Kendall Power Plant (a): Operational Options

Abstract

This case series brings together the evaluation of operating options and financial hedging. An electric power plant has several hourly operating alternatives, and a decision model is used to evaluate profit opportunities and decide among those options for fuel and operating level given the available capacity, which is in itself uncertain. The decision tree for these operating decisions is embedded in a Monte Carlo simulation that evaluates hedges of electricity, natural gas, and fuel oil, given uncertain prices.

Excerpt

UVA-QA-0597

KENDALL POWER PLANT (A): OPERATIONAL OPTIONS

Bill Lenin, the manager of Operations for Spiegel Energy Corporation's Kendall Power Plant, had decided to come to work early today in order to prepare for an afternoon meeting with several senior company officials. As he drove through early-morning rush hour traffic in Boston, Lenin considered the agenda for the afternoon meeting. Management at Spiegel Energy had recently expressed concern that their plant managers were failing to capture all the value out of Spiegel's power plants. But they had noticed that out of all their plant managers, Lenin typically had managed to outperform the others under similar market and operating conditions. Thus, management had asked Lenin to develop a model, based on the Kendall Power Plant, that could be used by other plant managers to duplicate his success by predicting the optimal configuration of a power plant. The model Spiegel's management had asked for needed to take into consideration the uncertainty of fuel and power prices as well as several of the physical constraints in the plant's operation. Lenin had worked over the last several days putting together the requested model. What he had come up with was a model that simulated the decision process Lenin went through on an hourly basis to determine how to optimally run the Kendall Plant.

Power generated at Kendall was typically sold on an hour-ahead basis, meaning the plant committed to deliver a certain amount of power to the market an hour before actually generating the power (base load generation units, which Kendall is not, commit into the day-ahead market as opposed to the hour-ahead market). Kendall power plant typically ran on an unpredictable and irregular basis and therefore had to purchase much of the fuel it used on the spot market. Thus, Lenin's model needed to simulate the process for committing and generating power (a two-step process) as well as the uncertainties that occurred during the commitment process. Lenin decided the first step in the model was to use a decision tree to simulate the various operating choices made by a plant manager prior to the beginning of each hour. In addition, this tree needed to make use of forecasts for market prices of electricity and the fuels that could be used in the plant. This first step of the model would represent the commitment of the power plant to produce power one hour into the future.

. . .

Keywords: Crystal Ball OptQuest Decision Tree TreePlan Utility Energy Forecast Hedge