Explainable Artificial Intelligence (XAI) for Fault Detection in Smart Grid Distribution Systems: A SHAP-Based CNN-LSTM Approach

Dilshad Shah; Maneesh Kumar Gupta

doi:10.36948/ijfmr.2026.v08i02.71195

Explainable Artificial Intelligence (XAI) for Fault Detection in Smart Grid Distribution Systems: A SHAP-Based CNN-LSTM Approach

Author(s)	Mr. Dilshad Shah, Mr. Maneesh Kumar Gupta
Country	India
Abstract	Reliable and rapid fault detection in smart grid distribution systems is critical for maintaining power quality, minimizing outage duration, and ensuring grid stability. While deep learning models have demonstrated remarkable accuracy in fault classification, their black-box nature limits adoption in safety-critical power systems where operators require transparent decision-making. This paper proposes a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) architecture augmented with SHapley Additive exPlanations (SHAP) to address both high detection accuracy and model interpretability simultaneously. The CNN layers extract spatial features from multi-channel voltage and current waveforms, while the LSTM layers capture temporal fault dynamics. SHAP values are subsequently computed to provide feature-level explanations for every prediction, enabling power system engineers to understand which electrical parameters contributed most to each fault classification decision. The proposed framework is evaluated on the IEEE 13-bus benchmark test system with five distinct fault types: single line-to-ground (LG), double line-to-ground (LLG), three-phase (3Φ), line-to-line (LL), and high-impedance fault (HIF). Experimental results demonstrate that the CNN-LSTM+SHAP model achieves an overall accuracy of 98.7%, precision of 97.2%, recall of 98.0%, and F1-score of 99.1%, with a real-time detection latency of 43 milliseconds. Comparative analysis against Random Forest, SVM, and standalone ANN models confirms significant performance improvements. The SHAP-based explanation layer further reveals that voltage sag magnitude and rate of change of current (di/dt) are the most decisive features in fault discrimination. The proposed XAI framework offers a practical pathway toward transparent, trustworthy AI deployment in next-generation smart distribution networks.
Keywords	Explainable Artificial Intelligence (XAI), Smart Grid, Fault Detection, CNN-LSTM, SHAP, Distribution System, Deep Learning, Power Systems
Field	Engineering
Published In	Volume 8, Issue 2, March-April 2026
Published On	2026-03-11
DOI	https://doi.org/10.36948/ijfmr.2026.v08i02.71195

View / Download PDF File

E-ISSN 2582-2160

doi

CrossRef DOI is assigned to each research paper published in our journal.

IJFMR DOI prefix is
10.36948/ijfmr

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 8 Isu 2 Cover Page Vol 8 Isu 1 Cover Page Vol 7 Isu 6

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Explainable Artificial Intelligence (XAI) for Fault Detection in Smart Grid Distribution Systems: A SHAP-Based CNN-LSTM Approach

Share this