D.I. Wilson (Sweden)
discretisation, chemical process models, sampling, stability
Estimation and control applications in the chemical pro cess industry require discretisation of continuous-time pro cess models that are usually available from £rst principles. Simple explicit techniques for discretisation, have rarely been employed for the fear of introducing excessive error in the discrete model. In our view such techniques suf£ce for most chemical engineering applications given the prior inaccuracy in the continuous-time model and the choice of sampling frequency dictated by the measurement hard ware, the plant dynamics and the needs of the estimation or control algorithms. In this work, the Euler and second order Runge-Kutta procedures are shown to provide suf £cient accuracy for a wide range of chemical engineering applications. Use of these techniques is recommended for practically all explicit discretisation needs, at least as a £rst step to establish possible need for more complex alterna tives.
Important Links:
Go Back
