Enhancing Power System Cyber-Security With Systematic Two-Stage Detection Strategy

In order to realize the position detection of false data injection, a position detection model based on bi-directional gated cycle unit optimized full convolution neural network is proposed. In view of the problem that the traditional false data injection detection is limited to judging whether the attack occurs or not, the processing of multi-level detection is proposed to realize the fine discrimination of the attacked data and effectively avoid the resource loss caused by direct location detection. According to the time series characteristics of the measured data, the bi-directional gated cycle unit is applied to the convolution layers of the full convolution neural network, which can effectively extract the spatiotemporal features between the data and improve the accuracy and efficiency of false data injection detection. The experimental results show that the optimized position detection model has a certain improvement in detection accuracy and efficiency compared with the traditional convolution neural network.

With the advent of smart grid concept, Internet of Things (IoT) and the deployment of smart meters, the cyber-attack threats on power networks have increased due to the use of communication systems that can be accessed by adversaries. Attackers will have the ability to manipulate the outcomes of smart meters which in turn influence the core application of Energy Management System (EMS): State Estimation (SE). Bad data analytic tools may fail to detect some attacks into measurements. Meanwhile, Machine Learning (ML) solutions have been proposed for detecting False Data Injection (FDI) attacks. However, there is a lack of ML time-series solutions presented in the state-of-the-art that is yet to be less complex. In signal processing, time-series solutions does not only consider the signal, but also the statistics of the signal over time. Therefore, in this paper, a machine learning for time-series solutions is presented as an application to model the measurements of the power grid that are used in SE. The presented model takes into account adaptive linear and non-linear filters: Finite Impulse Response (FIR), and Infinite Impulse Response (IIR). The presented models are implemented and performed on the IEEE-118 bus system. The results indicate the advantage of applying those filters over the state-of-the-art machine learning solutions.

This study investigates a false data injection detection method in an automatic data acquisition platform for electric energy measurement with the aim of ensuring the stability and security of the power system. A fake data injection detector model, including a predictor and discriminator, was constructed. The predictor was based on a gradient boosting decision tree to predict potential data injection anomalies. The discriminator used a multilayer perceptron (MLP) neural network, combined with difference analysis between the predicted and actual values, to determine false data injection. The improved Cauchy mutation grey Wolf optimization algorithm is used to optimize the model training to improve the detection accuracy. Experimental results show that the proposed method can efficiently process a large number of packets per second, the ROC curve performance is excellent, the detection accuracy is more than 99.97%, and the delay is less than 0.04 s. This research provides an effective false data injection detection technology for an electric energy measurement data collection platform, significantly enhances the security protection ability of the system, and has important practical application value.

The use of knowledge-based systems (KBS) in the development of energy management system (EMS) applications is addressed. Since the user interaction principles in today's EMSs are not considered good enough, a tentative specification for the next generation of EMS applications is proposed. The justification for the use of KBS technology in power system operation is discussed, and a prototype KBS development environment is presented. The goal of this project was to get security analysis tools into daily operation through a user-friendly and powerful man-machine interface. Dedicated hardware (LISP-machine) and sophisticated software tools (expert system shell) were used in the KBS development environment. >

International audience

In this paper, we consider the problems of state estimation and false data injection detection in smart grid when the measurements are corrupted by colored Gaussian noise. By modeling the noise with the autoregressive process, we estimate the state of the power transmission networks and develop a generalized likelihood ratio test (GLRT) detector for the detection of false data injection attacks. We show that the conventional approach with the assumption of Gaussian noise is a special case of the proposed method, and thus the new approach has more applicability. The proposed detector is also tested on an independent component analysis (ICA) based unobservable false data attack scheme that utilizes similar assumptions of sample observation. We evaluate the performance of the proposed state estimator and attack detector on the IEEE 30-bus power system with comparison to conventional Gaussian noise based detector. The superior performance of both observable and unobservable false data attacks demonstrates the effectiveness of the proposed approach and indicates a wide application on the power signal processing.

Blockchain technology provides considerable promise to enhance security and transparency in energy management systems. As energy systems grow decentralized and involve various stakeholders, challenges including data integrity, transaction security, and transparency are becoming crucial. Traditional energy systems have risks such as cyberattacks, data manipulation, and inefficiencies owing to centralized control. Blockchain’s decentralized, immutable, and transparent ledger system can overcome these difficulties by enabling safe, verifiable, and tamper-proof energy transactions. Smart contracts further automate and ensure trust in energy trades. This article analyzes blockchain’s application in energy management systems, concentrating on decentralized energy trade, real-time grid balancing, and transparent billing. It also addresses the security implications of blockchain, its limitations, and future potential. While scalability and energy efficiency challenges persist, blockchain's role in constructing more resilient, secure, and transparent energy infrastructures shows promise for the future of energy management.

A cyberattack detection model based on supervised deep neural network is proposed to identify random false data injection (FDI) on the tank’s level measurements of a water distribution system. The architecture of the neural network, as well as various hyper-parameters, is modified and tuned to acquire the highest detection performance using the smallest size of training data set. The efficacy of the proposed detection model against various activation functions including sigmoid, rectified linear unit, and softmax is examined. Regularization and momentum techniques are applied to update the weights and prohibit overfitting. Moreover, statistical metrics are presented to evaluate the performance and effectiveness of the proposed model in the presence of a range of measurement noise levels. The proposed model is tested for three attack scenarios composed for the battle of the attack detection algorithms. Results confirm that the size of the data sets required to train the neural network (NN) to accomplish the highest levels of accuracy and precision is significantly decreased as the number of hidden layers is increased. The trained 4- and 5-layer deep neural networks are able to detect the readings’ FDIs with 100% precision and accuracy in the presence of 30% background noise in the sensory data.

Deep learning-based model predictive control with exponential weighting strategy and its application in energy management systems

This paper reviews the work in the areas of machine learning applications for energy management in smart buildings, 5G technology’s role in smart energy management, and the use of machine learning algorithms in microgrid energy management systems. The first area focuses on the adaptability of building-integrated energy systems to unpredictable changes through AI-initiated learning processes and digital twins. The second area explores the impact of 5G technology on smart buildings, particularly in Singapore, emphasizing its role in facilitating high-class services and efficient functionalities. The third area delves into the application of various machine learning algorithms, such as supervised and unsupervised learning, in managing and monitoring microgrids. These broad areas collectively offer a comprehensive understanding of how machine learning can revolutionize energy management systems in smart buildings, making them more efficient, adaptable, and sustainable.

This paper describes a new method for enhancing the accuracy of network analysis applications in energy management systems. Topology processing, state estimation, power flow analysis, and contingency analysis play a key factor in the stable and reliable operation of power systems. In this respect, the aim of topology processing is to provide the electrical buses and the electrical islands with the actual state of the power system as input data. The results of topology processing is used to input of other applications. New method, which includes the topology error analysis based on inconsistency check, coherency check, bus mismatch check, and outaged device check is proposed to enhance the accuracy of network analysis. The proposed methodology is conducted by energy management systems and the Korean power systems have been utilized for the test systems.

Considering the global energy crisis and the need to reduce energy consumption while providing thermal comfort to occupants, building performance prediction using building simulation programs requires higher accuracy of output data. Therefore, it seems necessary to study the impact of occupant behavior, which is the main source of uncertainty in residential buildings. The traditional courtyard houses, which are recognized as a successful passive house model, respond to different climatic conditions. Therefore, this research focuses on this building type to analyze occupant window opening control scenarios and determine which control works better. For this purpose, several probabilistic controls and their effects on the adaptive thermal comfort of occupants in zones around a central courtyard were compared in the three cities of Yazd, Bandar Abbas, and Tabriz. Energy Plus was used as a simulation program for the application of Grasshopper's energy management system (EMS) along with the Ladybug and Honeybee environmental plugins. The results show that the window control algorithms can increase the adaptive thermal comfort of occupants by 25.7%, 32.2%, and 20.3% in each of the climates of Yazd, Bandar Abbas, and Tabriz cities, respectively. Indoor and outdoor temperature were the most significant variables for opening windows in the warm and cold seasons, respectively.

During this period, as distributed energy resources (DERs) are crucial for meeting energy needs and renewable technology advances rapidly, photovoltaic (PV)-powered energy islands (EIs) requiring a constant energy supply have emerged. EIs represent a significant milestone in the global energy transformation towards clean and sustainable energy production. They play a vital role in the future energy infrastructure, offering both environmental and economic benefits. In this context, reliance on information and communication technologies for system management raises concerns regarding the cybersecurity vulnerabilities of PV-supported EIs. In other words, since EIs transmit power through power converters—integral cyber-physical components of these systems—they are uniquely susceptible to cyber-attacks. To tackle this vulnerability, a cyber-attack detection scheme using a supervised machine learning (SML) model is proposed. The initial goal is to ensure the transfer and maintenance of energy demands without power loss for critical loads by detecting cyber-attacks to establish a defense mechanism. Two distinct artificial neural network (ANN) structures are implemented to identify cyber threats and support subsequent power demand, resulting in a complementary approach. The findings reveal the model’s effectiveness, demonstrating high accuracy (e.g., a cross-entropy loss of 12.842 × 10−4 for ANN-I with a 99.98% F1 score and an MSE of 1.0934 × 10−7 for ANN-II). Therefore, this work aims to open the fundamental way for addressing this issue, particularly concerning hijacking attacks and false data injection (FDI) cyber-attacks on PV-powered EIs. The success of this model and its outcomes confirm the effectiveness of the proposed approach method.

Smart grid represents a computerized platform for two-way communications between system utilities and customers through digital and information technologies. Information, which is crucial for cyber-physical systems, is exchanged between power system utilities and smart homes through communication infrastructures. The more digitization of power system, the more vulnerability to various security threats such as False Data Injection (FDI) cyberattacks. The significance of the FDI cyberattacks is inevitable for some sectors such as homes, medical services, industry, etc. In this paper, attacker is considered to 1) gain access to the information of smart home and 2) manipulate the electricity demand of the smart home such that its overall cost is not significantly affected. The objective function and the operation of the smart home with/without FDI cyberattack are compared. The results demonstrate that attacker is able to successfully affect the performance of smart home by changing the shape of demand.

This paper presents a framework to perform experimental validation of a remedial action for a lab-scale microgrid targeted by false data injection (FDI) cyberattack. The microgrid, which is located at Southern Illinois University (SIU), Carbondale, IL, USA, is managed to operate based on an electricity market mechanism. The corresponding renewable-based distributed generation (DG) owners submit their packages (i.e., power/bid) to be evaluated by market operator aiming at minimizing the total cost. Targeting the system by an FDI cyberattack, an attacker manipulates the submitted package associated with one of the DG owners at a specific time. After experiencing the attack, the operator reacts to the suspicious DG via excluding it from the market. Then, an invitation signal will be submitted to the non-attacked DG owners such that they can rapidly re-dispatch their generation in a secondary market to cope with the shortage of supply associated with the attacked DG.