Abstract:

This paper investigates the parameter identification and fitting methods for large-scale structures to predict the concrete creep in an exponential form. In the investigation, first, the international standard BP2 model of concrete creep is employed to construct a common degenerate-kernel Dirichlet series as the basis function. Next, in view of the fact that the actual parameters are more than the unknown variables in the fitting process, the least square method is used to transform the inconsistent equation group into a normal one. Then, the normal equation group is solved by means of the conjugate gradient method. Finally, based on the presented techniques, program is worked out to identify the parameters of Dirichlet series of BP2 model. It is found that the fitted results accord well with the original ones, the relative error being less than ± 3 %. The recommended Dirichlet series of BP2 model is suitable for the tracking and analysis of long-term creep behavior of large-scale structures. Also, it provides a reference for the performance evaluation of existing structures.

Key words: creep function, BP2 model, parameter identification, Dirichlet series, conjugate gradient method