Detection of False-Reading Attacks in Smart Grid Net-Metering System

Abstract

In the advanced metering infrastructure (AMI) network of the smart power grid, smart meters (SMs) are installed at the customers' premises to report their fine-grained power consumption readings to the utility for billing and load monitoring purposes. Moreover, to create a clean power system, customers install solar panels on their rooftops to generate power and sell it to the utility. However, malicious customers may compromise their SMs to report false readings to achieve financial gains illegally. Reporting false readings not only causes hefty financial losses to the utility but may also degrade the grid performance because the reported readings are used for energy management. This paper is the first work that investigates this problem in the net-metering system, in which one SM is used to report the difference between the power consumed and the power generated. First, we prepare a benign dataset for the net-metering system by processing a real power consumption and generation dataset. Then, we propose a new set of attacks tailored for the net-metering system to create malicious dataset. After that, we analyze the data and we found time correlations between the net meter readings and correlations between the readings and relevant data obtained from trustworthy sources such as the solar irradiance and temperature. Based on the data analysis, we propose a general multi-data-source deep hybrid learning-based detector to identify the false-reading attacks. Our detector is trained on net meter readings of all customers besides data from the trustworthy sources to enhance the detector performance by learning the correlations between them. The rationale here is that although an attacker can report false readings, he cannot manipulate the solar irradiance and temperature values because they are beyond his control. Extensive experiments have been conducted, and the results indicate that our detector can identify the false-reading attacks with high detection rate and low false alarm.

AI explainability and governance in smart energy systems: A review

“…The work has identified a range of specific and broad challenges including resource allocation in wireless sensor networks with multiple UAVs (Seid et al, 2021;, governance of AI in power-related general-purpose technologies (Niet et al, 2021;Przhedetsky, 2021;Nitzberg and Zysman, 2022), fault detection, fault diagnosis, and anomaly detection in smart energy systems (Sun et al, 2021;Badr et al, 2022), edge computing for detecting power demand attacks (Alagumalai et al, 2022;Haseeb et al, 2022;Zhu et al, 2022), blockchain-based reliability and security (Al-Abri et al, 2022;Jose et al, 2022), governance of energy markets and energy pipeline systems (Belinsky and Afanasev, 2021;Serna Torre and Hidalgo-Gonzalez, 2022;Sun et al, 2022), forecasting short-term energy demand (Xie et al, 2021;Gürses-Tran et al, 2022), energy trading using federated learning in smart cities (Bracco et al, 2022), energysaving edge AI applications (Khosrojerdi et al, 2022), performance optimization and stability of smart grid operations and nuclear power systems (Luo et al, 2021;Volodin and Tolokonskij, 2022), and others. All these areas are candidates for future research.…”

“…It captures various dimensions of "Anomaly Detection and Security," including applying XAI methods to identify conductive galloping in power grids (Sun et al, 2021), achieve transparency of fault diagnosis in electrical grids (Ardito et al, 2022), and monitoring data poisoning attacks in smart grids using white-box and black-box analysis (Bhattacharjee et al, 2022). Other dimensions include identifying erroneous measurements in the smart grid measuring system using trustworthy data sources (Badr et al, 2022), improving distributed denial-of-service (DDoS) security of software defined networks (SDN)-based smart grids to increase security and trustworthiness (Nagaraj et al, 2021), fault detection (Landwehr et al, 2022), anomaly classification for power consumption data in smart grid (Bhattacharjee et al, 2021), anomaly attacks detection in power networks, and ML and DL models for fault detection in power systems.…”

Real-Time Detection of False Readings in Smart Grid AMI Using Deep and Ensemble Learning

“…Unlike these papers, we focus on securing both billing and energy management application, so we designed our detector to detect false readings fast by using multiple models and methods, such as deep and ensemble learning, training models on different ratios of false readings, hyperparameter optimization, etc. The results of Experiment 3 indicate that our detector can effectively detect the false readings after sending a few false readings (15 readings) comparing to the daily detection approaches [3], [14]- [16] and the weekly detection approaches [11], [17] that need 144 and 1,008 readings, respectively, to detect the attack.…”

“…As will be explained in details in Section VI, the existing papers in the literature focus on securing the billing, so some detectors such as [3], [14]- [16] are designed to detect electricity theft daily by learning and processing the daily energy consumption, while other detectors such as [11], [17] are designed to detect electricity theft weekly by processing the weekly consumption. Unlike these papers, we focus on securing both billing and energy management application, so we designed our detector to detect false readings fast by using multiple models and methods, such as deep and ensemble learning, training models on different ratios of false readings, hyperparameter optimization, etc.…”

“…However, the existing solutions in the literature focus only on securing the billing, so they are not designed to detect fraudulent consumers in real time, and thus, the SO may use false readings for a long period in load monitoring and energy management until they are identified. Specifically, the existing solutions aim to detect electricity theft every day [3], [14]- [16] or week [11], [17] by processing the daily or weekly consumption patterns.…”