The Design of Solar Inverters
The photovoltaic (PV) industry is working to overcome several technical challenges in order to become a trusted and reliable energy provider on a truly large scale.
By Marc Johnson, PV Powered Inc.
The photovoltaic (PV) industry is working to overcome several technical challenges in order to become a trusted and reliable energy provider on a truly large scale. Among these challenges are the needs for improving PV system reliability, maximizing total energy harvest and solving the unique problems associated with high-penetration of PV generation on the grid. A key element of the solution is solar inverter technology, which is receiving intense focus as part of multi-disciplinary research and development programs.
This mobile solar cart has been constructed by PV Powered to test out different solar array technologies as part of the SEGIS R&D program.
Improving Inverter Reliability
Since inverter failures have been responsible for approximately 80 percent of all PV system downtime, increasing inverter reliability can dramatically improve the overall reliability of a PV system. Improved reliability not only improves the return on investment for the PV project by avoiding lost energy production, it minimizes the cost of repairs or replacement of failed inverters over the lifetime of the project.
Most PV inverters on the market today are based on product designs that originally came with a one-year warranty. Market forces have raised the bar to where a five-year minimum warranty is now expected by most customers, and most inverter manufactures have adapted by incorporating the cost of anticipated repair or replacements during a five-year (or in some cases, 10- or 15-year) warranty period into their pricing structure. PV Powered commercial inverters came into existence in a fundamentally different way. Starting with a clean sheet of paper, they designed a set of products that are intended to have a 20-year minimum productive lifetime. This required a fundamentally different approach to inverter design, and the resulting products are different in some important ways.
One of the most obvious differences in inverters designed by PV Powered can be appreciated by simply looking inside the cabinet. They have a very low component count. Although such a design requires substantially more engineering to develop, this practice is based on one of the most fundamental and well-understood features of design for reliability: The reliability of a system is the product of the reliabilities of all the individual components. Each component has an annual reliability of 99.99 percent and, therefore, has only a 0.01 percent chance of failing in one year. Now, suppose that an inverter is constructed from 100 such components, and that the inverter would fail if any one component fails. The resulting inverter would have an overall reliability of 0.9999100, or 0.99, (99 percent). To refine the familiar analogy, a chain is actually weaker than its weakest link.
Ninety-nine percent is a fairly typical availability figure for most PV inverters, and this is an impressive achievement, considering the challenging environment in which PV inverters must operate. Although PV systems with an availability of 99 percent have been considered quite good by consumer and commercial customers, utilities have a much higher expectation of reliability for components on their system, and PV systems must live up to those expectations to be taken seriously by utilities.
System availability, or uptime, also depends on the time required to make repairs when a failure does occur. In light of this, PV Powered has engineered its commercial inverters so that most of the major components can be replaced within 30 minutes.
Other factors that influence the reliability of inverters are equally important but difficult to see. Significant improvements in reliability are achieved in the design stage by applying techniques that have proven themselves in the aerospace industry. These include very rigorous qualification of components, extensive modeling at the design stage, and sophisticated thermal management.
Proper component selection begins with choosing appropriately rated devices and utilizing them in a manner that will lead to long lifetimes. Each manufacturer must be vetted for quality manufacturing processes and must provide the data to enable PV Powered to model reliability over the required 20+-year lifetime. Once a part has passed the initial qualification process, its performance is simulated as part of a real system, including the expected normal and worst-case stresses it will be expected to endure. Next, each component is tested in the laboratory—first on the bench and then as part of a completed inverter. During testing, measured values are compared with predicted values, and these data are fed back to refine the modeling process.
Effective thermal management is an essential part of the reliability of any electronic system and is even more critical for a PV inverter, which may be required to endure both extremely hot and cold ambient temperatures and daily temperature cycles of 30 C or more. Thermal management in PV Powered commercial inverters is accomplished with a fully integrated mechanical design that is simple, reliable and which delivers exactly the cooling that is required to each part in the system. Forced convection cooling is used because it provides superior cooling performance at a lower cost and mechanical complexity than with other types of cooling (e.g., liquid cooling).
Air is drawn in at the top of the inverter, filtered and exits the bottom through fine mesh stainless steel screens. This greatly reduces the introduction of dirt and corrosives into the sensitive electronics inside. In addition, the entire cabinet operates under positive pressure; so, air can only enter the cabinet after being filtered. The stainless steel mesh over the exhaust ports assures that animals will not be able to enter the cabinet and cause problems—a common cause of failure in electronics exposed to a natural environment. All cooling airflow is achieved using a set of redundant blower motors. Although the blower motors have been carefully selected, modeled and tested to assure a 20+-year lifetime, the redundancy means that the inverter can operate at full power at its maximum rated ambient temperature (50 C) with one blower not functioning. All cooling air is routed to each component by means of ducting and vents. In the event of a blower failure—or even a clogged filter—the inverter will send an alert, allowing a repair to be scheduled at a convenient time. The inverter will continue to operate at full power while waiting for the repair to be made, so that no energy production is lost.
Another essential component of any reliability program is performance monitoring and feedback. If an Internet connection is provided, PV Powered commercial inverters automatically transmit a rich set of performance data to their server. This data is monitored for fault codes, warnings and trends that may indicate a failure is likely to happen in the future. This service is provided at no cost as long as the inverter has Internet connectivity. Since many commercial PV systems are already monitored by third party performance monitoring systems, adding a network connection is often as simple as adding a LAN cable.
Using the above reliability-improving techniques, the data show that the availability for the installed base of PV Powered inverters is greater than 99.9 percent, reflecting a tenfold decrease in failure rates compared to legacy inverter designs. As PV Powered refines its techniques and gains experience, its total fleet availability is expected to continue to increase.
PV Powered's 260 kW commercial solar inverter is specially designed to maximize the energy harvest over a 20+-year lifetime in utility-scale solar power systems.
Enhanced Grid Support
As grid penetration of solar energy increases, technical hurdles scarcely envisioned by the pioneers of grid-tied PV are becoming commonplace, and solutions to the problems must be developed for PV production to achieve the scale needed to meet the renewable energy needs of the future. Power factor control, low-voltage ride-through, wider frequency tolerance, SCADA command and control, enhanced anti-islanding and optimized maximum power point tracking are some of the features PV Powered is developing through the Department of Energy Solar Energy Grid Integration System (SEGIS) program. SEGIS is a competitive program funded by the DOE to develop new technology to help enable high-penetration of solar generation on the grid. The team led by PV Powered received the largest second stage grant of the five teams remaining in the program.
Power factor control is one technical requirement rarely thought about by PV system designers until now. Distributed generation on the large scale that is being planned by PV system developers was never envisioned by utility companies when their lines were constructed—in some cases nearly a hundred years ago. Backfeeding several megawatts at the end of a rural distribution line can cause an unacceptable voltage rise on the utility system due to the impedance of the relatively small lines at that point in the power system. Reconductoring the feeder can cost several hundred thousand to over a million dollars per mile, and the cost would often destroy the financial viability of a PV project. One novel solution is to use the natural capabilities of the PV inverters to operate the plant at a specific lagging power factor—limiting voltage rise on the power lines. PV Powered is actively working with several utilities and PV project developers to use power factor control to limit voltage rise, thereby avoiding the need for new transmission lines. This approach not only avoids more costly methods of voltage control, it greatly accelerates the project schedule because the lead times on any modifications to the utility system are typically very long.
The PV Powered SEGIS team is also actively developing optimized maximum power point tracking (MPPT) algorithms to improve energy harvest with several different PV array technologies. For example, the PV community has discovered that the algorithms currently in use for finding and tracking a PV array's maximum power point are inefficient for many thin film modules, and vulnerable to sub-optimization in large arrays. Improvements in these algorithms will result in increased energy harvest—especially as PV systems age and the array power curves change from the idealized power curves for which current algorithms are designed.
Low voltage ride-through is another new frontier for PV inverters. Current utility rules for PV inverters were developed before it was envisioned that total generation from PV on some segments of the grid would be as high as 10, 20 or even 30 percent. Suppose a particular distribution line was receiving 30 percent of its power from local PV generation, and the line experienced a voltage sag down to 80 percent of nominal voltage. Current regulations would require that the PV inverters trip offline in not more than two seconds, which would exacerbate the voltage sag. Many utilities are now rethinking the conditions under which the inverters trip off, and PV Powered is demonstrating the capabilities of PV inverters to provide grid support during abnormal events such as unintended brownouts.
Related to this are developments being made to support enhanced anti-islanding (avoiding the condition where the PV system is producing power into a utility grid that is experiencing an unplanned outage). Along with PV Powered's other SEGIS partners, Schweitzer Engineering Labs and the utility Portland General Electric, the firm is developing a novel and greatly improved method of detecting and preventing unintentional power islands. An outgrowth of this work will give utilities the option to intentionally island a feeder in the event of an outage, thereby using the PV inverters to keep the power on in conditions where a total outage would have occurred in the past.
Utilities are accustomed to monitoring and controlling assets on their systems with their own SCADA systems. Although it will require some work for inverter manufacturers to adapt to providing this type of connectivity, it will enable PV systems to more peacefully coexist on the utility system. Under the SEGIS grant, PV Powered is developing the capability to not only interface with utility SCADA systems, but with the building management systems and energy management systems found in many commercial buildings today.
With ongoing research and development, including enhancement of the design of utility-scale inverter systems to resolve the issues described above, solar power will fulfill its promise to become a major component of the nation's renewable energy portfolio.