Modeling of distributed simulation of carbothermic aluminum process using recording controllers

In this thesis we propose a novel algorithm for distributed process simulation. The algorithm is used to develop the simulation of a new technology, the Carbothermic Aluminum Process. The stability analysis for this algorithm is not performed from a numerical point of view but from a physical one. This approach eliminates the need for synchronization and a coordinator nor does it require an iterative procedure. The use of recording-controller (RC) devices is introduced for this purpose, and is added to original process system. The functions of the recording-controller are to accumulate the simulation error between communication intervals, perform error checking and control, as well as update the interconnection variables that carry over to the next time horizon. This algorithm is used to develop dynamic and steady state simulations for two different simulation strategies: distributed simulations with fix time horizon and distributed simulations with adaptive time horizon.; Mathematical models are developed for the three different process units of the carbothermic aluminum process. The smelting stages (stage one and two) are assumed to operate at equilibrium. The mathematical models developed for these process units are Equilibrium Stirred Tank Reactors model (ESTR). For the Vapor Recovery, two different mathematical models are developed to represent the two operation modes of this unit: fixed bed and moving bed. The proposed process simulation model can be for the conceptual design of the carbothermic aluminum pilot plant, for control system development and for process optimization.