Soft-sensor approach for measuring pulverised coal flow

摘要

This study presents a methodology to estimate the instantaneous pulverised coal flow rate of the E-type vertical spindle coal mills. Modifications over a six-segment coal mill model have been proposed, and particle swarm optimisation technique with Kalman correction has been used to optimise the unknown parameters of the modified model by using real plant data. To account for the deviation of the mill parameters with time, an extended Kalman filter (EKF) using the modified model with the optimised parameters has been implemented that updates the pulverised coal flow rate to reflect the actual evolving measurements.