Frequently Asked Questions
FAQs on the Role of Residential AC Units in Contributing to the Risk of Power System Blackouts
- What is the problem?
- What causes the problem?
- Why hasn't this been a problem in the past, and is this really a national problem?
- So, how come there haven't been any blackouts as a result of this problem?
- What are utilities doing about it; shouldn't that be enough?
- Why should demand-side solutions also be considered?
- The path forward
What is the problem?
Following a system fault (e.g., lightning striking a transmission line or a flashover to a tree that has grown too close to a line), the electrical response of residential air-conditioning (AC) units (and, for that matter, most induction motors) can rapidly degrade the electrical integrity of the surrounding power system (in a matter of milliseconds). This degraded condition greatly increases the risk of a wide-spread, cascading blackout.
The August 14, 2003 Northeast Blackout, which was not caused by residential AC stalling, was a dramatic example of the underlying electrical phenomena, which is called a voltage collapse.1
What causes the problem?
Voltage is the electric system equivalent of blood pressure. When there is no pressure (voltage), blood (electricity) cannot flow. In electric power systems, falling voltages can reach a point of no return. At this point, the voltage "collapses" (i.e., drops immediately to zero) and all power ceases to flow (blackout).
Voltage is maintained by balancing reactive power (to be distinguished from "real" power) production with consumption. Reactive power is produced by generators or standalone devices, such as capacitors. Reactive power is consumed by induction motors and to a lesser extent by transmission and distribution lines, themselves.
The induction motors used in residential AC units "stall" following a system fault causing a momentary voltage drop of over 30%. Stalling means that they will greatly increase their consumption of reactive power (by a factor of 4 or more) virtually instantaneously (i.e., within milliseconds). If the system cannot respond with increased reactive power production, voltages will drop. If there are not adequate reactive power resources, voltages will decline to the point of collapse and a blackout will ensue.
The greatest risk occurs when a fault takes place on a high-voltage transmission line (e.g., 500 kV), during a time when there is also heavy AC use (e.g., hot summer afternoons). In such a situation, the voltage decline could affect a large geographic area (e.g., several adjacent communities within the L.A basin). If adequate reactive reserves are not available within this area, a voltage collapse could ensue. The concern is that a major voltage collapse in one portion of the system might then "cascade" and spread rapidly to adjacent portions of the system (e.g., all of southern California or more).
Why hasn't this been a problem in the past, and is this really a national problem?
There are two reasons utilities have not been as concerned in the past as they are now. First, unless residential AC units (or other induction motors) represent a large portion of electricity load, the stalling of affected units will not have a significant impact on overall system voltages.2 This is now a national issue because residential AC penetration across the US is at an all time high and growing rapidly.
Second, until the recent arrival of high-speed monitoring technologies, it has not been possible to detect or document the underlying electrical phenomena accurately. Some utilities, such as Southern California Edison (SCE), have now used this technology to develop very good documentation on a large and growing number of incidents, which had previously gone undetected. In addition to Southern California Edison (SCE), four additional utilities have reported similar, localized (not region-wide) incidents in the literature.
The incidents are troubling because they do not correspond with what the planning models used by utilities tell them should occur (and upon which they have based their reliability practices). They suggest, instead, that there is a previously unrecognized, not yet fully understood, and now growing risk of cascading blackouts. This topic is now under active study by the Western Electricity Coordinating Council's (WECC) Load Modeling Task Force.
As discussed below, the utilities believe they are taking appropriate steps to address the problem. However, based on preliminary investigations, they are concerned that supply-side approaches (on the utility side of the meter), alone, may not be adequate to fully address the problem.
So, how come there haven't been any blackouts as a result of this problem?
If you ask the utilities, I think they would answer that there have been no major blackouts — at least, so far — because we have been very lucky, in that recent incidents have been localized. As noted, there is already good documentation in the literature examining isolated instances of the phenomena. However, there is great concern that an initiating incident takes place on a high-voltage line serving a broad area (in which case the incident would not be a localized) might initiate a widespread cascading blackout. Hence, they view the growing record of localized incidents in the past with increasing alarm of what could happen in the future.
Utilities' heightened concern also stems from the fact that now having recognized high residential AC penetration as the underlying cause, they believe (as discussed next) that they are limited in what they can do on the supply-side to prevent a full-scale voltage collapse. They are concerned that, without coordinated actions — on both the supply- and demand-side — it is increasingly likely that the best efforts on the supply-side, alone, may be insufficient. They are concerned that a future triggering incident during a time of high AC use, will lead to a large voltage collapse and, once such a collapse gets started, there will be nothing they can do to stop it from cascading over a wide area.
What are utilities doing about it; shouldn't that be enough?
Utilities are taking concerted actions to evaluate and implement available supply-side solutions, as well as to better understand the problem. However, preliminary evidence suggests supply-side solutions, alone, may not be fully effective in addressing the underlying problems and, moreover, may be needlessly expensive.
The first and most effective supply-side response is to increase the reactive power output of generators. However, reactive power (unlike real power) cannot travel, so this response is limited to only those rare situations in which generators are already or can be located close to the loads where the AC units are located.
Following generators, the traditional (and least expensive) supply-side response to low system voltages is installation of capacitors close to loads. This solution can address "slow" voltage collapse, but it will not fully address a "fast" voltage collapse caused by AC stalling. This is because it is not possible to switch on capacitors fast enough and because the efficacy of the capacitor declines with lower voltage (i.e., at exactly the time when they would be needed the most).
More novel and expensive supply-side responses include installation of synchronous condensers and static VAr compensators (SVC). These devices can be installed locally adjacent to areas with high AC penetrations and they should be able to reduce the length and limit the area affected by all but the most rapid voltage declines.
Nevertheless, these devices are very expensive. They cost approximately $20-50 million per installation. Because the greatest concern is with faults on high-voltage transmission lines, the large numbers of these devices that would have to be deployed could become very costly. And even this large investment, may not be effective in preventing an extremely fast voltage collapse following a transmission fault.
As noted earlier, a significant effort is under-way through WECC, supported by the California Energy Commission PIER program, to improve the modeling tools that are used to analyze this problem. These efforts will help determine how fast voltage collapse might develop, how the risk of collapse escalates as AC penetrations grow, and finally the trade-offs in performance of various technical solutions (on supply- and demand-side).
Why should demand-side solutions also be considered?
The utilities believe that an important element of an effective and economic response to the problem will involve a coordinated response that includes both supply- and demand-side actions. On the demand-side, the principal action is to modify residential AC units with a function that would turn them off automatically and rapidly (within a 1/10 of a second) in response to a low voltage condition. Doing so, would eliminate the problem at its source.
In fact, the utilities believe that part of the reason a large-scale voltage collapse blackout has not been observed to date is that, after about 5 to 20 seconds in a stalled condition, residential AC units will turn themselves off automatically because they overheat. Once a large enough population of units has turned themselves off, system voltages recover (though other problems ensue). Nevertheless, utilities believe that even 5 to 20 seconds may be too long to wait for voltages to recover. In theory, a fast voltage collapse can develop in a matter of milliseconds. A much faster acting switch triggered directly by low voltage conditions is required to address these situations.
Implementing demand-side measures involves retrofitting existing residential AC units or modifying new residential AC units. The utilities have begun discussions with manufacturers on available retrofit technologies, which have been available for some time. However, available retrofit options may not be fully effective since they were designed originally with a different purpose in mind (addressing persistent low systems voltages that are characteristic of rural electric systems). For new units, the solution is to install faster-acting switches at the factory.
Given the large number and slow turnover of the existing population of residential AC units, there are, of course, practical and financial limitations on the rate at which either demand-side measure could be implemented. Hence, as a practical matter, utilities have no choice but to take appropriate supply-side actions, which can both be targeted precisely and be put in place more rapidly. As noted, however, there may be limits to the efficacy of these actions alone.
The path forward
Clearly, a coordinated strategy involving improved understanding of how the physical characteristics of the grid combined with those of residential AC units contribute to the risk of voltage collapse is required, followed by coordinated actions on both the supply- and demand-side taking into account cost, timeliness, and effectiveness is desirable. The DOE workshop on the role of residential AC units in contributing to fault-induced delayed voltage recovery represents an important public forum at which the electricity industry and residential AC manufacturing industry can share information on what is known, the options available, and the considerations that must be taken into account in assessing the best paths forward.
1 The 2003 blackout is an example of what is known as a "slow" voltage collapse caused by growing loads over the afternoon in the face of slowly degrading system conditions; the degradation took place over the course of nearly one hour. As noted in the Blackout Investigation, there was time for operators to respond with manual control actions to prevent its spread.
The voltage collapse caused by residential A/C stalling would be an example of what is known as a "fast" voltage collapse in which system conditions degrade rapidly in a matter of milliseconds to a few seconds. In this case, there is essentially no time for manual intervention to prevent its spread, so the responses must be pre-planned and automatic.
2 It is important to recognize that the problem of residential AC stalling cannot be blamed on the mandatory energy-efficiency appliance standards that apply to new units. Both newer energy-efficient and older less-efficient residential AC units are built with single-phase induction motors. It is the high penetration of these induction motors, not their efficiency, which creates an electrical environment in which stalling by these units can lead to system-wide problems.