- Conference Article
- 10.1109/iceconf65644.2025.11379387
Solar Energy Harvester and UNet Segmentation with CNN-Based MPPT Algorithm in an Efficient Photovoltaic Unit
- Oct 09, 2025
- Tumula Tirumala + 5 more +5
The Solar Energy Harvester and UNet Segmentation with CNN-Based MPPT Algorithm in Efficient Photovoltaic Unit (SEUCMP) combines PV design with advanced deep learning methodologies to enhance energy management. It aims to configure solar arrays to produce 11 kW daily, optimizing the arrangement of modules for maximum efficiency. A UNet-based segmentation framework improves predictive accuracy by analyzing images of solar plants. The system utilizes a distinctive Energy Monitor (EM) circuit for regulation, where minimal capacitance is designed to reduce future dependence on software promotional techniques. At the MPPT level, a hybrid LSTM-FNN strategy is employed to optimize energy forecasting, addressing the variability in solar conditions through the Cuckoo Search Optimizer (CSO). Using the following parameters, we calculated the SEUCMP model, power analysis, mean square error, loss & accuracy, and mean IOU calculations.
Read more