The Estimated Impact of Weather on Daily Electric Utility Operations

The Estimated Impact of Weather on Daily
Electric Utility Operations

Ronald N. Keener, Jr.
Senior Scientist
Duke Power Company
Charlotte, NC
rnkeener@dpcmail.dukepower.com

Introduction

The electric utility industry is currently undergoing a transition from a highly regulated industry to a competitive market driven industry which will allow customers to choose their electric provider much in the same way people choose their long distance phone company. The opening up of retail, and eventually residential markets, to allow choice will have a great impact on individual utility operations. Electricity providers will focus on lower cost generation and customer service as markets open up. This will put a premium on individual utility companies to decrease generation costs by increased operating efficiencies. While there will continue to be pressure to reduce costs through downsizing, other areas where increased operating efficiencies can reduce costs will be focused on by utility managers. One of these areas is better generation demand forecasting which relies on the ability to predict weather from several hours out to several weeks which determines the best and most cost efficient generation mix to meet electricity load demands. Another area where weather will play a role is customer service. The ability to predict severe weather that disrupts electric service to utility customers and act in a proactive manor will aid those companies that place excellence in customer service as part of their business market strategy.

This paper will discuss the way weather information plays in daily utility operations and attempt to provide some conservative economic estimates of the costs involved.

Weather Impact on Power Generation

The impact of weather forecast information on power generation can be substantial. Utilities base their daily generation on a load forecast of power demand for their perspective native load regions. This load forecast is a model of the relationship of power demand, time of day, season, and weather. Of weather variables, the load demand is most sensitive to ambient temperatures followed by dewpoint temperatures, cloud amounts, precipitation, and winds. Duke Power Company utilizes a load forecast model which takes hourly forecasts of ambient temperatures and dewpoints and projects an hourly generation load out through eight days. This load projection is used by the system operating center to dispatch the available generation resources in the most economical manor. The sensitivity in generation load to weather forecast can be as much as 600 to 1000 Mw of electricity demand on a forecast error of five degrees during peak winter and summer months. A typical cost of operating a midsize coal-fired unit of 500 Mw is $250k/day while 500 Mw of combustion turbines can be as much as $500k/day. The economic costs to utility operations are the costs associated with startup-shutdown of generation units which can be the result of error in the short-term hourly temperature forecasts. A conservative annual estimate of weather error costs associated with startup-shutdown of generation units is $8,000,000 for Duke Power. Improvements in short-term forecasts of temperatures can save millions of dollars annually in these unit startup-shutdown costs.

Daily weather forecasts also impact other areas of power generation such as bulk power marketing and hourly energy pricing. Utilities such as Duke Power market blocks of energy in the form of short term and long term contracts to other utilities or large industrial customers. The energy price must be competitive with the market with some risk built into the price that reflects uncertainty in the available generation mix (i.e., fossil, nuclear, or hydroelectric) and the weather forecast. Therefore a significant error in the 1 - 8 day forecast could result in a substantial economic loss to the energy provider. Improvements in the medium range forecasts could reduce the risks associated with bulk power marketing and hourly energy pricing. Consequently, increased confidence in seasonal forecasts for departures from normal for temperatures will aid in the energy sector's ability to price energy and schedule operation and maintenance of large generation resources more efficiently between seasons.

Utilities with hydroelectric operations utilize precipitation forecasts to optimize the availability of their water resources. Duke Power utilizes QPF forecasts daily to route water through 25 hydroelectric stations (1007 Mw net) and two pump storage reservoirs (1675 Mw net). While hydroelectric power generation is less than 14% of the total Duke Power net system capability it provides an important and economically inexpensive means of power generation to meet peak demands. By utilizing good QPF forecasts and operating the hydroelectric system in a proactive manor, it has been estimated that Duke Power could conservatively save $2,000,000 over a five year period in wasted water (water not available for hydroelectric generation). An important side benefit to efficient utilization of good QPF forecasts is to minimize shoreline impacts of high water to property owners along Duke Power reservoirs. Improvements in QPF forecasts in the 0 - 48 hour forecast period and extending the QPF forecasts through the 48 - 72 hour forecast period can provide additional economical savings to water resource and hydroelectric managers.

Severe Weather Impact on Daily Utility Operations

The most obvious impact of weather on electric utility operations are power outages as a result of some weather phenomena. Electric power outages are the most noticeable result of severe weather. Severe weather can be classified as weather events that are life threatening. In the utility industry severe weather is classified as a weather event that directly causes widespread outages to a utility's distribution system, or in a worst case, causes extensive damage to a utility's transmission system. Over the years Duke Power Company has experienced several severe weather events that have caused extensive damage to Duke's electrical distribution system. The table below lists the dates of several big storm events, type of storm, total customer outages (out of a total 1.8 million customers served throughout the service area), and the total cost of restoration which includes material and labor.

* 1990 is used as a typical year without any major storms. Dollar amounts are the total restoration costs in labor and materials for all weather related events throughout the year.
*Storm Date*	*Storm Type*	Total Customer Outages	*Cost*
May - 89	Tornadoes	228,341	$ 15,189,671
Sep - 89	Hurricane Hugo	568,445	$ 64,671,150
1990*	ALL STORMS		$ 753,805
Mar - 93	Wind, Ice and Snow	146,436	$ 9,176,203
Oct - 95	Hurricane Opal	116,271	$ 1, 655,350
Jan - 96	Western NC Snow	88,076	$ 872,585
Feb - 96	Ice Storm	660,000	$ 22,905,627
Sep - 96	Hurricane Fran	409,935	$ 17,471,826

Utilities utilize severe weather forecasts to plan and mobilize resources to meet the anticipated challenges of storm restoration. For example, prior to landfall of Hurricane Fran, Duke Power Company pre-staged materials and restoration crews in the northeastern portion of the service area. The result was that a concentration of available crews and materials were ready to begin restoration work immediately after Fran's passage through the region. This aided in reducing the total storm outage time for customers in the affected regions. Improvements in severe weather forecasts of major events such as Hurricanes, winter storms, and severe thunderstorms can aid utility managers in resource scheduling and materials management. Utilities that focus on excellence in customer service will retain customers as the electric utility industry moves into a more competitive market.

Finally, lightning related outages are estimated to cost the nation's utility industry over $100 million annually in materials and labor costs (from EPRI). For Duke Power alone, lightning causes nearly 90% of power outages during the summer months. Lightning and wind damage associated with severe thunderstorm activity can disrupt electrical service throughout the year. Forecasts of storms with high probability of intense electrical activity with just a few hours lead time would reduce customer outage time if decision makers were more proactive in utilizing a forecast of this type. The economic benefit to Duke Power has been conservatively estimated to be near $200k annually in reduced outage time. A much bigger savings to customers with large investments in standby power generation could be realized as well.

Summary

Improvements in weather forecasting from temperature forecasts to severe weather events can have a large economic impact to a utility's bottom line. The discussion above focuses on how weather impacts one southeast US utility. Extrapolating the estimated costs of unit startup - shutdown due to weather error in the temperature forecasts alone may well be in the hundreds of millions annually nationwide for the utility sector. Better QPF forecasts and improvements in how water resource managers utilize these forecasts can save millions annually nationwide and reduce shoreline impacts as a result of high water operations.

Societal Aspects of Weather
Workshop's Main Page
Table of Contents