Research on False Data Injection Attack Detection of S mart Grid Based on Machine Learning

The smart grid combines the classical power system with the information technology, leading to a cyber-physical system. In such an environment, the malicious injection of data has the potential to cause severe consequences. Classical residual-based methods for bad data detection are unable to detect well designed false data injection (FDI) attacks. Moreover, most of the works on FDI attack detection are based on the linearized DC model of the power system and fails to detect attacks based on the AC model. The aim of this paper is to address these problems by using the graph structure of the grid and the AC power flow model. We derive an attack detection method that is able to detect previously undetectable FDI attacks. This method is based on concepts originating from graph signal processing (GSP). The proposed detection scheme calculates the graph Fourier transform of an estimated grid state and filters the graph’s high-frequency components. By comparing the maximum norm of this outcome with a threshold, we can detect the presence of FDI attacks. Case studies on the IEEE 14-bus system demonstrate that the proposed method is able to detect a wide range of previously undetectable attacks, both on angles and on magnitudes of the voltages.

In this study, to enhance the system reliability under false data injection (FDI) attacks and DC-link voltage (DCLV) sensor failures, a hybrid control strategy for a DC microgrid (DCMG) based on the Lagrange extrapolation and voltage observer is proposed. Under normal conditions without FDI attacks or DCLV sensor failures, the DCMG system works in a distributed control scheme. To enhance the reliability of the system under the DCLV sensor failure or FDI attack, the DCMG system utilizes a hybrid control strategy that combines distributed control with decentralized control. The hybrid control strategy is achieved by the proposed detection algorithms for FDI attacks and DCLV sensor failures. The detection of FDI attacks is accomplished by comparing the predicted secondary controller output based on the Lagrange extrapolation with the actual one. When a power agent detects an FDI attack, its control mode is switched to decentralized control by using the proposed hybrid control strategy. The DCLV sensor failure detection algorithm to enhance system reliability against DCLV sensor failures is achieved by comparing the estimated DCLV with the measured one from the voltage observer. Upon detecting a DCLV sensor failure, the operation of the power agent is switched to the current control mode to sustain the system operation even under DCLV sensor failures. The proposed detection algorithms are simple, effective, and precise, operating without mutual interference that deteriorates the detection accuracy. Simulation and experiments are carried out under various uncertain test conditions to validate the reliability and effectiveness of the proposed control strategy.

In the recent years, cyberattacks to smart grids are becoming more frequent. Among the many malicious activities that can be launched against smart grids, the False Data Injection (FDI) attacks have raised significant concerns from both academia and industry. FDI attacks can affect the (internal) state estimation process—critical for smart grid monitoring and control—thus being able to bypass conventional Bad Data Detection (BDD) methods. Hence, prompt detection and precise localisation of FDI attacks are becoming of paramount importance to ensure smart grids security and safety. Several papers recently started to study and analyse this topic from different perspectives and address existing challenges. Data‐driven techniques and mathematical modelling are the major ingredients of the proposed approaches. The primary objective is to provide a systematic review and insights into FDI attacks joint detection and localisation approaches considering that other surveys mainly concentrated on the detection aspects without detailed coverage of localisation aspects. For this purpose, more than 40 major research contributions were selected and inspected, while conducting a detailed analysis of the methodology and objectives in relation to the FDI attacks detection and localisation. Key findings of the identified papers were provided according to different criteria, such as employed FDI attacks localisation techniques, utilised evaluation scenarios, investigated FDI attack types, application scenarios, adopted methodologies and the use of additional data. Finally, open issues and future research directions were discussed.

The work presented in this chapter is an extension of our previous research of bringing together the Cognitive Dynamic System (CDS) and the Smart Grid (SG) by focusing on AC state estimation and Cyber-Attack detection. Under the AC power flow model, state estimation is complex and computationally expensive as it relies on iterative procedures. On the other hand, the False Data Injection (FDI) attacks are a new category of cyber-attacks targeting the SG that can bypass the current bad data detection techniques in the SG. Due to the complexity of the nonlinear system involved, the amount of published works on AC based FDI attacks have been fewer compared to their DC counterpart. Here, we will demonstrate how the entropic state, which is the objective function of the CDS, can be used as a metric to monitor the grid’s health and detect FDI attacks. The CDS, acting as the supervisor of the system, improves the entropic state on a cycle to cycle basis by dynamically optimizing the state estimation process through the reconfiguration of the weights of the sensors in the network. In order to showcase performance of this new structure, computer simulations are carried out on the IEEE 14-bus system for optimal state estimation and FDI attack detection.

The introduction of information and communication technologies makes network environments increasingly open, leaving smart-grid control systems incredibly vulnerable to malicious attacks. False data injection (FDI) attacks stealthily tamper with measurement data, resulting in erroneous decisions made by the control center that greatly influence the normal operation of the power system. By taking advantage of real-time data acquisition with edge computing, in this article, we propose a scheme based on classification of predicted residuals (CPRs) for the FDI attack detection. The CPR scheme first predicts the acquired measurement data at the edge of the sensing network via developing an accurate prediction model. Followed the novel real-time classification method under the edge devices supporting, it classifies the predicted residuals independent of the false data to enhance the detection accuracy. Through these two steps, the detection rate of FDI attacks is greatly improved. The proposed scheme is validated in a real microgrid testbed. Experimental results show that the CPR scheme performs well in detecting FDI attacks and remains sensitive in injection attack probability and magnitude. The detection scheme even has effectiveness at low injection attack probability and magnitude (5% and 0.018 per thousand, respectively). Furthermore, it also proves that the proposed scheme has applicability in high real-time requirements at the edge of smart grids.

This paper mainly studies vector-related learning and pairwise optimization feature selection for false data injection attack detection in smart grid. In order to provide experimental detection of false data injection attacks, this article will conduct a detailed analysis of the power system segmentation. This paper comprehensively considers the similarity of nodes, security control and confidentiality strategies to complete the optimal partition. Next, we preprocess the measurement data. In order to improve the adaptability of the algorithm structure to the grid structure and further improve the accuracy and convergence speed of the algorithm output, a differential evolution algorithm with swarm intelligence is proposed. Obtain a higher-precision state estimate useful for detecting bad data. In the experiment of this article, if false data account for 20% of all data, the detection accuracy exceeds 75%. As the number of experimental groups increases, the detection accuracy of only one type of false data that does not meet the rules will continue to increase, but the detection accuracy of other types of false data will not change much, but the overall detection accuracy will become higher. Experimental results show that the detection framework can not only effectively detect and identify false data injection attacks on multiple bus nodes, but also has high detection accuracy and can effectively recover false data.

In the distributed load frequency control (DLFC) scheme, the compromised areas by attackers may endanger the stable and reliable operation of the entire power system. A credibility‐based secure DLFC strategy is in this work proposed to sustain the stable operation of the power system under false data injection attacks. In this secure DLFC method, the gradient‐variation‐based credibility metrics are defined to detect potential attacked measurements. Then, the communication isolation decisions are made based on the adaptive thresholds. To compensate the impact of the communication isolation of misbehaving areas, predictions of future measurements by using the average gradient variation of stored healthy data are feeded back to the area controller. Simulation results of a four‐area power system show the proposed secure DLFC can identify and cut off the communication of the misbehaving areas and still maintain the stability of the power system.

To detect false data injection attacks (FDIAs) in power grid reconstruction and solve the problem of high data dimension and bad abnormal data processing in the power system, thereby achieving safe and stable operation of the power grid system, this study introduces machine learning methods to explore the detection of FDIAs. First, through the utilization of the standard IEEE node system and the simulation of FDIAs under the condition of non-complete topology information, the construction of the attack data set is completed, and the MatPower tool is applied to simulate and analyze the data set. Second, based on the isolated Forest (iForest) abnormal score data processing algorithm combined with the Local Linear Embedding (LLE) data dimensionality reduction method, an algorithm for data feature extraction is constructed. Finally, based on the combination of the Convolutional Neural Network (CNN) and the Gated Recurrent Unit (GRU) network, an algorithm model for FDIAs detection is constructed. The results show that in the IEEE14-bus node and IEEE118-bus node systems, the overall distribution of the state estimated before and after the attack vector injection is consistent with the initial value. In the iFores algorithm, the number of iTree and the number of samples affect the extraction of abnormal score data. When the number of iTree n is determined to be 100, and the corresponding number of samples w is determined to be 10, the algorithm has the best detection effect. The FDIAs detection algorithm model based on CNN-GRU shows good detection effects under high attack intensity, with an accuracy rate of more than 95%, and its performance is better than other traditional detection algorithms. In this study, the bad data detection model based on deep learning has an active role in the realization of the safe and stable operation of the smart grid.

The integration of information and communication technologies (ICTs) has provided modern smart grids, such as direct current (DC) microgrids (MGs), with various advantages, e.g., the possibility of large-scale integration of renewable energy sources and complex control schemes. This extensive deployment of ICTs, however, makes DC MGs prone to a variety of cyber threats, e.g., false data injection (FDI) attacks. On this basis, this paper develops online detection and mitigation schemes to counter FDI attacks against DC MGs. Initially, in the offline training of the detection scheme, the density-based spatial clustering of applications with noise (DBSCAN) method is used for extracting prescriptive features of DC MG measurements and clustering them in healthy and under-attack operations under several uncertainties in DC MG operations. These features are used to train an adaptive network-based fuzzy inference system (ANFIS) architecture for the online identification of FDI attacks. Afterward, the trained ANFIS is used to detect FDI attacks and obtain the attack vectors benefiting from the real-time measurements estimated by an optimal state observer. Finally, an adaptive model predictive controller (AMPC) is developed to update the power-sharing commands—and mitigate the FDI attack—using the ANFIS’s obtained attack vectors. The effectiveness of the proposed framework is evaluated for several realistic FDI attack scenarios and different uncertainties. To show the scalability of the proposed detection and mitigation methods, a DC MG with eight converters is also studied in the presence of FDI attacks and uncertainties using real-time simulations.

The rapid adoption of the Industrial Internet of Things (IIoT) model has exposed methods to inadequate security measures. False data injection attacks (FDIAs) pose a significant security threat in IIoT, as they aim to mislead industrial platforms by manipulating sensor readings. In IIoT, owing to the data attacker’s changeability, the FDIA is the most critical intrusion network. As aggressive devices permitted in the network perform their typical data-collecting tasks, identifying an FDIA becomes a nontrivial effort due to potential attacks. Conventional attack detection models have proven insufficient in addressing FDIAs, and most current countermeasures focus on the need to legalise data, primarily in the context of data clustering services. Presently, deep learning (DL) systems are employed for detecting FDIA in real-time and supply personalised protective measures to the threat. This paper presents an Intelligent Management of False Data Injection Attacks Using Feature Selection and Voting Classifier (IMFDIA-FSVC) technique in IIoT systems. The primary purpose of the IMFDIA-FSVC technique is to develop a model for detecting and mitigating FDIAs to ensure safe and trustworthy operations of IIoT systems. Initially, the data pre-processing stage involves two stages: missing value analysis and normalisation to standardise the input data for an effective study. For an effective feature selection, the IMFDIA-FSVC model utilises a statistical and information-theoretic selection (SITS) technique to select optimal features from the input data. Finally, the classification process is mainly deployed by three models: the temporal convolutional network (TCN), the deep belief network (DBN), and the autoencoder (AE). An ensemble classifier is then performed using the voting classifier. The comparison study of the IMFDIA-FSVC method showed a superior accuracy value of 99.15% compared to existing models on the IIoT and FDIA datasets.

The modern-day designs of power system have become very complex hardware-software based systems as a result of the rapid development of the Internet of Things (IoT) technology. Designing a smart grid with different smart devices at different location contribute to increase the efficiency of generation, transmission and distribution. State estimation is the most fundamental components in a smart grid to determine the state of operation of the grid with the sensor data and grid topology. It is very important to ensure the authenticity of measurements of the sensor devices in smart grid clusters. The data from such devices that reflect the operation of the grid can be tempered by false data injection attack (FDIA). This paper presents a novel architecture for cluster-based false data injection attack detection and location identification in smart gird. The state vector of the smart grid is represented as a multivariate time series. A vector autoregressive processes are used to predict this FDIA detection and location identification process by exploiting the spatula and temporal correlation of states. Various schemes such as general observation schemes (GOSs) have been proposed to detect and isolate this kind of attacks on a single sensor, but they are inappropriate to detect confrontation when multiple sensors are under attack. Here, we propose a scheme called cluster-based iterative observer scheme (IOS), which uses a mathematical model for smart grid that identifies the confrontations on the smart grid and isolate the attacks on the sensors by clustering the sensors of the smart grid into a small group (subset) of sensors and perform rigorous tests on these group of sensors to isolate the sensors are attacked. At the end of this paper, performance analysis by comparing the results of the proposed algorithm with the existing methods. From the result, the performance of proposed methods is better than conventional methods in terms of false data detection rate, prediction accuracy rate and security. The efficiency of the proposed scheme to identify and quarantine the confrontations in the smart network is shown in simulations results.KeywordsFalse data injection attackSmart gridInternet of thingsRenewable energyState estimationPrediction general observation schemesIterative observer schemeCyber physical systemsKalman filter

False data injection attacks (FDIA) are a main category of cyber-attacks threatening the security of power systems. Contrary to the detection of these attacks, less attention has been paid to identifying the attacked units of the grid. To this end, this work jointly studies detecting and localizing the stealth FDIA in power grids. Exploiting the inherent graph topology of power systems as well as the spatial correlations of measurement data, this paper proposes an approach based on the graph neural network (GNN) to identify the presence and location of the FDIA. The proposed approach leverages the auto-regressive moving average (ARMA) type graph filters (GFs) which can better adapt to sharp changes in the spectral domain due to their rational type filter composition compared to the polynomial type GFs such as Chebyshev. To the best of our knowledge, this is the first work based on GNN that automatically detects and localizes FDIA in power systems. Extensive simulations and visualizations show that the proposed approach outperforms the available methods in both detection and localization of FDIA for different IEEE test systems. Thus, the targeted areas can be identified and preventive actions can be taken before the attack impacts the grid.

The load frequency control (LFC) system is an essential component of the smart grid and is located between cyberspace information and the physical infrastructure. Because the operation of LFC depends on the modern network communication technology, it is vulnerable to network attacks. The false data injection attack (FDIA) injects attack data into the data collected by the LFC system, affecting the control center's calculation and causing errors in the control signals, which affects the stable operation of the power system. For the LFC system, this article builds a cyber-physical model. Based on the model, an FDIA detection and defense mechanism based on the GAN network is proposed. By calculating, contrasting, and replacing the control signals, the proposed method can detect and defend against FDIAs. The experimental results on the two-area interconnected power system, the IEEE 39-bus, and IEEE 118-bus power system show that the proposed defense method can ensure the power system frequency and tie-line power deviation remain within the set value when suffered from FDIA.

Due to the integration of cyber–physical systems, smart grids have faced the new security risks caused by false data injection attacks (FDIAs). FDIAs can bypass the traditional bad data detection techniques by falsifying the process of state estimation. For this reason, this paper studies the detection and isolation problem of FDIAs based on the adaptive Kalman filter bank (AKFB) in smart grids. Taking the covert characteristics of FDIAs into account, a novel detection method is proposed based on the designed AKF. Moreover, the adaptive threshold is proposed to solve the detection delay caused by a priori threshold in the current detection methods. Considering the case of multiple attacked sensor nodes, the AKFB-based isolation method is developed. To reduce the number of isolation iterations, a logical decision matrix scheme is designed. Finally, the effectiveness of the proposed detection and isolation method is demonstrated on an IEEE 22-bus smart grids.

In recent studies, moving target defense (MTD) has been applied to detect false data injection (FDI) attacks using distributed flexible ac transmission system (D-FACTS) devices. However, the inherent conflict between the security goals of MTD (i.e., detecting FDI attacks) and the economic goals of D-FACTS devices (i.e., reducing power losses) would impede the application of MTD in real systems. Moreover, the detection capabilities of existing MTDs are often insufficient. This article proposes a multistage MTD (MMTD) approach to resolve these two issues by adding a group of designed security-oriented schemes before D-FACTS' economy-oriented scheme to detect FDI attacks. We keep these security-oriented schemes for a very short time interval and then revert to the economy-oriented scheme for the remaining time to ensure the economic requirements. We prove that a designed MMTD can significantly improve the detection capability compared to existing one-stage MTDs. We find the supremum of MMTD's detection capability and study its relationship with system topology and D-FACTS deployment. Meanwhile, a greedy algorithm is proposed to search the MMTD strategy to reach this supremum. Simulation results show that the proposed MMTD can achieve the supremum against FDI attacks while outperforming current MTD strategies on economic indicators.