Research On Static Reliability Assessment Method For Long-span Cable-stayed Bridges

Because of reasonable structure performance, remarkable mechanical property, beautiful shape and economic cost and other characteristics, cable-stayed bridges develops by leaps and bounds during more than half a century in the past and become the first option of large span bridges. However, large span cable-stayed bridge design and construction still follows the traditional deterministic method nowadays so that the guiding ideology does not conform with many random factors during cable-stayed bridges construction period and operation period in actual. Therefore, the deterministic analysis methods are hard to demonstrate random factors influence on reliability index of cable-stayed bridge, random factors include the material parameters, geometric dimensions, external load, and many other factors. Large span cable-stayed bridges are high order statically indeterminate flexible structure whose structural response is quite complex and nonlinear problem is even more significant, hence high-effective intelligent reliability analysis method, which research on structural reliability, is a quite popular topic today, the method could also evaluate the reliability development trend of the complicated bridge structure under the action of many random factors and provide a more reasonable basis for a long-span cable-stayed bridge design, construction and maintenance. On the basis of reliability theory, this paper introduces a hybrid algorithm applied to complex structure reliability analysis of a long-span cable-stayed bridge, the algorithm was utilized to analyze reliability index and stochastic static analysis of a long-span cable-stayed bridge under serviceability limit state and a parameter sensitivity analysis of multiple random variables influence on the structural reliability index. Get the following conclusion:(1) A hybrid algorithm of the concrete cable-stayed bridge reliability assessment and stochastic static analysis was put forward. This algorithm was combined with nonlinear finite element of Monte Carlo importance sampling method numerical simulation and Radial Basis Function (RBF) neural network, and Genetic Algorithm (GA) was drew into the response surface fitting, not only does decrease calculation quantity of Monte Carlo sampling directly but also improve accuracy of the response surface. The example result show that the hybrid algorithm could well simulate the nonlinear mapping relationship between random variables and structural response, also could obtain a satisfactory precision and computational efficiency. The hybrid algorithm is suitable for large-scale complex structure reliability evaluation and parameter sensitivity analysis with multiple random variables and offered a new way for structure reliability calculations.(2) To explore whether it could better applicable to calculate the reliability of a complex structure in practical engineering, the hybrid algorithm was applied to reliability index and stochastic static analysis of a long-span cable-stayed bridge under serviceability limit state. Comprehensive expounded the reliability assessment of stochastic variable statistics and selection, the construction of finite element analysis model, the establishment of the limit state equation, the selection of network parameters and so on which were prepared for the structure reliability assessment and parameter analysis.(3) The proposed hybrid algorithm was used to calculate reliability of a cable-stayed bridge under serviceability limit state, the reliability index of mid-span displacement over limits of the cable-stayed bridge main girder and the reliability index of strength failure of the tail cable were calculated. A parameter sensitivity analysis was utilized to random variables influence on the structural reliability with the background of the cable-stayed bridge which shows deep comprehension on the key of structure parameters control during the construction period and operation period of cable-stayed bridge. Controlling these parameters could improve the structure quality in the process of construction, grasp the initiative in the process of construction and operation and ensure that the reliability index of the bridge structure could meet the design requirements during the construction and operation period.(4) In view of random variables such as the structure of the material properties, geometrical size and external load influence on the variability of structural reliability index, the hybrid algorithm was used to non-linear stochastic static analysis of a long-span cable-stayed bridge. Variation tendency of a cable-stayed bridge structural response (mid-span displacement of the main girder, tail cable force) was studied when many random factors changed and it provided the theoretical guarantee for reasonable design of cable-stayed bridge.