Intelligent identification of multiple power quality (PQ) disturbances is very useful for pollution control of power systems. In this study, we propose a novel detection framework for complex PQ disturbances based on multi-fusion convolutional neural network (MFCNN). Our contributions focus on automatic extraction and fusion of features from multiple sources. First, an information fusion structure is introduced in which the time domain and frequency domain information of the PQ disturbance signal are used as inputs. Additionally, the one-dimensional composite convolution is proposed to improve the diversity of network features based on the standard convolution and dilated convolution. Then, to speed up the training and prevent overfitting, batch normalization is used to adjust the distribution of features. Second, we use several visualization methods to resolve the internal mode of MFCNN, and demonstrate the working mechanism of the proposed method. Finally, we conducted various experiments to verify the effectiveness of the MFCNN. Compared with the handcrafted feature design methods and the general convolutional neural network models, the simulation under different noises and hardware platform-based experiments verify the effectiveness of noise immunity, higher training speed, and better accuracy of the method.

Identifying critical nodes in the power grid is a crucial aspect of power system security and stability analysis. However, the current methods for identification fall short in fully accounting for the power transfer characteristics between nodes and the consequences of node removal on the security and stability of power grid operation. To enhance the effective and accurate identification of critical nodes in the power grid, a method is proposed. This method is based on improved PageRank algorithm and node-weighted power flow transfer entropy, referred to as IPRA-PFTE. Firstly, based on the power flow and equivalent impedance between nodes, and the introduction of virtual nodes, an improved PageRank algorithm is obtained. Then the node-weighted power flow transfer entropy is derived by considering the uniformity of the transfer power flow distribution in the system following the removal of a node. Finally, the importance of nodes is obtained by combining the improved PageRank algorithm with the node-weighted power flow transfer entropy. The method’s effectiveness and accuracy are validated through simulation using the IEEE 39-bus example and subsequent comparison with existing methods.

Accurate power load forecasting is crucial for ensuring grid stability, optimizing economic dispatch, and facilitating renewable energy integration in modern smart grids. However, real load forecasting is often disturbed by the inherent non-stationarity and multi-factor coupling effects. To address this problem, a novel hybrid forecasting framework based on adaptive mode decomposition (AMD) and improved least squares support vector machine (ILSSVM) is proposed for effective short-term power load forecasting. First, AMD is utilized to obtain multiple components of the power load signal. In AMD, the minimum energy loss is used to adjust the decomposition parameter adaptively, which can effectively decrease the risk of generating spurious modes and losing critical load components. Then, the ILSSVM is presented to predict different power load components, separately. Different frequency features are effectively extracted by using the proposed combination kernel structure, which can achieve the balance of learning capacity and generalization capacity for each unique load component. Further, an optimized genetic algorithm is deployed to optimize model parameters in ILSSVM by integrating the adaptive genetic algorithm and simulated annealing to improve load forecasting accuracy. The real short-term power load dataset is collected from Guangxi region in China to test the proposed forecasting framework. Extensive experiments are carried out and the results demonstrate that our framework achieves an MAPE of 1.78%, which outperforms some other advanced forecasting models.

Accurate assessment of the operation state for smart electricity meters (SEMs) is crucial for electrical metering and service. Nevertheless, the real operation state estimation often ignores the impacts of multiple environmental stresses. In this article, a novel model based on multiple fusion support vector regression (MFSVR) and improved simulated annealing (ISA) is proposed to assess the operation state of SEMs under multiple environmental stresses. First, the MFSVR is used to integrate different input information, in which different types of kernel functions are weighted for emphasizing different feature attributes including time, temperature, and humidity. Then, the ISA is further presented to optimize the model parameters in MFSVR, which can contribute to improving the assessment accuracy. In ISA, the adaptive temperature control strategy and modified Metropolis-based criteria are carried out to enhance the parameter search efficiency. The MFSVR model is set up using the libSVM platform in MATLAB, and the dataset of SEMs is collected from the actual high-cold region in China to test our model. Extensive experiments are conducted for verification analysis from the aspects of accuracy, robustness, and sensitivity. The result shows that the proposed model has the lowest average RMSE of 3.40 $\times\; 10^{-2}$ and MAE of 2.14 $\times\; 10^{-2}$ , respectively. Besides, the proposed MFSVR-ISA obtains an average R $^{2}$ of 97.8% under the 20 dB noise level, which is more accurate and robust than some popular data-driven prediction methods.

With the increasing integration of non-linear electronic loads, the diagnosis and classification of power quality are becoming crucial for power grid signal management. This paper presents a novel diagnosis strategy based on unsupervised learning, namely residual denoising convolutional auto-encoder (RDCA), which extracts features from the complex power quality disturbances (PQDs) automatically. Firstly, the time–frequency analysis is applied to isolate frequency domain information. Then, the RDCA with a weight residual structure is utilized to extract the useful features in the contaminated PQD data, where the performance is improved using the residual structure. A single-layer convolutional neural network (SCNN) with an added batch normalization layer is proposed to classify the features. Furthermore, combining with RDCA and SCNN, we further propose a classification framework to classify complex PQDs. To provide a reasonable interpretation of the RDCA, visual analysis is employed to gain insight into the model, leading to a better understanding of the features from different layers. The simulation and experimental tests are conducted to verify the practicability and robustness of the RDCA.