Reactive Power Control in Utility-Scale PV
Inside this Article
To facilitate higher levels of distributed PV penetration and meet transmission interconnection requirements, inverter-based generators need to participate in utility-voltage regulation by absorbing or injecting reactive power.
Reactive power control may be one of the more confusing topics in the PV industry, but it is an area of significant promise. In this article, I provide a high-level overview of how voltage is controlled in transmission and distribution systems. I also summarize the effect of inverter-based distributed energy resources on electrical power system voltage levels and describe how reactive power control schemes can be used to mitigate those impacts. I discuss the evolving technical and jurisdictional requirements for reactive power control, which account for much of the confusion about this topic. I present some useful information for PV plant designers related to qualifying inverters for use in projects with reactive power control requirements and quantifying the system performance implications associated with meeting these requirements. I compare inverter-level versus plant-level reactive power control schemes as well as reactive power control requirements for transmission-level versus distribution-level interconnections, and then conclude with a brief discussion of compensation for reactive power capabilities. (For an overview of reactive power in ac circuits, see “Reactive Power Primer,”)
The American National Standards Institute (ANSI) defines acceptable (Range A) and tolerable (Range B) service and utilization voltage levels in ANSI C84.1-2011, “American National Standard for Electric Power Systems—Voltage Ratings (60 Hertz).” Electrical equipment manufacturers design their products to operate within the voltage tolerance boundaries described in this standard. Electric power system operators design and operate the utility grid to maintain those voltage levels under normal operating conditions.
Reactive power management is an essential part of how voltage levels are controlled in the electric power system. In effect, reactive power can be generated as a means of raising voltage levels or absorbed as a means of lowering voltage levels. System designers and operators coordinate reactive power compensation devices with voltage regulators—tap-changing autotransformers located at the source of or along distribution feeders—and on-load tap changers, which allow for stepped voltage regulation by automatically adjusting the turns ratio of substation transformers.
In the article “Ancillary Service Details: Voltage Control” (see Resources), Brendan Kirby and Eric Hirst summarize some of the challenges associated with voltage control and reactive power management at the transmission level: “At very light loading the [transmission] system generates reactive power that must be absorbed, while at heavy loading the system consumes a large amount of reactive power that must be replaced. The system’s reactive power requirements also depend on the generation and transmission configuration. Consequently, system reactive requirements vary in time as load levels and generation patterns change.”
Though traditional synchronous generators are capable of injecting reactive power into the system or absorbing excess reactive power—and can do so quickly—the operating and opportunity costs associated with this method of voltage control are relatively high. Therefore utility operators may seek to minimize the exchange of reactive power between the transmission system and generation units, which operate according to a voltage schedule provided by the transmission system operator. In the event of a disruption, such as the loss of a generator or of a transmission line, the synchronous generator’s excitation system will dynamically inject or absorb reactive power as needed until system voltage stabilizes.
Dynamic and static devices. To maintain the reserve capacity of generator reactive power for contingencies, system operators may prefer to add reactance or capacitance to the transmission system when voltage adjustments are needed. On one hand, switching in inductors reduces voltage at the point of connection by absorbing reactive power from the system; on the other, switching in capacitors raises the bus voltage at the point of connection by supplying reactive power to the system. While power system operators use inductors to control high-voltage levels on transmission systems, they primarily use capacitors to control low-voltage levels on both transmission and distribution systems.
One disadvantage of using static devices like inductors and capacitors for voltage control is that the response speed is relatively slow (on the order of seconds or minutes) compared to that of dynamic voltage control devices, which are continuously variable and can respond much faster. Inductors and capacitors are discrete devices with binary control—they are either on or off. Further, the process of switching these devices on or off can inject voltage transients into the electric power system.
Electric power system designers deploy synchronous condensers, static var compensators (SVCs) and static synchronous compensators (STATCOMs) throughout the network where faster and smoother reactive power control is needed. Synchronous condensers are rotating machines that only provide reactive power support and do not produce any real power. SVCs combine conventional capacitors and inductors with high-speed switching, which allows for a continuous rather than stepped control range. STATCOMs are solid-state power electronic devices—like PV inverters, but without the PV array—that inject or absorb reactive power by synthesizing a voltage that is either greater or less than the voltage at the bus. Though they are not synchronous machines, SVCs and STATCOMs are categorized as dynamic reactive control devices due to their fast response time and variable output.
Note that many commercial and industrial customers manage power factor at the point of consumption to avoid incurring the power factor penalties that utilities assess. These penalties encourage utility customers to operate their facilities closer to a unity power factor, which improves the efficiency and reliability of the electric power system as a whole by minimizing the need for reactive power flows in the distribution system.
Distributed generation. Since the electric power system was not designed to take into account distributed generation and bidirectional flows, utility system operators and regulators have generally treated distributed PV generation as a variable negative load. Therefore, most PV inverters interconnected to the electric power system at the distribution level operate at a unity power factor due to regulatory requirements. While this practice has allowed integration of PV systems into the electric power system at low penetration levels, it exacerbates voltage control challenges for system operators as PV penetration levels increase. (See “PV Generation and Its Effect on Utilities,” SolarPro magazine, June/July 2013.)
As shown in Figure 1, voltage rises when PV power is injected into the distribution system. The relative impact of this voltage rise varies based on the location where PV power is injected, the amount of PV generation relative to minimum and maximum daytime loads, solar resource variability and so forth. Treating distributed PV generation as a negative load not only ignores these voltage impacts, but also limits the amount of PV capacity that can be interconnected to the grid without sacrificing power quality and grid reliability. However, PV inverters are intrinsically capable of absorbing reactive power to counteract voltage rise at the point of connection, just as they are capable of injecting reactive power to support voltage sags. Moreover, if inverters are required or allowed to participate in voltage control and reactive power management, utility operators can accommodate higher PV penetration levels.
In the IEEE article “Options for Control of Reactive Power by Distributed Photovoltaic Inverters” (see Resources), Konstantin Turitsyn and his coauthors point out that high PV penetration levels present both a challenge and an opportunity for distribution utilities: “Rapidly varying irradiance conditions may cause voltage sags and swells that [slow-responding utility equipment cannot compensate for,] resulting in a degradation of power quality. Although not permitted under current standards for interconnection of distributed generation, fast-reacting var-capable PV inverters may provide the necessary reactive power injection or consumption to maintain voltage regulation under difficult transient conditions. As a side benefit, the control of reactive power at each PV inverter provides an opportunity and a new tool for distribution utilities to optimize the performance of distribution circuits: [for example,] by minimizing thermal losses.”
While we can envision a future where utilities directly control PV inverters with a robust communications network, existing requirements are not yet in place so that utilities can use inverters as a new tool to manage old problems. John Gajda, a lead engineer at Duke Energy, explains: “The potential for coordinated use of inverters is incredible; nevertheless, the obstacles are unfortunately still rather daunting. In theory, properly coordinated control of reactive power at any generation site could benefit overall utility operation. However, this opens up a potential Pandora’s box of issues that would need to be addressed, including proper contract arrangements for ancillary services, regulatory allowances or mandates for coordinated operation, technical standards for coordinated operation, and methods for managing the security of a utility-to–third party control system interface. In the short term, utility operators can possibly use inverter capability to allay concerns about steep ramp rates or flicker concerns related to rapid changes in irradiance. However, these capabilities only mitigate issues related to the PV system itself, rather than providing an enhancement for distribution system operations.”
Reactive Power Standards
Much of the confusion related to reactive power control in PV inverters is due to the lack of clarity and coordination of technical requirements across various jurisdictions. The requirements of the North American Electric Reliability Council (NERC) and the Federal Energy Regulatory Commission (FERC) do not clearly and specifically address PV inverter-based generators, and many regional grid operators and some electric utilities have established their own power factor and voltage control standards for transmission-connected plants, which share some similarities but are not harmonized. Meanwhile, the interconnection standard that applies to distributed energy resources, IEEE 1547, does not currently allow for voltage control.
In “Reactive Power Interconnection Requirements for PV and Wind Plants” (see Resources), Abraham Ellis and his coauthors provide a thorough review of existing reactive power standards. While efforts are under way to reconcile transmission and distribution interconnection standards in North America, this is a long-term effort. However, significant progress has been made recently on two fronts: amending IEEE 1547 to allow distributed energy resources to participate in voltage regulation, and establishing new technical requirements in California for inverter-based distributed energy resources.
IEEE 1547 amendment. IEEE 1547 is in the process of being substantially revised and updated. However, developing and adopting new standards is a lengthy process. Recognizing a need for short-term changes, members of the IEEE 1547 standards community initiated a fast-track process to amend the IEEE standard for interconnecting distributed resources with the electric power system. Identified as IEEE 1547a, the resulting amendment is a permissive update to the existing standard that specifically allows distributed resources “to actively participate to regulate the voltage by changes of real and reactive power,” provided that voltage regulation is approved by and coordinated with the utility operator. Additional work is under way to develop new testing requirements and best practices related to voltage control and other grid support capabilities—such as short-duration voltage and frequency ride-through—allowed under IEEE 1547a.
California’s Rule 21. In January 2014, the Smart Inverter Working Group published “Recommendations for Updating the Technical Requirements for Inverters in Distributed Energy Resources,” its recommendations to the California Public Utilities Commission (CPUC) for updating the technical requirements for inverters in distributed energy resources (see Resources). The intent of the CPUC’s efforts is to proactively implement new inverter functions that have proven technically feasible and beneficial to grid operators in Europe. In so doing, the CPUC hopes to avoid some of the power system emergencies experienced in Europe associated with high-penetration levels of variable energy resources.
The Smart Inverter Working Group recommends that the CPUC require seven autonomous inverter functions, including two specifically related to reactive power: dynamic reactive power control and fixed power factor operation. As shown in Figure 2, the group suggests that inverters be capable of dynamically injecting or absorbing reactive power in response to local voltage measurements to help maintain voltage levels within their normal ranges or improve the efficiency of electric power systems. It further recommends that inverters be capable of operating with a fixed power factor—0.85 lag to lead for larger inverters—to enable utility operators to compensate for other loads that generate reactive power, as well as to improve circuit efficiency.
Though inverter-based reactive power control requirements are still evolving, they have immediate implications for many industry stakeholders. According to Tobin Booth, CEO of Blue Oak Energy, “Utilities have started to require power factor control to compensate for rapid changes in the weather, which can strain the utility grid power quality in certain locations.” As a result, Blue Oak Energy is now configuring projects to control power factor, to gain interconnection approval where utilities would otherwise have denied it.
Ryan LeBlanc, senior application engineer at SMA America, notes: “Our experience shows that most, if not all, major utilities are now addressing power factor in one way or another on PV plants from 500 kW and up. The list of utilities with reactive power requirements that we’ve worked with includes Hawaiian Electric Companies, Los Angeles Department of Water & Power, National Grid, Pacific Gas and Electric Company, San Diego Gas & Electric Company and Southern California Edison. Most of our experience is on the distribution network, and utilities typically require reactive capabilities of 0.95 or 0.90 lag-to-lead power factor at the point of interconnection.”
Today’s utility-scale inverters include many advanced features, including reactive power support, that allow better interaction with the grid. According to SMA America’s LeBlanc:
“Inverters are very sophisticated power electronic devices that can quickly take a command and adjust to the desired power factor. Inverters are also an increasingly economical source of reactive power. The cost of inverters is reducing at three or four times the rate of costs for traditional var compensation devices, like STATCOMs and capacitor banks.” LeBlanc points out that in many cases these capabilities are critical to project success: “By using advanced inverter functions to provide reactive power support, system designers and developers can eliminate the cost and complexity of providing power factor mitigation by other means, which can cause projects to be delayed or canceled.”
Utility-interactive inverters work by switching a dc voltage into the ac circuit at high frequency, typically by using a pulse width modulation technique. When inverters operate at a unity power factor, this is done in sync with the utility system’s ac voltage waveform. However, it is also possible for an inverter to offset its current waveform from the utility voltage waveform by controlling the delay in the switching devices. This offset creates a subcycle energy exchange between the inverter’s dc link capacitor and the grid, which increases the inverter’s apparent current output and causes a corresponding increase or decrease in grid voltage, depending on the direction of the reactive power flow. This additional current does not contribute to the real power the inverter generates, nor does it consume energy from the PV array. As the offset between the inverter’s current waveform and the utility voltage waveform increases, the exchange of energy and the current increases. In this manner, inverters adjust power factor, absorbing or generating reactive power depending on the direction in which the current waveform is offset from the utility voltage waveform.
While reactive power management does not consume energy per se, operating an inverter off unity power factor does slightly increase operating losses. Additional performance tradeoffs are relevant to both inverter manufacturers and PV system designers.
Operating efficiency. Kleber Facchini, product manager for Eaton’s Power Xpert solar inverters, explains: “If an inverter maintains real power output levels while producing reactive power, conductive and iron losses increase slightly due to the increase in current. For instance, while our inverter has an ac current output of 2,710 A at its real power rating, the output current is close to 3,000 A at that same real power level and a 0.91 power factor.” While the benefits of inverter-based reactive power capabilities clearly outweigh such minor efficiency losses, inverter manufacturers must account for these higher current levels in their product designs.
Real and reactive power limitations. In theory, the full current capacity from a PV inverter could be available for generating or consuming reactive power. Moreover, since reactive power management does not require energy from the PV array, inverters are theoretically capable of providing reactive power at night. In practice, however, manufacturers have historically limited the available power factor range and have not designed PV inverters to provide reactive power while producing no real power; further, control systems typically shut down the inverter at night to eliminate parasitic losses.
While typical utility-scale PV inverters can absorb or generate reactive power at partial power output, doing so at full power necessarily requires additional inverter capacity to handle the full active and reactive current. Ellis and his coauthors explain the cost implications of providing this additional capacity: “Considering that inverter cost is related to current rating, provision of reactive power at ‘full output’ means that the inverter needs to be larger for the same plant MW rating, which comes at a higher cost compared to existing industry experience.”
Because of these cost implications, PV inverter manufacturers have traditionally dedicated the device’s maximum current capacity to the production of real power. This is beginning to change, however, due to evolving technical requirements. To avoid having to derate inverters to meet interconnection requirements, manufacturers have begun to design PV inverters that have additional reactive power capability at their real power rating.
Inverters are current-limited machines, and manufacturers have a choice in how they design the control system. They can either make the maximum current capacity available for real power or hold some current capacity in reserve so that the inverter has some reactive capability at its rated real power, operating similarly to a conventional synchronous generator. Both design approaches have advantages and disadvantages. For example, holding capacity in reserve may make the interconnection application process more straightforward because it is easier to communicate the project’s capabilities to utility engineers, who are used to working with synchronous generators. However, when the PV inverter is operating at its peak real power rating at unity power factor, the inverter will lose more energy to clipping than if it had its full current capacity available.
Real vs. apparent power ratings. Due to a lack of industry and market standardization, there is significant variation among manufacturers in terms of the ratings conventions they use for real power (kW or MW) versus apparent power (kVA or MVA). On one hand, some inverters—including models manufactured by AE Solar Energy, Bonfiglioli and Eaton—have an apparent power rating that exceeds the real power rating. These inverters have a limited power factor capability at the nameplate real power rating and can operate beyond that power factor range by limiting real power. On the other hand, some inverters—including models from Schneider Electric and SMA—have an apparent power rating equal to the real power rating, indicating that they have no reactive power capability at their real power rating.
PV system designers need to pay extra attention when specifying utility-scale inverters to ensure that they meet all project needs. They must also account for any real power losses due to apparent power limitations. Consider, for example, two inverters with 1,000 kW real power ratings: Inverter A has an apparent power rating of 1,111 kVA with a power factor capability of 0.9 lag to lead at full power; Inverter B has an apparent power rating of 1,000 kVA and no power factor capabilities at full power. If the solar resource and array capacity are adequate to peak the inverter output, Inverter A is capable of producing 1,000 kW of real power in addition to 484 kvar of reactive power while operating at a power factor of 0.9, as shown in Figure 3a. Under the same circumstances, Inverter B is capable of producing only 900 kW of real power while operating at a power factor of 0.9 (1,000 kVA x 0.9 PF), as shown in Figure 3b.
Note that at partial loading, both inverters have capacity available for reactive power control. For example, at 50% loading, both Inverter A and Inverter B could produce 500 kW of real power, 310 kvar of reactive power and 588 kVA of apparent power—given that both are capable of operating at a 0.85 power factor. Since inverters operate below rated output the majority of the time, latent inverter capacity is generally available to produce or consume reactive power.
Power factor limitations. In their IEEE article, Turitsyn and his coauthors note: “Dispatching [reactive power] places additional duty on the inverters of individual PV generators, which may lead to reduced lifetime and increased life-cycle cost.” Increasing inverter output current increases thermal stresses. Therefore, most manufacturers limit the power factor range available to an inverter to limit the amount of wear on the device that is not directly related to producing real power from the PV array.
It is common for inverters to have a 0.8 lag-to-lead power factor range, which should meet most utility requirements. However, depending on the manufacturer, none of this range—or only some of it—may be available to an inverter operating at full rated real power. It is important for system designers and operators to understand when power factor limitations may cause a loss of real power, and to account for these losses when modeling system production or verifying system performance.
Input dc voltage limitations. When an inverter generates or consumes reactive power, there is a voltage drop across the inductor in the output filter of the inverter that reduces or increases the voltage at the output of the switching stage of the inverter. (For more information on inverter topology, see “Central Inverters for Utility-Scale Applications,” SolarPro magazine, December/January 2013.) The inverter’s ability to effectively track the maximum power point of the array will change as a result.
System designers need to be aware that these dc voltage impacts may affect source-circuit sizing. If the system operates an inverter at a lagging power factor to generate reactive power, and the lower voltage limit of the dc input window increases as a result, this could limit string size options. System designers must account for any input voltage window changes at different power factors when specifying an array. If they do not, high temperatures may compromise the inverter’s ability to track the maximum power point of the array, potentially leading to lost energy generation. Since these impacts are product specific, inverter manufacturers’ application engineers are an ideal source of design guidance.
Andrew Riegler, manager of product and application engineering at Schneider Electric, points out: “The dc input voltage window for Schneider Electric’s Conext Core XC-NA inverters does adjust depending on whether the inverter is absorbing or delivering vars [reactive power]. The window will get larger if the inverter is consuming vars and will get smaller if it is injecting vars. While the published dc input voltage ratings are based on unity power factor, we provide tables that system designers can use to understand how the dc input window changes depending on the project’s reactive power requirements.”
Production estimates. While supplying reactive power clearly has some effect on PV plant performance—in terms of both energy production and levelized cost of energy—these impacts are difficult to quantify. PVsyst, the de facto industry-standard production-modeling tool, cannot account for these effects. However, PV system designers or developers can always post-process PVsyst data if they wish to estimate the possible performance impacts associated with reactive power support.
For example, system designers could post-process hourly output data from PVsyst to estimate potential energy losses associated with providing reactive power support at times when the inverter is operating at full load. System designers can also account for losses associated with dc input voltage limitations by post-processing hourly PVsyst data and looking to see when the array VMP falls below the lower limit of the inverter’s dc input window as modified by operation at a lagging power factor.
While there are conductive and iron losses associated with using an inverter to supply reactive power, it is unlikely that the production model needs to account for these losses. When supplying reactive power at full inverter loading, the decrease in inverter efficiency is likely on the order of 0.1%. These losses decrease at lower power levels and are not an issue when the inverter operates at unity power factor. Nevertheless, it is a good idea for system designers to ask the inverter manufacturer to quantify these losses. Designers can then determine whether to account for them or whether they have an insignificant impact.
System designers may also be able to use plant-level modeling to analyze reactive power performance trade-offs. Eaton’s Facchini explains: “When producing capacitive reactive power, the inverter terminals may experience the maximum rated voltage, in which case the dc input window needs to be narrowed. However, one must remember that capacitive reactive power is required whenever the ac voltage level needs to be boosted, meaning that voltage levels are lower than rated. Therefore, a plant-wide study needs to be performed to understand the different trade-offs when supplying vars from the inverter.”
Inverter-Level vs. Plant-Level Control Options
Designers can accomplish reactive power control using PV inverters in one of two ways, either at the inverter level or at the plant level. At the inverter level, plant operators can set internal parameters in the inverter software to allow the inverter to operate at a fixed power factor or to vary reactive power depending upon the generation level or the voltage at the inverter terminals. At the plant level, a central plant controller controls the reactive power magnitude, power factor or voltage at the point of interconnection. With a plant-level controller, system designers also have the option of using other equipment, such as switched capacitor banks, for reactive power control.
Inverter-level control. Some inverters on the market today have the ability to control reactive power output in multiple ways. Representative power factor control and reactive power control modes include constant power factor, constant reactive power, real power–dependent operation, inverter-level voltage control, dynamic set point control and reactive power ramp rate.
Constant power factor. While inverters operate at unity power factor by default, system designers can also configure them using this control mode to operate at a constant non-unity power factor, provided that the set point falls within the inverter’s power factor range.
Constant reactive power. This control mode allows an inverter to produce a constant level of reactive power independent of real power production levels, which vary according to solar resource availability.
Real power–dependent operation. When operating in this control mode, an inverter varies its power factor or reactive power magnitude according to characteristic curves—kW versus power factor or kW versus kVa reactive—that plant operators define by setting internal parameters in the inverter software.
Inverter-level voltage control. When operating according to this control mode, an inverter varies its power factor or reactive power output based on the voltage at the inverter output terminals. Setting internal parameters in the inverter software and defining the desired characteristic curve for voltage versus power factor or voltage versus kilovolt-ampere reactive accomplish this. While some inverters support voltage control at the inverter level, most often a central plant controller manages PV plant voltage at the point of interconnection.
Dynamic set point control. In this control mode, the inverter dynamically adjusts its power factor set point or reactive power level according to a digital or analog signal sent from a central plant controller.
Reactive power ramp rate. This mode allows the system to control reactive power ramp rates in much the same way as real power ramp rates, whenever possible limiting the rate at which reactive power output ramps up or down.
System designers must understand what control modes best suit a given project and what control modes different inverter models make available. Inverter-level controls have varying capabilities, as illustrated in Table 1. However, plant-level controllers can typically handle all of the control modes described in this article. Designers must know whether inverter-level control can meet a project’s reactive power requirements or whether they must specify a plant-level controller.
Plant-level control. Utility-scale PV power plants invariably comprise multiple inverter-based PV generators, often measured in the tens or even hundreds. Plant-level controllers allow plant operators to coordinate these individual inverters and operate them in aggregate as a single large generator. This capability is key to integrating variable energy resources into the grid.
In “‘Grid-Friendly’ Utility-Scale PV Plants” (see Resources), Mahesh Morjaria and Dmitriy Anichkov elaborate on how to achieve plant-level control: “The plant controller implements plant-level logic and closed-loop control schemes with real-time commands to the inverters to achieve fast and reliable regulation. Typically there is one controller per plant that is controlling the output at a single high-voltage bus (referred to as the POI [point of interconnection]). The commands to the plant controller can be provided through the SCADA HMI [supervisory control and data acquisition system human-machine interface] or even through other interface equipment, such as a substation RTU [remote terminal unit].”
Figure 4 provides a representative block diagram of a plant-level control system as it relates to other devices in a PV power plant. Morjaria and Anichkov describe the interactions within the PV power plant as follows: “The power plant controller monitors system-level measurements and determines the desired operating conditions of various plant devices to meet the specified targets. It manages capacitor banks and/or reactor banks, if present. It manages all of the inverters in the plant, ensuring that they are producing the real and reactive power levels necessary to meet the desired settings at the POI.”
A plant controller is especially useful for transmission-interconnected PV power plants that must operate in voltage control mode. In this mode, the plant controller varies reactive power levels in response to the voltage at the point of interconnection. Schneider Electric’s Riegler notes: “If the solar plant is being used for dynamic voltage regulation, a plant controller can constantly change the reactive power output of the inverter to maintain constant voltage at the point of interconnection. The inverters can be controlled either through the communication system or with analog 4- to 20-milliamp control loops, which are typically used for high-speed voltage regulation response times.”
Without a plant controller, inverters functioning independently in voltage control mode operate at cross-purposes. Eaton’s Facchini explains: “Voltage control needs to be performed via the plant controller. Voltage control at the inverter level becomes virtually impossible to perform when there is more than one inverter in the plant, due to the different voltage levels around the collector bus. Therefore, the inverters will ‘fight’ each other and not achieve the POI voltage control.”
According to Prerit Agarwal, SCADA engineer at Ausenco, controller reliability is a critical design consideration: “When designing utility-scale PV plants, it is important to consider the type of device that is to be used for plant control. The level of reliability required for commercial plant operation necessitates that plant control be performed by a firmware-based device, such as a PLC [programmable logic controller] or a smart relay device.” In some cases, the SCADA system already includes a smart relay device, such as a Schweitzer Engineering Laboratories RTAC [Real-Time Automation Controller], that the designer can use to implement plant-level control. Also, some inverter manufacturers, including Schneider Electric and SMA, have developed proprietary PLC-based plant controllers that a plant-level SCADA solution can incorporate.
Other equipment options. Utility-scale inverters are not the only means for meeting reactive power interconnection requirements. Especially if the inverters are sacrificing real power production to meet the reactive requirements, system designers may want to consider utilizing switched capacitors or reactors. Depending on the project-specific energy revenue rate, the designer could make an economic case that saving inverter capacity for producing real power leads to a lower levelized cost of energy. If the energy is valuable enough, there is an economic incentive to dedicate all of the available inverter capacity to real power production.
Before considering whether to use equipment other than inverters, system designers first need to understand to what degree the utility will allow the use of static devices to meet reactive capability requirements. This generally requires direct coordination with the utility. Eaton’s Facchini explains: “One of the main points of interpretation concerns the use of dynamic versus static methods of reactive power control. Power converters and STATCOMs supply dynamic vars, whereas shunt devices produce static vars. However, some grid codes fail to define the range of control that can be accomplished with dynamic or static means. The best way to mitigate this is to engage with the utility or ISO [independent system operator] planning team to make sure they approve the support provided. It is usually after an interconnection study that all of their concerns are addressed.”
Switched static devices have inherent disadvantages compared to dynamic devices such as STATCOMs or inverters. Switching a large capacitor or reactor into the electric power system can cause large voltage disturbances and harmonics that can potentially damage equipment and decrease the quality of service for utility customers. Switched devices are also limited in terms of response time and variability. For example, a switched capacitor bank, even if switched in stages, can provide vars in discrete amounts only. Its response time is also limited, both by the mechanical constraints of the switching devices and by the fact that once capacitors are switched out of the system, they must be allowed to discharge before being switched back in. While the capital cost per kvar is lower with static devices, dynamic devices provide higher-quality reactive power control because they respond quickly and are continuously adjustable.
Ellis and his coauthors note: “The provision of dynamic reactive capability may have cost implications different from that of static reactive capability, and thus should be separately specified.” They continue: “Dynamic reactive capability from [inverters] can be provided almost instantaneously in a manner similar to that of synchronous machines, responding almost instantly (i.e., within a cycle) to system voltage variations, to support the system during transient events, such as short circuits, switching surges, etc. Fixed capacitors or reactors can be used to shift the dynamic reactive capability toward the lagging or leading side, respectively, as needed. If there is inadequate dynamic reactive capability available from the variable generation resources, it may be necessary to supplement the variable generation resources with an SVC or a STATCOM.”
While inverter-based wind power plants are required to provide reactive power capacity at full output, this requirement has not always applied to the PV industry. Ellis and his coauthors note: “In the case of PV, a requirement to maintain reactive power range at full output power represents a change with respect to historical industry practice. This cost impact could be substantial if the PV plant relies on the PV inverters to provide a portion or all of the required plant-level reactive power capability.” To illustrate their point, the authors present a representative transmission-level interconnection and conclude: “In this case the PV plant would not meet the [reactive power] requirement at full output without adding inverter capacity, derating the plant, or installing external reactive power support devices. In order to achieve a power factor range of 0.95 lag to lead at the POI at rated plant output using only the inverters, the total inverter rating would have to increase by as much as 10% considering reactive losses.”
The difference between the manner in which reactive power —and thus voltage—is controlled on the transmission system versus the distribution system leads to distinct differences in transmission and distribution interconnections requirements.
Transmission-level interconnections. The typical requirements for PV generators interconnected into transmission systems originate from FERC Order 2003, otherwise known as the Large Generator Interconnection Agreement (LGIA). According to the LGIA, generators of more than 20 MW must operate according to a voltage schedule and have a 0.95 lag-to-lead power factor capability at rated power. The LGIA models these requirements after traditional synchronous generator capabilities.
Therefore, PV power plants interconnected at the transmission level typically must be able both to operate in voltage control mode and to have a power factor capability of 0.95 lag to lead at the point of interconnection. For the PV power plant to have the required power factor range at the interconnection point, the inverters themselves must have an even wider power factor range. This is due to the reactive losses within the inverter step-up transformers and the substation transformer(s) that step up the ac collection system voltage to the transmission voltage. Because of these two transformer stages, PV inverters may need to operate with a power factor between 0.85 and 0.9 to realize a power factor of 0.95 at the point of interconnection.
To determine the inverters’ utility-required operating power factor, design engineers can conduct a reactive study with power flow modeling software. As part of the impact study process, utilities typically request a collector impedance model and do a similar modeling to determine whether the project can reasonably meet reactive power capability requirements at the real power capacity indicated on the interconnection request.
Distribution-level interconnections. Reactive power supplied by generators connected to the transmission network has a limited effect on the voltage and reactive power flows on the distribution system. Furthermore, voltage is managed differently in the distribution system. Thus, reactive power requirements for distribution interconnections are different than for transmission interconnections.
The typical requirements for PV generators interconnected into distribution systems originate from FERC Order 792, otherwise known as the Small Generator Interconnection Agreement (SGIA). According to the SGIA, generators of less than 20 MW must operate within a 0.95 lag-to-lead power factor range. The subtlety in the difference in the language between the LGIA requirements and SGIA requirements is important. For transmission interconnections, generators must have the capability to operate in the 0.95 lag-to-lead range. For distribution interconnections, the generator must operate within this range, but the SGIA does not state that the generator must be capable of operating across the entire range, nor does it state that the generator must operate in voltage control mode.
Typical distribution-level interconnection requirements fall into two categories: power factor control and reactive power control. Power factor control requirements usually mandate that the project operate at a fixed power factor or within a set power factor range. Reactive power control requirements are currently less common and could require the generator to provide constant vars at a defined level independent of real power production or, as recommended in the proposed California Rule 21 changes, to operate on a volt/var schedule.
Transmission- and distribution-level interconnection requirements for reactive power control differ considerably. For example, Ellis and his coauthors note that power factor and reactive power control are generally not used for large PV power plants connected to transmission systems because “they can result in inappropriate responses to system voltage fluctuations, and they generally detract from local system voltage stability.”
Monetizing Reactive Power Control
While a detailed discussion of the possibility of direct compensation for reactive power ancillary services is outside the scope of this article, it is worth pointing out that project owners should view meeting today’s reactive power interconnection requirements as a responsibility and not as a service for which they expect compensation. We all know that real power translates directly to project revenue, whereas dispatching reactive power capabilities increases wear and decreases the lifetime of inverters. That in turn increases life-cycle costs and, potentially, decreases real power production and revenue. However, it is incumbent upon system designers and developers to understand these costs and take them into consideration as much as possible.
Rudy Wodrich, vice president of Schneider Electric’s solar business unit in the Americas, observes: “If you look at the interconnection requirements in California today, they do not typically have a rider or an adder for var control, and they are not paying for kvar-hours. While kvar-hour metering is one possible method of monetizing reactive power control, I’ve never seen it implemented. The fact of the matter is that if utilities are asking plant operators to run a PV asset within a power factor range of 0.95 leading to 0.95 lagging, the impact on the kWh revenue stream is likely not significant enough to justify kvar-hour metering, which might begin to make sense if utilities start to go outside that window.”
Wodrich concludes: “Right now compensation for reactive power control is more of a natural market mechanism. When people are bidding into reverse auction PPAs, they’ll have the utility’s var support requirements in mind. If they are going to lose energy yield due to var support requirements, then they will increase the kilowatt-hour price slightly to compensate. If you look at Puerto Rico, they are paying $0.17–$0.19 per kilowatt-hour, and the value (or cost) of providing reactive power control is simply embedded in that rate.”
Paul Brucke, PE / Black & Veatch / Cary, NC / bv.com
Ellis, Abraham, et al., “Reactive Power Interconnection Requirements for PV and Wind Plants—Recommendations to NERC,” Sandia Report SAND2012-1098, Sandia National Laboratories, February 2012
Kirby, Brendan, and Eric Hirst, “Ancillary Service Details: Voltage Control,” Oak Ridge National Laboratory, December 1997
Morjaria, Mahesh, and Dmitriy Anichkov, “‘Grid Friendly’ Utility-Scale PV Plants,” First Solar white paper, August 2013
SMA America, “PV Grid Integration: Backgrounds, Requirements and SMA Solutions,” SMA Solar Technology, fourth edition, May 2012
Turitsyn, Konstantin, et al., “Options for Control of Reactive Power by Distributed Photovoltaic Generators,” Proceedings of the IEEE, vol. 99, no. 6, June 2011
Smart Inverter Working Group, “Recommendations for Updating the Technical Requirements for Inverters in Distributed Energy Resources,” California Public Utilities Commission, January 2014