강건 최적화 기법에 의한 고속철도 소수주형 강교량의 LCC 효율적인 실용적 최적 설계

Abstract

현재 고속철도 교량구간에서도 강합성 소수주형교를 적용한 교량의 건설이 진행되고 있다. 철도교량에서의 강교를 좀 더 효율적으로 활용하기 위해서는 공용기간중의 강교의 유지관리도 중요하지만, 교량의 생애주기 동안 비용절감효과를 얻을 수 있도록 초기 설계단계에서 유지관리 최소화를 반영한 LCC를 고려한 최적설계가 이루어져야 한다. 따라서 본 연구에서는 고속철도 강교량의 강성 및 처짐, 진동 등에 대한 제약조건을 고려한 최적 설계를 통해 LCC 효율적인 최적설계방법론을 제시하고자 한다. 본 연구는 2단계의 과정으로 구성되었다. 레이아웃 및 시스템 최적설계단계에서는 먼저 근사 레이아웃 결정을 위해 위상 최적설계에서 주어지는 정보를 활용하여 전체 구조 시스템의 강성의 흐름을 파악하고 이를 레이아웃 선정에 적용하고자 한다. 또한 시스템 최적설계를 위해 통계적 분석 기법인 강건 최적화 기법을 적용 이산공간에서의 잡음에 강건한 근사 최적해를 탐색한다. 이러한 과정을 거쳐 얻어진 최적설계와 기존설계와의 LCC 평가를 수행하여 레이아웃을 고려한 강건최적 설계의 LCC 효율성을 검증하고자 한다.; The thesis presents both practical and realistic optimization methodology of high-speed railway steel bridge with concrete slab-on-steel girders in terms of ROM(Robustness Optimization Method). In addition, the thesis suggests LCC methodology for the LCC-effective optimum design of the bridge. The thesis is mainly divided into two phases. Whereas one phase is with regard to the robustness and system optimization, the other phases is dealing with LCC assesment. The former is to find an approximate layout of the structure, which leads to obtain the system optimization in terms of both robust optimum algorithm and TOM(Topology Optimization Method). Robust design is the design methodology that makes a product or process insensitive against noise causing the performance variation of the system. Regarding the Robust design, Taguchi method is well-accepted but complex. The method makes the use of both orthogonal array for experiment plan and S/N ratio as a measurement for checking the robustness of the system. Meanwhile, RO algorithm developed in the thesis is easy to use in general and efficient when the performance evaluation function requires intensive computational efforts in particular. The algorithm does not need any derivative information and thus the near-optimum solution is obtained very quickly. The algorithm is even applicable to the case where discrete design variables are considered. Hence, it can be a powerful tool for the practical application of structural optimization cases. Efficiency of the developed algorithm is proved in terms of the comparative study with GA(Genetic Algorithm). Two different cases using the ROM are modeled for the high-speed steel -girder railway bridge. While Case 1 is the model without brace system in lower part of the bridge, Case 2 is the model with full brace system. Both cases are modeled to satisfy the limits of bending and shear stress, displacement and plate buckling. Additionally, the optimum model of the bridge should satisfy the natural frequency limit, which is caused by the train-moving load at the speed of 350km/h. Both optimal value of element and the optimal layout of the structure are estimated in the two cases in terms of the ROM. Once the above optimal features are obtained, LCC optimization design is carried out taking into account initial cost, maintenance cost and indirect/direct cost. Moreover, comparison has been made between the LCC optimization design and Initial Cost Optimal Design of the real model. Particular emphasis is placed on the expected rehabilitation cost which requires for the failure probability associated with crack, damage and etc. Accordingly, limit state functions taking into account the loss of both strength and workability are provided in the thesis. Once the probability of failure is evaluated by the limit state function, the expected rehabilitation cost has been determined. In all, achieved are both sizing and layout optimization for the high-speed steel-girder railway bridge in terms of TO and RO. Also obtained are LCC effective optimization design and cost. In view of the above, a realistic yet practical LCC effective optimization should be taken into account for the cost analysis and estimation of an enormous system. In addition, there is an urgent need that a new model taking into account the strength deterioration should be developed for the high-speed railway steel bridge.