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

Abstract

In the advanced metering infrastructure, smart meters are deployed at the consumers' side to regularly transmit fine-grained electricity consumption readings to the system operator (SO) for billing and real-time load monitoring and energy management. However, fraudulent consumers may compromise their meters to launch electricity-theft cyberattacks by reporting low-consumption readings to reduce their bills. These false readings not only cause financial losses but also degrade the grid's performance because they are used for energy management and load estimate. The existing solutions in the literature focus only on securing the billing, so they are not designed to detect the attacks in real time, and thus the SO may use false readings for a long period of time in load monitoring and energy management until they are identified. In this paper, we propose a general ensemble-based deep-learning detector that enables the SO to detect false readings in real time. To do that, we first train several deep learning models on samples generated from a sliding window of the readings. Then, we use the best-performing model to train several models on different ratios of false readings and use them in our ensemble-based detector. Extensive experiments are conducted, and the results indicate that comparing to the literature, our detector can detect the false readings after sending a few false readings (around 15) comparing to the existing daily and weekly detection approaches that need 144 and 1,008 readings, respectively.INDEX TERMS Security, false readings detection, electricity theft, AMI networks, and smart power grid. ABDULAH JEZA ALJOHANI (Senior Member, IEEE) received the B.Sc. (Eng.

Deep learning for cybersecurity in smart grids: Review and perspectives

“…The RF is an output classifier based on the obtained features to detect electricity theft by the consumer. Similarly, [42] developed a CNN-GRU-MLP-based detector to identify malicious customers; [43] designed an AE and bidirectional GRU model to identify anomalies in electricity consumption patterns; [44] proposed an ensemble-based DL detector to improve accuracy and quickly detect false readings.…”

RFE Based Feature Selection and KNNOR Based Data Balancing for Electricity Theft Detection Using BiLSTM-LogitBoost Stacking Ensemble Model

“…As previous both categories of ETD techniques are highlyexpensive, therefore, many researchers have moved towards the third category that is data mining based ETD techniques [11], [20], [21], [22], [23], [32], [33], [34], [35], [36], [37], and [38], to handle the problem of electricity theft in SGs. Authors in [11] propose a deep model for electricity theft and non-theft consumers' classification in SGs.…”

“…However, fine tuning of hyperparameters is ignored that leads a classifier to the local optimal stagnation and low ETD performance. Futhermore, in [36], an ensemble based DL electricity theft detector is developed. The proposed model is the combination of multiple DL models and the outputs of the multiple DL models are passed to the majority voting classifier to calculate the final classification result.…”

Cyber Threats to Smart Grids: Review, Taxonomy, Potential Solutions, and Future Directions

“…He et al [145], proposed the DL-based neural network model to detect FDI attack in terms of bypass the state estimation and causes for congestion of transmission lines in SG. In addition, researchers [146] exploited the ensemble-based DL method to identify the false readings. A couple of DL models are trained based on the samples derived from sliding window of the readings.…”