Model reference adaptive control of electric generating plants

Abstract

This paper presents an application of multi-input, multi-output (MIMO) model reference adaptive control (MRAC) to a typical power plant. Three MRAC designs are presented, two analog and one digital. With one analog algorithm, the system output errors (plant and reference model) are asymptotically stable, provided a positive real constraint is satisfied. A second analog algorithm results in a bounded output error but relies on a less restrictive constraint. The third algorithm is digital and is similar to the second. The power plant model is based on the Cromby model with the addition of a governor valve management algorithm, turbine dynamics and turbine interactions with extraction flows.