Input saturation in nonlinear multivariable processes resolved by nonlinear decoupling

A new method is presented for the resolution of the problem of input saturation in nonlinear multivariable process control by means of elementary nonlinear decoupling (END). Input saturation can have serious consequences particularly in multivariable control because it may lead to very undesirable system behavior and quite often system instability. Many authors have searched for systematic techniques for designing multivariable control systems in which saturation may occur in any of the control variables (inputs, manipulated variables). The method of elementary nonlinear decoupling (END) described can be applied directly to the case of saturating control variables by deriving as many control strategies as are there combinations of saturating control variables. The method is demonstrated by the multivariable control of a simulated fluidized catalytic cracker with satisfactory results