Multilevel Reliability-Based Optimization of Structural Systems

对于随机荷载作用下的不确定结构系统，目前的优化方法几乎总是将正常使用状态和极限状态分开处理。这些方法的一个主要缺点是，在一个极限状态下的设计决策必须在缺乏关于其在其他受关注的极限状态下的后果的明确信息的情况下做出。这要么导致过度设计，要么导致结构系统可靠性较低，造成相当大的资金浪费。该项目的基本目标是开发一种高效有效的方法，用于结构系统在正常使用状态和极限状态下同时施加性能约束的多目标可靠性优化。几乎没有进行结构优化的基础研究来开发这样的方法。综述了非确定性结构构件在服役和极限随机荷载作用下识别概率行为模型的最新研究成果。这些单独的模型将被合并到一个通用的系统模型中，该模型将用于开发基于多目标可靠性的优化方法，从而在两个服务端最终极限状态下同时满足性能约束。这种方法将被扩展，以适应最优解对定义载荷和构件阻力的随机参数的变化的一般响应灵敏度。预计研究结果将对多可靠性性能约束下结构体系的有效设计具有重要的参考价值。

Current optimization approaches to nondeterministic structural systems subject to random loads almost invariably treat the serviceability and ultimate limit states separately. A major drawback to these approaches is that design decisions at one limit state must be made in the absence of explicit information as to their consequences at other limit states of concern. This leads either to a considerable waste of money because of overdesign or to a low reliability of structural systems. The essential objective of this project is to develop an efficient and effective method for the multi-objective reliability-based optimization of structural systems subject to performance constraints imposed simultaneously at both serviceability and ultimate limit states. Almost no basic studies in structural optimization have been performed to develop such a method. A survey of recent research will be made for identifying probabilistic behavior models for nondeterministic structural members under service and ultimate random loads. These individual models will be incorporated into a general system model which will be used for developing a multi-objective reliability-based optimization method whereby performance constraints will be satisfied simultaneously at both service end ultimate limit states. This method will be extended to accommodate generel response sensitivities of the optimum solution to changes in the random parameters that define the loading and the member resistances. It is anticipated that the results will be of considerable value for the efficient design of structural systems under multiple reliability-based performance constraints.