Nonlinear component analysis as a kernel eigenvalue problem
Neural Computation
Least Squares Support Vector Machine Classifiers
Neural Processing Letters
Hi-index | 0.00 |
Sequencing batch reactor (SBR) processes, a typical batch process, due to nonlinear and unavailability of direct on-line quality measurements, it is difficult for on-line quality control. A MKPCA-LSSVM quality prediction method is proposed for dedicating to reveal the nonlinearly relationship between process variables and final COD of effluent for SBR batch process. Three-way batch data of the SBR process are unfolded batch-wisely, and then nonlinear PCA is used to capture the nonlinear characteristics within the batch processes and obtain irrelevant variables of un-fold data as input of LS-SVM. Compared with the models of LS-SVM, the result obtained by the proposed quality prediction approach shows better estimation accuracy and is more extendable. The COD prediction of sewage disposing effluent quality can be helpful to optimal control of the wastewater treatment process, and it has some practical worthiness.