Online prediction of unburned carbon content in fly ash with clustering LS-SVM models

In this paper, a novel model called clustering least squares support vector machine (CLS-SVM) is proposed to predict the unburned carbon content in fly ash on line. Prediction accuracy and fast response are major advantages of using CLS-SVM for estimating the unburned carbon content. Moreover, the optical factors influencing the unburned carbon content are selected by means of minimal redundancy maximal relevance (mRMR) criterion. An online updating algorithm is applied to the CLS-SVM model to achieve the online prediction. In the end, comparisons between the proposed CLS-SVM and the traditional LS-SVM are presented to demonstrate the effectiveness. Results are verified on practical data obtained from a 300 MW boiler of a thermal power plant.