Electricity Theft Detection In Smart Grids Based On Artificial Neural Network

Computer Engineering

The rise in electricity demand has spurred the adoption of smart grids, offering benefits like increased energy efficiency, fewer power outages, and heightened security. However, a significant hurdle in smart grid implementation is electricity theft, causing substantial revenue losses for utility providers. Detecting such theft is crucial. Hence, this project aims to devise an efficient method using Artificial Neural Networks (ANN) to identify electricity theft in smart grids. Utilizing electricity usage data sourced from Kaggle, we preprocess the data and input it into the ANN, enabling it to learn consumption patterns and anomalies. By training the model on legitimate consumption data and testing it on instances of theft, we evaluate its efficacy. Our proposed system boasts a Training Accuracy and Validation Accuracy of 99%. Performance metrics like accuracy, precision, recall, and F1-score will be used for evaluation. Additionally, we implement the system using the Flask Web framework for user-friendly interaction and better interface. The anticipated outcome is a robust approach for detecting electricity theft in smart grids, aiding utility companies in revenue enhancement and grid security. This project's methodology could extend to other domains requiring anomaly detection in large datasets, such as fraud detection in finance and intrusion detection in networks.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.