Application of a new dataset selection procedure for the prediction of the Syngas composition of a gasification plant

摘要

In this paper the model identification of a gasification process for the estimation of the Syngas composition is considered and a new procedure for the selection of the identification dataset is proposed. Estimations are needed to integrate the gascromathographic measurements of the Syngas composition which are often not available because of periodic calibrations. This work is part of a broader project for the development of a supervisory controller for process optimization and for fault detection and isolation scope. Improvements in the identification process from standard procedures have been obtained by means of a suitable selection of the input data set. The proposed input data selection procedure is based on the application of the Fuzzy C-means (FCM) algorithm for the generation of the main clusters. Results on a gasification process of a refinery plant show the effectiveness of the proposed FCM method in filtering a large dataset and the reliability of the model in the prediction of the Syngas composition.