Ricardo Ruther

LABSOLAR — Laboratorio de Energia Solar
Universidade Federal de Santa Catarina / UFSC
Caixa Postal 476, Florianopolis — SC 88040-900 Brazil
Tel.: +55 48 331 5174 FAX: +55 48 331 7615 Email: ruther@mbox1.ufsc. br

LABSOLAR is the solar energy research laboratory at Universidade Federal de Santa Catarina (UFSC) in southern Brazil. The laboratory’s activities in the field of photovoltaics (PV) started in 1997, with solar thermal and radiometry research going on since 1989. Over the last seven years, LABSOLAR has been engaged in promoting PV in Brazil in a number of scientific and dissemination projects. Training human resources with a sound scientific and technological background in the field has also been a major activity, carried out to multiply the still small number of PV specialists in the country. This paper reviews some of the most relevant scientific, dissemination and capacity building projects.

1. The first grid-connected, building-integrated, thin-film PV system in Brazil

In September 1997 the first grid-connected, building-integrated, thin-film PV system in Brazil was installed at LABSOLAR, on the main campus at Universidade Federal de Santa Catarina. The PV installation consists of a 2kWp thin-film amorphous silicon (a-Si) array, plus DC/AC inverter, irradiance (horizontal and plane-of-array), and temperature (ambient and back-of-module) measurement instrumentation, and a dedicated data logging system. The generator is comprised of 54 opaque and 14 semitransparent, double-junction (pin-pin), glass-glass 60 x 100cm2 a-Si modules from RWE-Schott, with a total power output rated at 2078Wp at STC, and a total surface area of ~40m2. The total power is distributed in four ~500Wp sub-systems, and fed to four independent single-phase, line-commutated sinewave inverters (from Wurth Elektronik GmbH, model WE 500 NWR, each rated at 650W). The PV array uses unframed modules designed for BIPV applications, that were installed onto a simple steel structure retrofitted as an overhang to the existing building, facing true north with latitude tilt (27o). Electrical parameters as well as irradiance and temperature data are continuously measured and stored at four-minute intervals. Figure 1 shows the BIPV system; further details on PV system design and configuration have been presented elsewhere [1,2].

The project’s main objectives are twofold: (i) demonstration and dissemination of the concept of PV in buildings in Brazil; and (ii) a long-term experiment on the seasonal effect affecting the performance of thin film a-Si, with emphasis on determining its suitability for operation under the higher temperatures common in building-integrated PV systems in warm and sunny climates. Seasonal performance shifts due to both Staebler-Wronski degradation/annealing [3] mechanisms enhanced by higher operating temperatures in summer, and to seasonal shifts in the spectral content of sunlight, as well as a small temperature coefficient of power, render a-Si a relatively better PV converter in summer than in wintertime, in contrast to the performance profile of crystalline silicon PV converters, which are more efficient in winter due to lower operating temperatures.

Performance results on the fully monitored BIPV installation operating continuously for over six years have indicated peculiarities in system sizing (PV array vs. inverter rated power) and have demonstrated that a-Si is a good performer at sunny sites and warm climates. Performance ratios (PR, defined as the ratio of the energy output and the rated efficiency, times the total solar radiation incident on the PV module’s surface) obtained during this period averaged 91.4% (DC) and 81.5% (AC), and annual AC energy yield was 1231kWh/kWp for a 1507kWh/m2 annual plane-of-array irradiation level at the site. While in the first year (1997 data) the a-Si modules showed a small but negative temperature coefficient of power (TcoeffPmax = -0.22%/oC), after stabilisation of the light-induced degradation inherent to the a-Si material (the Staebler-Wronski effect [3]) our most recent results show that TcoeffPmaX drops to negligible (and positive, TcoeffPmaX = +0.08%/°C) values (2003 data); i. e., in the stabilised level, the net performance of a-Si becomes somewhat independent of temperature, as shown in Figure 2.

One further peculiarity that can be shown by the analysis of the solar radiation data measurements over the period, is the issue of inverter x PV array sizing with respect to prevailing irradiation levels. Because there is a considerable amount of energy at high irradiation levels available at the site, inverter undersizing might be a considerable limitation to PV systems’ annual energy yields. As shown in Figure 3, some 15% of the total radiation reaching the PV modules is in the > 1000W/m2 range, and nearly 60% of the total available energy is in the > 700W/m2 range, while only less than 10% of the solar energy available lies in the < 200W/m2 range. Under conditions as such, and using thin-film a-Si PV modules, PV array vs. inverter rated power ratios < 1 will lead to smaller power losses. Furthermore, inverter efficiencies usually peak at power levels below maximum nominal power (i. e, highest efficiency takes place at partial loading), and high inverter loading levels lead to high inverter operating temperatures, reducing the DC to AC conversion performance even further.

Figure 3: Distribution of the plane-of-array irradiation in Florianopolis (48°W, 27°S), distributed at 100W/m2 bins, averaged for the six years of continuous operation. A considerable portion of the incident energy occurs at high (> 700W/m2) irradiation levels.

Following up on this project, two other BIPV systems have been further installed on campus to demonstrate to University students, our future decision-makers, the potential of this energy source. Figures 4 and 5 show these two most recent BIPV installations.