FURTHER DEVELOPMENTS IN CHEMICAL-PLANT COST ESTIMATING USING FUZZY MATCHING

Capital cost estimates for chemical plants are needed to assess economic viability. Simple methods for producing accurate estimates are needed at the beginning of a project. However, existing methods have limited applicability and disappointing accuracy. Plants with similar specifications (capacities, process conditions, etc.) have similar costs, therefore we have tried using the capital cost of the existing plant that is 'most like' the proposed plant as an estimate of the cost of the new plant. We have developed fuzzy matching, a method based on fuzzy logic, which finds the 'most like' plant from a database of existing plants by quantifying the closeness of their specifications. Fuzzy matching has since been enhanced by testing different functions for quantifying closeness, and 'optimising' parameters in these functions. The method has been extended to take account of plant materials of construction. Finally, the closeness of individual specifications has been weighted in the calculations of the quantity expressing overall closeness, to account for the different degree of influence of the specifications on the capital cost. Fuzzy matching offers accurate preliminary capital cost estimates, especially when compared with existing methods, and with minimal estimating effort.