Modeling of a drying process using subtractive clustering based system identification

This paper describes the modeling of an industrial drying process into a three-input one-output first order Sugeno system. An objective system model is identified from input-output data of the system by applying the subtractive clustering algorithm. The input-output data represents process parameters measured during the drying of starch in a jet spouted dryer. Minimum error models are obtained through enumerative search of clustering parameters. A set of checking data is used to verify the model output. The optimal model, as well as its output, is presented. The step size used in the clustering parameter search is varied and its influence on the modeling performance is presented. Models obtained by setting the same cluster radius for all data dimensions and models obtained by setting a cluster radius for each data dimension are computed and their performance is compared.