Supporting Structure Reliability Optimization Research Of Foundation Pit By The Method Of Response Surface

Reliability optimization design includes two aspects:one is reliability analysis; The second is to optimize the design. At present, there are still exists some problems to use of the method of reliability optimization design. First of all, it is difficult to obtain the explicit expression of the objective function, because of the theoretical calculation model and mechanical mechanism of complex foundation pit engineering is very complex and often showed highly nonlinear; As a result, it is difficult to use the reliability analysis of moment method and the optimization analysis of the mathematical programming method. Second, the optimization method of the heuristic method and reliability analysis of Monte Carlo method have the common problem is the low efficiency of calculation, but both the methods do not need to obtain the explicit expression of the objective function. Therefore, it is difficult to realize the reliability optimization design combine the both methods. In this paper, the response surface method is adopted to realize the reliability optimization design.The author collected a lot of geological survey data of foundation pit engineering in the area of Kunming, and analyzed the correlation between parameters by the method of orthogonal transformation, obtained 5 kinds of typical soils of Cross correlation coefficient of the statistical range, it is show that between the shear strength indexs c and φ are negative correlation. By uniform design and ACE regression techniques to establish efficient response surface. on this basis, we studied the soil parameters of cross correlation on the influence of reliability analysis, to provide the basis in the optimization design considering the cross correlation of the soil parameters.The test data obtained from the geological survey data is less, belong to the small sample data. So, the Bootstrap technique is introduced to solve the problem. For a specific project, the improved response surface method based on uniform design and ACE regression technology to implement under the condition of small sample reliability optimization design of foundation pit supporting structure. System reliability as constraint conditions, the quantities of the main process for the optimization goal; on the premise of guarantee the reliability of 95%, obtained the optimal parameters of supporting structure design. In order to compare, realized based on the assumed distribution of foundation pit supporting structure reliability optimal design. Undering the condition of large and small samples of reliability optimized design, considering the influence of the cross correlation between the soil parameters. The results show that under the condition of the large and small samples optimization results has certain difference, but under the condition of small sample optimization results can reflect the real situation more. Cross correlation between the soil parameters is the nature of the soil properties, considering the cross correlation between the parameters can obtain optimization design scheme of more economic.Through the method implements the reliability optimization design, the supporting structure can be ensured safe and reliable, and can obtain optimal design scheme. Using uniform design and ACE regression technique to establish the effective response surface method, optimization of high efficiency and short computing time. According to the project example analysis, it is proved that the feasibility and effectiveness of this method.