| A physical model of a roll and race coal pulverizer is developed. The processes of grinding, pneumatic transport, drying, and classification of a Babcock and Wilcox MPS-89 mill are modeled based on the fundamental physical laws. Roller pressure, coal particle strength, and the Gaudin and Meloy function are used to predict the coal size distribution in the grinding process. Drag force and momentum and energy conservation equations are combined to compute the pressure drop and the temperature profiles. Classification is determined by the particle trajectory which is calculated based on the centrifugal, drag, and gravity forces. Mill start-up, pulse tests, and steady-state tests at different load conditions were performed on a smaller MPS-32 mill. The model simulation steady-state results compare well with the test data, and the pulse and start-up results are mainly compared qualitatively and on a percentage-wise basis. A transfer matrix, derived from the frequency analysis of the computer simulation data, is used to represent the system having five inputs and four outputs. The Inverse Nyquist Array (INA) method is then applied to the 4 x 5 system to control all the outputs while producing less time delay in the coal output flow for a step increase in the coal output demand simulation. The INA application of a 2 x 2 system, which has the same structure as the current control design, is able to produce a faster response with the dominant time constant reduced by a half. |
