Abstract:Detection of abnormal power consumption is very important to reduce irregular power consumption and economic losses. To reduce power wastage, it is necessary to understand the power consumption habits of users and detect unusual usage behavior promptly. Researchers have developed various methods to detect abnormal electricity consumption behavior. However, it does not provide enough consideration to identify the relationships between different time stamps and multi-scale characteristics of the power utilization series. Therefore, a grey relation projection–based better clustering learning technique is proposed for assessing power consumption abnormalities. Power data from a power consumption information gathering system is analyzed with a diagnosis model for anomalous electricity use by customers constructed using big data techniques and machine learning-related algorithms. Density-based clustering is offered as a means of analyzing instances of excessive electricity use. First, consumers in the same area are grouped based on the similarities between their electricity usage patterns. Then, using density clustering, we identify the most extreme cases of electricity use amongst consumers who are otherwise comparable. Using a comparable user electricity consumption framework and the historical electricity usage approach, we can determine the degree to which these anomalies match the norm. Finally, the threshold value allows for discrimination of anomalous electricity usage with the level of comprehensive support. The proposed DBSCAN approach attains the PR-AUC values of 0.96, and AUC attains 0.911, results from simulations demonstrate that this approach can detect aberrant patterns of electricity consumption.

Fast Dynamic Density Outlier Detection Algorithm for Power Quality Disturbance Data.

Anomaly Detection Algorithm Based on Electric Equipment

Outlier Detection Method Based on Compressed Time Series and Voronoi Diagram for Power Data

A Fast Density Peak Clustering Method for Power Data Security Detection Based on Local Outlier Factors

Detection and Classification of Anomalies in Power Distribution System Using Outlier Filtered Weighted Least Square

An Improved Data-Efficiency Algorithm Based on Combining Isolation Forest and Mean Shift for Anomaly Data Filtering in Wind Power Curve

A New Outlier Detection Algorithm Based on Fast Density Peak Clustering Outlier Factor

An Method for Anomaly Detection of State Information of Power Equipment Based on Big Data Analysis

Power Equipment Anomaly Detection Based on Spatiotemporal Clustering

Analysis of short time power quality disturbances based on statistic feature vector symbolic and ensemble empirical mode decomposition

Anomaly Detection for Power Quality Analysis Using Smart Metering Systems

A New Classification Method for Transient Power Quality Combining Spectral Kurtosis with Neural Network

An Outlier Detection Technique Based on Spectral Clustering

Method for Identifying Abnormal Power Metering Data in Shared Power Information Collection System

Enhancing power utilization analysis: detecting aberrant patterns of electricity consumption

A Spectral Clustering Based Outlier Detection Technique.

Anomaly Subsequence Detection with Dynamic Local Density for Time Series

Dynamic Micro-cluster-Based Streaming Data Clustering Method for Anomaly Detection.

Autoencoder and Kernel Density Estimation-Based Anomaly Detection Method for Power Consumption among Distribution Network Customers

Research on Anomaly Detection Model for Power Consumption Data Based on Time-Series Reconstruction

Detecting Anomalous Users via Streaming Data Processing in Smart Grid