Open Access Version

Abstract:
This work analyzes photovoltaic performance as a function of irradiance and temperature in order to improve the energy yield of Cu(In,Ga)(S,Se)2 (CIGS) modules under operating conditions. The analysis utilizes various methodologies, algorithms and models to evaluate the performance of photovoltaic devices. The one diode model (ODM) analytical approach was utilized to obtain insights into the main IV parameters as functions of both irradiance and temperature and on how non-ideal components of the model affect these trends. Previous studies have exposed some of the influences of these characteristics on either irradiance or temperature behavior. However, this work shows the coupling of irradiance, temperature and non-ideal characteristics affecting the IV parameters, which is especially observable in the temperature coefficients of the open circuit voltage and the maximum power point and therefore, affecting the energy output. The improvement of these characteristics was targeted to 30x30 cm2 CIGS solar modules and their gradual improvement was evaluated according to the ODM. The improvements were categorized in three phases. The first phase included the scaling up from cells to modules as a starting point for evaluation. The second phase encompassed the efforts on improving the lateral homogeneity of the 30x30 cm2 CIGS material by modifying the rapid thermal processes. The third phase comprised variations in the precursor and module monolithic interconnections, where the Ga content of the precursor was increased and the P1 scribe was slightly modified to improve the fill factor. The increase of power output under various conditions, and therefore yield, was measured indoors as well as outdoors during several months. A multiple linear regression (MLR) model, developed in this work, was employed to estimate the main IV parameters (Isc, Voc, Pmpp and Impp). The MLR model is composed of a set of irradiance-temperature dependent functions, one per IV parameter, which were verified statistically by fitting simulated and empirical data achieving very high correlations. Applications of this model such as extraction of temperature coefficients or diode parameters were also explored. The model was validated for energy rating and output prediction. Using indoor and outdoor measurements of our CIGS modules to predict outdoor performance, the MLR model was compared to four known models, yielding similar or better results. The MLR model was used to predict module performance ratios at different weather conditions using indoor and outdoor measurements, which in combination with data from a geographical information system, was utilized to obtain insights on output forecast for modules installed at different locations.