By Ray Saka, Sales Manager, TMEIC
In 2013, the U.S. solar market is projected to make the 5-GW mark with new installations. Utility-scale projects are expected to account for almost 2.5 GW of that total. As more systems are installed, system reliability and performance becomes an increasing concern for owners, utilities and independent system operators (ISOs). When viewing reliability and performance of the PV power plant, there are three levels to consider including the component, system and plant levels.
Component Level
Inverter failures can significantly impact the overall energy harvest of the plant. System owners have modeled this impact through advanced calculation based on uptime by using failures rates such as mean time between failures (MTBF), and mean time to recovery (MTTR). Catastrophic failures on insulated-gate bipolar transistors (IGBTs) or the cooling system would result in entire inverter shutdown, decreasing the uptime of the system. Inverter manufacturers are challenged with high temperature and power stress of the IGBTs.
New technology such as multi-level topology and advanced cooling systems help mitigate these risks. Multi-level topology distributes stress among a higher number of IGBTs to reduce the voltage stress on individual IGBTs. Cooling technology simplifies and reduces the number of components to lower failure rate and so increase uptime. There are several types of cooling technology including liquid cooled, air cooled and convection cooled. Both liquid-cooled and air-cooled systems are challenged with high failure-rate moving mechanical parts (fans, pumps, etc.) On the flipside, convection cooled systems are simple cooling systems with lower failure rates; however, they are not as efficient at cooling compared to liquid-cooled. An example of hybrid cooling by using air-cooled and convection cooled technology leads to mitigation of reducing mechanical moving part count, and at the same time providing an efficient cooling system. It’s the best of both worlds, so to speak. Redundancy of cooling system components is also a mitigation factor to reduce downtime; however, this needs to be carefully evaluated due to the increase in cost.
System Level
It is also important to consider the performance of the inverter as a system. This means considering inverter performance and reliability when the inverter is connected to a DC source (the PV panels) and AC (typically a medium-voltage step-up transformer). Most developers use performance-modeling software such as PVSYST, which models efficiency and the DC voltage MPPT window. However, design risks associated with interactions between the DC array and the inverter are not modeled in PVSYST.
With the recent panel price degradation, many PV designers are starting to load more DC power above the inverter’s AC nameplate to capture as much energy harvest as possible. Since inverters are current-limiting in nature, as more DC power is loaded, DC voltage slides upward. This becomes more critical as the system encounters situations such as the cloud-edge effect and when discussing low voltage ride through (LVRT) functions.
The cloud-edge effect is when a thin distributed cloud passes over the DC array and the edges of the cloud intensify the irradiance moving instantaneous transient power to the inverter. This can potentially cause an over-voltage issue in the inverter reaching above its nominal MPPT range thereby shutting down the inverter. Almost all inverters use an IGBT switching device to convert DC (PV Panel) to AC (grid) power. These devices have limitation based on design factor. Some designs would not allow you to safely track MPPT beyond 800V for a 100-V system for instance. When cloud-edge effect happens in the field, the inverters typically limit the current, thereby raising the DC voltage above nominal condition reaching above MPPT range. This will cause the inverters to shut-down; thus, accumulating downtime.
The MPPT range also comes into play when discussing LVRT function in the inverter. LVRT is a new requirement in the industry, and so it has not been studied as much. While riding through a fault event using the LVRT feature, the voltage could be as low as 10% or 20% of nominal condition. This would mean, restriction of AC power down to 10% to 20% of nominal condition, and it would mean the same for the DC side. While limiting and riding through of the overall output of the inverter due to grid fault event, the DC side voltage will rise. Depending on the DC array design and DC loading factor, the voltage may shoot beyond the allowed MPPT window of the inverter, causing inverter shut-down. Inverter and system design should be carefully discussed and assessed in terms of real-life scenarios, IGBT design and tolerance to these types of conditions. Having a wider MPPT window is quite important to withstanding cloud-edge effects and LVRT situation.
Plant Level
A PV power plant consists of two or more power blocks paralleled together and interconnected to the grid at the substation. As many inverters are connected, intelligent power optimization using one central power controller becomes increasingly important. For example, inverters that are capable of generating above their nominal kW typically have reserved capability for reactive power output, which can be used as active power when needed. While cloud coverage is causing energy loss in some of the plant’s blocks, the other healthy part of the plant would be able to compensate for the lost energy by using this feature and therefore optimizing plant energy harvest.
Intelligent power optimization is also helpful during inverter failures and maintenance. While some inverters are shutdown during maintenance, other inverters are online and able to compensate for those losses. The reactive power and voltage can also be dynamically controlled with one centralized controller to match what is required at the point of common coupling (POCC). As the larger bulk of PV power assets are being installed in the grid system, being able to interact and support the grid system becomes increasingly important.
By Ray Saka, Sales Manager, TMEIC
Want more? Try these articles:
Do We Need Solar Inverter Reliability Standards? Here’s What Manufacturers Think
2014 Trends: Central Solar Inverters
How To Calculate Solar Reliability