Explainable Deep Learning for PV-EV Microgrids: A Hybrid RNN-LSTM-SHAP Model for Outage Prediction and Operational Optimization

International Journal of Electrical and Electronics Engineering

© 2026 by SSRG - IJEEE Journal

Volume 13 Issue 2

Year of Publication : 2026

Authors : S. Selvi, M.R. Mano Jemila, Aruna S, S. Nooray Sashmi, R. Vijaya Kumar Reddy, M. Sabarimuthu

10.14445/23488379/IJEEE-V13I2P111

How to Cite?

S. Selvi, M.R. Mano Jemila, Aruna S, S. Nooray Sashmi, R. Vijaya Kumar Reddy, M. Sabarimuthu, "Explainable Deep Learning for PV-EV Microgrids: A Hybrid RNN-LSTM-SHAP Model for Outage Prediction and Operational Optimization," SSRG International Journal of Electrical and Electronics Engineering, vol. 13,  no. 2, pp. 145-152, 2026. Crossref, https://doi.org/10.14445/23488379/IJEEE-V13I2P111

Abstract:

In recent years, in the Indian power system, there has been a transition from centralized grid systems to distributed energy sources. This transition has reshaped how electricity is produced and managed. A key role in this transition is played by microgrids that combine distributed generation sources such as solar, wind, and energy storage with electric vehicle charging infrastructure. These systems enable localized energy generation and flexible demand management. Despite these advantages, microgrids face operational challenges due to unpredictable outages, fluctuating renewable generation, and nonlinear load behaviour. Further, the rapid growth of EV charging increases load uncertainty and stresses microgrid energy management and stability. However, Machine Learning And Deep Learning Techniques can mitigate these challenges by accurately forecasting renewable generation and charging demand, enabling intelligent and adaptive control A Hybrid Deep Learning approach is introduced in this work, where Long Short Term Memory (LSTM) and Recurrent Neural Network (RNN) model are integrated for multi-class outage prediction and PV–EV microgrid optimization to deal with operational challenges in microgrid. In this study, a 15-year Power Outage Dataset along with a Microgrid PV–EV Charging Dataset was used. Learning is carried out from sequential data patterns such as irradiance levels varying across the day, changes in demand, and the corresponding grid supply responses. Based on these evolving patterns, outages are anticipated in advance, and decisions on energy distribution or storage are also determined to maintain system stability. An explainability layer based on SHAP highlights the most influential operational parameters affecting outage probabilities and PV–EV power dynamics. An overall accuracy of 88.4% was attained by the Hybrid Model for outage prediction, and a R² value of 0.93 was achieved in PV energy forecasting. The proposed model performance is compared with baseline approaches such as CNN, XGBoost, and static baseline models.

Keywords:

Outage Prediction, Microgrid Optimization, Solar Power Forecasting, Electric Vehicle Charging, Time-Series Analysis, Multi-Task Learning.

References:

[1] Muhammad Naveed Akhter et al., “An Hour-Ahead PV Power Forecasting Method based on an RNN-LSTM Model for Three Different PV Plants,” Energies, vol. 15, no. 6, pp. 1-21, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[2] Edna S. Solano, Payman Dehghanian, and Carolina M. Affonso, “Solar Radiation Forecasting using Machine Learning and Ensemble Feature Selection,” Energies, vol. 15, no. 19, pp. 1-18, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[3] Nur Liyana Mohd Jailani et al., “Investigating the Power of LSTM-based Models in Solar Energy Forecasting,” Processes, vol. 11, no. 5, pp. 1-21, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[4] Wang Zeyu et al., “Random Forest based Hourly Building Energy Prediction,” Energy and Buildings, vol. 171, pp. 11-25, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[5] Ahmed Faris Amiri et al., “Faults Detection and Diagnosis of PV Systems based on Machine Learning Approach using Random Forest Classifier,” Energy Conversion and Management, vol. 301, pp. 1-15, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[6] Bita Ghasemkhani et al., “Machine Learning Model Development to Predict Power Outage Duration (POD): A Case Study for Electric Utilities,” Sensors, vol. 24, no. 13, pp. 1-29, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[7] Zhibin He et al. “A Comparative Study of Artificial Neural Network, Adaptive Neuro-Fuzzy Inference System and Support Vector Machine for Forecasting River Flow in the Semiarid Mountain Region,” Journal of Hydrology, vol. 509, pp. 379-386, 2014.
[CrossRef] [Google Scholar] [Publisher Link]
[8] L. Prieto-Godin et al., “Predicting Weather-Related Power Outages in Large Scale Distribution Grids with Deep Learning Ensembles,” International Journal of Electrical Power and Energy Systems, vol. 170, pp. 1-13, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[9] Ekin Ekinci, “A Comparative Study of LSTM-ED Architectures in Forecasting Day-Ahead Solar Photovoltaic Energy using Weather Data,” Computing, vol. 106, no. 5, pp. 1611-1632, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[10] Honglin Xue et al., “Power Forecasting for Photovoltaic Microgrid based on Multi-Scale CNN-LSTM Network Models,” Energies, vol. 17, no. 16, pp. 1-13, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[11] Su-Chang Lim et al., “Solar Power Forecasting using CNN-LSTM Hybrid Model,” Energies, vol. 15, no. 21, pp. 1-17, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[12] Alberto Dolara et al. “A Physical Hybrid Artificial Neural Network for Short-Term Forecasting of PV Plant Power Output.” Energies, vol. 8, no. 2, pp. 1138-1153, 2015.
[CrossRef] [Google Scholar] [Publisher Link]
[13] Xiaosheng Peng et al., “Short-Term Forecasting of Photovoltaic Power Generation based on Feature Selection and Bias Compensation-LSTM Network,” Energies, vol. 14, no. 11, pp. 1-16, 2021.
[14] Ali Ahmed Alguhi, and Abdullah M. Al-Shaalan, “LSTM-based Prediction of Solar Irradiance and Wind Speed for Renewable Energy Systems,” Energies, vol. 18, no. 17, pp. 1-18, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[15] Jianhua Yuan et al., “Prediction Method of Photovoltaic Power based on Combination of CEEMDAN-SSA-DBN and LSTM,” Science and Technology for Energy Transition, vol. 78, pp. 1-9, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[16] Shahad Mohammed Radhi et al., “Machine Learning Approaches for Short-Term Photovoltaic Power Forecasting,” Energies, vol. 17, no. 17, pp. 1-23, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[17] Berny Carrera, and Kwanho Kim, “Comparison Analysis of Machine Learning Techniques for Photovoltaic Prediction using Weather Sensor Data,” Sensors, vol. 20, no. 11, pp. 1-26, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[18] Gunapriya Balan et al., “An Improved Deep Learning-Based Technique for Driver Detection and Driver Assistance in Electric Vehicles with Better Performance,” International Transactions on Electrical Energy Systems, vol. 2022, no. 1, pp. 1-16, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[19] Jae-Gon Kim et al., “Daily Prediction of Solar Power Generation Based on Weather Forecast Information in Korea,” IET Renewable Power Generation, vol. 11, no. 10, pp. 1268-1273, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[20] Paweł Piotrowski, and Marcin Kopyt. “Short-Term Forecasts of Energy Generation in a Solar Power Plant Using Various Machine Learning Models, along with Ensemble and Hybrid Methods,” Energies, vol. 17, no. 17, pp. 1-24, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[21] Fachrizal Aksan et al., “A Comparative Study of Machine Learning Models for PV Energy Prediction in an Energy Community,” Energies, vol. 18, no. 22, pp. 1-17, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[22] J. Antonanzas et al., “Review of Photovoltaic Power Forecasting,” Solar Energy, vol. 136, pp. 78-111, 2016.
[CrossRef] [Google Scholar] [Publisher Link]