Resistance Design Value Inference Method Based On Probability Model

The structure design method based on tests is a new international structure design method. It introduces structure tests into the specific design process and directly determines the strength design value of the design objects through load test and statistical inference. It has prominent positiveness and is suitable for mass designs of similar components, complex structure and new structure. It is one of the important directions for the development of structural design method.The inference method based on analysis model is an important part of the design method based on test, and is a key method to infer resistance design value. It makes use of the probabilistic properties of known influence factor in force analysis model during inference, can reduce the uncertainty of the inference in a large extent and obtains better inference results. However, the method in the international standard ISO2394: 1998 is an indirect inference method. It adopts the analysis model of design and cannot reflect the influence of probabilistic properties of various influencing factors directly. So it fails to ensure resistance design value with a consistent guarantee rate during inference,which goes against the control of structure reliability. The method in the European standard EN1990:2002 can infer the design value of resistance under a consistent guarantee rate, but it is an experiential method with theoretical defects whose inferences tend to be conservative, even with resistance design value to be zero. At present, there is no good solution to infer resistance design value which is under a consistent guarantee rate and based on a strict theory.Aiming at the shortcomings of the current international resistance inference method, this thesis presents an inference method of resistance design value based onprobability model, which is directly according to resistance probability model and totally on the basis of statistical method. Aiming at the two situations of “unknown variable coefficient of computing model uncertainty coefficient” and “known variable coefficient of computing model uncertainty coefficient”, this method puts forward the inference formula of resistance design value under a consistent guarantee rate. The method has a strict theoretical basis and overcomes the shortcomings of former resistance inference methods. To test the applicability of the method, this thesis introduces an engineering example, and uses the interval estimation method and the method in this thesis respectively to infer the design value of load-bearing capacity of bending components and analysis the results. In the end, under the same guarantee rate,this thesis compares the inferences of the method in the European standard and the method in this thesis. The result shows that resistance design value inference method based on probability model not only comply with the general law of statistics, but also can give an appropriate inference result.Resistance inference method based on probability model improves the resistance inference method based on analysis model, and is a complement of the present structural component resistance inference method.