Application of fuzzy logic algorithms for predicting the pipe-culvert technical condition on railways

Abstract

This article is part of the author’s dissertation research in the field of developing a methodology for predicting the pipe-culvert technical condition on railways using the mathematical apparatus of fuzzy logic. One of the common methods for assessing engineering structures is the point rating method. During the examination, each identified defect is assigned a certain number of points. Then the number of points is summed up according to the developed methodology and, based on the obtained result, a conclusion is drawn about the condition of the object. However, in conditions of significant subjectivity and limited input data, the issue of predicting the structure technical condition is quite labor-intensive. As an addition to existing methods for predicting the structure technical condition, in particular pipe-culvert, it is proposed to apply a method based on the fuzzy logic principles, which is a generalization of fuzzy logic and set theory. In fuzzy logic, unlike classical logic, instead of the values true and false, the truth value degree is used, which takes any value from an infinite set from 0 to 1 inclusive. To predict changes in condition and determine the deterioration rate of the technical characteristics of a pipe-culvert, it is necessary to create a model in which scoring methods are transformed into estimates presented in terms of fuzzy logic. The article presents a methodology for using fuzzy logic algorithms to predict the pipe-culvert technical condition on railways. As a result of the research, numerical gradations of the term «pipe-сulvert Age» were obtained. With limited and subjective input data, expressed numerically in terms of fuzzy logic — «pipe-culvert Age» and «pipe-culvert Condition», it is possible to calculate the values of the term «pipe-culvert Degradation Rate». The author presents a method for predicting the technical condition of railway pipe-culverts, based on the principles of fuzzy logic, founded on fundamental set theory. The developed mathematical model makes it possible to calculate the structure transition time from one technical state to another. To implement fuzzy logical inference of the presented model, the Fuzzy Logic Toolbox package is used in the MATLAB software environment.