Estimation of measurement error variances/covariance in data reconciliation

Data reconciliation is the important method for reconciling or adjusting process measurement data for mass and energy balance. All reconciliation strategies start with the hypothesis that the measurement errors are random and statistical distributed and the assumption of a known covariance matrix Q by operators, which results non-correct without considering without considering the variety of instrument's precision. The estimation of measurement error covariance matrix Q based on constraint residual of spatial redundancy is presented. It is available for improving the present method that is depended on experience completely And to nonlinear or bilinear constraint balance, the resolve scheme is proposed. The estimation of matrix Q is tested by the example of certain coking sub-plant. The corresponding modification to the method is provided based on actual purification and shift reactor process. The result shows that the method in this paper is available to obtain the initial estimates of Q