Power Play: A More Reliable U.S. Electric System
U.S. utilities have a lot to learn about avoiding power outages. They can benefit from the experience of foreign utilities, other U.S. industries, and even their own nuclear power plants.
The United States ranks toward the bottom among developed nations in terms of the reliability of its electricity service. Catastrophic events, such as the August 14, 2003, blackout that put 50 million people in the dark, are well known, but that is only the most visible evidence of a problem that is pervasive in the U.S. electric system. Frequent small outages are endemic throughout the country. Although these might seem to be relatively minor inconveniences to homeowners, they can create serious problems for businesses. (See sidebar, “The Effects of Power Outages”). Other countries demonstrate that much greater reliability is achievable, and the U.S. nuclear power industry has demonstrated over the past three decades how vast improvements can be made in the United States.
The average U.S. customer loses power for 214 minutes per year. That compares to 70 in the United Kingdom, 53 in France, 29 in the Netherlands, 6 in Japan, and 2 minutes per year in Singapore. These outage durations tell only part of the story. In Japan, the average customer loses power once every 20 years. In the United States, it is once every 9 months, excluding hurricanes and other strong storms.
Despite decades of sober technical reports written by investigation teams in the aftermath of blackouts, the frequency of electric power outages in the United States is no less today than it was a quarter-century ago. Whether measured in terms of city-sized blackouts or smaller events, the statistics show that reliability has not improved. Indeed, if the data show any trend in the past few years, it is toward lower reliability.
The causes of outages in the United States show there is considerable room for improvement. If outages from major storms are excluded, the causes of each hour of outage include equipment failure (24 minutes), as in the 1965 Northeast blackout; untrimmed trees near power lines (6 minutes); and mistakes by power company personnel (4 minutes), as in the 1977 New York blackout and the 2005 Los Angeles outage. This history of blackouts creates ample public demand to increase reliability, opening a window of opportunity for the industry.
Congress made an effort to boost reliability with a provision in the Energy Policy Act of 2005 that calls for the creation of an Electricity Reliability Organization (ERO), but the details of the plan make it unlikely that the new ERO will be capable of doing all that is needed. It is more likely that it will merely lock in place the status quo. The United States does not have to look far to find a better model for enhancing reliability. U.S. nuclear power producers have developed an extremely effective mechanism for improving the performance of the entire industry, and at least some of the lessons from that effort can be applied to the entire power industry
Where we’ve been
In 1962, as the scattered power systems in the eastern United States were about to be interconnected, 10 voluntary regional reliability councils were established to coordinate the planning and operation of generation and transmission facilities owned by their members. After the 1965 blackout, the U.S. Federal Power Commission recommended that a national reliability coordinating council be created, and in 1968 the North American Electric Reliability Council (NERC) was formed to coordinate the regional councils. One of NERC’s primary functions has been to develop voluntary reliability standards for the regional generation and transmission of power.
In January 1997, recognizing that the familiar landscape of rate-of-return regulation was about to be replaced by a competitive market for electricity, a NERC panel proposed federal legislation that would establish an electric reliability organization with power to establish and enforce mandatory standards. The U.S. Department of Energy endorsed that recommendation in 1998.
Seven years later, the Energy Policy Act of 2005 followed that recommendation, creating a new section of the Federal Power Act that gives FERC responsibility for reliability and the authority to certify an ERO. On March 30, 2006, FERC issued its final rule establishing the criteria that an entity must satisfy to qualify to be the ERO, including the ability to develop and enforce reliability standards. The Commission intends to certify one such ERO, which may (upon FERC approval) delegate its enforcement responsibilities to regional entities.
If all this sounds a bit like NERC and the regional reliability councils (of which there are now eight), that is not a coincidence. Four days after the final rule was published, NERC filed an application seeking Commission certification as the ERO. NERC hopes to be certified in time to implement mandatory reliability standards early in 2007.
NERC is the only organization proposing to become the ERO. It has requested that FERC approve the existing NERC voluntary standards as the first mandatory reliability standards adopted under the new legislation.
The current program is administered through the eight regional councils. As an example, the regional council responsible for reliability in Florida directs confidential annual self-audits and performs or directs periodic confidential triennial and spot audits, random checks, and investigations (the latter in response to a complaint or notice of a suspected violation). Monthly reports must be made on such items as transmission protection system misoperations. The entire ERO compliance audit program itself will be evaluated every three years by an outside group.
NERC proposes to continue its program of triennial reliability readiness audits (begun after the 2003 blackout). NERC wants these to “ensure that operators of the bulk electric system have the facilities, tools, processes, and procedures in place to operate reliably under future conditions.” The readiness reports, stripped of business-sensitive data, are to be made public. NERC also plans to compile and publish examples of excellent reliability practices noted during these audits.
The proposal states, “NERC’s budgeting and business plan development processes will be open and will extensively consider industry views. NERC’s independence in this area will be maintained by virtue of the board being the ultimate body to vote on and approve NERC’s and the regional entities’ budgets and business plans, prior to submission to the Commission for approval.”
This sounds promising, but the fact remains that independence is precarious for a body funded by the industry it is supposed to regulate, even with the caveat that FERC must approve the ERO’s submitted budget. The degree to which the ERO will be able to act to increase reliability depends on the seriousness with which reliability is taken by FERC and the ERO’s members, and on the balance they strike between profit and reliability expenditures in an often cutthroat competitive environment.
On March 28, 1979, Reactor 2 at the Three Mile Island nuclear power plant suffered the meltdown of approximately half its core. “TMI shook the industry to its foundation, ending an age of innocence,” according to the chairman of the board of the Institute of Nuclear Plant Operators (INPO), which was formed within the year.
INPO’s mission is “to promote the highest levels of safety and reliability—to promote excellence—in the operation of nuclear electric generating plants.”
The nuclear electric power industry has achieved major improvements in its reliability. At the time of the TMI accident, U.S. nuclear plants were online 58% of the time. By 2004, they were producing electricity 91% of the time.
INPO is a big part of the reason for the improvement in reliability. When the nuclear industry was rocked by TMI and seven years later by Chernobyl, U.S. nuclear industry executives feared that their plants would be closed. They agreed on a major effort to avoid another mishap. They were given a not-too-gentle push when the Nuclear Regulatory Commission (NRC) shut down reactors operated by the Tennessee Valley Authority, Philadelphia Electric, and other companies until operations and equipment were improved.
INPO’s board of directors is made of up 12 CEOs and presidents of power companies. As the institute states, “The industry’s recognition that all nuclear utilities are affected by the action of any one utility motivated its commitment to and support for INPO.” The commitment has also made the plants much more reliable—and profitable.
INPO’s regular evaluations of nuclear electric generators are centered around comparing plant performance to metrics that emphasize safety and reliability. These metrics include the online time percentage, unplanned automatic interruptions, safety system performance, chemistry and fuel defects, industrial safety, and plant emissions. The metrics are developed jointly with the World Association of Nuclear Operators (WANO), and goals are set for each type of plant.
Not only does the use of metrics provide targets that can be incorporated into a plant manager’s compensation, they also allow the identification of early signs of performance decline in time to avoid service interruptions or mishaps. In exit meetings at the conclusion of plant audits, the sustainability of plant performance on the metrics is addressed explicitly. Members are provided with comparisons of their plants’ performance with metrics for the industry as a whole. The insurance industry has linked premiums to scores on INPO performance metrics.
INPO makes a distinction between regulations promulgated by bodies such as the NRC and performance objectives measured by metrics. The institute has found that reliability excellence can be achieved by a combination of the two. Industrywide performance objectives are difficult to meet every year, but provide goals and measurable outcomes; the NRC regulations provide a minimum floor for operations.
Despite competitive pressures in the industry, easy access to plant and equipment performance and operating experience is available on INPO’s secure Web site for members. As one of the conditions for institute membership, organizations agree to “share information, practices and experiences to assist each other in maintaining high levels of operational safety and reliability.” They agree to assist each other in benchmarking best industry practices.
INPO has recognized that the electric power–generating industry may not have all the answers to safe and reliable operations. One of their stated principles is to “use expertise and experience from outside the U.S. nuclear industry.” INPO has formed an advisory council with experts on aviation, insurance, finance, human performance, and organizational effectiveness from the commercial world and universities. They review institute activities and advise the board on objectives, and on methods to meet the objectives.
INPO involves equipment manufacturers and plant designers, who make up a supplier participant advisory committee. Through WANO, INPO brings experience from many countries to bear when performing its plant audits.
To ensure that recent industry experience is embedded within the institute, plant operators loan personnel to INPO and use INPO personnel on reverse loans.
INPO does not rely on its audit program alone. Special assistance is given to any member who requests it, or whose metrics are trending in a poor direction. These between-evaluation programs help to reverse undesirable trends and are prioritized to devote more resources to plants whose metrics show that help is needed. The team includes peers from other utilities who have handled similar problems well.
We see the key to INPO’s success as the agreement among all nuclear plant operators that one poorly performing plant presents a threat to the continuing operation of all nuclear operators. As the U.S. system becomes more interdependent, electric power producers using all fuels, not just uranium, are no longer masters of their own destiny. A shortage of generation in Akron can plunge New York into darkness.
The best large coal plants (1000 megawatts and above) operate 92% of the time (the same as the average nuclear plant), whereas the least reliable large coal plants operate less than 30% of the time. The average coal plant operates 60% of the time. Surely there is room for improvement. Although the future performance of the system is often dependent on the weakest link, the failure of a fossil fuel plant has nothing like the impact on public opinion that would result from a nuclear accident, so it is more difficult to command high levels of attention and concern from others across the industry. However, generator unavailability can still dramatically affect the grid. In the rolling blackouts that hit Texas in April 2006, roughly 20% of the generators in the state were unavailable because maintenance was being performed.
In its notice of proposed rulemaking for the ERO, FERC sought comment on which aspects of INPO’s programs would serve as useful models for the ERO and what lessons can be drawn from INPO’s complementary role with the NRC.
A third of the responders felt strongly that FERC had no business even discussing the idea. One went so far as to state that FERC was exceeding the scope of its authority by suggesting the establishment of an organization that deals with safety (the respondent ignored the 2005 congressional mandate for reliability).
A majority of those who commented had positive things to say about the INPO model. One group of large users of electricity pointedly advised FERC, “The Commission needs to overcome the tendency of economic regulation to tolerate mediocre behavior.”
A common theme among supporters of the INPO model is that enforcement of compliance with reliability standards should be separated from the collaborative functions that an INPO-like organization would undertake. Several felt that such a separation would be feasible within the ERO: Audits for compliance have a very different purpose than audits for excellence.
The periodic site-visit assessment of performance was thought to be a key to the success of such an organization, along with the sharing of equipment failure and operational error and event data. However, most felt that performance ratings and reports should be kept confidential. Several organizations noted that the imposition of sanctions by the ERO would have a chilling effect on information sharing within the ERO. The rotation of personnel and senior management involvement were both felt to be important in a best-practices organization.
To summarize, the industry responses to FERC’s question about possible lessons for the future ERO from the INPO experience included some who felt that the status quo in reliability is fine, some who felt that the regional reliability councils should (in some undefined manner) act as best-practices organizations, and some who felt that national best-practices groups for various segments of the electric power industry are necessary, but that compliance with minimum reliability standards and the achievement of excellence in reliability are two very different functions that must be kept separate.
Enshrining the status quo
NERC and the regional reliability councils began life 40 years ago in an environment of public outrage after the 1965 blackout. Outrage returned after the 1977 New York blackout, particularly when it was revealed that the root cause was a utility practice that left a single critical operator without the tools and training to stop a fairly normal occurrence from snowballing. After the outcry following the 2003 blackout, NERC adopted some of the techniques pioneered by INPO, taking steps in the direction of becoming a best-practices organization. For example, it has performed “readiness audits” (that are planned to be triennial) of generation plants, transmission operators, and independent systems operators. These audits have led to publicly posted examples of excellent practice, such as “The Salt River Project provides highly redundant and independent systems and power supplies at its control center that result in an extremely reliable and secure set of tools for its operators.”
However, in proposing to become the ERO, NERC is morphing into a standards-setting and compliance organization, a role that is filled in other industries by various arms of the government. ERO and the NERC regional councils will thus be funded by the companies that will have to meet the ERO standards. We can thus expect to find that the ERO standards will be set by industry consensus, since a two-thirds vote of NERC members is required for adoption. The standards will vary by region in response to both regional technical differences and to the different characters of companies operating in the regions.
Some facets of NERC’s proposal are admirable. The record of the past quarter-century has shown that NERC and the regional councils have helped to slow the slide in reliability. By making NERC’s reliability standards mandatory, the ERO should be more effective. However, we worry that because the NERC standards were regional industry-consensus standards, their stringency has been limited by the influence of members with substandard performance and that such influence could continue in the future.
The TMI and Chernobyl incidents convinced nuclear plant owners that they were in immediate danger of having their plants closed and losing billions of dollars unless they could convince a skeptical public and Congress that they could operate safely. The INPO experience showed them that tough standards and cooperative efforts could make their assets more profitable and valuable.
The owners of coal- and natural gas–fired generators and of transmission and distribution lines have no reason to fear that a mishap would shut their plants. They might be tempted by the notion that tough standards and cooperative efforts would make their assets more valuable. The outage statistics for fossil-fuel plants and for transmission and distribution indicate that significant improvements can be made and that utilities may get important insights from pooling data. But the most significant reason for optimism is that the grid is getting more tightly integrated every year, so that a problem at a distant generator can cause a cascading failure that blacks out millions a thousand miles away.
The proposed ERO triennial compliance audits are a good and necessary function, although they might be performed more frequently. The procedure for outside evaluation of the compliance audit process is an excellent idea. However, NERC’s proposed penalties ($1,000 to $200,000 for violations of its reliability standards) are low. The U.S. Environmental Protection Agency has levied fines of $25,000 per day for infractions, and total penalties have been as high as $30 million. The 2003 blackout’s cost was estimated at $6 billion, which is 30,000 times the largest ERO penalty. Although provisions are made in NERC’s ERO proposal for fine multipliers in egregious cases, the typical fine will probably be too low to change company behavior.
Having recognized that human and organizational performance is often the root cause of incidents, INPO evaluates the performance of personnel with exercises in high-fidelity simulators during biennial plant evaluations. The checklist for evaluation includes organizational effectiveness and performance improvement. Corporate support of operating plants is evaluated explicitly during plant audits. Members agree to certain organizational expectations, such as making the senior nuclear executive in the line organization accountable in an unambiguous way for safe and reliable plant operation.
In contrast, the proposed ERO blackout and disturbance response procedures state that during investigations, the focus will be on technical aspects. The guidance given to investigation writers is that the conclusions and recommendations section should address “from a technical perspective, what are the root causes of this blackout? What additional technical factors contributed to making the blackout possible?” No mention is made of human or organizational factors.
The ERO should modify its investigation guidelines to stress human factors and corporate support of operational personnel. The disastrous consequences of a wrongly set relay in 1965, an overloaded and underinformed human operator in 1977, a sequence of operator errors and inaction due to lack of data in 2003, and a wrongly cut wire in 2005 should tell us that reliability improvements do not rest on engineering alone but also on social and organizational science.
The cost of the unreliable U.S. electric system is demonstrated through buying decisions: One out of every six dollars spent on electric power generation and delivery equipment goes for emergency backups. Significant savings could be achieved by making the primary system more reliable. Improving the reliability of the U.S. electric power system to the levels achieved in Europe and Japan requires a more stringent approach than compliance with consensus standards.
The ERO as proposed is a necessary, but not sufficient, condition for improvement. After the restructuring of the electric power industry, it is difficult to convince a company to invest in reliability. The ERO will raise the bar modestly by requiring compliance with existing voluntary standards. But converting weak standards from voluntary to mandatory is not likely to lead to the reliability improvement that is needed to raise the United States to parity with its competitors abroad. To get started quickly, we agree with making the current standards mandatory. However, these initial standards should be reviewed critically by FERC to ensure that they significantly improve reliability. FERC should also require the ERO to create a mechanism that would review all the standards over the next three years and propose modifications. These changes should then be approved in a single vote to prevent underperforming companies from weakening selective provisions. FERC should also require the ERO to revisit standards on a three-year schedule to avoid freezing standards at today’s level.
An alternative (and foolhardy) course of action is to wait for the next large blackout to stimulate emergency congressional action that is likely to be hasty, ill-conceived, crude, and ultimately ineffective or counterproductive. This is what happened when industry failed to take responsible action to control vehicle emissions in the 1960s. The 1970 Clean Air Act finally took action, but in a way that was needlessly costly to the industry.
We recommend that FERC provide leadership by acting on the knowledge that in a global economy, lack of reliable power puts the United States at a competitive disadvantage. Likewise, states should recognize that reliable power may put them at an advantage as compared to their neighbors. Reliable power is a public good, no less than excellent highways.
The road to reliability
Americans need better information on reliability. Power generators in New Zealand provide these statistics on the Internet. In the United States, a Freedom of Information Act request is required in many states to acquire these vital data. FERC (and the states for intrastate companies) should mandate that reliability data be available on the Internet for everyone.
Providing reliability data is an example of a transparent and easily understood metric. INPO has found such metrics to be critical to leading the nuclear industry out of the swamp of mediocre reliability. NERC has displayed only a sporadic commitment to making data on failures available to the public or to industry. For example, the failure database is out of date by more than three years as this is being written. A timely public database of all major disturbances is essential, as are data shared among the industry on equipment and operational failures.
A number of the industry’s own comments indicate that the roles of a standards-compliance organization and a best-practices organization are incompatible. We agree. FERC should require the formation of nationwide function-specific best-practices organizations that are not a part of the ERO’s standards and compliance organization. These could be created within the ERO (in a separate compliance function, just as INPO is separate from the NRC), or they could be part of an entirely separate organization. We favor the latter.
The best-practices organization should be responsible for continuing the readiness audits begun by NERC, because these are a collaborative function rather than a regulatory function. It should ensure that these site visits have as their purpose benchmarking each facility against the best metrics found in the industry.
The best-practices organization should follow INPO’s example of seeking advice from reliability experts in other industries as well as their own staff experts. In addition, representatives from the technical staff of state public utility commissions and from the best non-U.S. utilities should be invited to participate, as should representatives from equipment manufacturers, ranging from those who make relays and transformers to those who design software.
In addition to sharing experience about effective maintenance and operational practices, the organization should also share experience and insights about new technologies, such as underground equipment, automated failure recovery systems, and instruments to monitor, display, and control the flow of power.
The ERO will fail to improve reliability significantly unless generators, transmission and distribution owners, and equipment makers are convinced that they face large penalties for substandard performance. In the current deregulated environment, generators battle for even a slight cost advantage over their competitors and are reluctant to contribute to best-practice lists. Thus, any best-practice activity will need a firm regulatory incentive to compel all parties to cooperate.
FERC was right in September 2005 to ask what lessons INPO can teach the power industry. The commission should not accept the view that “nuclear is different” and should not be content with the easy course of simply designating NERC as the ERO. As presently constituted, such an ERO can do only half the job. The nation needs two organizations: one to enforce standards and the other to promote best practices.
J. Apt, L. B. Lave, S. Talukdar, M. G. Morgan, and M. Ilic, “Electrical Blackouts: A Systemic Problem,” Issues in Science and Technology 20, no. 4 (Summer 2004): pp. 55–61.
Rules Concerning Certification of the Electric Reliability Organization; and Procedures for the Establishment, Approval, and Enforcement of Electric Reliability Standards, 114 FERC ¶ 61,104 (February 3, 2006) (“Order No. 672”); order on rehearing, 114 FERC ¶ 61,328 (March 30, 2006) (“Order No. 672-A”).
Secretary of Energy Advisory Board, Maintaining Reliability in a Competitive U.S. Electricity Industry: Final Report of the Task Force on Electric System Reliability (Washington, DC: U.S. Department of Energy, 1998; available at .
Kristina Hamachi LaCommare and Joseph H. Eto, Understanding the Cost of Power Interruptions to U.S. Electricity Consumers (Lawrence Berkeley National Laboratory report LBNL-55718, available at http://certs.lbl.gov/ pdf/55718.pdf).
Effects of Power Outages
In a 2004 study, the Lawrence Berkeley National Laboratory (LBNL) estimated that the annual costs of U.S. power outages are at least $22 billion and may be as high as $135 billion. Most of the losses (72%) are borne by commercial customers, whereas industrial customers shoulder 26% of the loss and residential users only 2%. In the LBNL study, customer losses were estimated by consolidating a large number of independent utility outage-cost surveys. Interestingly, this work found that the costs of short interruptions of five minutes duration or less caused two-thirds of the economic losses.These short interruptions are probably the ones that can be most easily reduced by attention to reliability.
The rolling blackouts during 2001 in California provide concrete examples of what parts of our society are adversely affected by the loss of power. Only 40% of petroleum refineries in the state had any onsite power generation capability, and those affected by a blackout can require one to two weeks to resume operation. Spot shortages of gasoline occurred. At a large Internet retailer, one 20-minute outage deleted roughly 20,000 product orders and $500,000 in revenue when a backup power system failed during a rolling blackout. Some smaller companies have no backup power; the Wall Street Journal reported that Integrated Device Technology, a semiconductor manufacturer, estimated $50,000 in lost production and damaged products from a 2-hour blackout. Forbes wrote that SDL, a fiber-optic component maker, lost $3 million in product before it could purchase a backup system.Traffic lights were turned off in parts of San Francisco, causing fender-benders and gridlock for two hours. Apple and Hewlett-Packard engineers worked without desktop computers and crowded into offices with windows. Intel froze its California hiring temporarily, and Miller Brewing laid off 260 workers in Southern California for the duration of the rolling blackouts.