The proposed knowledge-based system (expert system) will primarily provide recommendations as to the control actions that an alumina calciner operator should implement to enable the existing control system of the calciner to react better towards dynamic changes in the internal and external operating parameters of the alumina calcination process. This will result in direct thermal as well as electric energy savings amounting to 10% and 5% of their nominal consumption respectively. The knowledge base of the proposed expert system will have hybrid structure i.e. it will consist of both "shallow" and "deep" knowledge. The "shallow" knowledge shall be formulated on the basis of explicit heuristic control rules directly extractable from the knowledge of the expert personnel responsible for the operation of the calciner. On the other hand, the "deep" knowledge will be founded on a dynamic model that will simulate the dynamic behaviour of the calciner. The proposed system will be tested as aPC-based process control decision-aid "tool" in the operation of the flash calciner of Aluminium de Grece S.A at its alumina producing plant in Boietia, Greece. It will have on line process data acquisition capabilities and on the basis of the real-time status of the calcination process, the expert system will provide the process operator with the recommended control actions as well as with the new tuning parameters of the process controllers in order to assist the operator in his or her effort to improve the stability and thus the energy efficiency of the alumina calciner.
Funding SchemeCSC - Cost-sharing contracts
32003 Paralia Distimou