Best Practices: Predictive model minimizes exposure to market risks – Combined Cycle Journal

Best Practices: Predictive model minimizes exposure to market risks

Category: Management and Predictive Analytics

Wolf Hollow I
Owned by Stark Investments
Operated by NAES Corp

Challenge. The current economic conditions, with depressed market demand and pricing in the power generation industry, require innovative means to improve earnings. Predicting the maximum plant capacity a day ahead has limited the ability of gas-turbine-based powerplants to realize their maximum earnings potential as the output is directly impacted by ambient conditions. The plant can be short and required to purchase real-time replacement power at costly prices or can be operating with additional available capacity during high realtime market conditions and miss the opportunity for additional revenue. Providing real-time maximum plant capacity limits to the dispatching entity to capture maximum earning potential every hour in the market is essential. Assessing real-time (or as close to real-time as market rules allow) plant capability has become a key parameter in maximizing plant earnings.

Solution. The plant maintains and updates a weekly model of capacity and performance based on ambient temperature, relative humidity, and compressor efficiency. The model incorporates the effects of inlet air cooling and HRSG duct firing. The plant’s dispatching entity utilizes National Oceanic and Atmospheric Administration (NOAA) weather prediction data, ambient conditions, and the current plant model for day-ahead capacity submittal to the Electric Reliability Council of Texas. NOAA provides weather predictions on an hour-by-hour basis for the upcoming 48 hours, which is updated every hour. Developing an interface and programming between NOAA and the plant model, the software provides an hourly update of the plant’s maximum capacity based on the current

model and NOAA data. The system has been automated to update each hour and include alarm capabilities at the plant and dispatch entity locations to notify both parties when the current submitted schedule and plant predicted capacity vary by set limits.

Results. This system was implemented in early 2009, and the plant has realized a reduction in exposure to market risks associated with inability to meet scheduled maximum capacity because of scheduling errors as a result of changing ambient conditions.    The ability to predict and schedule the plant’s capacity on an hourly basis has minimized negative earnings. The accurate prediction and scheduling has also provided additional opportunities to increase earnings through optimizing ancillary service earnings and out-of-merit capacity dispatch. Additionally, this system has led to the development of a system to optimize the plant’s capacity and performance model on a real-time basis. This is a continuing effort to ensure the plant’s earnings potential is maximized.


730-MW, gas-fired, two-unit, 1 × 1 combined cycle located in Granbury, Tex
Plant manager: Kelly Fleetwood
Key project participants: Adam Jackson, Plant engineer
Alan Harding, Operations supervisor
Dimitri Anichkov, Berke Solutions

Scroll to Top