Electricity theft has caused enormous damage to grids safety and economy globally, bringing plentiful attention to electricity theft detection. However, the inherent problems of data imbalance, data sparsity, and data shift due to the residents will or the seasons remain challenges. Besides, there is a lack of appliance-level information provided by non-intrusive load monitoring to improve the data graininess in electricity theft detection. To address these problems, an electricity theft detection algorithm based on contrastive learning and non-intrusive load monitoring is proposed. Firstly, a semi-supervised learning architecture composed of Gramian angular field encoding for sequence visualization and contrastive learning architecture featuring few-shot learning is established for load monitoring and initial detection considering the inherent operating characteristics of appliances. Furthermore, after the filtration of typical regular-switched appliances by Kendalls coefficient of concordance, in-depth detection of abnormal operation routines of appliances is conducted for suspicious residents, aiming to improve the fault-tolerant ability and resist the emergence of unknown electricity theft methods. Finally, the electricity theft probability is computed to confirm fraudulent users. Both the public and practical datasets are utilized to verify the effectiveness of the proposed study, and the results show an overall better performance compared to other state-of-the-art algorithms.
With the rapid growth of photovoltaic (PV) power in recent years, the stability of system operation, the performance of system contingency analysis as well as the power quality of the power grid are threatened by the inherent uncertainty and fluctuation of PV output. It is necessary to have the knowledge of PV output characteristics for reliable power system dispatching. Day-ahead PV power forecasting is an effective support for achieving optimal dispatching. Probabilistic forecasting can describe the uncertainty that is difficult to depict by deterministic forecasting, and the forecasting results are more comprehensive. An ensemble nonparametric probabilistic forecasting model of PV output is proposed based on the traditional deterministic forecasting method. Quantile regression averaging (QRA) is used to ensemble a group of independent long short-term memory (LSTM) deterministic forecasting models for obtaining the probabilistic forecasting of PV output. Real measured data are used to verify the effectiveness of this nonparametric probabilistic forecasting model. Additionally, in comparison with the benchmark methods, LSTM-QRA has higher prediction performance due to the better forecasting accuracy of independent deterministic forecasts.
Recently, a large number of distributed photovoltaic (PV) power generations have been connected to the power grid, which resulted in an increased fluctuation of the net load. Therefore, load forecasting has become more difficult. Considering the characteristics of the net load, an ultrashort-term forecasting model based on phase space reconstruction and deep neural network (DNN) is proposed, which can be divided into two steps. First, the phase space reconstruction of the net load time series data is performed using the C-C method. Second, the reconstructed data is fitted by the DNN to obtain the predicted value of the net load. The performance of this model is verified using real data. The accuracy is high in forecasting the net load under high PV penetration rate and different weather conditions.
As the coupling and integration of multi-energy flow in the integrated energy system (IES) deepens increasingly, the cascading failure will develop across different energy systems more easily and widely through the energy hub (EH). And it brings great challenges to the security and reliability of IES. The defects of present cascading failure model of IES have been summarized and a novel search strategy of fault chains in IES combined heating and power network was proposed in this paper. Firstly, the initial risk assessment index of each energy branch is proposed to form the initial fault sets. Then combined heat and power control strategies (CHPC) is introduced to deal with the branch overload conditions during the cascading failure. What’s more, in order to reduce the workload and overcome the limitation of present methods, we analyzed the relevance of the branches to be predicted by using Kernel Fuzzy C-means (KFCM) clustering algorithm and selected the branches with the highest value of relevance as the subsequent failure. Based on the predicted fault chain, vulnerability analysis is presented to locate critical component and find out the correlation between cascading outages. Comprehensive evaluation index is also established to effectively evaluate the impact severity of the cascading failure. Finally, the case studies are carried out on the combined heating and power systems to demonstrate the effectiveness of the proposed method.