| |
Objective:
To quantify and improve the measurement methods used to rate solar photovoltaic (PV) modules and to validate algorithms that predict the installed performance of PV products.
Problem:
Photovoltaic (PV) modules are sold based on their rated power output – the higher the rated output, the higher the cost of the photovoltaic product. The rated output is evaluated at one specific set of conditions: 1000 W/m2 irradiance, a 25°C module temperature, a zero angle of incidence, and a specific radiative spectrum (AM 1.5) that is representative of outdoor solar conditions. These conditions are never exactly reproduced during testing, with the spectrum match being the most difficult criterion to meet. To varying degrees, corrections are available for translating measurements from the tested conditions to the rated conditions.
Most PV module ratings are determined using one of two test methods. Both methods are typically employed on a solar flash simulator. One method is based on using a reference module to calibrate the solar simulator. For the second method, a reference cell is tested alongside the test specimen or substituted with the test specimen over repeated tests. The former test method is typically used for measurements made as part of an automated manufacturing process. The latter test method is used in cases where time is not a constraint and more accurate measurements are sought. Ideally, the reference module should be of the exact same construction as the particular production line modules and its short-circuit current at rated conditions should have a low associated uncertainty. In practice, the actual reference module can depart significantly from the ideal. Finally, the reference cell based test method is dependent upon the degree that the reference cell’s performance matches that of the cells used in the test module.
As a point of reference, a recent international module inter-comparison project reported rated powers for the same mono-Si, 3a-Si, and CIS modules (among others) that were distributed over ranges of 7.4%, 16.9%, and 11.6%, respectively. This variability is significant given that this study represents a best case scenario with regard to participants and that a 1 percent ratings change equates to an annual economic impact of 72 million dollars (for the current market conditions).
The goals of this NIST project are to identify both shortcomings with and improvements to current industry practices, publish findings that may be helpful to applicable ASTM and IEC working groups, help the industry progress towards better solar flash simulators, and either expand the range of applicability for using a single reference PV cell and a single reference module or eliminate the need for the reference devices altogether.
Two other impediments within the PV industry are the lack of performance data and the lack of qualified predictive tools. “Qualified”, in this case, equates to documenting the trade-offs from using different modeling algorithms – the complexity of the required inputs relative to the accuracy of the predicted annual energy generation. These impediments cause key decision makers to have insufficient resources on which to judge the merits for incorporating solar components within a residential or commercial building.
Approach:
In FY07, NIST installed and characterized its state-of-the-art solar flash simulator. The simulator’s spatial uniformity and spectral characteristics have been determined, in large part via measurements using a high-speed spectroradiometer. A test protocol that is consistent with but goes beyond the requirements of IEC Standard 60904-1 (which covers the reference cell test method) was established and then used to measure the performance of several PV modules. Two different reference cells were used for these first generation flash tests. The rated performance of the selected modules was also determined from tests conducted outdoors using a well-instrumented, two-axis solar tracker. The comparison of indoor versus outdoor measurements provided a baseline that will hereafter be used to evaluate the impact of changes, both positive and negative, to the solar simulator measurement process. The indoor data will also provide a starting point for evaluating measurement repeatability and lamp degradation over time.
During FY08, NIST will document the rated power outputs from a series of tests using both the reference cell and reference module test methods. In both cases, test specimens will span from those offering a near ideal match with the reference cell or reference module to those that are completely different from the reference devices. Test variables that are part of the test protocol – i.e., location of the reference cell and test specimen within the test plane, the time delay before sampling IV data, the number of data points, single long-pulse versus multi-pulse mode, etc. – will be revisited to provide a comprehensive listing of how much the rated power output can change. A few tests will also be conducted to evaluate if the IEC reference module standard (Standard 60904-6) should be altered such that the solar simulator is calibrated based on matching the module’s rated power output rather than its rated short-circuit current. Finally, NIST will pursue making its high-speed spectroradiometer an integral part of the solar simulator test and measurement system. The spectroradiometer will be triggered to measure the light’s spectrum coincident with the IV curve of the test specimen. The measured spectrum will be used to translate the measured performance of the test specimen to the performance it would obtain under any prescribed solar spectrum.
NIST will document the range of performance and associated measurement uncertainties from using various test protocols that otherwise comply with the IEC standards 60904-1 and 60904-6. The results are likely to differ among PV cell technologies (crystalline silicon based versus thin film materials) and the size/time response of the PV module. Performance mapping of the solar simulator’s spectral output and spatial uniformity will be repeated, along with a few of the test specimen measurements first performed in FY07. These tests will help determine the degradation rate of the solar simulator lamp and help quantify the repeatability of the overall measurement system over time.
In FY07, NIST completed long term testing on 8 BIPV roofing products. Testing of one additional product will be completed in March 2008. Short-term characterization tests on most of these nine BIPV roofing products was also completed in FY07; a few remaining tests will be conducted during the first half of FY08. BIPV roofing technologies include 4 products designed to integrate within a residential concrete tile roof, 1 product that is designed for a residential slate roof, 1 product that is designed for an asphalt shingle roof, and three products that are designed for flat/low-sloped commercial building roofs. NIST will investigate the ability of at least two computer models to predict the electrical performance of these BIPV roofing products.
At the end of FY07, NIST entered into a Cooperative Research and Development Agreement (CRADA) with BP Solar, a leading manufacturer of solar photovoltaic products. The CRADA calls for the sharing existing test facilities at both NIST-Gaithersburg and BP Solar in Frederick, Maryland and expanding the NIST outdoor test facilities. Both short-term indoor and outdoor long-term testing of different PV products are planned. The CRADA-related work will complement the solar simulator test protocol and modeling work described above.
In FY07, a Phase I SBIR project that looked into the feasibility of developing a Light Emitting Diode (LED) based solar simulator light source was completed. The SBIR contractor addressed several of the technical hurdles of such a design and even delivered a small-scale pre-prototype assembly to NIST. Overall, the company made substantial progress. The company was thereafter approved for Phase II funding to continue its work. NIST researchers covered under the STRS-funded Photovoltaic Measurement Techniques project will continue to monitor and assist the progress of the SBIR project. Involvement in this project complements other efforts planned for FY08 on investigating completely new ways to go about testing PV modules indoors.
|
