The performance of a solar cell is based on its ability to convert sunlight into electricity. Because much of the sunlight is reflected or absorbed by the silicon-based material that makes up the cell, a typical commercial solar cell has an efficiency of 15%. Photovoltaic (PV) cells generate between 2W and 3.5W — not even enough to power a digital watch or calculator — so they must be combined into modules consisting of about 40 cells.

A typical flat-plate PV array consists of 10 modules that can measure up to several meters on a side and can be mounted at a fixed angle facing south or on a tracking device that follows the sun. Residential applications use 10 to 20 PV arrays to provide enough power for a household. Commercial and industrial applications may require hundreds of interconnected arrays.

The typical warranty for the cells within a PV module comes from the manufacturer and ranges between 20 and 25 years. “The modules, covered under a 25-year warranty, are going to produce electricity for that time period,” says Josh Haney, manager of operation and maintenance for El Cajon, Calif.-based commercial and public-sector grid-connected solar electric power contractor Borrego Solar, a certified solar installer under the North American Board of Certified Energy Practitioners (NABCEP), Malta, N.Y.

The solar cells themselves are durable. “It's not normal glass, as many people think,” says Anne Schneider, spokesperson for cell manufacturer Solar World Industries America, Camarillo, Calif., explaining that the cells require very little maintenance. “It's a very special glass. You can even jump on them. Just wait until it rains — then the maintenance is done pretty much. Just wash the dust off. That's it.”

PV cell failure can occur, however, from blind spots caused by a constant shadow on a certain portion of the module. But failure of a module rarely comes from the PV cells themselves. Reliability issues for PV technology usually stem from the inverter. (For more information on the life cycle of PV modules as well as recycling considerations, see Recycling Requirements Going Forward on page C20.)

Inverter logic

In comparison to the decades-long warranty for the cells, mean time to first failure (MTFF) for inverters is estimated to be between five and 10 years (see Reasons Inverters Fail below). “The inverters are really the item that's going to fail from the normal life of the system,” Haney says. “They have the moving parts.”

A PV inverter converts the direct current (DC) from a PV array into alternating current (AC), so the energy generated by the modules can be used to power appliances or be connected to the electric utility grid. Grid-tied inverters match phase with an electric utility-supplied sine wave and currently are designed to shut down automatically when they're out of contact with the utility supply. “Inverters are equipped to sense the voltages of the grid,” says Haney. “If the voltage spikes from the grid outside of the operating window of the inverter, the inverter will shut off and won't turn back on until the voltage from the grid is back within its operating voltage range.”

Inverters are comprised of: passive electronic components, such as capacitors, inductors, transformers, wires, and PCBs; active electronic components, such as power transistors, diodes, and integrated circuits; and electromechanical components, such as relays, switches, connectors, and fans. The electromechanical components are more likely to fail sooner than the other component types. However, capacitors and the inverter bridge are not immune to aging, stress, usage beyond their operational limit, and thermal shock and overload. “The inverter can also have computer components that can fail,” Haney says.

High maintenance

Luckily, there are maintenance procedures that can extend the life span of the inverters. “Electricity performs better under cool conditions,” Haney says. “By cleaning off the heat sinks, you allow it to run cooler, which is more efficient.”

Other items that may extend the life of an inverter, according to Xantrex Technology, Burnaby, Canada, are relay activation under no-load conditions, variable-speed fans, operating the DC disconnect under no-load condition, avoiding contamination at contacts, placing the components within a controlled environment, proper seals, not installing the module in a horizontal position, proper torque level at terminal blocks, clean wires, and avoidance of extreme temperature conditions.

In addition, preventing failures of inverter components can be written into the system design. “In the design phase, we look at how long a system is going to perform,” Haney says. “One thing you need to make sure is that the modules fit well with the inverter. So over time, as the module degrades, the voltage output of the module will actually be lower. If, after 20 years, it drops down below the operating voltage of the inverter, that can be a problem. So by identifying that in the design phase, we can prevent those problems years down the line.”

For the installations under maintenance contracts, Borrego Solar connects a data acquisition system to the inverters. “The inverters are really smart and have set error codes they will send out so I can identify from a computer exactly what's wrong with the inverter without even going to the site,” Haney explains.

This way, the company can check voltage, current, system production, and local weather conditions. “We make sure everything's functioning correctly,” Haney says. “We monitor online how the system is performing. Every other week, we compare how the system is performing to how we expect it to perform so we can identify possible issues before the customers see them on their electric bill. Along the same lines, a lot of the data acquisition systems are set up so that if there is some sort of fault, it instantly sends an e-mail out, and then I would send someone out to troubleshoot any potential problems.”

In addition, good communication with the electric utility can allow the inverter to work as specified. “If an inverter frequently has an error because there's a lot of surges, that will help us identify if there's a problem,” Haney says. “Typically, we'll try to adjust it first by contacting the electric utility and having them check their power lines to make sure that they're sending the appropriate voltages to the facility. Sometimes, we've identified problems where the electric utility has gone out and they needed to do some work on their end to clean up the power. In other cases, we can do some things on our end. But for the most part, if there's clean power from the electric utility, the inverter will match that clean power — and it'll work as intended.”

Using certified and listed components is another way to maximize maintenance efforts. “All the inverters we use are listed to be connected to the grid so they don't interfere or work inefficiently with the existing grid,” Haney says.

Finally, choosing an inverter that doesn't contain capacitors may extend the life of your system. Three-phase inverter bridges have continuous input current and don't need bulk storage capacitors. However, they do need caps for ripple smoothing. Therefore, it's best to choose one with polycarbonate or other dry dielectric materials for slow aging instead of electrolytic material, which may age faster.

At what cost?

Published in April 2008, the Multi-Year Program Plan for 2008-2015 from the DOE's Energy Efficiency and Renewable Energy (EERE) Solar Energy Technologies Program (SETP) establishes a goal of reducing the Levelized Energy Cost (LEC) by 2015 to between $0. 06/kWh to $0.08/kWh for commercial PV systems and $0. 08/kWh to $0.10/kWh for residential PV systems (click here to see Table). Meeting the solar market cost goals, according to the program plan, will result in 5 GW to 10 GW of PV installed by 2015 in the United States and 70 GW to 100 GW by 2030.

However, in order to meet this goal, PV inverter prices would need to decline to $0.25 to $0.30 W (peak) by 2020. Many industry experts believe that lower prices for inverters will lead to further sacrifices in performance and reliability in a market that has already caused manufacturers to focus on lowering first cost over improving reliability.

Some design choices, made on the manufacturer level, could raise the cost and lessen performance of inverters in the name of reliability. For example, to replace electrolytic caps with dry film caps, Xantrex Technology claims the cost per unit of capacity would be 30 times higher and the volume 15 times larger. In this climate, higher price is not an option for manufacturers.

Yet, first cost isn't the main expense for the facility when it comes to inverters. The inverter accounts for 10% to 20% of the initial system cost, according to a report from the National Renewable Energy Laboratory (NREL), Golden, Colo., in conjunction with Navigant Consulting, Inc. (NCI), Burlington, Mass. However, because inverters generally need to be replaced every 5 to 10 years, they result in investment in a new inverter three to five times over the life of a PV system.

In “History of Accelerated and Qualification Testing of Terrestrial Photovoltaic Modules: A Literature Review,” authors C. R. Osterwald and T. J. McMahon, NREL, write of the compromise between reliability and cost, claiming that protection for elements of degradation is “imperfect, and a design tradeoff between cost and protection exists. Typically, the costs of the materials used for protecting the internal solar calls (packaging costs) are roughly 50% of the total materials cost.”

According to Russell H. Bonn, Sandia National Laboratories, Albuquerque, N.M., author of “Developing a ‘Next Generation’ PV Inverter,” two of three ways to improve reliability of a new product — over-design and implementing redundancy — result in increased cost. The third way — obtaining a detailed knowledge of the product working environment — requires large numbers of fielded product, something PV systems don't have yet. “For inverters, low cost and high reliability are conflicting goals,” writes Bonn. However, Bonn's “next-generation” inverter will have improved performance, high reliability (10-year MTFF), and improved profitability, which may be the longest industry experts expect inverters to go.

Inverter lifetimes greater than 15 years appear difficult to achieve, says the NREL/NCI report, which indicates that the benefits of designing inverters that can operate for 10 to 15 years or more before replacement are limited. The inverter offerings of the two manufacturers that offer 10-year warranties have a lower part count than other inverters. Fewer parts mean fewer failures, and higher-quality parts can be used while still meeting inverter cost targets. Both manufacturers use a Digital Signal Processing (DSP) chip, which translates the waveform into digital format, allowing it to be processed, evaluated, and modified in very short time periods. As a result, the inverter can respond quickly to a wide range of situations, which improves reliability and efficiency.

Ideally, inverters would last as long as other PV system components — 25 years — but according to NCI, many question whether such improvements will ever be reached at a reasonable cost. In the near- to medium-term, an MTBF of >10 years is more likely, reached through improving quality control, better heat dissipation, and reducing complexity.

Sidebar: Reasons Inverters Fail

Inverters fail due to transients from the grid or photovoltaic (PV) generator, component aging, and operation beyond the designed limits. Following are some common reasons specific components of inverters age quickly or fail:

  • Electrolytic materials age faster than polycarbonate and other dry dielectric materials
  • Voltage stress
  • Continuous operation under max voltage conditions
  • Frequent short-term voltage transients
  • Current stress
  • High current increases the internal temperature
  • Thermal stress on component terminals
  • Charge/discharge rate
  • Ambient conditions (temperature)
  • Mechanical stress
  • Vibrations, hot/cold expansion
  • Usage beyond its rated operating limit
  • Voltage, current, overshoots, inrush
  • Other malfunctioning components
  • Thermal shock
  • Thermal overload
  • Extremely cold operating temperature
  • Component stress
  • Contamination at contacts
  • Extreme temperature conditions

Source: Xantrex Technology

Sidebar: Recycling Requirements Going Forward

Although grid-tied photovoltaic (PV) technology was available as early as the 1980s, the majority of PV systems in the United States have been installed within the last five to 10 years. “We've been in business since 1980, so we do have some older systems, but the majority of the systems have been installed in the last eight years,” says Josh Haney, manager of operation and maintenance for El Cajon, Calif.-based commercial and public-sector grid-connected solar electric power contractor Borrego Solar, a certified solar installer under the North American Board of Certified Energy Practitioners (NABCEP), Malta, N.Y.

Annual grid-connected PV installations increased from 10MW per year in 2001 to 180MW per year in 2006, according to a 2007 study by the U.S. Department of Energy (DOE), resulting in a cumulative installed base of 480MW grid-connected PV in the United States at the end of 2006. In addition, the grid-tied PV segment experienced an 81% growth rate for the amount of installed power in 2008 (292MW) over the amount installed in 2007 (161MW), according to the annual report from the Solar Energy Industries Association (SEIA), Washington, D.C.

As a result of this recent boom in PV system installations, some industry experts are predicting a substantial increase in the amount of maintenance and recycling needed for end-of-life PV modules in the near future. In Europe, for example, end of-life modules are predicted to jump from 290 tons in 2010 to 33,500 tons in 2040, prompting voluntary take-back programs as well as several new European Union (EU) directives focused on waste avoidance, recycling, and eco-design requirements. The aim of these efforts is to achieve lower energy payback times and save resources by reuse and recycling of wafers, silicon, and other material.

Contrary to these programs, in the United States, the EPA's RCRA (hazardous waste definitions) offer only general requirements. Some states provide additional regulations, such as California's limits for hazardous materials, but only a few PV manufacturers currently offer module recycling.

The SolarWorld Group, Freiberg, Germany, recently expanded its solar module recycling program to include the U.S. market and now offers customers the opportunity to return used and end-of-life solar modules. SolarWorld Industries America, Camarillo, Calif., will accept any crystalline solar module — regardless of the manufacturer, cell shape, or size — for recycling.

Newly developed technologies enable the company to recycle materials that have so far not been economically usable. The process is particularly effective with older modules that used thicker cells, consisting of two or even three times as much silicon compared to today's thinner cells.

In the recycling process, the frame is first separated from the module, and then the cells are removed from the laminate and treated in an etching process, resulting in clean wafers, which enables them to be used in the production of new cells and modules. “We have developed a very specific recycling process so we don't put our modules in landfills,” says Anne Schneider, spokesperson for Solar World Industries America. “We take the frames, the glass, and back sheets off, and then we put those materials back into recycling. We take the solar cells, etch them, get the solar wafer again, and then we process these wafers into new cells. Or, if those wafers are broken, we melt them down and grow new crystals out of them.”

With the average lifetime of PV modules expected to exceed more than 20 years, it may seem too soon to think about maintenance and recycling. “This is something that will be an issue in the next five to 10 years,” Schneider says.