On-Line Data Reconciliation By Minimization Of A Weighted Sum Of Squared Residuals And Node Imbalances

A method is proposed to perform on-line data adjustment and inference of unmeasured variables, using measured flowrates and on-stream assays in mineral processing plants. The algorithm minimizes a weighted sum of squared residuals of the measured variables and squared species imbalances around the flowsheet nodes. The method is based on a trade-off between the material transient accumulation, evaluated by the imbalances, and the data inaccuracies, evaluated by the residuals (i.e. the adjustments brought to the measured values). A recursive algorithm is developed for on-line implementation of the method. It is a real-time extension of the hierarchical method for balancing steady-state data for bilinear mass conservation constraints. An illustration for a simulated flotation circuit is provided.