During the extreme heat and humidity of summer 1999 in the USA, several power outages and disturbances disrupted the lives of millions of people and thousands of businesses. The public concern these incidents raised caused the US Department of Energy (DoE) to study the events, determine what caused them, and recommend what steps to take to avoid recurrence of the situation.

An independent Power Outage Study Team (POST) was set up to carry out this survey, taking into account the restructuring of the US power industry. The team consisted of 19 members, predominantly academics, with some DoE staff. POST concluded, unsurprisingly, that the reliability incidents that took place during summer 1999 clearly demonstrated that the necessary operating practices, regulatory policies, and technological tools to give a reliable level of performance were not yet in place.

POST studied a number of the outages, and reported on lessons to be learnt.

Description of incidents

New England – June 7-8 On June 7 and 8, 1999, record heat and humidity spread across the region. At the same time, many generating units were out of service for maintenance and refuelling in anticipation of high demand later in the summer.

As a result, the region had a reserve capacity shortage. Various prompt emergency measures enabled the operator to maintain a continuous supply of electricity under demanding circumstances.

New York City – July 6-7 On July 6, 1999, Washington Heights network in northern Manhattan had to be de-energised for 19 hours after the distribution network lost 8 of its 14 feeder cables. The loss of feeders occurred because of heat-related failures in connections, cables and transformers. Power to 68 000 customers was interrupted. The network was re-energised the next day.

Long Island – July 3-8 Service to 100 000 customers was interrupted for varying periods. A system peak load record of 4340 MWe was set on July 5, but was broken the next day when peak load reached 4590 MWe. On July 6, the New York Power Pool (NYPP) ordered a systemwide 5 per cent voltage reduction. Customers were asked to voluntarily reduce their use of electricity. Voltage collapse was only prevented by this voluntary load reduction, a 5 per cent voltage reduction, and load decreases associated with overloaded wire burndowns.

Mid Atlantic – July 6 and 19 The eastern half of the grid of the region suffered sudden and steep voltage drops twice during July. On both occasions, system integrity was maintained by voltage reductions, curtailling contractually interruptable customers, starting maximum emergency generation, appealing for voluntary load reductions, and curtailing power exports. The more rapid recovery on July 19 was due to the grid operator implementing lessons learned on July 6.

New Jersey – July 5-8 Three outages took place affecting two utilities in New Jersey during the heat wave in early July. PSE&G (Public Service Electric & Gas) suffered outages related to recurrent cable and switchgear troubles at several substations. Up to 10 000 customers were affected during the course of the troubles. Multiple terminator and cable failures affected three of four transformers at a substation in Newark. PSE&G shut down the entire substation to protect the remaining transformer. Service was restored to all customers within 11 hours.

On July 5, GPU Energy suffered outages resulting from problems with the four transformers at its Red Bank substation. Two transformers were severely damaged and had to be replaced. Scheduled and unscheduled outages affected over 100 000 customers.

Delmarva peninsula – July 6 The Delmarva Power and Light (DPL) system was severely affected by the July heat wave. High loads combined with various generation outages resulted in a capacity shortfall. On July 6, DPL implemented rotating outages. Around 138 000 customers experienced outages of varying duration and frequency. Studies indicated that the DPL concerns regarding imminent voltage collapse were well founded.

South-Central states – July 23 On the basis of load forecasts and expected generator availability, Entergy anticipated that it would be able to meet its reserve requirements, with the curtailing of some interruptible contracts. However, the situation deteriorated; load forecasts were higher than expected, while generator availability was lower than anticipated. Throughout the day, equipment problems further reduced available capacity. Entergy made a public appeal for conservation, and received emergency assistance from the Southwest Power Pool for an hour. When this aid ended, Entergy recalled nonfirm capacity purchases. As additional generating capacity was forced out of service, Entergy began curtailing 900 MWe of firm load. Customer outages occurred on a 20- to 30-minute cycle. Ultimately, over 550 000 customers experienced at least one outage.

Chicago – July 30 – August 12 Between July 30 and August 12, three major distribution system outages occurred in Commonwealth Edison’s (ComEd’s) service territory in Chicago. ComEd recorded an all-time high peak demand on July 30. The system began to experience difficulties. Two separate cable faults at the Northwest substation de-energised transformers, overloading nearby interconnected transformers and causing them to shut down to prevent equipment damage. Seven cable faults in and around the Lakeview substation also caused outages. Over 100 000 customers suffered temporary losses of power for up to several hours.

ComEd’s Jefferson substation plays a critical role in the topology of ComEd’s system because it is the sole path of power for six downstream substations. Power interruptions on August 12 resulted from intentional load shedding to protect overloaded equipment.

Report findings

POST’s study of the events resulted in 38 findings specific to the events. A number of these findings are likely to be generally applicable. Some of these include: l Electricity suppliers respond to market signals, but there is a lag of several years before new generation can be placed into service.

Retail customers have limited mechanisms and incentives to conserve energy.

Cable condition is not accurately assessed by conventional diagnostics and practices, which may accelerate cable failure.

Real-time data on cable temperature are not available.

Cables are located in harsh environments, which contributes to reliability problems.

Load predictions are often inadequate.

Transformer failure can require lengthy repair times.

Traditional methods of supplying power to load pockets were not able to keep up with load growth.

Unit ratings may not be consistent with operating performance.

There may not be adequate incentives for reactive power production.

Planning tools may not predict significant voltage degradation during periods of high loads.

There are no reliable tests to identify incipient failure in feeder cables.

Utilities may experience lengthy delays in replacing failed critical equipment.

Notice requirements in load management contracts do not permit an efficient response to emergencies.

Reliability criteria for generation reserves may not be sufficient.

Summer ratings and capability may differ.

Public appeals for conservation have a limited, temporary effect.

Reliance on non-firm purchases to meet operating reserve can result in inadequate reserves. Relying on other members of a reserve-sharing agreement saves on operating reserve costs under most conditions. However, during unusual situations, availability of shared reserve may be curtailed.

Load forecasting techniques may not be able to predict the effects of extreme weather conditions.

Substation protection is critical. Some substation protection and equipment configuration practices lacked flexibility in emergency response situations.

Planned distribution system upgrades were not implemented on schedule in time for the summer peak.

Maintenance planning did not consider transformer overload analysis.

Substation maintenance programmes did not anticipate component weaknesses.

Maintenance management contributed to the severity of the outages. Management of maintenance activities was weak; tracking of inspection and maintenance processes was poor; and employee training and skill levels were inappropriately matched to their duties. A large backlog of corrective and preventative maintenance activities had accumuluated.

Transmission and distribution maintenance expenditure declined dramatically from 1991 to 1998.


As a result of the study, POST made several recommendations to improve system reliability. These assume that the markets and industry should address issues related to reliability, while the federal government is responsible for representing the public’s interests.

POST proposed increased federal leadership in selected areas, such as comprehensive restructuring legislation, reliability-enhancing activities by federal utilities, and facilitating regional solutions for siting.

POST made 12 main recommendations: l Promote market-based approaches to ensure reliable electric services.

Restructuring is based on the principle that competition and markets result in better investment and operating decisions. Mechanisms to ensure adequate supplies of electricity should be designed on this principle. The federal government should support the implementation of fair, efficient and transparent markets.

Enable customer participation in competitive electricity markets.

POST believes that meaningful customer participation is a prerequisite for achieving full reliability benefits of restructuring.

Remove barriers to distributed energy resources.

POST believes that distributed generation will help utilities respond more rapidly to an increased demand in areas where demand is already high.

Support mandatory reliability standards for bulk-power systems.

The power system was originally designed to serve full-service utilities, integrated across the generation, transmission and distribution functions. It is being transformed to one that will support a competitive market. This change makes a system of voluntary compliance with reliability standards inadequate for ensuring reliability.

Support reporting and sharing of information of best practices.

Enhance emergency preparedness activities for low-probability, high-consequence events on bulk power systems.

Emergencies on bulk power systems affect large geographic areas, involve many stakeholders, and affect millions of customers. Effective communication and co-ordination among many parties are critical.

Promote best reliability practices at federal utilities.

Conduct public interest reliability-related R&D consistent with the needs of a restructuring power industry.

Industry investments in reliability-related R&D have declined over the past few years. Federal investment is required during this period of transition. POST recommends development of real-time system monitoring, communication and control technologies.

Facilitate and empower regional solutions to the siting of generation and transmission facilities.

Promote public awareness of electric reliability issues.

General public awareness of the issues associated with maintaining reliable electricity supply is low. However, public interest in reliable electricity supply is high. Greater public understanding can result in better response to appeals for voluntary conservation.

Monitor and assess vulnerabilities to system reliability.

The outages studied by POST resulted from unpredicted events that hit specific weakenesses. Other system vulnerabilities were identified during POST’s studies.

Encourage energy efficiency to enhance reliability.

Increased energy efficiency can enhance system reliability by reducing demand growth in areas experiencing power shortages.

Summer 2000

Looking to the future, POST is confident that opening up the US power market to full competition will improve the reliability of the power system. To date, 24 states and the District of Columbia are in the process of passing legislation to permit customers to choose their power supplier.

The National Electric Reliability Council (NERC) is soon due to issue its estimate of electric power demand for the USA for the coming summer. That will give a clearer indication as to how much difficulty the system will face in the near future.

However, the first test, in early May, was not encouraging. A heat wave struck New England at the same time as twice as many plants were down for maintenance as scheduled. As a result, spot prices went soaring up to 20 times their normal level. In a particularly embarrassing incident, ISO New England had to ask clients to restrict their power demand on the same day that it forecast that power supplies would be adequate to meet the summer peak demands. Similar prices hikes hit Texas, Ohio and Pennsylvania. Prices even trebled in the south east, unaffected by the heat wave, as traders bought up power to sell in the heat-affected regions.