eess.SYFeb 16, 2026

Fault Detection in Electrical Distribution System using Autoencoders

Authors: Sidharthenee Nayak, Victor Sam Moses Babu, Chandrashekhar Narayan Bhende, Pratyush Chakraborty, Mayukha Pal

Organizations: ABB Ability Innovation Center, Asea Brown Boveri Company, Hyderabad, 500084, India. · School of Electrical Sciences, Indian Institute of Technology Bhubaneswar, Bhubaneswar, 751013, India. · Department of Electrical and Electronics Engineering, BITS Pilani Hyderabad Campus, Hyderabad, 500078, India.

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

In recent times, there has been considerable interest in fault detection within electrical power systems, garnering attention from both academic researchers and industry professionals. Despite the development of numerous fault detection methods and their adaptations over the past decade, their practical application remains highly challenging. Given the probabilistic nature of fault occurrences and parameters, certain decision-making tasks could be approached from a probabilistic standpoint. Protective systems are tasked with the detection, classification, and localization of faulty voltage and current line magnitudes, culminating in the activation of circuit breakers to isolate the faulty line. An essential aspect of designing effective fault detection systems lies in obtaining reliable data for training and testing, which is often scarce. Leveraging deep learning techniques, particularly the powerful capabilities of pattern classifiers in learning, generalizing, and parallel processing, offers promising avenues for intelligent fault detection. To address this, our paper proposes an anomaly-based approach for fault detection in electrical power systems, employing deep autoencoders. Additionally, we utilize Convolutional Autoencoders (CAE) for dimensionality reduction, which, due to its fewer parameters, requires less training time compared to conventional autoencoders. The proposed method demonstrates superior performance and accuracy compared to alternative detection approaches by achieving an accuracy of 97.62% and 99.92% on simulated and publicly available datasets.

Figures & tables

Figure 1: Flow chart of the proposed fault detection process
Figure 2: Architecture of autoencoder used in the proposed algorithm
Figure 3: Detailed layers of the CNN-based autoencoder model
MetricFormula
AccuracyAs=TP+TNTP+TN+FP+FNA_{s}=\dfrac{TP+TN}{TP+TN+FP+FN}
PrecisionPs=TPTP+FPP_{s}=\dfrac{TP}{TP+FP}
RecallRs=TPTP+FNR_{s}=\dfrac{TP}{TP+FN}
SpecificitySFs=TNTN+FPSF_{s}=\dfrac{TN}{TN+FP}
F1-scoreF1s=2×Ps×RsPs+RsF1_{s}=\dfrac{2\times P_{s}\times R_{s}}{P_{s}+R_{s}}
Table 1: Evaluation metrics to assess model performance
Figure 4: Single Line Diagram of simulation model consisting of a 400 kW solar PV farm connected to the grid with three loads and two feeders.
Figure 5: Current and Voltage signals of all three phases from the simulation model
Figure 6: Current and Voltage signals of all three phases from the public dataset
Figure 7: Distribution of reconstruction error of the model on train data. Each bin represents the number of segments of data having a given MAE loss. The highest reconstruction error is chosen as the threshold( α\alpha ) for detecting the faults.
Figure 8: Distribution of reconstruction error of the model on test data of all three phases.
Figure 9: Current signals of three phases with highlighted fault segments detected by the model
Figure 10: Confusion matrices
MetricSimulated dataPublic data
Accuracy97.6299.92
Precision96.8599.89
Recall91.351.00
Specificity99.2399.74
F1-score94.0299.94
Table 2: Evaluation metrics for simulated and public dataset
Figure 11: Current signals of three phases with highlighted fault segments detected by the model on public dataset
ModelsAccuracy
Logistic Regression [ 32 ]60.11
Support Vector Machine [ 32 ]99.66
K Neighbors Classifier [ 32 ]99.34
Proposed Model99.92
Table 3: Comparision of our model performance with other ML models

Explore similar work

CardsList

  1. FaultXformer: A Transformer-Encoder Based Fault Classification and Location Identification model in PMU-Integrated Active Electrical Distribution System

    Feb 27, 2026Kriti Thakur, Alivelu Manga Parimi, Mayukha PalFault LocalizationFault Diagnosis

    ABB Ability Innovation Center, Asea Brown Boveri Company, Hyderabad, 500084, Telangana, India · Department of Electrical and Electronics Engineering, Birla Institute of Technology and Science, Pilani- Hyderabad Campus, Hyderabad, 500078, Telangana, India

  2. A Hybrid Two-Stage Machine Learning Pipeline for Fault Detection and Classification in Power Transmission Systems

    Aug 24, 2026Sahil Manikshete, Atharva Gujarathi, Thanh Long Vu +2Fault DetectionFault Diagnosis

    Department of Computer and Information Science, University of Michigan-Dearborn, Dearborn, MI, USA · Pacific Northwest National Laboratory, Richland, WA, USA · Department of Electrical and Computer Engineering, Universit´e Laval, Qu´ebec City, QC, Canada +1

  3. Deep Learning-Based Detection of Electrical Faults and Power Quality Disturbances in Aerospace Power Systems

    Sep 9, 2026Ian C. Guzmán, Radu Babiceanu, Berker PeközTime Series ClassificationEdge ML

    Embry-Riddle Aeronautical University, Daytona Beach, FL 32114, USA · Western Michigan University, Kalamazoo, MI 49008, USA