MONITORING AND DIAGNOSIS OF A GRINDING PROCESS

A rational and reliable (i.e. experimentally grounded) approach to the advanced monitoring and diagnostics of coal pulverizers is presented in this paper. Two different forms of the dynamic model of the grinding process are proposed: a simulation model, exactly accounting for the distributed nature of the grinding process and for the various nonlinearities, and an estimation model, which “condences” the relevant process dynamics properties into a couple of expressive parameters: the grinding time delay and a storage time constant. A further important aspect considered in the paper is how to avoid “false alarms”. Finally, to check the practical applicability of the method, the whole monitoring and diagnostic system has been implemented in a Personal Computer by a real time process control package.