Computational investigation of dust settlement effect on power generation of three solar tracking photovoltaic modules using a modified angular losses coefficient.

• The effect of dust accumulation on the performance of three solar tracking PV modules is investigated. • The modified angular losses coefficient (α r) is calculated at different tilt angles and a dust deposition density. • The α r coefficient is used by a computational method to predict the power generation of solar tracking modules. • The proper performance of the α r coefficient is examined using the results of previous experimental studies. • The reduction in electricity generation and temperature distribution of the modules are calculated. Providing a suitable method for predicting the effect of particle aggregation on photovoltaic (PV) performance can be one of the best ways to make better use of solar energy. In this study, a mathematical method was used and modified to provide a better prediction of angular losses coefficient (α r). The modified coefficient as well as a computational method were used to find the performance of dusty modules equipped with one-axis (East-West (EW), North-South (NS)), and dual-axis (DAT) solar tracking systems. The modified α r , which here is called the degree of surface dirtiness, can be used to estimate the area covered by particles per cm 2 , i.e. A ∗ C (θ). A better prediction of α r causes a better prediction of power loss due to dust settlement. Results showed that A ∗ C (θ) is the factor that shows the role of incidence angle in visualizing the type of dust cover, while α r shows the effect of slope angle on dust deposition density. Furthermore, combination of α r with the computational model revealed that the dirty DAT module has the least daily power loss which is 9% of the clean counterpart power. The power losses for the NS and EW dusty modules are 13% and 20.5% lower than the clean counterparts, respectively. It should be noted that the obtained results in this study are in a good compromise with the experimental and computational data available in the literature. [ABSTRACT FROM AUTHOR]