State Grid Corporation Research Articles

Detection Method for Three-Phase Electricity Theft Based on Multi-Dimensional Feature Extraction.

The advent of smart grids has facilitated data-driven methods for detecting electricity theft, with a preponderance of research efforts focused on user electricity consumption data. The multi-dimensional power state data captured by Advanced Metering Infrastructure (AMI) encompasses rich information, the exploration of which, in relation to electricity usage behaviors, holds immense potential for enhancing the efficiency of theft detection. In light of this, we propose the Catch22-Conv-Transformer method, a multi-dimensional feature extraction-based approach tailored for the detection of anomalous electricity usage patterns. This methodology leverages both the Catch22 feature set and complementary features to extract sequential features, subsequently employing convolutional networks and the Transformer architecture to discern various types of theft behaviors. Our evaluation, utilizing a three-phase power state and daily electricity usage data provided by the State Grid Corporation of China, demonstrates the efficacy of our approach in accurately identifying theft modalities, including evasion, tampering, and data manipulation.

A self-decision ant colony clustering algorithm for electricity theft detection

The load data features of some electricity-theft consumers during the theft period are similar to those of normal consumers, making these electricity-theft consumers outliers from the cluster of electricity-theft. The current classification method, which uses the mean value to determine the cluster centers, is vulnerable to the influence of outliers. Therefore, this paper proposes a self-decision ant colony clustering algorithm for electricity theft detection method that is targeted to self-decision which samples are used to update the cluster centers. The method constructs a dynamic weighting approach to determine the cluster centers based on the idea of Backpropagation, and updates the weights of each sample in the clusters to reflect the different importance of different samples, thus reducing the influence of outlier samples. A new activation function, Odd, is proposed to enhance the ability of the proposed method to solve linearly indistinguishable problems. A self-decision dropout mechanism is proposed which evolves the mechanism of randomly stopping the work of samples in clusters into a targeted and self-decision mechanism that stops the work of redundant or non-active samples as well as improves the contribution of outlier samples with positive effects. In this paper, the proposed method is tested by the electricity consumption data provided by the State Grid Corporation of China (SGCC) and the Smart* Data Set for Sustainability (SDSS) provided by the UMass Trace Repository, and the experimental results show that the proposed method effectively solves the above problems with higher detection accuracy, it has certain advantages over other current studies.

“Comprehensive Integration” Method Based on the Bayesian Network and the Big Data Using in the Construction of New Energy System

With the continuous application of the big data technology, “Comprehensive Integration” theory is an approach to management based on Rooted theory, which has been successfully applied to the management of large enterprises and research organizations of all kinds. State Grid Corporation of China in the management process of the construction of a new energy system, with a strong innovative, large-scale investment, many participating units, urgent needs and other characteristics of its management capabilities put forward higher requirements. In this paper, a system integration principle based on the Bayesian probabilistic approach and the big data is proposed to realize the engineering application of the theory of “Comprehensive Integration”. By splitting the management system of the new energy system construction, the unit composition of the system can be interpreted to form a complete system and node interaction Bayesian network structure. The algorithm proposed in this paper is implemented in Netica, a commercial statistical analysis software, and its effectiveness is verified, which provides an important theoretical basis for the engineering practice of the theory of “Comprehensive Integration”. Finally, the effectiveness of the Dynamic probabilistic network method was quantitatively verified in a lithium battery pack energy storage management example.

Research on the Application of Intelligent Multi-branch Version Management Strategy of Electricity Marketing Service System Software

In recent years, the Market Marketing Department of State Grid Corporation of China, in collaboration with the Internet Department, has initiated the construction of an intelligent energy internet marketing service system. This paper aims to improve the development efficiency and quality of the power marketing service system. To achieve this goal, an intelligent multi-branch version management strategy has been introduced. Firstly, the necessity of multi-branch version management is analyzed in relation to the characteristics of the power marketing service system software and the potential issues that may arise in its development and version management process. Then, based on the concept of productizing the general functions of the entire network and projecting the functions of each provincial network, a multi-branch version management strategy is designed, which includes the core components of intelligent multi-branch version management and intelligent multi-branch version protection. Finally, an application case of power marketing service system development is presented. This solution not only standardizes the version management process, improves software development efficiency, ensures the order and control of versions, but also demonstrates the enormous potential of intelligent technology in enhancing system development and management efficiency.

Stochastic Models of Electricity Consumption

The article considers a stochastic model of electricity consumption based on the convolution theory. It is assumed that the processes of switching on and off energy consumption on a city scale depend on a large number of independent random factors and have the form of a normal probability distribution. The functions of switching on and off the load are presented, graphs are plotted. The comparison of the intelligent power supply system with the existing power supply system in the Russian Federation is carried out. The reasons for the slow introduction of smart grids in the Russian Federation are revealed. At the moment, smart grids are not very popular in the Russian Federation, while in the countries of the European Union, the USA and China, projects related to the development, production of components and the widespread introduction of smart grids are actively receiving state financial support. Thus, the European Union allocates more than a billion dollars annually for projects in the field of electric power and smart grids. These funds are allocated according to the European Green Deal strategy. At the same time, the State Grid Corporation of China has introduced the concept of the "Global Energy Internet", which, in accordance with the government's expectation, will stimulate the smart grid market in the country. As in other countries of the world, the US electric power industry is also facing problems related to rising utility prices, fluctuations in peak load and the need to reduce carbon dioxide emissions. The consequence of this was the adoption at the end of 2021 of an "infrastructure" law providing for large-scale investments in green energy projects.

A novel two‐stage method to detect non‐technical losses in smart grids

AbstractNumerous strategies have been proposed for the detection and prevention of non‐technical electricity losses due to fraudulent activities. Among these, machine learning algorithms and data‐driven techniques have gained prominence over traditional methodologies due to their superior performance, leading to a trend of increasing adoption in recent years. A novel two‐step process is presented for detecting fraudulent Non‐technical losses (NTLs) in smart grids. The first step involves transforming the time‐series data with additional extracted features derived from the publicly available State Grid Corporation of China (SGCC) dataset. The features are extracted after identifying abrupt changes in electricity consumption patterns using the sum of finite differences, the Auto‐Regressive Integrated Moving Average model, and the Holt‐Winters model. Following this, five distinct classification models are used to train and evaluate a fraud detection model using the SGCC dataset. The evaluation results indicate that the most effective model among the five is the Gradient Boosting Machine. This two‐step approach enables the classification models to surpass previously reported high‐performing methods in terms of accuracy, F1‐score, and other relevant metrics for non‐technical loss detection.

Modeling and Optimization for Cost Optimization Model of Material Distribution System in Power Distribution Network--A Case in Hunan

This paper proposes a method to address several issues in the distribution system of distribution network materials, including non-transparent distribution costs, low punctuality of distribution, and frequent misreporting and misreporting of distribution costs by local warehouses. The method involves constructing a distribution hierarchy for distribution network materials and developing a mathematical model to calculate the costs associated with distribution, loading and unloading, and warehouse finishing. Additionally, a node identification mechanism is established based on the demand for distribution network materials to accurately identify nodes that can be removed from distribution routes. Furthermore, a distribution route optimization algorithm is developed to optimize distribution costs. The effectiveness of the method is then verified through its application to the distribution of distribution materials for the Hunan State Grid Corporation.

Mechanisms Accounting for the Formation of the Strong Winds that Caused the Tripping Incident of Transmission Line in Eastern Inner Mongolia

Meteorological disasters pose a serious threat to the State Grid Corporation of China, which covers ~ 88% of Chinese national territory. Of these, strong winds deserve a special attention, as they often induce windage yaw discharge of transmission lines and even toppling of transmission towers, resulting in serious economic losses. On 28 June 2023, a severe tripping incident of transmission line appears in Eastern Inner Mongolia due to strong winds. In this study, we conduct comprehensive analyses to clarify the favorable background conditions and governing mechanisms for producing the strong winds. Main results are shown as follows. Synoptic analysis indicates that, the favorable background environments for the event are characterized by a strong upper-level jet associated upper tropospheric divergence; an intense middle-level warm advection ahead of a shortwave trough; and a long-lived lower-tropospheric mesoscale vortex. The strong winds that cause the tripping incident mainly occur in the southeastern quadrant of the vortex. Vorticity budget presents that the period from the mesoscale-vortex’s formation to 4 h before is crucial to the mesoscale vortex, as cyclonic vorticity increases rapidly mainly due to the lower-level convergence-related vertical stretching. In contrast, the horizontal transport mainly results in a net export of cyclonic vorticity, which is the most detrimental factor. Kinetic energy (KE) budget shows that, after the mesoscale vortex forms, the strong winds within its southeastern quadrant enhance rapidly. Overall, the positive work done by the pressure gradient force associated with the mesoscale vortex dominates the enhancement of strong winds; the horizontal transport of KE is the second dominant factor, and the vertical transport of KE (i.e., the downward momentum transportation) shows the least contribution.

A Novel Electricity Theft Detection Strategy Based on Dual-Time Feature Fusion and Deep Learning Methods

To enhance the accuracy of theft detection for electricity consumers, this paper introduces a novel strategy based on the fusion of the dual-time feature and deep learning methods. Initially, considering electricity-consumption features at dual temporal scales, the paper employs temporal convolutional networks (TCN) with a long short-term memory (LSTM) multi-level feature extraction module (LSTM-TCN) and deep convolutional neural network (DCNN) to parallelly extract features at these scales. Subsequently, the extracted features are coupled and input into a fully connected (FC) layer for classification, enabling the precise detection of theft users. To validate the method’s effectiveness, real electricity-consumption data from the State Grid Corporation of China (SGCC) is used for testing. The experimental results demonstrate that the proposed method achieves a remarkable detection accuracy of up to 94.7% during testing, showcasing excellent performance across various evaluation metrics. Specifically, it attained values of 0.932, 0.964, 0.948, and 0.986 for precision, recall, F1 score, and AUC, respectively. Additionally, the paper conducts a comparative analysis with mainstream theft identification approaches. In the comparison of training processes, the proposed method exhibits significant advantages in terms of identification accuracy and fitting degree. Moreover, with adjustments to the training set proportions, the proposed method shows minimal impact, indicating robustness.

Practical experience of low-carbon development evaluation and low-carbon transformation of international energy enterprises

Since the issue of climate change, it has become a general consensus that greenhouse gas emissions will trigger global warming, reduce greenhouse gas emissions is imperative, and actively promote the low- carbon transformation. Moreover, economics can provide a solid theoretical support with a large number of relevant theories. With the continuous dissemination and penetration of the concept of low-carbon development around the world, the internationally renowned multinational large enterprises generally take the initiative to assume their social responsibilities, actively advocate and practice the concept of low-carbon development, and actively take relevant low-carbon measures in practice. This chapter takes state Grid Corporation as an example, and combs and analyzes its low-carbon development evaluation indicators selection and development measures, in order to gain beneficial enlightenment.

Multidimensional data processing and tensor analysis for large-scale power grids in a parallel computing framework

Abstract Grid data is compressed and stored without processing. There are problems of large compression errors and long running time, which affect the compression and storage effect. Therefore, this paper proposes a method for storing grid data using parallel computing frameworks. While drawing on the task scheduling strategy of MapReduce distributed computing framework, combined with the current situation of centralized deployment of data in the information system of the State Grid Corporation, an efficient tensor-based parallel computing method is proposed. The massive multivariate and heterogeneous smart grid data is compressed using the Tucker decomposition method for big data compression of smart grids. This method’s line resistance and reactance dynamic parameters have average values of 0.033 and 0.520, respectively, which are very close to the actual values. The method used in this paper has a high degree of accuracy in estimating resistance dynamic parameters. It possesses certain practical application performance. By predicting the daily average load of the North China Power Grid, the accuracy of the prediction data of the gridded load analysis is close to the actual value, indicating the superiority of the grid multidimensional data processing method designed in this paper.

Deep Active Learning-Enabled Cost-Effective Electricity Theft Detection in Smart Grids

In industrial informatics-enabled smart grids, machine learning approaches have exhibited high potential in data-driven electricity theft detection (ETD), whereas none of the existing studies pay sufficient attention to the high costs of manually labeling massive sensing data during learning data preparation. To address this defect, this paper develops a cost-effective data-driven ETD approach that significantly reduces the data labeling costs without sacrificing the reliability of ETD. Specifically, the approach is systematically realized via an intelligent deep active learning (DAL) scheme. By seamlessly incorporating convolutional neural network (CNN) learning with Monte Carlo dropout-based Bayesian active query, the DAL scheme efficiently selects the most valuable instances for ETD model training. In this way, the proposed approach is able to derive a reliable CNN-based ETD model with limited labeled learning instances, thus largely reducing the data labeling costs. Experimental test results on an actual ETD dataset provided by the State Grid Corporation of China extensively illustrate the efficacy of the proposed approach.

Layout optimization of irregular storage areas under class storage strategy based on clustering and multi-bin size packing problem.

This paper proposes an optimization scheme for the layout of irregular warehouse spaces based on a class-based storage strategy. Firstly, we transform the irregular warehouse space into several regular rectangular areas. Next, through the class-based storage strategy, we develop an algorithm that converts the non-linear clustering problem of homogeneous shelves into a linear selection problem of different sized regular shelf areas. Finally, a comprehensive shelving clustering algorithm and packing problem with different box sizes selection were constructed, and empirical analysis was conducted based on actual data from Xiangtai Warehouse of State Grid Corporation of China. The results show that the new model not only effectively solves the irregular warehouse layout optimization problem under the class storage strategy but also reduces the complexity of the model and shortens the solution time. It is a universally applicable method with significant value for generalization.

Detecting anomalous electricity consumption with transformer and synthesized anomalies

Non-technical losses are consistently a troubling issue for power suppliers. With the application and popularization of smart grid and advanced measurement systems, it has become possible to use data-driven methods to detect anomalous electricity consumption to reduce non-technical losses. A range of machine learning models have been utilized for detecting anomalous electricity consumption and have achieved promising results. However, with the evolution of techniques like electricity theft, coupled with the exponential increase in electricity consumption data, new challenges are constantly being posed for anomalous electricity consumption detection. We propose a Transformer-based method for detecting anomalous electricity consumption. The Transformer is composed of multi-head attention, layer normalization, point-wise feed-forward network, etc., which can effectively handle electricity consumption time-series data. Meanwhile, to alleviate the problem of imbalanced training data between anomalous and normal electricity consumption, we propose a method for synthesizing anomalies. The experimental results demonstrate that our proposed Transformer-based method outperforms the state-of-the-art methods in detecting anomalous electricity consumption, achieving a precision of 93.9%, a recall of 96.3%, an F1-score of 0.951, and an accuracy of 95.6% on a dataset released by the State Grid Corporation of China.

Research on electromagnetic vibration energy harvester for cloud-edge-end collaborative architecture in power grid

With the deepening of the construction of the new type power system, the grid has become increasingly complex, and its safe and stable operation is facing more challenges. In order to improve the quality and efficiency of power grid management, State Grid Corporation continues to promote the digital transformation of the grid, proposing concepts such as cloud-edge-end collaborative architecture and power Internet of Things, for which comprehensive sensing of the grid is an important foundation. Power equipment is widely distributed and has a wide variety of types, and online monitoring of them involves the deployment and application of a large number of power sensors. However, there are various problems in implementing active power supplies for these sensors, which restrict their service life. In order to collect and utilize the vibration energy widely present in the grid to provide power for sensors, this paper proposes an electromagnetic vibration energy harvester and its design methodology based on a four-straight-beam structure, and carries out a trial production of prototype. The vibration pickup unit of the harvester is composed of polyimide cantilevers, a permanent magnet and a mass-adjusting spacer. The mass-adjusting spacer can control the vibration frequency of the vibration unit to match the target frequency. In this paper, a key novel method is proposed to increase the number of turns in a limited volume by stacking flexible coils, which can boost the output voltage of the energy harvester. A test system is built to conduct a performance test for the prototype harvester. According to the test results, the resonant frequency of the device is 100 Hz, the output peak-to-peak voltage at the resonant frequency is 2.56 V at the acceleration of 1 g, and the maximum output power is around 151.7 mu W. The proposed four-straight-beam electromagnetic vibration energy harvester in this paper has obvious advantages in output voltage and power compared with state-of-the-art harvesters. It can provide sufficient power for various sensors, support the construction of cloud-edge-end architecture and the deployment of a massive number of power sensors. In the last part of this article, a self-powered transformer vibration monitor is presented, demonstrating the practicality of the proposed vibration energy harvester.

Study on China’s New Energy Development Trend through Achieving the Consumption Proportion Target for Non-fossil Energy

Recently, the Statistical Communiqué of the People’s Republic of China on the 2022 National Economic and Social Development (hereinafter referred to as "Communiqués") was published by the National Bureau of Statistics. According to the Communiqués, China’s total primary energy consumption in 2022 amounted to 5.41 billion tons of standard coal. It is projected that by 2025, the total energy consumption will reach approximately 5.84-6.04 billion tons of standard coal. In order to meet the target of non-fossil energy consumption accounting for 20% of total energy consumption by 2025, the installed new energy capacity within operating areas of State Grid Corporation of China is anticipated to exceed the previous estimate by 100-140 million kilowatts. From a risk assessment standpoint, the research team has constructed an evaluation model to gauge China’s new energy development trend, taking into account the accomplishment of the consumption proportion target for non-fossil energy. They delve into the issues brought about by the unexpected development of new energy and put forth pertinent measures to tackle them.

A risk evaluation model of electric power cloud platform from the information perspective based on fuzzy type-2 VIKOR

With the swift progression of digital technology, the construction of electric power cloud platform is of great significance for the transformation of the electricity industry. Based on the demonstration of the State Grid Corporation of China's cloud platform, power companies should proactively encourage the establishment of cloud platforms, and the research from the perspective of information risk is imperative before finalizing any decisions related to cloud platforms. Therefore, an index system containing 15 information risk factors is established at first. Secondly, the weight of each index is determined based on GCWM (game combination weighting method). Thirdly, an MCDM (Multi-Criteria-Decision-Making) model based on the fuzzy type-2 VIKOR (from Serbian: VIseKriterijumska Optimizacija I Kompromisno Resenje) algorithm is established. Then, the validity of the model is verified through case study, which reveals that scheme 2 should become a better choice for power enterprises. Finally, sensitivity analysis, comparative analysis, and corresponding management suggestions are also put forward based on the research content. Drawing upon the foundation of the MCDM model, this paper studies the risk evaluation problem of electric power cloud platform from the information perspective, which can provide a certain reference for the development of power enterprise cloud platform in the digital technology environment, as well as bearing relevance for the implementation of associated technologies.

A novel data balancing approach and a deep fractal network with light gradient boosting approach for theft detection in smart grids

Electricity theft is the largest type of non-technical losses faced by power utilities around the globe. It not only raises revenue losses to the utilities but also leads to lethal fires and electric shocks at distribution side. In the past, field operation groups were sent by the utilities to conduct inspections of suspicions electric equipments stated by the public. Advanced metering infrastructure based recent development in the smart grids makes it easy to detect electricity thefts. However, the conventional supervised learning techniques have low theft detection performance mainly due to imbalance datasets available for training. Therefore, in this paper, we develop a novel theft detection model with twofold contribution. A unique hybrid sampling technique named as hybrid oversampling and undersampling using both classes (HOUBC) is proposed to balance the dataset. HOUBC first performs undersampling and then oversampling using both the majority (normal) and minority (theft) classes. A new deep learning method, fractal network is applied with light gradient boosting method to extract and learn important characteristics from electricity consumption profiles for identifying electricity thieves. The proposed model relies on smart meter's data for theft detection and hence, a rapid and widespread adaption of this model is feasible, which shows its main advantage. The performance of the model is evaluated with real-world smart meter's data, i.e., state grid corporation of China. Comprehensive simulation results describe the effectiveness of the proposed model against conventional schemes in terms of electricity theft detection.

Electricity theft detection for energy optimization using deep learning models

AbstractThe rapid increase in nontechnical loss (NTL) has become a principal concern for distribution system operators (DSOs) over the years. Electricity theft makes up a major part of NTL. It causes losses for the DSOs and also deteriorates the quality of electricity. The introduction of advanced metering infrastructure along with the upgradation of the traditional grids to the smart grids (SGs) has helped the electric utilities to collect the electricity consumption (EC) readings of consumers, which further empowers the machine learning (ML) algorithms to be exploited for efficient electricity theft detection (ETD). However, there are still some shortcomings, such as class imbalance, curse of dimensionality, and bypassing the automated tuning of hyperparameters in the existing ML‐based theft classification schemes that limit their performances. Therefore, it is essential to develop a novel approach to deal with these problems and efficiently detect electricity theft in SGs. Using the salp swarm algorithm (SSA), gate convolutional autoencoder (GCAE), and cost‐sensitive learning and long short‐term memory (CSLSTM), an effective ETD model named SSA–GCAE–CSLSTM is proposed in this work. Furthermore, a hybrid GCAE model is developed via the combination of gated recurrent unit and convolutional autoencoder. The proposed model comprises five submodules: (1) data preparation, (2) data balancing, (3) dimensionality reduction, (4) hyperparameters' optimization, and (5) electricity theft classification. The real‐time EC data provided by the state grid corporation of China are used for performance evaluations via extensive simulations. The proposed model is compared with two basic models, CSLSTM and GCAE–CSLSTM, along with seven benchmarks, support vector machine, decision tree, extra trees, random forest, adaptive boosting, extreme gradient boosting, and convolutional neural network. The results exhibit that SSA–GCAE–CSLSTM yields 99.45% precision, 95.93% F1 score, 92.25% accuracy, and 71.13% area under the receiver operating characteristic curve score, and surpasses the other models in terms of ETD.

Benchmarking Generations of You Only Look Once Architectures for Detection of Defective and Normal Long Rod Insulators

Effective infrastructure monitoring is a priority in all technical fields in this century. In high-voltage transmission networks, line inspection is one such task. Fault detection of insulators is crucial, and object detection algorithms can handle this problem. This work presents a comparison of You Only Look Once architectures. The different subtypes of the last three generations (v3, v4, and v5) are compared in terms of losses, precision, recall, and mean average precision on an open-source, augmented dataset of normal and defective insulators from the State Grid Corporation of China. The primary focus of this work is a comprehensive subtype analysis, providing a useful resource for academics and industry professionals involved in insulator detection and surveillance projects. This study aims to enhance the monitoring of insulator health and maintenance for industries relying on power grid stability. YOLOv5 subtypes are found to be the most suitable for this computer vision task, considering their mean average precision, which ranges between 98.1 and 99.0%, and a frame per second rate between 27.1 and 212.8, depending on the architecture size. While their predecessors are faster, they are less accurate. It is also discovered that, for all generations, normal-sized and large architectures generally demonstrate better accuracy. However, small architectures are noted for their significantly faster processing speeds.