Novel data reconciliation method based on Kernel principal component analysis

This paper proposes a novel data reconciliation method based on Kernel Principal Component Analysis (KPCA). In the proposed method, essential information of nonlinear system is extracted by KPCA. Then the data is reconstructed by small count of important nonlinear principle components, which represent essential information of nonlinear system, and noise is removed from reconciliated data. The proposed method is applied to nonlinear data reconciliation of ternary distillation column. Effective results show that proposed method provides a new method for nonlinear data reconciliation.