and fuzzy GHI data in an end-to-end manner. Our method has been comprehensively validated on data provided by the National Renewable Energy Laboratory, demonstrating its effectiveness, and achieving the highest forecasting accuracy compared to state-of-the-art methods.

False data injection attack (FDIA) can pose a severe threat to the distribution networks (DN), and the accurate detection of FDIA plays a key role in the safe and reliable operation of the DN. In this article, an end-edge-cloud collaboration-based detection framework is proposed to detect FDIA in the DN. First, in order to effectively preserve the privacy of different stakeholders in the DN and solve the problem of data island, a federated-learning-based edge-cloud collaboration mechanism is designed according to the proposed end-edge-cloud collaboration framework to jointly train the local FDIA detection models and eventually build a comprehensive FDIA detection model. Then, considering the temporal–spatial correlation of measurement data, a local data-driven FDIA detection model is proposed based on a novel temporal–spatial graph convolutional network, which can extract temporal–spatial features of the measurement data and improve the FDIA detection performance. In general, compared with the traditional centralized FDIA detection methods, the proposed method can make full use of the computational capacity of distributed edge devices and reduce the pressure of computation on the control center. Finally, simulation results based on the modified IEEE 14-bus and IEEE 118-bus distribution systems indicate that the proposed method can effectively improve the accuracy of FDIA detection compared with other methods.

Although using image strip sum, an orthogonal Haar transform (OHT) pattern matching algorithm may have good performance, it requires three subtractions to calculate each Haar projection value on the sliding windows. By establishing a solid mathematical foundation for OHT, this paper based on the concept of image square sum, proposes a novel fast orthogonal Haar transform (FOHT) pattern matching algorithm, from which a Haar projection value can be obtained by only one subtraction. Thus, higher speed-ups can be achieved, while producing the same results with the full search pattern matching. A large number of experiments show that the speed-ups of FOHT are very competitive with OHT in most cases of matching one single pattern, and generally higher than OHT in all cases of matching multiple patterns, exceeding other high-level full search equivalent algorithms.

Based on the "convexly separable" concept, we present a solid geometric theory and a new general framework to design piecewise linear classifiers for two arbitrarily complicated nonintersecting classes by using a "multiconlitron," which is a union of multiple conlitrons that comprise a set of hyperplanes or linear functions surrounding a convex region for separating two convexly separable datasets. We propose a new iterative algorithm called the cross distance minimization algorithm (CDMA) to compute hard margin non-kernel support vector machines (SVMs) via the nearest point pair between two convex polytopes. Using CDMA, we derive two new algorithms, i.e., the support conlitron algorithm (SCA) and the support multiconlitron algorithm (SMA) to construct support conlitrons and support multiconlitrons, respectively, which are unique and can separate two classes by a maximum margin as in an SVM. Comparative experiments show that SMA can outperform linear SVM on many of the selected databases and provide similar results to radial basis function SVM on some of them, while SCA performs better than linear SVM on three out of four applicable databases. Other experiments show that SMA and SCA may be further improved to draw more potential in the new research direction of piecewise linear learning.