This study presents a novel feature-engineered–natural gradient descent ensemble-boosting (NGBoost) machine-learning framework for detecting fraud in power consumption data. The proposed framework was sequentially executed in three stages: data pre-processing, feature engineering, and model evaluation. It utilized the random forest algorithm-based imputation technique initially to impute the missing data entries in the acquired smart meter dataset. In the second phase, the majority weighted minority oversampling technique (MWMOTE) algorithm was used to avoid an unequal distribution of data samples among different classes. The time-series feature-extraction library and whale optimization algorithm were utilized to extract and select the most relevant features from the kWh reading of consumers. Once the most relevant features were acquired, the model training and testing process was initiated by using the NGBoost algorithm to classify the consumers into two distinct categories (“Healthy” and “Theft”). Finally, each input feature’s impact (positive or negative) in predicting the target variable was recognized with the tree SHAP additive-explanations algorithm. The proposed framework achieved an accuracy of 93%, recall of 91%, and precision of 95%, which was greater than all the competing models, and thus validated its efficacy and significance in the studied field of research.
19 August 2020 in International Transactions on Electrical Energy Systems
This paper presents a novel, sequentially executed supervised machine learning‐based electric theft detection framework using a Jaya‐optimized combined Kernel and Tree Boosting (KTBoost) classifier. It utilizes the intelligence of the XGBoost algorithm to estimate the missing values in the acquired dataset during the data pre‐processing phase. An oversampling algorithm based on the Robust‐SMOTE technique is utilized to avoid the unbalanced data class distribution issue. Afterward, with the aid of few very significant statistical, temporal, and spectral features extracted from the acquired kWh dataset, the complex underlying data patterns are comprehended to enhance the accuracy and detection rate of the classifier. For effectively classifying the consumers into “Honest” and “Fraudster,” the ensemble machine learning‐based classifier KTBoost, with Jaya algorithm optimized hyperparameters, is utilized. Finally, the developed model is re‐trained using a reduced set of highly important features to minimize the computational resources without compromising the performance of the developed model. The outcome of this study reveals that the proposed theft detection method achieves the highest accuracy (93.38%), precision (95%), and recall (93.18%) among all the studied methods, thus signifying its importance in the studied area of research.
24 September 2022 in International Transactions on Electrical Energy Systems
This paper proposes parameter-free two best-worst optimizers (BWOs) that combine the searching capabilities of Jaya and Rao-1 algorithms to intensify their exploration and exploitation capabilities. For the proposed first optimizer, BWO-1, parallel taxonomy has been adopted for the Jaya and Rao-1 combination to obtain dual sets of updated solutions for a given current solution set. For the proposed second optimizer, BWO-2, along with the parallel taxonomy, a subloop of random scaling factors has been introduced in the solution updating mechanism of both standard Jaya and Rao-1 algorithms to generate multiple sets of updated solutions in a single iteration. The best solutions from obtained multiple solution-sets will survive only and lead to the next iteration. Hence, the proposed solution updating mechanisms for BWO-1 and BWO-2 increase the probability of getting a quality solution under set conditions. The ability and scalability of BWO-1 and BWO-2 were assessed, solving the optimization problem of power loss reduction and voltage deviation minimization in the IEEE 33-bus and 69-bus test systems. The results show that the BWO-2 exhibited a lead of 2.97%, 1.83%, and 0.72% over the Rao-1, Jaya, and BWO-1 techniques, respectively. Besides, the BWO-2 achieves up to 38.76% more reduction in power losses against the existing standard, improved, and hybrid optimization techniques.
1 June 2022 in International Journal of Applied Power Engineering (IJAPE)
Available transfer capability (ATC) is an important metric used to measure the techno-economic viability of the transmission networks. Several methods have been presented in literature for ATC assessment, however, only some few articles incorporate CBM in ATC calculation and those few papers only considered conventional power generation sources in CBM evaluation. CBM is a function of the reliability of generating units. This paper presents the inter-area CBM calculation in the presence of wind energy source using graph theory technique and the results are incorporated in ATC computation using repeated power flow. CBM is incorporated in ATC computation as non-recallable and recallable power transfers. The results with and without CBM incorporation, in the presence of wind power system, are compared. The contribution of the paper is to study the influence of wind power (WP) integrated CBM on ATC evaluation. The results showed that the incorporation of CBM, in the presence of renewable energy, has a significant influence on ATC, without which ATC would be inaccurately estimated. Simulations using IEEE 24-Bus RTS is used to implement the proposed approach. This approach can be employed by transmission operators to assess the technoeconomic viability of the power network for possible power transactions.
About Me
Saddam Hussain is currently pursuing a Ph.D. at University Teknologi Malaysia (UTM) Johor Bahru, Malaysia.
Saddam Hussain is currently pursuing a Ph.D. at University Teknologi Malaysia (UTM) Johor Bahru, Malaysia.