Simple Model for Predicting Time Series Soiling of Photovoltaic Panels

Results from a model to predict soiling losses of photovoltaic (PV) arrays are presented. The model uses ambient airborne particulate matter concentrations, both PM10 and PM2.5, the tilt of the PV array including if the array is tracking, and rain data to estimate soiling losses over time. The model uses relationships between average airborne PM concentrations and dust accumulation, rain removal of accumulated dust, and the accumulated dust and transmission loss that have been developed in the literature to estimate soiling losses in a time series. Multiple model runs were performed and were compared with measured soiling data at seven locations. The model was run with both typical Meteorological Year rain data and with recorded rain data from Oregon State University's PRISM database. Each model was run with variable deposition velocities, static deposition velocities, and static settling velocities. When the model was run with recorded rain data and static settling velocities the results matched the general slopes, frequencies, and magnitudes of soiling losses when compared with measured soiling data. Overall, the simple model demonstrates the ability at accurately simulating soiling.