Dirk Jordan

Outdoor PV degradation comparison

As photovoltaic (PV) penetration of the power grid increases, it becomes vital to know how decreased power output may affect cost over time. In order to predict power delivery, the decline or degradation rates must be determined accurately. At the Performance and Energy Rating Testbed (PERT) at the Outdoor Test Facility (OTF) at the National Renewable Energy Laboratory (NREL) more than 40 modules from more than 10 different manufacturers were compared for their long-term outdoor stability. Because it can accommodate a large variety of modules in a limited footprint the PERT system is ideally suited to compare modules side-by-side under the same conditions.

Testing and analysis for lifetime prediction of crystalline silicon PV modules undergoing degradation by system voltage stress

Acceleration factors are calculated for crystalline silicon photovoltaic modules under system voltage stress by comparing the module power during degradation outdoors with that in accelerated testing at three temperatures and 85% relative humidity. A lognormal analysis is applied to the accelerated lifetime test data, considering failure at 80% of the initial module power. Activation energy of 0.73 eV for the rate of failure is determined for the chamber testing at constant relative humidity, and the probability of module failure at an arbitrary temperature is predicted. To obtain statistical data for multiple modules over the course of degradation in situ of the test chamber, dark I–V measurements are obtained and transformed using superposition, which is found to be well suited for rapid and quantitative evaluation of potential-induced degradation. It is determined that shunt resistance measurements alone do not represent the extent of power degradation. This is explained with a two-diode model analysis that shows an increasing second diode recombination current and ideality factor as the degradation in module power progresses. Failure modes of the modules stressed outdoors are examined and compared with those stressed in accelerated tests.

Analytical improvements in PV degradation rate determination

As photovoltaic (PV) penetration of the power grid increases, it becomes vital to know how decreased power output may affect cost over time. In order to predict power delivery, the decline or degradation rates must be determined accurately. For non-spectrally corrected data several complete seasonal cycles (typically 3–5 years) are required to obtain reasonably accurate degradation rates. In a rapidly evolving industry such a time span is often unacceptable and the need exists to determine degradation rates accurately in a shorter period of time. Occurrence of outliers and data shifts are two examples of analytical problems leading to greater uncertainty and therefore to longer observation times. In this paper we compare three methodologies of data analysis for robustness in the presence of outliers, data shifts and shorter measurement time periods.

The Dark Horse of Evaluating Long-Term Field Performance—Data Filtering

This paper addresses an issue of long-term performance that has seen relatively little attention in the industry, yet we will show that it can be of vital importance, not only for obvious financial reasons but, technically, because of its linkage to field failure as well. We will discuss how different data filtering on one particular system can lead to a variety of different degradation rates compared with indoor measurements and how it may change the field failure interpretation for a single module. A method based on the variation of the uncertainty in the determined degradation rates is proposed to aid the data filtering process when no baseline measurements exist. Finally, based on this experience, we propose a set of guidelines as a basis for a standardized approach to long-term performance assessment.

Performance and Aging of a 20-Year-Old Silicon PV System

This paper examines the long-term performance of a 20-year-old crystalline silicon system. The degradation was found to be about 0.8%/year and is consistent with historical averages. In contrast with the majority of historical degradation, the decline for this particular system is more attributable to fill factor (FF) than to short-circuit current (Isc). The underlying cause was determined to be an increase in series resistance. We found no evidence that this system degrades significantly differently from its worst performing strings and modules. Maintenance events during the 20 years were dominated by inverter issues with a total of four replacements.

Photovoltaic Investment Risk and Uncertainty for Residential Customers

The revenues generated by rooftop photovoltaic (PV) systems have several sources of uncertainty. We use a Monte Carlo framework to explore the sensitivity of PV investment returns to three categories of PV investment uncertainty: 1) interannual solar variability, 2) PV technical performance and maintenance costs, and 3) market risks including future electricity rates and the possibility that retail electricity rates will be restructured for PV customers. We find that PV investment risk and uncertainty is driven by market factors in some U.S. regions (California and Massachusetts) and by the PV technical performance in other U.S. regions (Missouri and Florida). We explore the relative impacts of three methods for reducing PV investment uncertainty: research-and-development-driven performance improvements, system performance guarantees that are common for third-party owned systems, and long-term power purchase contracts. We find that the effectiveness of each risk reduction option varies by region, depending on which factors drive regional PV investment uncertainty.

Thin-film reliability trends toward improved stability

Long-term, stable performance of photovoltaic (PV) modules will be increasingly important to their successful penetration of the power grid. This paper summarizes more than 150 thin-film and more than 1700 silicon PV degradation rates (Rd) quoted in publications for locations worldwide. Partitioning the literature results by technology and date of installation statistical analysis shows an improvement in degradation rate especially for thin-film technologies in the last decade. A CIGS array deployed at NREL for more than 5 years that appears to be stable supports the literature trends. Indoor and outdoor data indicate undetectable change in performance (0.2±0.2 %/yr). One module shows signs of slight degradation from what appears to be an initial manufacturing defect, however it has not affected the overall system performance.

Measuring degradation rates without irradiance data

A method to report photovoltaic (PV) system degradation rates without using irradiance data is demonstrated. First, a set of relative degradation rates are determined by comparing daily AC final yields from a group of PV systems relative to the average final yield of all the PV systems. Then, the difference between relative and absolute degradation rates is found using a Bayesian statistical analysis. This approach is verified by comparing to methods that utilize irradiance data. This approach is significant because PV systems are often deployed without irradiance sensors, so the analysis method described here may enable measurements of degradation using data that were previously thought to be unsuitable for degradation studies.

Acceleration factor determination for potential-induced degradation in crystalline silicon PV modules

Potential-induced degradation in conventional p-type silicon-based photovoltaic solar cell modules is described as a failure mechanism involving positive ion migration, understood to be primarily Na+, drifting from the glass to the cells in negative-voltage arrays. Acceleration factors for this mechanism are determined for silicon photovoltaic modules by comparing the module power during degradation outdoors to that in accelerated testing at three temperatures and 85% relative humidity. A lognormal analysis is applied to the accelerated lifetime test data considering failure at 80% of the initial module power. Activation energy of 0.73 eV for the rate of failure is determined for the chamber testing at the constant relative humidity, and the probability of module failure at an arbitrary temperature is predicted. Estimation of module power in-situ in the environmental chamber is achieved using dark I-V measurements transformed by superposition. By this means, the power of the degrading module can be semi-continuously determined so that statistical data for multiple modules undergoing potential-induced degradation can be easily and accurately obtained.

Electrical bias as an alternate method for reproducible measurement of copper indium gallium diselenide (CIGS) photovoltaic modules

Light-to-dark metastable changes in thin-film photovoltaic (PV) modules can introduce uncertainty when measuring module performance on indoor flash testing equipment. This study describes a method to stabilize module performance through forward-bias current injection rather than light exposure. Measurements of five pairs of thin-film copper indium gallium diselenide (CIGS) PV modules indicate that forward-bias exposure maintained the PV modules at a stable condition (within 1%) while the unbiased modules degraded in performance by up to 12%. It was also found that modules exposed to forward bias exhibited stable performance within about 3% of their long-term outdoor exposed performance. This carrier-injection method provides a way to reduce uncertainty arising from fast transients in thin-film module performance between the time a module is removed from light exposure and when it is measured indoors, effectively simulating continuous light exposure by injecting minority carriers that behave much as photocarriers do. This investigation also provides insight into the initial light-induced transients of thin-film modules upon outdoor deployment.