Abstract:Anomaly detection is a critical task that involves the identification of data points that deviate from a predefined pattern, useful for fraud detection and related activities. Various techniques are employed for anomaly detection, but recent research indicates that deep learning methods, with their ability to discern intricate data patterns, are well-suited for this task. This study explores the use of Generative Adversarial Networks (GANs) for anomaly detection in power generation plants. The dataset used in this investigation comprises fuel consumption records obtained from power generation plants operated by a telecommunications company. The data was initially collected in response to observed irregularities in the fuel consumption patterns of the generating sets situated at the company's base stations. The dataset was divided into anomalous and normal data points based on specific variables, with 64.88% classified as normal and 35.12% as anomalous. An analysis of feature importance, employing the random forest classifier, revealed that Running Time Per Day exhibited the highest relative importance. A GANs model was trained and fine-tuned both with and without data augmentation, with the goal of increasing the dataset size to enhance performance. The generator model consisted of five dense layers using the tanh activation function, while the discriminator comprised six dense layers, each integrated with a dropout layer to prevent overfitting. Following data augmentation, the model achieved an accuracy rate of 98.99%, compared to 66.45% before augmentation. This demonstrates that the model nearly perfectly classified data points into normal and anomalous categories, with the augmented data significantly enhancing the GANs' performance in anomaly detection. Consequently, this study recommends the use of GANs, particularly when using large datasets, for effective anomaly detection.

Online Anomaly Detection in DC/DC Converters by Statistical Feature Estimation Using GPR and GA

Online Anomaly Detection in Switching-Mode Power Module Using Statistical Property Features Comparison and Gaussian Process Regression

Power Transformer Fault Diagnosis Using Svm With Genetic Algorithms Optimization

Anomaly Detection Algorithm Based on Electric Equipment

A Statistical Feature-Based Anomaly Detection Method for PFC Using Canonical Correlation Analysis

A Methodology for Transformer Fault Diagnosis Based on the Feature Extraction from DGA Data

A Self-Commissioning Edge Computing Method for Data-Driven Anomaly Detection in Power Electronic Systems

A Methodology to Diagnose Transformer Faults Based on Principal Components Analysis and Artificial Neural Network

Dtasd: A Novel Online Detection Method for Anomalous State of Dry-type Transformer

Research on Anomaly Detection Method of Online Monitoring Data of Dissolved Gas in Transformer Oil.

Fault Detection and Classification of DC-DC Converters Based on Permutation Entropy

Deep learning anomaly detection in AI-powered intelligent power distribution systems

Anomaly Detection in Power Generation Plants with Generative Adversarial Networks

Incipient Fault Diagnosis Method for DC–DC Converters Based on Sensitive Fault Features

A Semi-Supervised Learning Framework Based on EMD and SGAN for Online Stability Monitoring of DC Power Electronic Systems

An Unsupervised Anomaly Detection Approach Based On Industrial Big Data

An Identification Method for Anomaly Types of Active Distribution Network Based on Data Mining

A Robust and Explainable Data-Driven Anomaly Detection Approach For Power Electronics

Anomaly Detection Based on Random Matrix Theory for Industrial Power Systems.

Integrated Generative Adversarial Network and XGBoost for Anomaly Processing of Massive Data Flow in Dispatch Automation Systems