The Challenge of Non-Technical Loss Detection Using Artificial Intelligence: A Survey

Glauner, Patrick; Meira, Jorge Augusto; Valtchev, Petko; State, Radu; Bettinger, Franck

doi:10.2991/ijcis.2017.10.1.51

Cited by 187 publications

(148 citation statements)

References 53 publications

(34 reference statements)

Supporting

Mentioning

146

Contrasting

Unclassified

Order By: Relevance

“…For instance, [9] focuses on feature engineering with random forest, logistic regression and support vector machine as classifiers. A survey on this field is presented in [5], citing also other approaches based on fuzzy systems, genetic algorithms, etc.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Recurrent Dynamical Projection for Time Series-Based Fraud Detection

Antonelo

State

2017

Artificial Neural Networks and Machine Learning – ICANN 2017

Self Cite

View full text Add to dashboard Cite

Abstract. A Reservoir Computing approach is used in this work for generating a rich nonlinear spatial feature from the dynamical projection of a limited-size input time series. The final state of the Recurrent neural network (RNN) forms the feature subsequently used as input to a regressor or classifier (such as Random Forest or Least Squares). This proposed method is used for fraud detection in the energy distribution domain, namely, detection of non-technical loss (NTL) using a real-world dataset containing only the monthly energy consumption time series of (more than 300K) users. The heterogeneity of user profiles is dealt with a clustering approach, where the cluster id is also input to the classifier. Experimental results shows that the proposed recurrent feature generator is able to extract relevant nonlinear transformations of the raw time series without a priori knowledge and perform as good as (and sometimes better than) baseline models with handcrafted features.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Fraudsters alter (or bypass) the eletricity meter in order to pay less than the right amount. Many different methods have been used for devising fraud detection models [4,5]. These methods usually require feature engineering on the consumptions time series data in order to train classifiers for fraud detection.…”

Section: Introductionmentioning

confidence: 99%

Recurrent Dynamical Projection for Time Series-Based Fraud Detection

Antonelo

State

2017

Artificial Neural Networks and Machine Learning – ICANN 2017

Self Cite

View full text Add to dashboard Cite

show abstract

“…Non-technical losses are an important variable in the impairment of the economy of a country, producing loss profits, stability in the power network and environmental damage [4].…”

Section: Introductionmentioning

confidence: 99%

“…In addition, this is a scourge that not only affects developing countries. Developed countries such as the United States or the United Kingdom present estimates of non-technical losses ranging from 1-6 million euros [4]. In India, losses in transmission and distribution systems were around 15% in the period 1966-67, and have increased gradually to 28.36% between 2011-2012 [9].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Identification of the characteristics incident to the detection of non-technical losses for two Colombian energy companies

Sánchez-Zuleta

Fernández-Gutiérrez

Piedrahita-Escobar

2017

Rev. Fac. Ing. Univ. Antioquia

View full text Add to dashboard Cite

ABSTRACT:The study of non-technical losses affecting energy trading companies has guided the researchers' perspective on different techniques and tools that allow them to detect, and why not, to forecast such losses. In the search for a solution to the problem, the different researchers rely on variables that, in many cases, the same marketing companies, from their practical experience, have been considered as incidents in the identification of the problem. However, most of the studies carried out do not support their solutions with the fact that each trading company retains particular data in which both, technical and socio-economic characteristics recorded, are not necessarily shared in their databases. In this work, we follow up on some of the characteristics registered by two Colombian energy trading companies, which serve two different regions of the country in terms of topography and idiosyncrasy. In particular, attention is focused on two characteristics measured in both companies, which by their nature, will always be on the data of any energy trading company: Consumption in kWh, and the period, measured in months. For this purpose, Benford curves analysis, MultiDimensional Scaling (MDS), and hierarchical cluster will be implemented. Finally, it will be studied if the incidence of the variables visualized in the studies presented is reflected in the decision tree model. RESUMEN:El estudio de las pérdidas no técnicas que afectan a las empresas comercializadoras de energía ha orientado la mirada de los investigadores hacia diferentes técnicas y herramientas que les permitan detectar y pronosticar dichas pérdidas. En la búsqueda de una solución al problema los diferentes investigadores se apoyan en variables que, en muchos casos, las mismas empresas comercializadoras, desde su experiencia práctica han determinado como incidentes en la identificación del problema. Sin embargo, la mayor parte de los estudios realizados no anteponen a sus soluciones el hecho de que cada empresa comercializadora registra en su conjunto de datos una serie de características, tanto técnicas como socioeconómicas, que no necesariamente comparten entre ellas. En este trabajo se hace seguimiento a algunas de las características registradas por dos empresas comercializadores de energía colombianas, las cuales atienden dos regiones diferentes del país en cuanto a topografía, e idiosincrasia. De manera particular, se centra la atención en dos características medidas en ambas empresas, y que, por su naturaleza, siempre estarán en los datos de cualquier empresa comercializadora de energía, El Consumo en kWh, y el Periodo, medido en meses. Con este propósito se implementarán análisis de curvas Benford, escalamiento multidimensional MDS, y clúster jerárquico, para finalizar estudiando finalmente si la incidencia de las variables visualizada en los estudios planteados se refleja en el modelo de árboles de decisión.

show abstract

Gradient boosting and Shapley additive explanations for fraud detection in electricity distribution grids

Santos¹,

Yamouni²,

Albiero³

et al. 2021

Int Trans Electr Energ Syst

View full text Add to dashboard Cite

Fraud in electrical energy consumption represents a critical economic burden for utility companies around the world. Despite systematic efforts to mitigate electricity theft, this practice persists mostly in developing countries where companies rely on traditional detection methods. In Brazil it is estimated that around 7% of the total electrical energy available for consumption in 2016 was lost due to frauds. Here we describe an efficient and scalable system to predict fraudulent behavior and guide in loco inspections. We compared the performances of several machine learning algorithms using consumption and inspection data provided by CPFL Energia. We show that proper feature engineering and boosted classification trees trained with XGBoost are able to extract patterns related to fraud occurrence and to achieve predictive power of practical consequences. Moreover, we demonstrate how Shapley additive explanation (SHAP) values can be employed to build user friendly explanations. Together, the proposed model and its explainers contribute not only to reveal potentially fraudulent behavior but also to understand root causes, what can be used to devise robust mitigation strategies.

show abstract

The Challenge of Non-Technical Loss Detection Using Artificial Intelligence: A Survey

Cited by 187 publications

References 53 publications

Recurrent Dynamical Projection for Time Series-Based Fraud Detection

Recurrent Dynamical Projection for Time Series-Based Fraud Detection

Identification of the characteristics incident to the detection of non-technical losses for two Colombian energy companies

Gradient boosting and Shapley additive explanations for fraud detection in electricity distribution grids

Contact Info

Product

Resources

About