PV Generation and Its Effect on Utilities
Inside this Article
The PV industry depends on a stable electric power system. Fortunately, PV systems can facilitate an electric grid that is more reliable and intelligent than today's.
Given that PV installers and designers focus on optimizing system efficiency, ensuring electrical safety, evaluating structural adequacy and so forth—while constantly looking for opportunities to improve the value proposition to clients without compromising reliability—it is no surprise that they tend to treat the impact distributed power generation has on the electric power system as an afterthought.
From the point of view of PV project developers, engineers and contractors, 100-page interconnection agreements, long review periods and much-feared impact studies—which can delay or even derail a proposed PV system—may seem like unnecessary obstacles. However, if industry professionals ignore the challenges that increasing amounts of PV penetration present for grid operators, it is at their own peril. Grid stability is essential to the reliable operation of the systems they design and deploy, and the long-term success of the PV industry depends on its ability to provide value to grid operators rather than create new problems.
In this article, we discuss the basic architecture of the electric power system in the continental US. We present some of the challenges utility engineers will likely face in the future and explain why the growth of distributed renewable energy systems could exacerbate some of these. Finally, we explain why we expect PV systems to defy conventional wisdom and become an essential part of the solution for grid operators by providing benefits that ultimately lead to a more reliable and intelligent grid.
The basic architecture of the existing electric power grid is essentially the same today as it was 100 years ago, with centralized generation sources pushing power out to a wide variety of loads. While the first electric power grid used low-voltage direct current and served only a few city blocks, today’s electric power system is many orders of magnitude larger and is standardized around alternating current.
As the electric grid expanded from its origins in downtown metropolitan areas, utility service areas increased. Eventually, adjacent utilities were able to interconnect previously autonomous grid systems and provide more-reliable service by sharing excess generation capacity. Today, the electric power system is highly interconnected, consisting of three independently synchronized grids: the Eastern Interconnection, the Western Interconnection and the Electric Reliability Council of Texas.
As summarized in part one of an MIT study, “The Future of the Electric Grid” (see Resources), “The electric power system is composed of four interacting physical elements: energy generation, high-voltage transmission, lower-voltage distribution and energy consumption.” The electric power system originates with large centralized power plants, located far away from most electricity consumers. These generation sources feed step-up transformers that output transmission-level high voltages, ranging from 34.5 kV to 800 kV, which facilitate long-distance transmission paths to utility substations located in the vicinity of the electric load. Distribution transformers located at these substations step transmission-level voltages down to the medium-voltage levels used in distribution lines, typically in the 12 kV to 25 kV range.
At this point, the structural design of the electric power system generally changes from a mesh of redundant bidirectional transmission paths to radial circuits fed from distribution substations, as shown in Figure 1. These lower-voltage lines are much easier for utilities to manage. They allow for the use of relatively inexpensive customer-sited transformers that step the voltage down further to low-voltage levels (under 600 V) for commercial and residential consumers. Under normal operation, power flows in one direction, from the point of generation through transformers, switching devices and miles of conductors before finally reaching the load.
In addition to describing these physical elements, the authors of the MIT report also identify “operational, regulatory and governance structures” that are essential to the management of the electric grid. They explain, “Two less tangible elements are also important: the operational systems that protect and control the physical elements, and the regulatory and governance structures that shape the system’s evolution.”
While the interconnected nature of the electrical power system greatly improves system reliability, it can also allow a fault in one area to cascade through other areas. For example, the August 2003 blackout that affected the northeast portion of the Eastern Interconnection began with the unexpected shutdown of a power plant in Ohio. Losing this single plant had a cascading effect, and neighboring plants tripped off-line as they failed to meet the new load requirements.
Thanks to automatic and manual protective schemes, outages like this are extremely rare. Devices like high-power circuit breakers and frequency and overload relays ensure impressive grid stability and reliability. Among other capabilities, these devices can shed loads to stabilize the grid when power generation drops. (Better load-shedding capabilities would have limited the 2003 blackout to a minor and brief outage in Ohio alone.)
The US electric power grid consists of at least 10,000 traditional generator units providing energy across more than 6 million miles of transmission and distribution lines to greater than 3,200 distribution organizations and on to more than 140 million customers. Yet its basic architecture dates from the late 1800s. The fact that grid operators can achieve 99.9% system reliability—equivalent to 8.8 hours of power outage per customer per year—while maintaining service voltages within 3% of the service rating, as required by ANSI standards, is a major engineering accomplishment.
Understanding the Culture of Caution
PV professionals are often frustrated by the regulations and reviews they must navigate to shepherd a project from proposal through commissioning. While some of these requirements clearly need review, revision and consolidation, as an industry we also need a better understanding of the challenges facing utility operators and the nature of their concerns regarding distributed renewable energy resources.
Power system protection engineers are charged with designing the sensors, fault detectors and control strategies that must protect equipment from damage and prevent injury to people—all while reliably providing high-quality power. Like all engineers, they must make tradeoffs between cost and reliability, sensitivity and security, fail-safe and fail-often, and so forth. However, the consequences of their decisions have particular weight, and the resulting culture of caution may create a resistance to change, even after technological improvements have been shown to effectively eliminate certain risks.
Imagine the emotional stress associated with the electrocution of a passenger or a rescue worker after a car accident. Being in some way responsible for a death caused by a downed and energized power line after a storm would be even worse. This concern is the impetus for the anti-islanding features built into interactive inverters. Now imagine the pressure associated with the potential for economic damage to a business due to a circuit outage. While less traumatic than personal injury, economic damage is far more common. A big box retailer that just lost power is unlikely to accept “reverse overload” as an excuse, even if its rooftop PV system contributed to the outage. Customers simply expect more from utility grid operators.
In addition to bearing the heavy responsibility of maintaining a safe operating environment for both utility personnel and the general public, utility operators are also expected to achieve 99.999% reliability— equivalent to 5 minutes of power outage per customer per year— in spite of the fact that the US and Canada both have extremely low population densities relative to other developed countries, and power lines in rural areas are more difficult to maintain. Though distributed renewable energy resources provide societal benefits, these do not keep the lights on by themselves.
While the distribution system is generally robust enough to accommodate low levels of PV penetration, it was not designed with distributed generation and bidirectional flows in mind. This means that grid operators need more sophisticated monitoring, communication and control systems to accommodate the higher levels of PV penetration and advanced overcurrent protection schemes expected in the future. The need is twofold: Both the electrical power system and the PV equipment connected to it need to evolve.
Distribution System Changes Needed
In 2007, the US Department of Energy launched a renewable systems interconnection study, consisting of 15 separate reports, intended to identify the technical and analytical challenges associated with a high-penetration renewable energy future. Included is “Renewable Systems Interconnection Study: Advanced Grid Planning and Operations” (see Resources), a Sandia National Laboratories report that looks at “issues and options for increasing the penetration of renewable generation.” The authors focused on identifying conditions that limit the interconnection of PV resources, as well as approaches for overcoming any inherent structural limitations. They identified four key issues for grid operators: voltage regulation, overcurrent protection, grounding, and switching and service restoration. The industry can overcome the majority of these only by developing new communication and control infrastructures and implementing strategic and tactical modifications to the structural design of the distribution system.
Communications and control. In a summary of the structural limitations associated with today’s distribution systems, the Sandia report authors observe that “voltage control is achieved with devices (voltage regulators and capacitors) that have localized controls.” While these local schemes work well on radial circuits with a single power source, they cannot solve the voltage regulation problems associated with distributed power generation, two of which Figures 2a and 2b illustrate. Other observations in the report emphasize this need for improved communications and control:
- “Minimal communication and metering infrastructure is in place to aid in restoration [of service] following faults in the system.”
- “No communication infrastructure exists to facilitate control and management of distributed resources that could include renewables, other distributed generation and storage. Without communication and control, the penetration of distributed generation on most circuits will be limited.”
- “There is no communication to customer facilities to allow customers and customer loads to react to electricity price changes, emergency conditions or both.”
To facilitate a high penetration of renewable resources in the future, grid operators need a smarter grid. There are 10,000 traditional generators and approximately 200,000 PV systems interconnected to the utility grid today. Going forward, grid operators need centralized control of distributed generation resources and more-sophisticated control capabilities than they are currently employing.
Overcurrent protection practices. Overcurrent protection for electric power systems depends on the coordinated operation of many devices, including circuit breakers, various types of fuses, relays, reclosers and sectionalizing switches. Unfortunately, the equipment and practices in use today were never intended to accommodate distributed generation. Radial distribution systems are designed to have power flow from the substation out to the load. As the Sandia report authors explain, “The presence of distribution-connected PV introduces new sources of fault currents that can change the direction of flow, introduce new fault-current paths, increase fault-current magnitudes, and redirect ground-fault currents in ways that can be problematic for certain types of overcurrent protection schemes.”
Power protection system engineers worry that these new fault-current paths may cause sympathetic tripping (a type of nuisance tripping) of reclosers or circuit breakers in the distribution system. Another concern relates to the coordination of overcurrent protection devices, which are often designed so that reclosable circuit breakers trip before fusible links elsewhere in the circuit melt, a practice known as fuse saving. New sources of fault currents could undermine these fuse-saving schemes, resulting in longer or more frequent outages.
While a high-penetration renewable energy future clearly requires more-advanced protection schemes, utility engineers often fail to understand that interconnected PV systems are inherently current limited. Whereas a conventional rotating alternator can generate six to 10 times its rated current in the event of a fault, a PV system can produce only a little over its normal rated current, and then only for a short time before shutdown. Utility concerns that PV infeed upsets the coordination and effectiveness of overcurrent protection are largely unfounded.
Concerns about bidirectional flows are also often overstated. Bidirectional power flows on a radial distribution feeder occur when distributed PV begins to form a significant percentage of the peak load on the feeder. For example, it is easy to imagine how rooftop PV energy might make it back to the substation bus in a residential area with high PV penetration levels on a cool, sunny weekday, when customers are away and loads are low. However, feeder protection schemes are normally based on the magnitude of current only, with fuses protecting branch circuits and overcurrent detection controlling the substation circuit breakers. The direction of power flow is not a factor. Reverse currents should not cause technical problems with most feeder protection schemes so long as they do not approach the thermal limits of the feeder conductors.
Grounding compatibility. In the US, a four-wire multi-grounded neutral grounding scheme predominates, accounting for more than 70% of the distribution system circuit miles. This practice has served as a utility design standard for low-load-density areas since the 1930s because it provides the most cost-effective way of servicing rural and suburban areas. While grounding compatibility is a concern with rotating induction types of distributed generators, inverters do not present the same set of problems. Inverters are current limited, and their relatively high-impedance connection to the feeder eliminates problems associated with protection coordination, fuse-saving and fault detection. The authors of the Sandia report conclude that advanced inverter controls and transformer schemes will accommodate “high-penetration scenarios without changing protection strategies.”
Challenges Associated with PV Systems
Just as today’s electric power system will need to be smarter in the future, when high penetration of renewable resources becomes the norm, smarter PV systems are also needed. In March 2012, Sandia published the “Solar Energy Grid Integration Systems” report, produced by representatives of the Florida Solar Energy Center (FSEC) and its team members. According to the FSEC team report (see Resources), there are several disadvantages associated with today’s PV systems when considered from a utility system operator’s point of view. These disadvantages are summarized in the FSEC team’s introduction to its report as follows:
- “A PV system is an intermittent power source, dependent on the fluctuating sunlight local to the area in which it is installed.”
- “Conventional PV systems operate at unity power factor, regardless of reactive power needs of the utility network.”
- “Due to concerns regarding unintentional islanding, current interconnection standards require distributed PV resources to cease to export power during voltage and frequency disturbances, thereby reducing generation at times it is needed most.”
While these are significant challenges, the FSEC team’s report demonstrates that PV system designers can overcome all of them by incorporating advanced inverters under utility control. Ultimately, the goal of designing and deploying inverters with advanced power management features is to ensure that utility operators view PV systems “as valuable resources, rather than liabilities.”
Variability. Since loads are inherently variable, it is not immediately evident to many in the solar industry why PV system variability is a big concern for grid operators. However, the MIT report points out that European utility operators have firsthand experience with compromised grid stability due to the increased penetration of variable energy resources. The net load that generators need to support at any given time is equal to demand minus the contribution of variable energy resources. As the amount of grid-interconnected solar and wind increases, the net load “becomes noticeably more variable and difficult to predict than demand alone.”
The good news for PV installation professionals is that one of the ways to overcome some of the potential problems associated with variable energy resources (VERs) is to install more of them. According to the authors of the MIT report, “By aggregating a geographically diverse collection of VERs, rapid changes in the outputs of individual VERs are replaced by the slower output variations of the aggregated resource.”
Power factor. The utility must satisfy two types of loads: those that consume real power and those that consume reactive power. Real power, which is measured in watts, does useful work. Classic examples of real power loads include incandescent light bulbs and other electric resistance heating elements. Reactive power, which is measured in VAR, is unique to ac circuits and is a result of temporarily storing energy in inductive or capacitive elements. Many of the most useful ac devices are reactive loads with a significant inductive component: HVAC compressors, elevator motors, manufacturing equipment, fluorescent ballasts, power supplies and so forth.
The term power factor refers to how much of the total power is real power. For example, a power factor of 1.0—also known as a unity power factor—indicates that only real power is present and that the current and voltage waveforms are in phase with one another. Meanwhile, a power factor of 0.8 signifies that real power makes up 80% of the total apparent power, which is expressed in VA. If the power factor is not at unity, then the current waveform and the voltage waveform are not in phase. This power factor can be described as leading or lagging based on the relative inductance or capacitance in the circuit. Inductive components cause the current waveform to lag the voltage waveform; capacitive components cause the current waveform to lead the voltage waveform.
To keep the lights on, a utility must provide both real and reactive power, and balance both of these in relation to real and reactive loads. An imbalance between real generation and real load affects frequency. Excess generation increases the frequency above 60 Hz, while excess load decreases the frequency below 60 Hz. The sudden loss of a generating unit causes an immediate decline in the speed of motors, and the loss of a large-enough unit can cause a blackout. An imbalance in reactive generation and reactive load causes an increase or decrease in system voltage.
Ideally, everything would operate at unity. To encourage this, some utilities penalize commercial customers if their power factor drops below a certain threshold. For example, the Lakeland, Florida, municipal utility has set 0.9 as the minimum power factor for large customers as a condition of service. While this does improve the power factor across the service territory, it does not eliminate the need for the utility to provide reactive power to its customer.
Rather than oversizing generators and distribution lines, utilities install power factor correction capacitors on distribution lines and in substations across their service areas. Just as generators must supply real power loads, reactive power generators must supply reactive power. Utilities have traditionally delivered reactive power to the distribution feeder with capacitors located at the medium-voltage level, as well as capacitors in industrial plants at the service-voltage level. This keeps the reactive sources close to the reactive loads, which is the most economical and stable situation.
Most inverters installed today are set to produce real power only. This can be a problem for commercial sites that have a poor power factor. If a PV inverter supplies only real power in this situation, the power factor at the site may drop significantly. This can not only cause voltage regulation problems at the site, but can also trigger utility penalties or surcharges.
Islanding. Utilities have always set very stringent requirements for grid-connected PV systems, sometimes imposing the same requirements on a 5 kW inverter as for a 5 MW generator. In 1994, for example, a utility in Florida required $60,000 in extra hardware for a 10 kW PV system to interconnect to its grid. These measures were taken to ensure that the system would not keep operating after the loss of utility power and create an unintentional island.
Concerns about islanding drove the development of the first utility interface standard for PV systems, IEEE 929, in 1998 and are evident in the UL 1741-2010 and IEEE 1547-2003 standards in use today. These standards apply to utility-interactive inverters used for distributed energy resources and share a basic premise: If the grid becomes unstable, then the interconnected distributed generation source needs to disconnect from the grid so that grid operators can fix the problem.
While this relatively simple binary anti-islanding scheme has served its intended purpose—the protection of the general public and utility service personnel—it is not a viable scheme in the long term. In the future, grid operators will increasingly rely on distributed renewable energy resources as a means for providing peak power. For example, California is on track to generate more than 33% of its electricity from renewable resources by 2020. To reach this goal, grid operators cannot have all of their PV power assets dropping off-line because of a grid disturbance elsewhere in the system.
However daunting these challenges may seem, the FSEC team concludes, “Thanks to relatively recent improvements in power electronics, including advances in fast semiconductor switching devices and real-time, computer-based control systems, PV inverter technology actually has the potential to overcome these barriers and provide significant added value beyond the simple kilowatt-hour production of energy.”
Why PV Systems Are Part of the Solution
Utilities face many challenges going forward, some of them associated with increasing levels of distributed generation. According to the authors of the MIT report, “Even though the US electric grid is not broken today, emerging challenges, if not met, could substantially degrade the system’s reliability and efficiency over the next few decades.” Some of these challenges include voltage regulation on heavily loaded feeders, power quality issues, losses between distant generators and customer loads, and the need to meet new demands, such as increased charging of electric vehicles using the existing infrastructure. Fortunately, the widespread use of advanced inverters under utility control can address all of these challenges while providing grid support services that ultimately enable utilities to better meet the needs of interconnected customers.
Voltage regulation on feeders. Utilities have to manage voltage drop along feeder circuits that can extend tens of miles. According to industry standards, a utility is obligated to keep the voltage at each service within 5% of its rated values. It usually accomplishes this using distributed capacitor banks and voltage regulators—automatically adjustable load-tap–changing transformers—placed at a few locations along each feeder.
If a feeder has many large distributed PV systems on it, voltage regulation can prove difficult. Instead of voltages dropping along a feeder as current flows out from the substation, bidirectional power flows can increase voltages unexpectedly at certain points in the circuit. These voltages fluctuate more frequently with large amounts of variable PV generation.
Advanced inverters can help with this. An upcoming amendment to the IEEE 1547 standard is expected to permit all grid-interactive inverters to regulate service voltage, when approved by and coordinated with the utility operator. This will allow inverters under utility control to produce reactive power as needed to regulate voltage.
Many central inverters on the market can generate or absorb reactive power on demand. This is a potential benefit to utilities because distributed power electronics can provide or consume local reactive power more efficiently than utility-owned capacitors can. Some inverters may even be able to provide power factor correction services at night. These reactive power generation capabilities can mitigate variability. For example, if an inverter is operating close to its rated output and a passing cloud suddenly affects real power production, the inverter can increase its reactive power production to maintain a constant voltage.
Improved power quality. Certain types of reactive loads—like an arc furnace or a large motor at start-up—can cause voltage flicker on feeders and other power quality problems that are unacceptable to utilities. While remote capacitors or tap changers cannot manage some of these problems, customer-sited inverters can respond to these events instantaneously, effectively suppressing harmonics and phase imbalances.
Many of the complex power quality issues that utilities contend with are the result of poorly designed power electronics—motor drives, switching power supplies and so forth. Inverters, which belong to the same family of power electronics, can actually mitigate these problems. Furthermore, they can provide this type of voltage support autonomously, responding more quickly and effectively than the utility’s own dispatchable resources. Customer-sited voltage regulation capabilities may ultimately lengthen the useful life of voltage tap changers and in turn reduce utility maintenance costs. This is just one of the ways that advanced inverters can enhance grid reliability.
Inverter ride-through. Advanced inverters can also improve grid stability. Whereas traditional inverters must disconnect from the grid in the event of a disturbance, advanced inverters under utility control have an extended tolerance for short-duration deviations in grid voltage and frequency. This tolerance, generally referred to as ride-through, results from suppressed anti-islanding measures that allow the inverter to stay on line during a disturbance. The goal of ride-through is to allow an inverter to ride out a specific type of grid disturbance—such as low voltage, high voltage, low frequency or high frequency—and produce reactive power as needed in an effort to stabilize the grid. Note that the utility must give its permission to enable these capabilities.
Several examples of inverter ride-through are described in a case study published by Advanced Energy (AE) regarding PG&E’s 15 MW Westside Solar Station (see Resources). PG&E requested that the 30 AE 500 kW inverters installed at the site be able to ride through a 0 V fault for a full 2 seconds—well outside the IEEE 1547 limits. In March 2012, these inverters successfully rode through four “low voltage events caused by momentary phase-to-phase faults” on an adjacent feeder, each of which resulted in a 50% voltage depression for approximately 0.12 seconds. Rather than being forced off-line per IEEE 1547, all of the AE inverters withstood the momentary grid disturbance and ramped back up to their previous power production, limiting the impact on the grid.
Since grid-support capabilities conflict with conventional anti-islanding protocols, industry stakeholders are developing new protection schemes to detect unintentional islands. One such scheme is a permissive high-frequency signal coupled into the distribution feeder, which a simple receiver in each PV inverter detects. This signal verifies the continuity of the conductors for each phase of the distribution line. If the distribution line is broken, then the signal disappears. The system thereby denies the inverter permission to remain on line, causing it to shut down before it forms a dangerous unintentional island.
Dynamic control. Whereas traditional inverters output power based solely on the solar energy on the array at any given moment, utility operators can control advanced inverters, either individually or in bulk, much as they control traditional power plants. These capabilities include ramp-rate control and retail power curtailment control, as well as remote adjustment of over- or under-voltage trip settings. In a future where renewable resources have reached high penetration, these dynamic controls will allow utility operators to avoid transmission system overloads, manage power flow constraints and regulate power system frequency.
These control capabilities also benefit project developers. Many good sites for large-scale, ground-mounted PV installations are located in rural areas, where service may depend on small feeders vulnerable to overload. Since conventional inverters operate autonomously, a utility could deny a proposed PV project an interconnection agreement if a study determines that the potential for feeder overload exists—even if it is only for a few hours per year. The same project becomes feasible with advanced inverters, since dynamic control allows utility operators to limit real power output on demand.
Many commercially available central inverters can facilitate VAR regulation, disturbance ride-through and dynamic control. Utility operators in the future will have access to improved forecasting methods and models, as well as increasing amounts of interconnected energy storage capacity. Resource forecasting and energy storage will be important when renewable resources reach high penetration, because both can mitigate variability. Utilities can do forecasting at the project level or at the fleet level, based on ground or satellite imagery. These forecasts will allow them to predict how interconnected PV generators will perform 1 minute, 1 hour or even 1 day in advance and coordinate other generators accordingly. The benefit of energy storage is that utilities can dispatch the energy as needed to mitigate temporary mismatches between load and generation, provide frequency regulation or assist in the restoration of service.
While realizing all of the benefits associated with advanced inverters and energy storage systems will require improved control systems, these systems will also serve other purposes. For example, future utility operators will face challenges associated with load growth and demand variability as electric vehicles go mainstream. The same control systems that dynamically manage renewable resources and distributed generators can also manage demand-response systems. The ability to shed loads strategically will allow utility operators to accommodate increasing peak loads without building additional fossil-fired reserve capacity. Improving capacity utilization will ultimately provide a net benefit to consumers, who pay the costs associated with idle capacity.
Dave Click / Florida Solar Energy Center / Cocoa, FL / fsec.ucf.edu
Bob Reedy / Florida Solar Energy Center / Cocoa, FL / fsec.ucf.edu
Bower, Ward, et al., “Solar Energy Grid Integration Systems: Final Report of the Florida Solar Energy Center Team,” Sandia report SAND2012-1395, energy.sandia.gov, March 2012
Hionis, Anastasios, and Steven Ng, “Case Study: Advanced Energy PV Inverters Ride-Through PG&E Low Voltage Events,” Advanced Energy Industries, solarenergy.advanced-energy.com, 2012
McGranaghan, Mark, et al., “Renewable Systems Interconnection Study: Advanced Grid Planning and Operations,” Sandia report SAND2008-0944 P, February 2008
“The Future of the Electric Grid: An Interdisciplinary MIT Study,” MIT Energy Initiative, December 2011