| Recently, many scholars from china and foreign countries built all kinds of displacement models of Dam's measurement data. And made a serials of analysis of Dam's condition on the basis of these models. Moreover many analytical methods are used on Dam's safety monitoring field. For example wavelet analysis , genetic algorithm , artificial neural network, and so on.In fact, the models built by stepwise regression method usually exist unfit problem. Genetic algorithm is a kind of new optimization method in recent years and used widely many fields. It has some merits such as intelligence hunting, parallel mode and global optimization, moreover no drawback calculation. Classical optimization method base on gradient calculation. Discussing this problem from mathematic theory, genetic algorithm may be led on the basis of regression model. Dependence on its effective self adapting global searching optimization. This result reduces optimization error of regression.This article applies genetic algorithm to mixing model of dam's safety monitoring for optimization it. Taking example for some arch dam, specific methods that displacement mixing models of some measuring points converse physical parameters is introduced, and influence of observation accuracy, water height and measuring points' position are discussed.This article still introduced how to built mixing models of multipoint displacement and conversion methods whether there is reversed pendulum or not. The same arch dam as an example, one dimension mixing model of multipoint displacement of some monolith and two-dimension one of different monoliths in dam crest are built, and reckoning average elastic ratio of dam body and dam foundation. Moreover fore and aft results is compared and analyzed, and then genetic algorithm is used to optimize multipoint mixing model. The result is relatively good. At last, all the research is connected in series by a procedure.
|
PDF Full Text Request