Risk based robust Model Predictive Controland its application to power plant optimal scheduling

摘要

This paper proposes a Model Predictive Control method with stochastic set point and costparameters. It is applicable to optimal scheduling and load dispatching problem with multiple unitsystems such as group control of distributed electric generators. "Value at Risk" type constraintcondition, which means a worst-case index, is introduced to realize robustness of optimization againstrandom parameters. Additionally a computation reduction method using orthogonal matrix is newlyapplied. An example for power system schedule optimization problem is evaluated to show theeffectiveness of the proposed method.