Nowadays, electricity theft has been a major problem worldwide. Although many single-classification algorithms or an ensemble of single learners (i.e., homogeneous ensemble learning) have proven able to automatically identify suspicious customers in recent years, after the accuracy of these methods reaches a certain level, it still cannot be improved even if it continues to be optimized. To break through this bottleneck, a heterogeneous ensemble learning method with stacking integrated structure of different strong individual learners for detection of electricity theft is presented in this paper. Firstly, we use the grey relation analysis (GRA) method to select the heterogeneous strong classifier combination of LG + LSTM + KNN as the base model layer of stacking structure based on the principle of the highest comprehensive evaluation index value. Secondly, the support vector machine (SVM) model with relatively good results of the stacking overall structure experiment is selected as the model of the meta-model layer. In this way, a heterogeneous integrated learning model for electricity theft detection of the stacking structure is constructed. Finally, the experiments of this model are conducted on electricity consumption data from State Grid Corporation of China, and the results show that the detection performance of the proposed method is better than that of the existing state-of-the-art detection method (where the area under receiver operating characteristic curve (AUC) value is 0.98675).
10 October 2024 in IEEE Transactions on Instrumentation and Measurement
User-side carbon emission accounting is a means to sort out the amount and source of user carbon emissions. Real-time carbon estimation methods based on non-intrusive load monitoring (NILM) tend to miss the fact that there is not a negative correlation between equipment identification accuracy and carbon emission estimation error. Higher accuracy does not mean lower error. A two-stage training NILM network, LRMSE-ResNeSt, is proposed to reduce the error of real-time carbon emission monitoring results. It emphasizes the low Root-Mean-Square Error (RMSE) of the carbon emissions monitoring. The feature extraction network is first constructed using convolution and split attention mechanism, and then, the model is trained using a two-stage training approach. In the first stage, backpropagation is used to update the network parameters for accurate device identification. In the second stage, the parameters of the fully connected layer are tuned using Particle Swarm Optimization to make the classifier more focused on the identification accuracy of devices with high carbon emissions. Finally, the proposed LRMSE-ResNeSt is validated using the IAID industrial dataset. The experimental results show that the LRMSE-ResNeSt successfully reduces the RMSE of real-time carbon estimation by an average of 14.94%, which addresses the problem that the NILM method only focuses on the accuracy of the equipment identification but ignores the error of the carbon emission estimation results.
3 July 2023 in IEEE Transactions on Instrumentation and Measurement
Electricity theft has been a major concern all over the world. There are great differences in electricity consumption among residents from different regions. However, existing supervised methods of machine learning are not in detecting electricity theft from different regions, while the development of transfer learning provides a new view for solving the problem. Hence, an electricity-theft detection method based on Convolutional and Joint Distribution Adaptation(CJDA) is proposed. In particular, the model consists of three components: convolutional component (Conv), Marginal Distribution Adaptation(MDA) and Conditional Distribution Adaptation(CDA). The convolutional component can efficiently extract the customer’s electricity characteristics. The Marginal Distribution Adaptation can match marginal probability distributions and solve the discrepancies of residents from different regions while Conditional Distribution Adaptation can reduce the difference of the conditional probability distributions and enhance the discrimination of features between energy thieves and normal residents. As a result, the model can find a matrix to adapt the electricity residents in different regions to achieve electricity theft detection. The experiments are conducted on electricity consumption data from the Irish Smart Energy Trial and State Grid Corporation of China and metrics including ACC, Recall, FPR, AUC and F1Score are used for evaluation. Compared with other methods including some machine learning methods such as DT, RF and XGBoost, some deep learning methods such as RNN, CNN and Wide & Deep CNN and some up-to-date methods such as BDA, WBDA, ROCKET and MiniROCKET, our proposed method has a better effect on identifying electricity theft from different regions.
The industrial flotation process plays a crucial role in metallurgical production. Accurate monitoring of the industrial flotation process is particularly important. However, the industrial flotation process is intricate, involving numerous variables, making it challenging to achieve precise control over quality parameters. To address these issues, distributed dynamic graph embedding (DDGE) for quality-relevant process monitoring is proposed. This method incorporates the quality index and enhances local information by integrating data timestamps and neighborhood information. Firstly, the mutual information and distributed partition methods are employed to segment into several quality-related blocks and quality-unrelated blocks. Subsequently, the distributed dynamic graph embedded model is established and integrated into a global decision process based on Bayesian inference along with statistics and control limits. The feasibility and effectiveness of the proposed method are demonstrated using the Tennessee Eastman (TE) benchmark process and the industrial flotation process.
19 September 2025 in IEEE Transactions on Automation Science and Engineering
This paper proposes a novel energy interaction management strategy for off-grid microgrids, addressing the challenges posed by the uncertainty in renewable energy generation and the variability of consumer willingness to participate in demand response (DR). This study introduces an adaptive load type conversion (ALTC) model to quantify consumer willingness and a dynamic coefficient of responsibility (DCR) mechanism for the fair allocation of costs arising from renewable energy prediction errors. A consumer-free interaction algorithm based on the Stackelberg game is further proposed to manage energy scheduling to balance supply-side (SS) and demand-side (DS) objectives. The proposed method is validated using real-world data from an off-grid microgrid in Guizhou, China. Experimental results demonstrate a 26.2% improvement in the profit-to-cost ratio (PCR) compared to the baseline method, and a 37% reduction in peak-to-valley load ratio (PVR). By integrating privacy-preserving load scheduling and adaptive tariff policies, our approach enhances the adaptability of hourly scheduling operations and promotes equitable DR participation, offering a robust solution for modern microgrids. Note to Practitioners—Off-grid microgrids, which power remote communities, industrial sites, or emergency facilities, often struggle with two key challenges: unreliable renewable energy generation and unpredictable consumer electricity usage. These uncertainties can lead to wasted energy, higher costs, and unfair pricing for users. Our work provides a practical solution by introducing an automated system that estimates how willing consumers are to adjust their electricity use without violating privacy, fairly assigns extra costs when renewable energy forecast errors affect the balance between supply and demand, and optimizes energy pricing and supply without requiring direct communication between producers and consumers. In real-world testing, this approach reduced peak energy demand by 37% and improved cost efficiency by 26%, making microgrids more stable and economical. While the system works best with historical usage data and may need adjustments for larger or more complex grids, it can be applied today in remote villages. Future improvements could expand it to networks of multiple microgrids or adapt to rapidly changing weather conditions. This method offers a smarter, fairer way to manage off-grid energy systems while keeping costs and complexity low.