Abstract
Photovoltaic (PV) system performance is significantly influenced by soiling—the deposition of dust and particulate matter on module surfaces—which reduces solar transmittance and energy output. While numerous models have been developed to predict soiling losses, most are calibrated for arid or semi-arid climates and do not accurately reflect the conditions in tropical sub-humid regions. In Ghana, and particularly in the city of Kumasi, the unique combination of high humidity, frequent rainfall, and locally generated aerosols from biomass burning and urban emissions results in soiling patterns that diverge markedly from those assumed in conventional models. This disconnect presents a critical challenge for reliable energy yield forecasting and operational planning of PV systems in the region.
This study addresses the gap by developing a predictive model specifically adapted to Kumasi’s climatic and environmental conditions. Field experiments were conducted over a 12-month period using PV modules exposed to natural weathering. Optical transmittance measurements were used to quantify soiling losses, while corresponding meteorological data—including rainfall, temperature, humidity, wind speed, and aerosol optical depth—were collected using both ground-based and satellite sources. Advanced statistical analysis and machine learning algorithms, including multivariate regression and random forest regression, were employed to identify key predictors and construct a robust model capable of forecasting soiling rates and associated energy losses.
Preliminary findings indicate that dry spell duration, relative humidity, and aerosol concentration are the dominant climatic variables influencing soiling accumulation in Kumasi. Unlike in arid regions where wind-driven dust predominates, localized emissions from biomass and vehicular sources play a significant role. The developed model demonstrates a mean absolute percentage error (MAPE) of less than 10%, significantly improving upon existing generalized models. This climate-specific approach offers a valuable tool for solar asset operators, enabling more accurate energy output predictions, optimized cleaning schedules, and informed financial modeling.
This work represents a significant contribution to the field of PV performance modeling in tropical regions. It emphasizes the need for localized soiling models that account for regional climatic dynamics, thereby supporting the efficient integration of solar energy into Ghana’s renewable energy mix and strengthening the long-term viability of PV systems across sub-Saharan Africa.
Keywords:
Photovoltaic soiling, climate-specific modeling, Kumasi, Ghana, tropical environment, solar energy, machine learning,
PV performance, energy yield forecasting.
Biography
Mr. Joseph Agbogla is an accomplished mechanical engineer with a strong interdisciplinary background that bridges industrial experience and academic expertise. With over a decade of professional engagement, he has applied engineering principles to drive innovation and efficiency in both industry and higher education. In the industrial sector, Joseph has led performance-focused initiatives, including a notable operational improvement project that resulted in a 20% increase in productivity and system efficiency. Currently serving as a lecturer in Mechanical Engineering, Joseph teaches core modules including Thermodynamics, Heat Transfer, Fluid Mechanics, Instrumentation Science and Technology, and Plant Maintenance. His teaching is informed by practical experience and a deep understanding of energy systems, which he continues to advance through rigorous research. As a PhD candidate, his work focuses on the mathematical and experimental modeling of photovoltaic (PV) soiling, with a particular emphasis on the impact of regional climatic conditions—especially within tropical sub-humid environments such as Kumasi, Ghana.
Joseph has published in leading peer-reviewed journals in the field of renewable energy. His recent works include “Soiling estimation methods in solar photovoltaic systems: Review, challenges and future directions” (https://doi.org/10.1016/j.rineng.2025.104810), which presents a comprehensive analysis of current soiling estimation approaches and outlines future research needs; and “Techno-environmental
assessment of the fuel properties of a variety of briquettes for biomass boiler applications” (https://doi.org/10.1016/j.cles.2025.100185), which evaluates the energy performance and sustainability implications of biomass briquettes in industrial heating systems. Driven by a commitment to sustainable energy development and capacity building, Joseph continues to mentor the next generation of engineers while contributing to the advancement of context-sensitive energy solutions in sub-Saharan Africa. His combined experience in technical practice, applied research, and engineering education positions him as a valuable contributor to discussions on renewable energy strategy and implementation.