Employing a Machine Learning Boosting Classifiers Based Stacking Ensemble Model for Detecting Non Technical Losses in Smart Grids

Abstract

In the modern world, there are numerous opportunities that help in the detection of electricity theft happening in the realm of electricity grids due to the widespread shifting of people from old metering infrastructure to advanced metering infrastructure (AMI). It is done by studying the consumers' energy consumption (EC) readings provided by the smart meters (SM). The literature introduces a variety of machine learning (ML) and deep learning (DL) strategies to use EC data for identifying power theft in smart grids (SGs). However, the existing schemes provide low performance in terms of electricity theft detection (ETD) due to the usage of imbalanced data and using schemes in an individual manner. Moreover, the existing detectors are validated using a limited number of performance evaluation measures, which are not suitable for conducting model's comprehensive validation. To tackle the above mentioned problems, an ML boosting classifiers based stacking ensemble model (MLBCSM) is proposed followed by adaptive synthetic sampling technique (ADASYN) in the underlying work. Data preprocessing followed by data balancing and classification are the three major parts of the model introduced in this work. Besides, the EC data acquired from the consumers' SMs is used for detecting electricity theft. Moreover, the simulation results reveal that MLBCSM combines the benefits of adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), histogram boosting (HistBoost), categorical boosting (CatBoost), and light gradient boosting (LGBoost). Additionally, the model's validation is ensured via different metrics. It is deduced via extensive simulations that the proposed model's outcomes are superior to those produced by the individual models in terms of ETD.

Electricity theft detection for energy optimization using deep learning models

“…In boosting ensemble category, AdaBoost was the first model 58 . Furthermore, in gradient boosting, weak classifiers are trained employing a gradient descent optimizer and differentiable loss calculation function 56 . XGBoost is computationally efficient as compared with the gradient boosting classifier 58 .…”

“…The boosting rounds are equal to the number of weak learners (base estimators) in a boosting technique. In the first boosting iteration in AdaBoost, the first weak classifier is trained and prediction results are generated 56 . In addition, the error is computed for the first round.…”

“…58 Furthermore, in gradient boosting, weak classifiers are trained employing a gradient descent optimizer and differentiable loss calculation function. 56 XGBoost is computationally efficient as compared with the gradient boosting classifier. 58 Recently, it is proposed by Yan and Wen, 13 to detect the electricity theft using the metering data present in the Irish data set.…”

“…In the first boosting iteration in AdaBoost, the first weak classifier is trained and prediction results are generated. 56 In addition, the error is computed for the first round. In the second iteration, the misclassified records are weighted more as compared with the correctly classified observations in the previous round.…”

Exploiting machine learning to tackle peculiar consumption of electricity in power grids: A step towards building green smart cities

“…There is a serious economic loss associated with electricity theft in the existing power grids due to the NTLs. Electricity theft is a significant problem in many countries [3][4][5]. Annually, the United States loses $6 billion to electricity theft.…”