|
By Suzanne S. Biddle, Proceedings of ASCE World Water & Environmental
Resources Congress, Orlando, Florida, May 2001 Abstract: For the past several years, the Tennessee Valley
Authority (TVA), U. S. Bureau of Reclamation (USBR), and the University
of Colorado’s Center for Advanced Decision Support for Water and
Environmental Systems (CADSWES) have partnered to develop a software package
called RiverWare for use in modeling multipurpose reservoir systems. RiverWare
now offers three different modeling approaches: interactive simulation,
rulebased simulation, and optimization. TVA implemented the interactive
simulation software in April 1996 for scheduling both power and nonpower
reservoirs in the TVA system. In May 2000, TVA implemented optimization
modeling for scheduling the reservoir system. Prior to this, optimization
modeling had been used in parallel with interactive simulation modeling
in scheduling the system for about two years.
The optimization modeling approach takes into account not only the power
economics of the hydro system, but also other demands placed on the reservoir
system. These demands, which include flood control, navigation, recreation,
and water quality, as well as special operations for maintenance and community
events, are defined in RiverWare as prioritized constraints. The optimization
routine uses goal programming to satisfy all of the constraints in order
of priority to the extent possible, and then with any remaining flexibility,
optimizes hydropower operations to reduce the total integrated system power
supply cost. Hydropower operations are optimized by maximizing the combined
value of energy generated during the forecast period and water reserved
in storage for future generation.
TVA’s RiverWare optimization model is typically run using 6 hour
timesteps for a forecast period of one week. The optimization model can
be run several times per day as forecasts of power demands or operational
constraints change.
There are several advantages to using RiverWare for optimizing TVA’s
reservoir system. First, the optimization model ensures that all of the
operational objectives of the reservoir system are met to the extent possible.
Then, with the physical and policy constraints satisfied, the timing of
hydro generation can be improved to maximize the dollar value of the hydro
system to TVA. Finally, RiverWare allows the user great flexibility in
scheduling the system. The user can interactively define operating policy
in the form of constraints, quickly and easily manipulate input data, and
use the objective function value to compare the cost difference of various
operating scenarios.
|