Dust accumulation on photovoltaic (PV) modules and its effect on their performance are of
high concern for regions with a high rate of dust, low frequency and intensity of rain. In this
thesis, the effect of dust on PV modules is investigated with respect to dust concentration and
spectral transmittance. The measured spectral transmittance of the dust sample shows spectral
attenuation effect that varies at different wavelengths. This effect is explained by the particle
size distribution of the dust samples: At shorter wavelengths more light is scattered due to the
effect of the smaller particles. This effect has a major impact on the PV module as it affects
PV technologies with a wider band-gap more than those of a narrower band-gap.
The effect of dust is accumulative, i.e. PV module performance is reduced by increasing
deposition over time or until it's cleared manually or by rain. The tilt angle of the PV
installation plays a major role in the amount of dust accumulated on the devices, where higher
tilt angles result in decreased dust concentrations. This effect is demonstrated in outdoor
measurements where tilted modules had lower losses in daily as well as total array yield. It is
also shown that tilted modules benefit from precipitation more than horizontal modules.
However over the exposure period the modules did not show any clear aging effect caused
specifically from dust accumulation or exhibit any seasonal variation.
Different tilt angles can produce varying non-uniform dust patterns on the device surface.
This effect and its pattern over long and short periods of exposure are investigated by means
of spatial three dimensional modelling. The simulations compare two dust accumulation
patterns that represent a short exposure to a single dusty day (one day) and a long exposure of
dust (3 months). Out of the two patterns, the long exposure patterns showed higher losses of
19.4% in comparison to 14.8% for the short exposure. The simulation also showed that dust
accumulation that promotes high concentration of dust at the bottom of the PV modules where
it covers a full cell has a high risk of triggering hot spots and thus risks permanent module
A dust correction model for energy prediction is developed. The model takes into
consideration dust concentration, spectral attenuation effect of dust, PV technology, and
various meteorological variables. The modified spectral transmittances of the dust were
incorporated into the model in the form of pre-measured data. This means in this work
samples collected in Kuwait were measured and used to generate the input. The model is
compared against the outdoor measured data and a good agreement between measurements
and simulations is demonstrated. Using this model two procedures were developed. The first
evaluates the uncertainties associated with dust over long periods of time. The second is to
find the optimised cleaning schedule and frequency of cleaning based on acceptable yield loss
margins over the simulated period of time. The optimisation of the cleaning schedule showed
that for Kuwait setting the daily energy losses in PV modules at less than 10% will set the
cost of cleaning higher than the cost of energy lost due to dust.
A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.