Skip to content
DSML.KZ
Publications

Talk · International Siberian workshops and tutorials on electron devices and materials · June 1, 2026

AI-Based Multi-Objective Optimization of Semi- Transparent Photovoltaic Greenhouses Considering Crop Light Constraints

Mukhriddin O. Eshkulov, Sherzod D. Qushakov, Asror Mustafakulov, Elnur A. Abdullaev, Oybek Matchonov, Mansur Qodirov

Mukhriddin Eshkulov
Open via DOI

Abstract

The integration of photovoltaics into greenhouse systems offers a promising pathway to simultaneously address growing energy demand and agricultural land constraints. However, conventional photovoltaic greenhouses often suffer from excessive shading and non-uniform light distribution, which can adversely affect crop growth and productivity. This study proposes an AI-based multi-objective optimization framework for semi-transparent photovoltaic (STPV) greenhouses, explicitly balancing photovoltaic energy generation, crop-relevant light availability, and spatial light uniformity. Experimentally validated STPV panels based on monocrystalline silicon cells encapsulated in an EVA-tempered glass multilayer were characterized through spectrophotometric measurements and in-situ greenhouse experiments. High photosynthetically active radiation (PAR) transmittance (85-95%) and effective ultraviolet attenuation were confirmed, while optimized trapezoidal-textured transparent regions enabled diffuse light redistribution and suppression of persistent shading. Internal illuminance measurements demonstrated that crop-level Daily Light Integral (DLI) values remain within optimal ranges for major greenhouse crops. An artificial intelligence-assisted surrogate model combined with evolutionary multi-objective optimization was employed to generate Pareto-optimal solutions, revealing continuous trade-offs between energy yield, DLI, and light uniformity. Crop-specific feasible regions were identified, enabling crop-aware selection of STPV configurations. The results demonstrate that STPV greenhouse design is inherently a trade-off problem rather than a single-solution task and show the potential for energy-autonomous or grid-independent greenhouse operation. The proposed framework provides a physically grounded, decision-oriented methodology for intelligent agrivoltaic greenhouse design and supports the development of next-generation sustainable, AI-enabled protected agriculture systems.

Greenhouse Technology and Climate ControlPhotovoltaic Systems and SustainabilitySolar Thermal and Photovoltaic Systemssemitransparent solar cellsagrivoltaicsgreenhouse design

More by Mukhriddin Eshkulov