Fuzzy logic techniques in bridge management systems

Abstract

Bridge-management systems are suggesting the required repair-work algorithmically. Sufficiency is derived from "IF-THEN" type logic decisions, or in more sophisticated applications as a group of parametric decisions. The contradiction between the "crisp" nature of algorithmically definable standards, and "soft" technical benchmarks derived from professional practice can be resolved by the introduction of "fuzzy logic" to the particular field. Applying fuzzy logic the compliance conditions/measures are not crisp values, but defined by fuzzy membership functions. The complex sufficiency index of bridges incorporating the load bearing-; width-; and clear height sufficiency are given on a 100 grade scale with 10 categories. The connecting element between the input (load bearing - width – clear height sufficiency data) and the output (complex sufficiency ratings) is the rules of logical preconditions based on the professional knowledge. For every combination of input ratings the rules designate separate complex sufficiency ratings. For executing the fuzzy deductions, in this paper the Mamdani reasoning method is suggested. For applying fuzzy reasoning, the measured load bearing - width - clear height values have to be converted to membership functions, by defining the membership values for the corresponding bridge data.