Multivariable Direct Adaptive Control of Thermal Mixing Processes

Abstract

Multivariable direct adaptive control is tested on a nonlinear thermal mixing process and is compared with state space based nonadaptive controllers. The linear quadratic optimal control approach is used to design two nonadaptive controllers: one without integral action (ordinary LQ) and the other with integral action (LQI). The operating point is changed over a wide region in the experiment. The adaptive controller is verified to perform most consistently under the tested conditions.