Long-term fuel scheduling addresses the problem of optimum allocation of fuels to various generating units using mixed and shared fuels subject to yearly and monthly constraints, inventory constraints, generation availability constraints and load require ments (both energy and capacity).This problem is formulated as a network flow optimization problem. Feasibility of the resulting fuel consumption plan with respect to fuel constraints and its optimality with respect to cost of fuel purchases are assured by the network flow algorithm. Feasibility with respect to the load duration and generation availability are assured by defining appropriate fictitious branches in the network. The limits on these branches are calculated by the method of cumulants.The unit priorities in each month are modified by an iterative procedure, on the basis of both fuel price and fuel availability as seen by the network flow algorithm.Results for a 17 unit/17 contract test system over one year are presented to illustrate various aspects of the long-term fuel resource scheduling problem.

No abstract