It is known that nearby lightning strikes can induce overvoltages in photovoltaic arrays. However, although many laboratory studies have been reported, relatively few field studies have appeared in the literature. This paper reports on a field study conducted at South Dakota State University and the Florida Solar Energy Center in which a high-speed datalogger was used to monitor the voltage at the terminals of two photovoltaic arrays. The design of the datalogger is described, and its measurements are given and discussed. Two important conclusions were reached: one, that the field measurements seem consistent with laboratory measurements and suggest that DC-side protection for PV power electronics is important; and two, that it is not always possible to tell from anecdotal observations whether a lightning storm is dangerous for PV power electronics.

Investigators at South Dakota State University and the Florida Solar Energy Center installed a high-speed datalogger on a high-voltage PV series string as part of an experiment to quantify lightning-induced transients in PV arrays. However, during the measurement period, an unexpected result was observed: large voltage spikes appeared at the PV array terminals during the inverter's morning start-up and evening shut-down procedures. This paper describes those observations, probable causes and effects, and mitigation options.

The Florida Solar Energy Center (FSEC) has partnered with Sandia National Laboratories (SNL), the Southwest Technical Development Institute (SWTDI), and the California Energy Commissions Public Interest Energy Research (CEC PIER) in an effort to characterize the performance of PV inverters operating over extended periods of time. As part of this characterization, SNL executed an initial performance characterization in the laboratory to be repeated at one or two year intervals. These inverters were then sent distributed to the above test facilities for installation and long-term testing. To perform this long-term testing, FSEC built an inverter test facility made up of a configurable 10.1 kW PV array and four test beds. Each test bed was set up to monitor the following parameters: DC voltage, DC current, AC Voltage, AC Power and inverter temperature. Solar irradiance, ambient temperature, and several PV array temperatures were recorded by a separate, synchronized datalogger. Three inverters are currently being tested: the Fronius IG 3000, the Xantrex GT 3.0, and the SMA Sunny Boy 2500U. The three inverters were loaded with a PV array close to the maximum power limit allowed by the inverter, and the array power supplied to each inverter was within 100 W of one another. Since the PV array for each inverter used identical PV modules mounted in the same orientation, it was easier to develop a model to compare inverter performance based on energy output. In addition to energy output, other parameters were obtained such as inverter efficiency. An attempt was made to compare the results obtained at FSEC to those results compiled by the California Energy Commission.

In January of 2003, the Florida Department of Environmental Protection/Florida Energy Office (DEP/FEO) allocated $600,000 in hardware funds toward the installation of photovoltaic (PV) solar systems on Florida schools. As a result of this program, grid-connected PV systems less than six kilowatts in size were installed on 29 schools in the State of Florida. The Florida Solar Energy Center (FSEC) has monitored these systems for approximately one year of operation(1). The performance of 28 of these systems was analyzed using standard performance parameters such as the performance ratio, PV array efficiency, inverter efficiency, and PV system efficiency(2). In addition, a life-cycle cost analysis was conducted using new cost data values and updated market assumptions. These data will serve as a benchmark to compare against future systems with respect to performance vs. installed system cost.

This analysis expands the photovoltaic (PV) life cycle cost (LCC) results presented at ASES 2004. That paper presented the model and concept used to develop PV LCC, and it showed the results of the analysis of over one hundred systems monitored by the Florida Solar Energy Center (FSEC). FSEC began tracking cost, performance and reliability data for systems installed in Florida in 1998, with data now available through a web-accessible database. For the majority of the 124 systems, installed cost information was collected as part of the state’s PV rebate and PV for schools programs. Results presented previously [1] indicated that over an assumed 20–30 system life time a PV system will have a positive life cycle cost. That is, a negative total return on investment. These results were based on actual cost, performance, maintenance, and reliability data. In the baseline case, average total system costs over the lifetime were 32.4¢/kWh while electricity savings totaled 3.7¢/kWh netting a life cycle cost of 28.7¢/kWh. While based on actual data from over 100 installed systems — some installed for over 6 years — a number of conservative assumptions also drove the analysis, such as the exclusion of the state’s rebate programs (varying from $2 to $5 per DC Watt) which impacted nearly all of the systems in the analysis. Since the first presentation of these results the PV LCC model has been further developed to incorporate additional performance information and expands the sample of systems incorporated. This paper will thus provide further insight into the relative importance of various up-front and on-going costs to the overall lifetime economics of a system. The paper will also address additional sensitivity analysis performed. Particular attention is paid to inverter mean time between failure (MTBF), the impact of incentives, and basic financial assumptions used in the model such as the discount rate and electricity rates. Various scenarios are considered in asking the question of what is necessary for the system LCC to break-even.

It has become increasingly apparent that the photovoltaics industry is in need of a resource for performance, reliability and economic information on fielded PV systems. This information would help manufacturers, distributors, utilities and other relevant parties in determining trends in operation, maintenance, and essential costs. By recognizing these recurring patterns, changes can be made in product design, manufacturing processes, installation procedures, or purchasing patterns to address, mitigate and/or correct these issues. Sandia National Laboratories has been developing a photovoltaics reliability database, with the Florida Solar Energy Center providing development and support for a database housing information on grid-tied systems. This paper contains a progress report on the status of the database and some preliminary results of analyses based on this data.

The lack of consistently high performance from inverters has created numerous issues in the widespread application and confidence in grid-tied photovoltaic systems. Additionally, the lack of a metric by which inverter manufacturers can evaluate their products presents a challenge to manufacturers by not allowing them to distinguish their products from those of their competitor, and to users since there is no way of comparing seemingly-similar models. This problem is further compounded by the dearth of inverters that have high performance and reliability [1]. This document is the first draft of a test protocol under development at the Florida Solar Energy Center to assist in this area. A need for such a test protocol has been recognized nationally, and it is hoped that this document may provide a foundation for future test procedures while providing significant and reasonable findings in the interim.