ISSN 2594-5327
55º Congresso anual — Vol. 55 , num. 1 (2000)
Título
Autoria
DOI
Downloads
Resumo
VAI offers a powerful blast furnace automation system, which was developed in close cooperation with the blast furnace division of VOEST-ALPINE Stahl Linz on the basis of existing models and operational know-how. This blast furnace automation system which has been named VAiron is now in operation at several blast furnaces worldwide with excellent results. The most recent commissioning has been at the 9000 t/d Blast Furnace L of Bethlehem Steel, Sparrows Point, USA. One of the highlights of the blast furnace automation solution is an expert system for process diagnosis which is based on a user-expandable knowledge-base. The system interprets the process, provides closed-loop control as well as proposals for corrective actions and describes in detail how the results were obtained. Expert systems are often denominated as process advisory systems, i.e. they only give proposals for actions to optimize the process. The VAI expert system is the next step into the future of optimized hot metal production. It executes coke rate control, basicity control, steam injection control etc. in closed-loop without any interaction of the operator. The expert system is successfully running on blast furnace A of VOEST ALPINE Stahl Linz. The decision for stepping to closed-loop operation was made in December 1996, after operating the system in advisory mode since 1992. For the future, extensions for optimize the system with machine learning capabilities in the area of the burden distribution control will be made. In the paper the structure of the expert system will be described in detail as well as the results of the closed-loop operation including the benefits reached. Additionally the success factors by moving from an advisory expert system to a closed-loop expert system are pointed out.
VAI offers a powerful blast furnace automation system, which was developed in close cooperation with the blast furnace division of VOEST-ALPINE Stahl Linz on the basis of existing models and operational know-how. This blast furnace automation system which has been named VAiron is now in operation at several blast furnaces worldwide with excellent results. The most recent commissioning has been at the 9000 t/d Blast Furnace L of Bethlehem Steel, Sparrows Point, USA. One of the highlights of the blast furnace automation solution is an expert system for process diagnosis which is based on a user-expandable knowledge-base. The system interprets the process, provides closed-loop control as well as proposals for corrective actions and describes in detail how the results were obtained. Expert systems are often denominated as process advisory systems, i.e. they only give proposals for actions to optimize the process. The VAI expert system is the next step into the future of optimized hot metal production. It executes coke rate control, basicity control, steam injection control etc. in closed-loop without any interaction of the operator. The expert system is successfully running on blast furnace A of VOEST ALPINE Stahl Linz. The decision for stepping to closed-loop operation was made in December 1996, after operating the system in advisory mode since 1992. For the future, extensions for optimize the system with machine learning capabilities in the area of the burden distribution control will be made. In the paper the structure of the expert system will be described in detail as well as the results of the closed-loop operation including the benefits reached. Additionally the success factors by moving from an advisory expert system to a closed-loop expert system are pointed out.
Palavras-chave
blast furnace, expert system, closed-loop control, automation, process optimization
blast furnace, expert system, closed-loop control, automation, process optimization
Como citar
Druckenhammer, H.; Scholz, B.; Brunnbauer, G.; Isfanger, F..
Automatic Blast Furnace Operation by Closed-Loop Expert System,
p. 645-656.
In: 55º Congresso anual,
Rio de Janeiro, Brasil,
2000.
ISSN: 2594-5327, DOI 10.5151/2594-5327-C00809
