The time series soft-sensor modeling based on Adaboost_LS-SVM

With regards to the petrochemical processes with various operating states and dynamic performance which will affect estimation precision for the static soft sensor, a time series soft sensor model which uses the time series of process variables to estimate the dynamic performance of quality variable was proposed. Meanwhile, the integrated Adaboost learning algorithm is introduced. With the help of this method, training samples and modeling for several times, according to the modeling error to renew the next sample data, in order to obtain a series of different basic models. Every basic model will be weighted in the last step; as a result, a more precise combined LS-SVM model will be established. According to the prediction of benzene content of column reactor in the azeotropic rectification tower, the effectiveness of the method is demonstrated.