A study of the effect of different installation angles, panel orientation, underlying roof coating and ventilation on a photovoltaic array

Abstract

Malta's limited land size has meant that roof-top photovoltaic installations comprise the bulk of installed photovoltaic capacity on the island. This creates a unique situation, as contrary to countries where land space is freely available, the focus in Malta cannot be solely on optimising the hardware generation, but also needs to optimise the available roof space. One can conceivably run contrary to mainstream methodologies that aim to maximise the generation potential of the hardware, as we begin contemplating installations where hardware is purposely not installed at its best generation configuration. Installations can be considered that aim to increase panel density on roofs, such as with the creation of smaller shading gaps between panels, by using lower than optimum tilt angles, or by having panel arrays following the axis of buildings - even when these latter are not facing directly south. The aim of this study is to quantify the effect of different panel inclination angles, azimuth, ventilation and roof surface coatings on a fixed, roof-mounted polycrystalline array that was constructed with roof optimization as one of its objectives. Mathematical predictor models for generation potential of the various configurations were then derived to assist in optimization of such PV installations. A study has been performed on a 308.7 kWp photovoltaic array at the Simonds Farsons Cisk brewery, where the author was employed as Technical Services Manager tasked with sourcing and installing such an array in 2013. The array comprises 1260 panels each of rated output 245Wp. Panels are wired in DC strings of 15, with three strings to each inverter. Expected annual generation required from the array was 1678 kWh / kWp ±4.5%, and this was planned to reduce the brewery ele All panels and inverters are of the same type, thereby creating an ideal scenario for statistical performance comparisons for hardware installed at different inclination angles, azimuth, ventilation and background roof coatings. As the array falls within a contained area, any localised geographic variations in radiation intensity, cloud cover and temperature could be reasonably taken to be the same for all the panels and inverters, thereby rendering comparisons more relevant. Panel type was JA Solar - JA-P6-60-245 and Inverters SMA - SMC 10000TL RS485. All data was captured through the SMA web portal using a 'Sunny Web box' instrumentation package comprising of panel temperature, ambient temperature modules, radiation pyranometers, and wind anemometers. To limit data capture errors and confirm measurement consistency, two separate instrumentation sets were used. When cases of doubt existed, these were compared with a control array maintained by the vendor. The pyranometers were also used to indicate when cleaning became necessary, with one being cleaned regularly, and the 6 other only during panel cleaning. This procedure then indicated the relative drop in radiation capture being experienced due to dirt accumulation. The array was constructed over 4500m2 of roof space over a number of existing buildings. Variations in panel installation resulted due to the existing physical limitations in the roofs. To maximise installation density, the panels were installed at inclinations of 20o and 25o and aligned to follow the building azimuth of 15o west of true south. As a control, one set of 45 panels was installed at what is considered the optimum 30o inclination for Malta and aligned exactly due south. Data capture was done for two different ventilation conditions and above different roof textures of: (i) Normal concrete, (ii) Concrete coated with white (roof compound) paint. (iii) Grey grit roof membrane coating - two membrane types were in use, one with a fine grit coating and other with a coarser mineral grit coating. Data over a twelve month calendar period was evaluated to quantify the effect of the different roof coatings, inclination angles and ventilation on generation. Regression analysis to indicate performance expectancy for these different roof and installation types was then performed. This study will be of relevance to better estimate the generation characteristic for fixed installations based on: 1) Ventilation 2) Roof surfaces textures (membrane, plain concrete or white roofing compound) 3) Panel inclination 4) Panel orientation The results presented are of relevance to Malta. These show that: a) For photovoltaic panels mounted at a 20o inclination above a white painted roof, the net effect observed was that for panels with a leading edge at 1.12m above roof surface the background type had no effect on generation potential. Slight improvement was seen in winter months, but this was then offset by a decrease in generation capability in summer months; probably due to back reflection of infra-red radiation from the more reflective white roof surface. b) Panels mounted in the same ventilation conditions and above a concrete roof surface at 25o inclination performed better when compared to panels at 20o inclination, c) Panels in better ventilated conditions regularly out performed panels in poorly ventilated conditions during summer months when panel average temperatures exceeded standard test conditions (STC).