轧制差厚板（TRB）是一种轻量化结构用材，具有较大应用潜力。本项目针对传统分析方法无法对复杂板壳结构细微几何特征和误差进行有效描述，不确定分析和可靠性设计难以兼顾精度和效率导致工程普适性不强等难点问题开展研究。首先，基于等几何分析，推导考虑厚向应力应变的实体壳单元；并研究消除其自锁现象，提高计算精度实现几何误差精确描述。其次，研究基于非概率模型的不确定等几何分析方法，实现快速不确定传播评估。再次，基于等几何分析理论，进行灵敏度解析推导，并开展优化解耦方法研究，解决可靠性设计求解的嵌套优化问题；同时，研究新型高维代理模型技术以及相应的模型管理策略，结合优化解耦方法，完成可靠性优化设计高效求解算法。最后，将相关模型、算法、准则进行集成。本项目最终得到的新型非概率等几何分析方法和可靠性优化设计高效求解算法，有望为复杂板壳结构的设计提供新思路，并有效增加了算法的工程实用性。

Tailor rolled blanks (TRBs) are lightweight shell structures. It is of great significance to study its mechanical behavior and reliability optimization under complex uncertain circumstances. This project intends to overcome the shortages of the deficiency of modeling the geometric error in complex shell structures, difficult to enhance accuracy and efficiency concurrently in uncertainty propagation analysis, and less applicability in real-world engineering. First, the solid-sell element considering the transverse stress and strain is established and its locking effect is investigated and alleviated under the framework of isogeometric analysis. Second, a novel non-probabilistic uncertain isogeometric analysis method is developed, based on which the uncertainty propagation will be investigated. Third, sensitivity analysis is conducted analytically based on isogeometric analysis. Study of decoupling methods could be carried out using the sensitivity information to solve the problems of nesting optimization processes in reliability design optimization. As a result, a highly efficient algorithm for reliability design optimization is developed combining with techniques of high dimensional model representation and corresponding model management strategy. Finally, the aforementioned models, algorithms will be modularized for integration. The new non-probabilistic isogeometric analysis methods and efficient reliability design optimization algorithm obtained in this project are expected to provide a novel approach for the design of complex shell structure, and effectively increase the practicality in real-world engineering.