2025
doi: 10.12720/jait.16.6.809-818
Solar Power Forecasting on Smart Micro Grid Using CNN-LSTM Ensemble Method
Nur Iksan1,2,*
,
Purwanto Purwanto3 ,
Heri Sutanto4
- 1Doctoral Program of Information System, School of Postgraduates Studies, Diponegoro University, Semarang, Indonesia
- 2Department of Electrical Engineering, Faculty of Engineering, Semarang State University, Semarang, Indonesia
- 3Department of Chemical Engineering, Faculty of Engineering, Diponegoro University, Semarang, Indonesia
- 4Department of Physic, Faculty of Science and Mathematics, Diponegoro University, Semarang, Indonesia
- Manuscript receivedDecember 10, 2024
- revisedDecember 27, 2024
- acceptedMarch 4, 2025
- publishedJune 12, 2025
Abstract
The use of solar Photovoltaic (PV) energy as an alternative renewable energy source has increased worldwide. Problems in solar PV systems are caused by intrinsic intermittency in solar radiation and other meteorological factors so the power generated from PV is uncertain and unstable. Therefore, a forecasting model using an Artificial Intelligence (AI) approach is needed to overcome the instability in the utilization of energy generated by solar PV, especially for distributed residential PV, which is operated and maintained independently. The purpose of this study is to develop an ensemble model on Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) to provide predictions on the generation of energy in solar PV. Performance evaluation is carried out by comparing several forecasting models, namely Deep Neural Network (DNN), CNN, LSTM, and the proposed ensemble model using scenarios of three different input data combinations (S1, S2, S3). Based on the results of experiments and performance measurements, the model that is proposed to have very good overall performance, especially superior in the combination of {S1, S3}, {S2, S3} and {S1, S2, S3}. The combination of {S1, S3} gives the best results in this model, namely Mean Square Error (MSE) = 0.0934; Root Mean Square Error (RMSE) = 0.3057 better than LSTM. The combination of {S2, S3} with MSE = 0.1359; RMSE = 0.3687. The combination of {S1, S2, S3} is also quite good with MSE = 0.0951.
Keywords
- Convolutional Neural Network (CNN)
- Long Short Term Memory (LSTM)
- ensemble
- forecasting
- Smart Micro Grid (SMG)
How to Cite
Nur Iksan, Purwanto Purwanto, and Heri Sutanto, "Solar Power Forecasting on Smart Micro Grid Using CNN-LSTM Ensemble Method," Journal of Advances in Information Technology, vol. 16, no. 6, pp. 809-818, 2025. https://doi.org/10.12720/jait.16.6.809-818
Copyright & License
Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).
