A Naive Bayes Classifier for Detecting Unusual Customer Consumption Profiles in Power Distribution Systems - APSPDCL

Murthy, T. Satyanarayana; Gopalan, N. P.; Ramachandran, Vedantham

doi:10.1109/icisc44355.2019.9036460

Cited by 10 publications

(5 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These detectors can be categorized into two main groups. The first group comprises detectors that utilize shallow ML detection algorithms such as decision trees (DTs), logistic regression (LR), Naïve Bayes, autoregressive integrated moving average (ARIMA), and support vector machines (SVM) [13], [14], [17], [33], [41], [42]. The second group consists of detectors that leverage DL detection algorithms [18], [20], [31].…”

Section: ) Machine Learning-based Methodsmentioning

confidence: 99%

Countering Evasion Attacks for Smart Grid Reinforcement Learning-Based Detectors

El-Toukhy,

Mahmoud,

Bondok

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Fraudulent customers in smart power grids employ cyber-attacks by manipulating their smart meters and reporting false consumption readings to reduce their bills. To combat these attacks and mitigate financial losses, various machine learning-based electricity theft detectors have been proposed. Unfortunately, these detectors are vulnerable to serious cyber-attacks, specifically evasion attacks. The objective of this paper is to investigate the robustness of deep reinforcement learning (DRL)-based detectors against our proposed evasion attacks through a series of experiments. Firstly, we introduce DRL-based electricity theft detectors implemented using the double deep Q networks (DDQN) algorithm. Secondly, we propose a DRL-based attack model to generate adversarial evasion attacks in a black box attack scenario. These evasion samples are generated by modifying malicious reading samples to deceive the detectors and make them appear as benign samples. We leverage the attractive features of reinforcement learning (RL) to determine optimal actions for modifying the malicious samples. Our DRL-based evasion attack model is compared with an FGSM-based evasion attack model. The experimental results reveal a significant degradation in detector performance due to the DRL-based evasion attack, achieving an attack success rate (ASR) ranging from 92.92% to 99.96%. Thirdly, to counter these attacks and enhance detection robustness, we propose hardened DRL-based defense detectors using an adversarial training process. This process involves retraining the DRL-based detectors on the generated evasion samples. The proposed defense model achieves outstanding detection performance, with a degradation in ASR ranging from 1.80% to 9.20%. Finally, we address the challenge of whether the DRL-based hardened defense model, which has been adversarially trained on DRL-based evasion samples, is capable of defending against FGSM-based evasion samples, and vice versa. We conduct extensive experiments to validate the effectiveness of our proposed attack and defense models. INDEX TERMSSecurity, electricity theft, evasion attacks, reinforcement learning, adversarial training, and smart power grids.

show abstract

Section: ) Machine Learning-based Methodsmentioning

confidence: 99%

Countering Evasion Attacks for Smart Grid Reinforcement Learning-Based Detectors

El-Toukhy,

Mahmoud,

Bondok

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Murthy et al [15] conducted a study using data mining techniques to investigate non-technical losses in the power distribution system. Their model consists of two stages; the first stage employs Fuzzy C-Means clustering to group consumers with similar con-sumption profiles, and the second stage applies a finely-tuned Naïve Bayes classification technique to identify potential fraudsters.…”

Section: Supervised False Data Detectormentioning

confidence: 99%

“…The supervised detectors undergo training and testing using both benign and malicious samples. These encompass shallow classifiers such as Naïve Bayes [15] and multiclass SVM [16], as well as deep classifiers like FF-RNN [22] and CNN-LSTM [26]. Conversely, the unsupervised detectors are trained exclusively on benign samples and subsequently tested on datasets comprising both benign and malicious instances.…”

Section: Benchmark Detectorsmentioning

confidence: 99%

Efficient One-Class False Data Detector Based on Deep SVDD for Smart Grids

Habbak,

Mahmoud,

Fouda

et al. 2023

Energies

View full text Add to dashboard Cite

In the smart grid, malicious consumers can hack their smart meters to report false power consumption readings to steal electricity. Developing a machine-learning based detector for identifying these readings is a challenge due to the unavailability of malicious datasets. Most of the existing works in the literature assume attacks to compute malicious data. These detectors are trained to identify these attacks, but they cannot identify new attacks, which creates a vulnerability. Very few papers in the literature tried to address this problem by investigating anomaly detectors trained solely on benign data, but they suffer from these limitations: (1) low detection accuracy and high false alarm; (2) the need for knowledge on the malicious data to compute good detection thresholds; and (3) they cannot capture the temporal correlations of the readings and do not address the class overlapping issue caused by some deceptive attacks. To address these limitations, this paper presents a deep support vector data description (DSVDD) based unsupervised detector for false data in smart grid. Time-series readings are transformed into images, and the detector is exclusively trained on benign images. Our experimental results demonstrate the superior performance of our detectors compared to existing approaches in the literature. Specifically, our proposed DSVDD-based schemes have exhibited improvements of 0.5% to 3% in terms of recall and 3% to 9% in terms of the Area Under the Curve (AUC) when compared to existing state-of-the-art detectors.

show abstract

“…In [24], the authors employed a gradient-boosting theft detector with feature engineering-based preprocessing. The work in [27] developed the APSPDCL NB classifier for detecting unexpected customer consumption trends in power distribution networks. Random forests [28] and AdaBoost [29] were utilized as ETD.…”

Section: A Electricity Theft In Consumption Domainmentioning

confidence: 99%

“…The classification approach [24,25,26,27,28,29], provides a high attack detection rate when a complex learning algorithm, such as deep learning (DL), is utilized to learn data patterns. However, a key bottleneck of this approach is the limited benign and malicious labeled dataset, which limits the ability to test how well the developed models generalize in larger or diverse malicious datasets.…”

Section: Introductionmentioning

confidence: 99%

Fine-Tuned RNN-Based Detector for Electricity Theft Attacks in Smart Grid Generation Domain

Eddin

Albaseer

Abdallah

et al. 2022

IEEE Open J. Ind. Electron. Soc.

View full text Add to dashboard Cite

In this paper, we investigate the problem of electricity theft attacks on smart meters when malicious customers (i.e., adversaries) claim injecting more generated energy into the grid to get more profits from utility companies. These attacks can be applied by accessing the smart meters monitoring renewable-based distributed generation (DG), and manipulating the reading. In this paper, we propose approaches that: (1) rely on data sources with only a single generator (i.e., solar only) and multi-fuel type and (2) address the crucial effects of slight perturbations that the attacker can add, which can deceive the detector. In particular, this paper introduces an efficient multi-task DL-based detector that offers a higher detection rate, copes with different fuel types, and uses only single data sources. The proposed detector incorporates months and days as two additional features to boost the performance and properly guide the model to successful detection. The proposed method is then extended to consider small perturbations that attackers may use to apply successful attacks. We conduct extensive simulations for two different detectors, one for solar DG and the other for multiple fuel types (i.e., solar and wind).Using a realistic dataset, the results reveal that the proposed RNN-based detectors identify adversaries at a higher rate than the existing solutions, even with minimal perturbations and different fuel types.

show abstract

A Naive Bayes Classifier for Detecting Unusual Customer Consumption Profiles in Power Distribution Systems - APSPDCL

Cited by 10 publications

References 8 publications

Countering Evasion Attacks for Smart Grid Reinforcement Learning-Based Detectors

Countering Evasion Attacks for Smart Grid Reinforcement Learning-Based Detectors

Efficient One-Class False Data Detector Based on Deep SVDD for Smart Grids

Fine-Tuned RNN-Based Detector for Electricity Theft Attacks in Smart Grid Generation Domain

Contact Info

Product

Resources

About