Fuzzy Expert System to Predict Maintenance Intervals in a Continuous Caster

SUMMARY During short-term scheduling in a steel plant, one problem is to know beforehand how long equipment will be usable and when it will have to be maintained. The proposed system answers this question for a part in a continuous caster in order to inform another system that schedules short-term production. It uses fuzzy inference rules to process input data and to compute the life-expectancy of the part. Wheras human operators tend to use pessimistic values in order to be on the safe side, the system performs better in predicting the life-expectancy since it considers more types of innuence and reasons with up-to-date information. For these reasons, the nal schedule matches reality closer, avoiding vasted raw materials and production delays, while at the same time increasing product quality. Both systems are currently developed in a joint project between the Austrian Industries Holding and the Christian Doppler Laboratory for Expert Systems.