Multi-scale failure analysis with polymorphic uncertainties for optimal design ofrotor blades

该项目的目标是识别转子叶片结构设计中的多态不确定性，并开发多尺度（空间和时间）非确定性模型和数值方法，这些模型和方法能够将这些不确定性整合到设计，优化，制造，测试和寿命维护的典型链中。第一个资助期的主要重点是转子叶片中的粘合剂，第二个资助期的重点是关键部件的多种失效机制，包括开裂、脱粘、屈曲和低周疲劳。第一个目标是全面研究代表性子部件的相应不确定性，以便能够验证预测。这个子组件本身将是经典的预先设计和制造。相应的不确定性将通过使用非破坏性测试（NDT）技术进行识别和测量。 第二个目标是开发具有多态不确定性的非确定性模型，用于开裂、脱粘、屈曲和低周疲劳损伤，还包括它们的相互作用。这些模型将在一个宏观尺度的参数结构模型中实现，该模型能够模拟低周疲劳载荷下的响应。这种循环准静态时间函数将从转子叶片的已知代表性载荷集合中导出，其模拟使用寿命期间的操作载荷。相同的准静态载荷将通过实验施加到子组件上，直到失效，包括通过光学、光纤和传统技术进行的全面响应测量。第三个目标是开发具有多态不确定性的数据同化方法，并在开发的模型和获得的测量结果上证明它们的合理性。显然，主要的不确定性可以适当地量化和最小化。 第四个目标是优化拓扑结构，形状和尺寸的给定的子组件，通过使用这个验证参数模型与多态的不确定性。开发适用于多态数据和约束的方法是另一个挑战。 最后将子部件的稳健优化设计与原始的经典预设计进行比较。在此基础上，多态不确定性在整个设计，建模，测试和优化链中的作用可以变得直接可见和可测量。这种独特的可能性是该项目的主要亮点。

The goal of the project is to identify polymorphic uncertainties in structural design of rotor blades and to develop multi-scale (in space and time) non-deterministic models and numerical approaches, which are able to integrate these uncertainties into a typical chain of design, optimization, manufacturing, testing and lifetime maintenance. Whereas the main focus of the first funding period was dedicated to adhesive bonds in rotor blades, the emphasis of the second period is directed to the multiple failure mechanisms of critical components including cracking, debonding, buckling and low-cycle fatigue.The first goal is to study the corresponding uncertainties on a representative sub-component comprehensively, in order to be able to validate predictions. This sub-component itself will be classically pre-designed and manufactured. The corresponding uncertainties will be identified and measured by use of non-destructive testing (NDT) techniques at disposal. The second goal is to develop non-deterministic models with polymorphic uncertainties for cracking, debonding, buckling and low-cycle fatigue damage, also including their interactions. These models will be implemented in a macro-scale parametric structural model able to simulate response under low-cycle fatigue loading. Such a cyclic quasi-static time function will be derived from the known representative load collectives for rotor blades, which simulate operation loads during service life. The same quasi-static loading will be applied to the sub-component experimentally until failure, including comprehensive response measurements by optical, fiber-optical and traditional techniques. The third goal is to develop data assimilation approaches with polymorphic uncertainties and to justify them on the developed models and obtained measurements. Evidently, the main uncertainties can be then properly quantified and minimized. The fourth goal is to optimize the topology, shape and size of the given sub-component by use of this validated parametric model with polymorphic uncertainties. The development of suitable methods for polymorphic data and constraints is an additional challenge. The robust optimal design of the sub-component will be finally compared to the original, classical pre-design. On this basis, the role of the polymorphic uncertainties in the entire chain of design, modelling, testing and optimization could become directly visible and measurable. This unique possibility is the main highlight of the project.