SFB 805: Control of Uncertainties in Load Carrying Systems in Mechanical Engineering

不确定性发生在产品和过程的开发过程中，以及产品生命周期的所有阶段。2014年，每一辆进入美国市场的新车，大约有四辆已经注册的汽车必须被制造商召回。在德国，同年有300万辆新车登记，150万辆汽车必须进行整改。根据合作研究中心（CRC） 805的科学家的研究，如果在产品生命周期的三个阶段（产品开发、生产和使用）中相应的工艺链不能确定到完全延伸，则产品和潜在产品缺陷的不确定性就会出现。因此，CRC 805处理的方法和技术来描述，评估和控制不确定性的传播分阶段的方法。CRC 805制定的方法基于三个支柱。第一个支柱包括由CRC 805开发和测试的术语。方法构成了第二个支柱。重点研究了不确定性的非参数化联动、鲁棒优化、鲁棒设计和不确定性的尺度化。第三个支柱包括描述和控制生产和使用中的不确定性的技术。在这里，研究的重点是过程和使用的不确定性与感官成分的综合识别。CRC 805的第三个资助期侧重于控制无知，这是一个常见和重要的不确定性类别。工程、数学和法律相互合作，分析数据引起的冲突作为模型不确定性的指示，模型不确定性作为无知的原因，弹性作为控制不确定性的原则。因此，系统和工艺链的弹性评估系统和工艺链特性的偏差，超出计划的公差范围。由于弹性需要功能监控、反应、学习和预测，因此出现了流程链和系统的适应以及模型的适应及其在产品生命周期不同阶段的重用的问题。

Uncertainty occurs during the development of products and processes, as well as in all phases of the product life cycle. In 2014, for each new car, which was brought to the US market, approximately four already registered cars had to be recalled by the manufacturer. In Germany, three million new cars were registered and 1.5 million cars had to be rectified in the same year. Following the scientists of the Collaborative Research Centre (CRC) 805, uncertainty in the product and potential product defect arises if the corresponding process chains in the three product life cycle phases product development, production and usage cannot be determined to full extend.Thus, the CRC 805 deals with methods and technologies to describe, evaluate and control the propagation of uncertainty in a phased approach. The approach developed by the CRC 805 is based on three pillars. The first pillar consists of the terms, which have been developed and tested by the CRC 805. The methods form the second pillar. The focus is on non-parameterized linkage of uncertainty, robust optimization, robust design and scaling of uncertainty. The third pillar includes the technologies to describe and control uncertainty in production and usage. Here, the research focuses on a process- and usage-integrated identification of uncertainty with sensory components. The third funding period of the CRC 805 focuses on the control of ignorance as a common and important category of uncertainty. Engineering, mathematics and law cooperate interactively to analyze data-induced conflicts as an indication for model uncertainty, model uncertainty as cause of ignorance and resilience as a principle to control uncertainty. Thereby, resilience of systems and process chains is evaluated for deviations of system and process chain characteristics, which exceed the planned tolerance ranges. Since resilience requires the functions monitoring, reacting, learning and anticipating, the question on adaption of process chains and systems as well as the adaption of models and their re-use in the different phases of the product life cycle arises.