Work

A hybrid VMD-BFO-LSTM method for accurate photovoltaic power forecasting: enhancing grid integration of renewable energy

Year: 2026

Type: article

Abstract: Accurate short-term photovoltaic (PV) power forecasting is vital for grid stability and renewable energy integration. This study proposes a novel hybrid model (VMD-BFO-LSTM) that combines Variational Mode Decomposition (VMD), Bitterling Fish Optimization (BFO), and Long Short-Term Memory (LSTM) networks to enhance prediction accuracy. VMD decomposes the original PV power data into intrinsic mode functions (IMFs), with BFO optimizing the decomposition parameters. These components, along with meteorological inputs, feed into multiple BFO-optimized LSTM submodels. Final predictions are generated by aggregating outputs from all submodels. The model was validated using real-world data from a 1 MW PV plant in Kahramanmaraş, Turkey. Compared to benchmark models, the proposed method achieved superior performance, with an MAE of 15.247 kW, RMSE of 19.753 kW, MAPE of 4.401%, and R 2 of 99.661%. The model demonstrated up to 89% and 84% reductions in MAE and RMSE, respectively, proving its robustness under varying solar conditions.

Source: Ain Shams Engineering Journal

DOI: 10.1016/j.asej.2026.104254

Authors: Javad Rahebi, Vedat Esen

Institution: Istanbul Topkapi University

Language: English

Topic: Solar Radiation and Photovoltaics

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDGs: Affordable and clean energy, Climate action, Industry, innovation and infrastructure

Open Access status: gold

Estimated APC paid: $1,200