Error Analysis And Probabilistic Approach For Parameter Identification Of Structures Based On Data Under Static Load Test

Structural model parameters identification using response under static load test for structural safety assessment is of great significance.. On the structural parameter identification research, identification of structural parameters faced various constraints, many scholars have done a lot of work. Based on the work of our predecessors, this article focuses on the following work:(1) Structural parameter recognition errors in a detailed classification, definition and description, errors are analyzed.(2) Test by cantilever beam, through the analysis of cantilever beam structure, based on model assumptions to identify alternative models and their parameters. Model assumes that the application, as well as provide some guidance on quantification of model errors.(3) The conception of model is expanded, propose the concept of structural model identification instead of structural parameter identification; Measurement quality equivalent to the ability to differentiate among models, prepared in accordance with the principle of maximum entropy procedure, select the maximum distinction recognition model of measurement systems, identifying the optimal set of points of the structure. Further, based on the maximum entropy principle determine unknown structures of optimal distribution of measurement systems and loading system.(4) Choose to use PGSL algorithm based on Bayesian theory and stochastic search Monte Carlo method as the model identification algorithm. The original algorithm is modified to meet the practical needs. By the complete algorithm, solve extreme value of a function containing multiple solution problem.(5) The algorithm is applied to the test on the identification of a cantilever beam. In accordance with the following steps: a. Model assumes that the concept of an alternative model and its parameters, parameters of the ranges are given; b. Based on the maximum entropy principle obtain the best measuring point system and the parameter sensitivity analysis; c. Calculate the measurement uncertainty of each measurement point, with detailed analysis of estimation error of the model assumptions, calculated item to allow residuals; d. Comprehensive steps b and c, identify several identification methods, respectively, for the structure of the model identification; e. Evaluation results to determine the optimal method, make conclusions.