This thesis presents a new type of solar simulator and new measurement
methods that allow for fast power rating of photovoltaic devices and for fast
performance measurements for energy rating and energy yield predictions
indoors under controlled, and more realistically simulated outdoor conditions.
A novel indoor measurement system for photovoltaic device characterisation
based on light emitting diodes (LEDs) as the light sources is described. The
solar simulator is capable of reproducing spectral changes seen in natural
sunlight, with its intricacies of variable air mass and weather conditions, to a
better match than previously possible. Furthermore, it allows measurements
under varying light intensity and device temperature.
The prototype LED-based solar simulator developed is characterised and its
measurement quality is analysed. The system achieves a class BAA solar
simulator classification with a class B spectral match, class A light intensity
uniformity and a class A temporal stability. It is the first system of its kind that
meets the standards of a solar simulator in spectral match to the standard
sunlight spectrum and in terms of minimum light intensity. An uncertainty
analysis shows that calibration uncertainty for crystalline silicon solar cells is
5% in maximum power with a 95.45% level of confidence. Recommendations
for further versions of the solar simulator are given and show potential of
reducing this uncertainty down to 2.9% across all measurement spectra
(1.8% with a primary calibrated reference cell).
A new method for automated power-rating of single- and multi-junction
devices is developed. The method uses a unique spectral response
measurement and fitting method. It eliminates the need of external
measurement equipment for determining spectral response. A simulated
characterisation of an amorphous silicon single- and double-junction solar
cell show accuracy of better than 0.5% in maximum power. First
measurements on the LED-based solar simulator show a measurement error of 4.5% in maximum power, which is due to a lack of measurement feedback
of spectral output and measurement irradiance.
The first three-dimensional performance matrix for use in photovoltaic energy
rating is reported, utilising the LED-based solar simulator. Device
characteristics are measured indoors under varying irradiance, temperature
and spectrum. A measurement method is detailed and utilised on a
crystalline and amorphous silicon solar cell. It allows for the first time a direct
investigation of spectral effects on photovoltaic devices under controlled
conditions. Results show that amorphous silicon devices are very sensitive to
changes in spectrum. Thus, spectral effects should not be neglected in energy yield predictions for such devices.
A Doctoral Thesis. Submitted in partial fulfillment of the requirements for the award of Doctor of Philosophy of Loughborough University.