Estimation of steel plant performance using a knowledge based system

Electric arc furnaces (EAFs) are used to produce steel billets mainly from scrap for use in rolling mills. The rolling mills may be owned by the same group that own the EAF. The operation of the mills in terms of costs, quality, lead times and service are significantly affected by these furnaces. Therefore estimating the performance of competitors' EAF's is vital for the strategy of continuous improvement. The first performance metric looked at internally is usually the cash or operating costs. There is often some information published regarding competitors' operations and experts at the steel plants apply rules of thumb to this information along with site visit reports and other sources of information to make estimates of the operation of the competitors' furnaces. These results are not necessarily consistent and the knowledge of how these costs are generated is limited to those experts. This paper describes the design of an advisory knowledge based system for estimating the costs, based upon the rules of thumb used by the experts, from the limited information obtainable. The raw information is in both qualitative and quantitative terms. This system takes the raw information and applies rules used by the experts to calculate the costs in a consistent, readily available and understandable manner. This also means that more time can be spent on examining the different EAF practices used rather than on calculating the quantitative measures.