Intelligent Pv-Battery Energy Management Using TFT Forecasting and MPPT Optimization

View at publisher Summarise with VIRAT
Authors: Renu V, A. Ramkumar
Cited by: 0
References: 13

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Solar energy, a clean, renewable, and abundant resource, plays a critical role in addressing growing global energy demands sustainably. Efficient utilization and storage of solar energy are vital to ensure continuous power supply despite the intermittency of solar irradiance due to daily and seasonal fluctuations. PV systems with battery storage are effective, but their efficiency relies on smart control strategies for managing energy flow. Traditional approaches such as Perturb & Observe (P&O) and Incremental Conductance (INC) for Maximum Power Point Tracking (MPPT) often fall short in rapidly changing environmental conditions, leading to suboptimal energy harvesting, power oscillations, and battery inefficiencies. To address these challenges, this study introduces a forecast driven control framework that integrates the Temporal Fusion Transformer (TFT) with Bayesian Optimization. The proposed system predicts photovoltaic output, load demand, and battery state of charge to optimize MPPT control and energy scheduling in real time using a dataset collected from a 5 kW photovoltaic array with 22 monocrystalline panels rated at 228 W each, connected to a kWh battery system. Environmental and electrical parameters including solar irradiance, temperature, power output, and battery state of charge were recorded every 10 minutes over a 7-day period using precision sensor instrumentation. Simulation results demonstrate the effectiveness of the proposed model, achieving high forecasting accuracy (MAPE$

Fields of Study & Subjects

Photovoltaic System Optimization Techniques Power Systems and Renewable Energy Smart Grid Energy Management

Publication details

Journal / source
International Conference on Communication, Computing and Electronics Systems (ICCCES)
Pages
1–9
Publication date
21 January 2026
Language
EN