Valve friction quantification and nonlinear process model identification

This paper extends existing methods that jointly estimate the process and friction model parameters, so that a nonlinear process model structure is considered. In addition, nonlinear optimization is applied to estimate the friction model parameters. The developed estimation algorithm is tested with data generated by a hybrid setup (composed of a real valve and a simulated pH neutralization process), in which the influences of the excitation signal magnitude and of the controller tuning on estimated models are investigated. The results demonstrate that the friction is accurately quantified, as well as “good” process models are estimated in several situations. In addition, the proposed extension presents the advantage of providing reasonable estimates of the nonlinear steady-state characteristics.