Research on modeling methods for solvent oil separation based on LSSVM

As traditional support vector machine and single modeling have some inevitable shortcomings, an integrated learning method of least squares support vector machine (LSSVM) is studied by using two-side line flow in the solvent oil separation of a refinery as the modeling object. The adaptive coefficient weighted fuzzy (AWFCM) clustering algorithm has been used to cluster the training samples. LSSVM is used to establish sub-models for each category of data. The LSSVM integrated models is then obtained by using PLS synthesis function. Finally, simulation experiments are carried out to verify the accuracy of the LS-SVM integrated model. The results show that the proposed algorithm has a great improvement in prediction accuracy and has important guiding significance for the prediction of separation effect in process control system.