Data reconciliation based on robust estimator and MT-NT method

Data reconciliation can provide accurate and consistent data for chemical processes. However, the presence of gross errors can severely bias the reconciled results. In this paper, we combine a new robust estimator and MT-NT method to decrease the effect of gross errors in data reconciliation. Firstly, a new robust estimator is proposed and the influence function is given to show its robustness. Secondly, the solution of the robust estimator is given. Finally, a recursive strategy is proposed to identify gross errors by using MT-NT method and give weights by using robust estimator which can avoid the decrease of coefficient matrix rank. The simulation results verify that the proposed method can decrease the bad effect of gross errors and get good reconciled results.