Thursday, November 12, 2015: 4:05 PM
Salon D (Salt Lake Marriott Downtown at City Creek)
The method of Economic Model Predictive Control (EMPC) is conceptually attractive in that it replaces the traditional quadratic objective function of MPC with an economic based objective function. However, many have indicated concerns about the trade-off between stability and economic / computational performance of EMPC. In a previous effort the notion of approximate infinite-horizon EMPC (AIH-EMPC) was introduced and shown to improve all of the above properties when applied to an exceedingly simple reaction system. In the current effort, the proposed AIH-EMPC method is applied to an electric power network system. In addition to being a much more complicated system (in terms of size as well as the inclusion of integer variables to address unit commitment aspect of grid operation), this example possess an economic based quadratic objective function. (This objective is distinct from that of a regulation type quadratic objective function.) The resulting controller is shown to possess much improved characteristics with respect to computational effort while sacrificing very little in terms of economic performance.