Work

Application of machine learning in life cycle assessment of photovoltaic systems: a systematic review

Year: 2026
Type: review
Study design: Systematic Review
Abstract: Assessing the environmental impact of photovoltaic (PV) systems requires life cycle assessment (LCA). However, conventional LCA is constrained by data scarcity, uncertainty, and high computational demands. Although machine learning (ML) is increasingly applied to PV LCA, its effectiveness remains unclear. This study systematically evaluates the role of ML in enhancing PV LCA by identifying current applications, methodological trends, limitations, and research gaps. A systematic literature review was conducted following the PRISMA guidelines, resulting in the analysis of 32 studies that integrate ML with LCA frameworks for PV systems. The results show that ML applications primarily focus on carbon emissions, which account for 72% of the reviewed studies, while other important environmental impacts, such as water use, toxicity, and land use, remain largely underrepresented. Artificial neural networks dominate the methodological landscape (15 studies), whereas deep learning appears in only seven studies and is primarily used for time series forecasting and image based geospatial analysis. Most models rely on simulation-generated data (12 studies), whereas only eight studies use real world datasets. The review demonstrates that ML can significantly enhance PV LCA by improving predictive capabilities, reducing computation time, and supporting better decision making. However, challenges related to data quality, model generalizability, and the limited integration of comprehensive LCA indicators persist. Addressing these challenges is essential for advancing the application of ML in PV LCA.
Language: English
Field: Energy
Open Access status: gold
Estimated APC paid: $3,630
OpenAlex ID: W7204254627
FWCI:
Cites: 87
Cited by:
Related to: 10