Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM62131
Volume:
10 |
4 |
April 2026
Pages:
1-8
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
School of Engineering and Technology
Sushant University
School of Engineering and Technology
Sushant University
School of Engineering and Technology
Sushant University
School of Engineering and Technology
Sushant University
So here’s the thing—renewable energy is everywhere now. Especially solar. But there's a catch. It’s not really reliable. You never know how much you’re going to get because it all depends on the weather—sunlight, temperature, clouds, you name it. That unpredictability is a headache if you’re trying to actually use solar power seriously. In this project, we basically tried to build something cheap and simple that could predict solar energy output in real time, without needing expensive lab equipment. We used an ESP32 microcontroller, a DHT11 sensor for temperature and humidity, and a simple LDR (light sensor) to estimate sunlight. The whole thing costs less than ten dollars. That’s it. Then we trained three machine learning models—Linear Regression, Random Forest, and Gradient Boosting—on a small solar dataset. And funny enough, the simplest one, Linear Regression, worked best for us. Not what we expected, but that’s how it turned out. We also made a backend using FastAPI so you can get predictions, see past data, and even look at a 24-hour forecast using ARIMA. Plus, we added a feature that shows how much CO₂ you’re saving. All of this is shown on a basic web dashboard. Honestly, the main takeaway is that you don’t need fancy hardware to build something useful. Even basic stuff can give you decent predictions.
KEYWORDS
We built a complete, low-cost IoT system for predicting solar energy output. It uses an ESP32, DHT11, and LDR—total cost under ten dollars. The backend is FastAPI, with three ML models trained and compared. Linear Regression performed best with R² = 0.971. We also added a 24-hour ARIMA forecast, live weather auto-fill, CO₂ savings estimation, and a web dashboard. The system is not perfect. The LDR is crude, the dataset is small, and the ARIMA order is hardcoded. But it works. And more importantly, it shows that you don’t need expensive gear to get into renewable energy prediction. This is especially useful for students, hobbyists, and rural areas where cost is a barrier. In the future, we want to expand the dataset, add more features (wind speed, UV index), use auto-ARIMA, and experiment with LSTM models. Cloud deployment and a mobile-friendly dashboard would also be great. And maybe blockchain for data integrity. Overall, this was a fun, hands-on project that proved a simple point: cheap hardware + basic ML can still be surprisingly effective.
[1] Yadav, A. K., & Sharan, H. (2014). Comparative study of different models for solar radiation prediction. Renewable and Sustainable Energy Reviews, 38, 472–481. [2] Sharma, N., Bhargava, R., & Purohit, G. (2019). Comparison of SVR, ANN, and decision tree for solar power prediction using rooftop PV data. Procedia Computer Science, 167, 2096–2104. [3] Eze, I. C., & Maduka, J. (2021). ARIMA and hybrid ARIMA-ML models for short-horizon solar energy forecasting. Energy Reports, 7(S6), 1214–1225. [4] Kapoor, S., & Bedi, J. (2020). IoT architectures for renewable energy monitoring: A survey of low-cost embedded systems. IEEE Internet of Things Journal, 7(10), 9741–9752. [5] Kalogirou, S. A. (2001). Artificial intelligence for the modeling and control of combustion processes: A review. Progress in Energy and Combustion Science, 29(6), 515–566. [6] Hossain, M., Mekhilef, S., Danesh, M., Olatomiwa, L., & Shamshirband, S. (2017). Application of extreme learning machine for short term output power forecasting of three grid-connected PV systems. Journal of Cleaner Production, 167, 395–405. [7] Central Electricity Authority, Government of India. (2022). CO₂ Baseline Database for the Indian Power Sector — Version 17.0. Ministry of Power, New Delhi. [8] Kingma, D. P., & Ba, J. (2015). Adam: A method for stochastic optimization. International Conference on Learning Representations (ICLR 2015). arXiv:1412.6980. [9] Dewan, S., Duhan, L., & Gupta, N. (2021). Secure electronic voting system based on mobile app and blockchain. Recent Advances in Computer Science and Communications, 14(9). [10] Dargan, J., & Gupta, N. (2024). Integration of blockchain for IoT communication. In Adaptive power quality for power management using smart technologies (pp. 319–343). CRC Press, Taylor & Francis.
Publication Date
4/30/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
4
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...