A strategy for modeling hydroelectric plants and improving their performance

Abstract

A plan for integrating modeling techniques and improving hydroelectric plant performance is presented. The plan begins with defining and establishing basic plant modeling parameters from the physical plant description. Guidelines for forming linear or nonlinear mathematical models are developed and the plan culminates in the determination of settings for P-I-D control which achieve optimum plant performance.

Mathematical models are developed for the hydroelectric plant components -- the penstock, the hydraulic turbine (specifically a Francis turbine), the generator with its connected electrical system, and the control system. A unique method for characterizing turbine performance from a hill diagram is presented.

Stability regions for linear models are determined and control settings which result in optimum plant performance are established. Settings associated with high derivative gains are indicated as giving optimum performance when a linear plant model is used. Nonlinear simulations at various operating conditions reveal that less lively gains must be used for the required gate motion to be achievable.

For the representative plant model studied, the optimum governor settings resulted in a 7 percent improvement over other recommended P-I-D settings while satisfying gate speed constraints which the other recommended settings violated. The benefit of this work is in an improved ability to develop and utilize appropriate mathematical models for a hydroelectric plant and in the determination of control settings which improve plant performance.