Detection Approach To Joint Damage In High-rise Tower Structures

Damage diagnosis of structures has recently become a hot problem in the field of engineering. Compared with other structures, steel-tower structures are exposed to awful environments, lashed by wind and corroded by water the whole year. In consequence, they are apt to damage. And for their special environments, ordinary examining theories hardly satisfy us. Taking the Xiangjiang River tower and the Hanjiang River tower as engineering background, the paper carry out an in-depth study on the practical application of Improved Genetic Algorithms (IGAs) and Fussy-pattern recognition (FPR) in the field of structural damage detection on constructions.A t first, the fuzzy pattern recognition theory is introduced, structure intelligent healthy monitoring based on the fuzzy pattern recognition is analyzed in detail. Based on the damage of high-rising steel-tower, a damaged finit element static modelhas been established, and A One-step detection approach of Fussy-pattern recognition is presented, It locates a specific damaged joint using the Fuzzy-pattern recognition based on strain mode theory. Performance of the proposed damage detection approach is demonstrated with analysis of a tower structure. The result turns up trumps.The genetic algorithm is a kind of global optimization method, which has performed great advantages over any other algorithms in many kinds of fields. However, it has two flaws: stagnation and premature convergence during the optimal process. Therefore, some improved strategies including multi-parents crossover, fine-tuning of variables and niche technique are introduced to improve the convergence efficiency ofGAs.The usual finite element analysis (FEA) for beam structures is introduced in the paper. And then a finite element static model is presented for the damage detection of high-rising steel-tower structures based on static response. Module procedures of great applicability are constructed for damage diagnosis with the APDL (ANSYS Parametric Design Language) and FORTRAN. And these modules can be accessible each other through API (the Application Programming Interface).A GA decoding means named local decoding technique (LDT) is developed for thedamage detection of large steel-tower structures in the paper. Traditional decoding means used in the field of structural damage diagnosis is to give a design variable (damage factor) to each element, which is only applicable to small structures of fewer elements for the reason that it will be time-consuming and of bad convergence properties when GAs confronted with large numbers of design variables. In contrast to traditional decoding means, the LDT gives two design variables, namely damage factor and damage location, to possibly damaged elements, as not all the elements will be damaged. Of course, the maximum number of damaged elements should be estimated before. The research indicates that both the convergence velocity and efficiency are evidently improved.The research has paved the way to the practical application of IGAs and FPR in the field of structural damage detection on constructions.