基于混合度量的不确定性结构动态载荷识别的理论与算法研究

确定结构所承受的动态载荷是保证工程结构设计安全可靠的重要前提。载荷识别技术为无法直接测量载荷的结构提供了一种有效的动态载荷获取方法。工程实际中不确定性广泛存在，客观有效地估计和评价不确定性对载荷识别的影响有着重要的学术研究价值和工程实际意义。本项目拟综合考虑系统模型的不确定性、载荷识别反问题的不适定性和嵌套求解计算量大等技术难点问题，研究和开发不确定性结构动态载荷识别的计算反求技术。首先，从概率和区间混合的角度对不确定性进行建模，实现不确定参数更为准确的定量描述；其次，分析不确定载荷识别反问题的不适定性，研究有效的克服载荷识别病态问题的正则化方法；同时，针对不确定量和时间量开发高效的混合不确定载荷识别算法，解决不确定载荷反求过程中复杂的两层嵌套求解问题，实现结构所受动态载荷的概率区间的快速求解。本项目的完成能够为实际结构提供更为合理的载荷统计特性的边界，有效提高载荷识别技术的工程实用性。

To determinate the dynamic load acting on the structure is the essential precondition for assuring the safety and reliability of structural design. The technique of load identification can provide an effective means of dynamic load obtainment for the structural load which can not be measured directly. In engineering practice, the uncertainty is widely existed. To objectively and effectively estimate the influence of uncertainty for load identification has important academic value and practical meanings. In this project, the technical difficulties about the uncertainty of system parameters, the ill-posedness from load identification and the large nested computation will be comprehensively considered, and the computational inverse techniques of dynamic load identification for uncertain structures will be investigated and developed. Firstly, the hybrid model of uncertainty will be established based on probability and interval, which can realize more accurate quantitative description about the uncertain parameters. Secondly, the ill-posedness of the inverse problem about uncertain load identification will be analyzed, and the regularization method to overcome the ill-posed problem will be researched. Furthermore, the high efficient hybrid algorithm for uncertain dynamic load identification will be developed, which can solve the complicated two layer nested problem in the load identification, and realize the speedily solving for the statistic interval of the dynamic load acting on uncertain structures. The completion of this project can provide more compelling statistic boundaries of the dynamic load, and efficiently improve the practicability of the load identification technique in engineering.