Transformed Manipulated Variables for Linearization, Decoupling and Perfect Disturbance Rejection

摘要

The objective of this work is to find new transformed manipulated variables (MVs) for nonlinear systems which linearize and decouple the system, and give perfect disturbance rejection (at least at steady-state). The proposed new input transformation is more general than feedback linearization in that it also allows for multiple-inputs multiple-outputs (MIMO) systems, disturbances, a more general class of models, and introduces a tuning parameter τ0. The key idea is to use decentralized SISO controllers for the outputyusing the new transformed inputsvas MVs. The SISO controllers givev, and a nonlinear calculation block solves algebraic equations which explicitly gives the original inputuas a function of the controller outputv, outputyand disturbancesd.The calculation block also handles decoupling, and feedforward action from the disturbanced.This new procedure can be applied both for static and dynamic processes, which is typical in process control.