Methodology for generating cross-technology metamodels (IKTINO)

对于制造公司来说，以最小的时间和成本投资制造所需组件特性的方式设计组件生产是高度相关的。到目前为止，制造组件的各个过程通常是彼此分开考虑的，并且只在更高的计划级别上合并。然而，工艺序列的跨技术考虑提供了很高的优化潜力，因为单个工艺和工艺参数因此以面向目标的方式进行协调。同时，越来越短的产品开发周期意味着必须在短时间内设计工艺序列。因此，需要基本的模型和方法来支持快速确定经济优化的工艺参数。到目前为止，还没有开发出一种方法来快速确定工艺序列的经济优化工艺参数，同时考虑到单个制造工艺和部件特性之间的依赖关系。为了缩小这一研究差距，将在拟议的研究项目中开发一种方法，以跨技术元模型为基础，确定工艺序列的经济优化工艺参数。元模型映射单个制造过程的输入-输出相关性，专门针对应用程序选择最重要的参数，从而在短时间内实现对过程序列设计的相关见解。因此，提出的研究项目的一个重点是过程序列的元建模。单个制造过程的元模型通过先前确定的单个过程之间的传递变量与跨技术元模型相关联。随后，研究选取最重要的参数所产生的参数约简对整个模型预测精度的影响，从而进行参数的优化选择。研究项目的第二个重点是跨技术元模型的经济优化。一方面，将开发一种方法，使元成本模型能够生成，以确定单个工艺参数对工艺序列成本的影响。另一方面，将开发一种基于遗传算法的新型优化算法，该算法能够基于跨技术元模型和所需部件特性确定经济优化的工艺参数。然后，开发的方法将在软件中实施，并使用两个案例研究进行验证。

For manufacturing companies, it is of high relevance to design the production of components in such a way that the required component characteristics are manufactured with minimized time and cost investments. Until now, the individual processes for manufacturing a component were often considered separately from each other and only merged at higher planning levels. However, the cross-technology consideration of process sequences offers a high optimization potential, because the individual processes and process parameters are thus coordinated in a target-oriented manner. At the same time, increasingly shorter product development cycles mean that process sequences have to be designed in a short time. Therefore, fundamental models and methods are needed that support the rapid determination of economically optimized process parameters for process sequences.So far, no approach has been developed to quickly determine economically optimized process parameters for process sequences, taking into account the dependencies between the individual manufacturing processes and component characteristics. In order to close this research gap, a methodology will be developed in the proposed research project to determine economically optimized process parameters for process sequences on the basis of cross-technology metamodels. Metamodels map input-output correlations for individual manufacturing processes, exclusively resorting to a selection of the most important parameters for the application and thus enabling relevant insights into the design of process sequences within a short period of time. One focus of the proposed research project is therefore the metamodeling of process sequences. Metamodels of individual manufacturing processes are linked to cross-technology metamodels via previously identified transfer variables between the individual processes. Subsequently, the effects of the parameter reductions resulting from the selection of the most important parameters on the prediction accuracy of the overall model will be investigated so that an optimized parameter selection can be made.The second focus of the research project is the economic optimization of cross-technology metamodels. On the one hand, a method will be developed that enables meta cost models to be generated in order to determine the effects of individual process parameters on the costs of the process sequence. On the other hand, a novel optimization algorithm based on genetic algorithms will be developed, which enables the determination of economically optimized process parameters based on cross-technology metamodels and required component characteristics. The developed methodology will then be implemented in software and validated using two case studies.