Soft Sensing Model of Underflow Concentration for Thickener Process based on Data Reconciliation

Measured data in industrial production inevitably contain stochastic errors that are influenced by the environment and equipment, and the stochastic errors will reduce the accuracy of the measured data. Thickeners are widely used in mineral processing and sewage treatment, but the accuracy of the variables of the thickener process cannot meet the requirements, caused by the stochastic errors, and the underflow concentration cannot be real-time detected. So, pressure sensors were installed in the thickener to establish the soft sensing model of the underflow concentration, but the model accuracy is low owing to the stochastic errors of pressure values. This paper presents a soft sensing model of underflow concentration based on data reconciliation. The data reconciliation method is used to improve the accuracy of variables in thickener process, and then the soft sensing model of underflow concentration is established using the reconciled pressure. Applying the data of impact thickener process, the validity of the model is confirmed.