Although steady-state optimization is performed, the simulator is designed to be dynamic. A steady-state (static) simulator should be computationally less intensive to perform this type of optimization; however, static simulators cannot be truly validated on the basis of mill data since the mill is always showing a dynamic behavior. One could argue that from plant data a steady-state condition can be determined and then compared against simulated one, but this would require identifying at least more than one steady-state to properly talk about validating a steady-state simulator. On the contrary, a dynamic simulator can be straightforward validated by comparison of measured and simulated dynamic signals. What is to be compared remains a critical factor that must be addressed during the mathematical modelling.