With the push to expand the amount of energy produced from renewable sources and the maturing of the solar energy industry, utilities are now beginning to add power from multi-megawatt photovoltaic projects into the utility grid. To make these installations valuable and cost-competitive with other energy sources at a utility scale, solar technology must continue to get less expensive and perform better.
The U.S. Government has noted the challenges and, in 2008, the Dept. of Energy commissioned the $24 million Solar Energy Grid Integration System (SEGIS) project as part of the Solar America Initiative (SAI). Its task is to encourage the development of new technology and products that help put solar photovoltaic (PV) systems into American utility grids.
Now five SEGIS teams across the country are working on various aspects of the project. PV Powered, Inc., maker of solar power inverters in Bend, Ore., leads one of them. Its work focuses on improving system economics through optimized energy harvest, reliability, and efficiency, developing technologies for utility-scale monitoring and control system integration and researching ways of reducing the impact of weather changes on solar power output. The hope is that SEGIS will bring advances that will lead to less downtime and more predictable operation, improving the overall economic equation for solar power plants.
There are many factors involved in the efficiency of energy harvest, but one of the most important is MPPT, or Maximum Power Point Tracking. This refers to how the PV electronics adjust the voltage/current ratio delivered by a solar array to maximize power as the output from the array changes. The array output might change substantially as clouds go by, the cells get dirty, and so forth.
Tests of this system will take place on a wide variety of configurations. These will range from high fill-factor technologies like concentrated solar or mono-crystalline silicon to low fill-factor thin film technologies, including copper indium gallium arsenide and cadmium telluride.
Large-scale solar power plants are typically modular arrangements. They may include several arrays, inverters, and other components configured somewhat in parallel. The idea is to prevent the whole system from shutting down because of a single-point failure or through a propagating fault. Instead, when a single subsystem fails, there is only a fractional loss in power output.
In systems of this nature, it is tough to make accurate reliability predictions. Reliability data can, however, be collected on the different subsystems (power inverters, modules, and tracking devices) and combined for a system-level analysis. Designers can use this information to project preventive maintenance schedules, budgets for spare parts, estimates of downtime, and the PV plant's overall financial viability. These projections are generally quite accurate. The resulting subsystem downtime estimates can be used to calculate system-level availability, power loss, and energy harvest.
Designers need to understand the installation environment to estimate the reliability of a PV system. PV systems are unlike hydroelectric, nuclear, and coal- or gas-fired power equipment because they aren't typically located in controlled surroundings such as a building. PV systems outdoors see environmental stresses such as humidity, dust, and so forth, as well as temperature extremes.
Continue to next page.
These factors depend on the geographic location of the installation and can significantly affect the reliability analysis. It takes complex time-dependent modeling to accurately predict component stresses and wear caused by natural sun cycles. Simple linear failure rate calculations won't work because temperature cycling contributes to device wear-out. Time-dependent prediction tools and analytical methods are a must to accurately predict component reliability.
It is essential to monitor some of the key individual components, not just the ac output of the entire system. For example, some systems add a “smart combiner” between the arrays and system inverter to monitor current from individual strings or groups of strings. An architecture is being developed through the SEGIS program that can better manage the system and produce more energy at an acceptable cost.
Weather changes and fluctuations in cloud cover can lead to instability of the power a PV system sends to the grid. Irradiance forecasting can mitigate the impact of these transients. The SEGIS program is furthering irradiance forecasting at both the inverter and utility level. Forecasting made at the inverter level can smooth variations in power going to the utility using intelligence on cloud position, movement, and transparency gathered from several sources. It also brings the possibility of strategically combining small power storage units into the overall PV system to smooth the power output. Meanwhile, utility-level forecasting factors transient effects into real-time dispatch processes that utilities employ.
Two factors complicate the addition of of solar power into the grid. First, the intermittent nature of sunlight means power is not always available on demand. Second, generating stations are widely distributed geographically, and this brings communications and control challenges.
Both of these issues are complex. They demand a collaborative effort between utilities and suppliers. The final solution will only come when a smart grid communications infrastructure is in place and interconnection standards have been updated to embrace PV as a key energy source. As first steps toward this goal, our SEGIS team is developing enhanced two-way utility communications and control.
SEGIS partner Portland General Electric (PGE) is working with the team to set up two-way communications between PGE's GenOnSys command and control (Scada) system and solar power plants. This will let the utility receive status date and assert control commands as necessary. PGE will also have the ability to dispatch or disconnect distributed resources remotely if needed.
The teams are also working on a means to detect a power island condition. That's when the grid experiences an outage while the PV system pumps power into it. It is important to prevent unintentional power islands. A set of algorithms usually identifies these conditions while also actively controlling and monitoring solar inverters.
This is an imperfect solution, however. Inverters cannot differentiate between a grid disturbance, when the PV system could assist in stabilizing things (e.g., in the case of a brownout); and a true utility outage, when it would have to disconnect. Without smart anti-islanding, an inverter might disconnect from the grid just when its additional power was needed most.
The PV Powered SEGIS team is preparing to demonstrate a new way of detecting islands with synchrophasor measurements taken between the solar power plant and a utility reference. This technique involves multiple phasor measurement units (PMUs). PMUs basically measure line-voltage and line-current amplitude and phase angle at different locations in a power system, using a single absolute time base. By correlating values from different locations with the time they take to arrive at a common collection point, the inverter can differentiate between a true unintentional island and a case when the PV plant needs to support the grid.
All in all, the development of PV systems is moving along at a rapid pace. Technology and standards coming out of partnerships such as the SEGIS program will help solar power fulfill its promise as a critical renewable energy component on the national grid.
PV Powered Inc., Bend, Ore., www.pvpowered.com/
FAQ for the Solar Energy Grid Integration Systems program (SEGIS), www.sandia.gov/SAI/