Using viscous resistance system based on machine learning in engineering concrete structures

Abstract

Through the combination of two approaches to evaluating structure change, a structural model and an unstructured model, a constructed model has been proposed in this article that evaluates structural change through the expansion of a linear model following the Hooke’s Law principle. The study has relied on the pure compression model of a structure’s concrete beam with elastic modulus ( E) and has added the coefficient of viscosity resistance ( C) to suggest a new evaluation method. By defining the aggregation of values of both coefficients C and E through the experimental model, the input parameters are the amplitude values of the vibration spectra and the values of frequencies based on machine learning, through which Z EC values are generated. The Z EC values determine a regression plane accumulated from the aggregation of values for both C and E. The article has introduced the Z EC concept as a useful parameter for the assessment of the quality of concrete structures by the nonlinear model with the appearance of the coefficient C. The results show that the Z EC values have expressed the distribution validity according to the structure’s differing degrees of change. Depending on the texture type and the structure status, these Z EC values will form different shapes. By implementing the actual surveys from many bridges with two types of beam structures, prestressed concrete and conjugated concrete, the Z EC values show the same development trend. On the contrary, in the case of a change in mechanical structure, the Z EC values tend to increase. This evidence proves, in regard to the process of structural change, that the larger the changes in the structure, the more pronounced the distribution of Z EC values, and the wider the distribution range. This shows that the ratio of the damping coefficient C to the elastic modulus E will become increasingly unstable as the structure becomes weaker and weaker. In the future, the results from this study can be applied in the assessment of many types of actual structures.