Optimal scheduling of long-term fuel purchase, distribution, storage and consumption

Abstract

The long-term fuel scheduling problem for optimizing the purchase, distribution, storage, and consumption of fuel is considered. This problem is formulated as a large-scale linear optimization problem with the objective of minimizing the total fuel purchase and storage cost while observing generation and fuel-related constraints. The fuel constraints include different fuel contracts and their associated limits, fuel delivery simple and cumulative limit profiles, and fuel storage limitations. Generation constraints include satisfaction of the thermal load, and area, plant, and unit generation constraints. Constraints on thermal unit peak/off peak generation output for weekday and weekend are modeled for representing peak/off-peak commitment decisions. A fuel price prorating and fuel contract/offer aggregation logic is used to control the size of the resulting linear program. The modeling and solution of the long-term fuel scheduling problem has been implemented in a production-grade computer program which will be used as one of the operations scheduling and planning tools by a large electric utility. Preliminary results of using the program are presented.<>