A Practical Method for Identification of Dynamic Processes

Abstract A practical approach to the identification of process dynamics is presented. The method assumes that experiment on the process due to input disturbance is possible and the output measurements can be recorded by an appropriate choice of time and duration. The basic structure of the model is in transfer function form. A nonlinear optimisation programme is developed for estimating the best possible set of coefficients of the model transfer function, such that the error between the measurement and model output is a minimum. The possibility of using Powell and Davidon-Fletcher-Powell minimising algorithms is discussed, and both are found to be suitable for the application. This programme is applicable to many experimental situations and is illustrated by means of examples. The advantage of the programme is that it is easy to operate, and it can be implemented for control system design or system modelling.