Fuzzy-arithmetical modeling of processes with uncertain prarameters

点击翻译按钮获取中文摘要

The goal of the subproject remains the development of a robust process modeling using metamodels for improved and efficient prediction of the effects of process fluctuations on function-critical measures. For the second phase the focus will be on the connection between metamodel and process simulation. A further approach of the simulations to real processes requires more detailed models and thus higher computing times. This makes it more difficult to evaluate enough simulations for the metamodel. It is therefore important to optimize simulations in such a way that the relationship between the input parameters and the result variables can be mapped as well as possible using the metamodel. Since both output and input variables play a decisive role, they should be coordinated with the respective TPs. From a manufacturing point of view, model assumptions such as material model or contact model as input variables for the simulation and result variables such as strength and force distribution are interesting. Process settings such as temperature or pressure as well as the geometry of the component are more important for the tolerance assignment, since these variables can be tolerated. However, these conflicts of interest are always related to each other and are represented by the simulation. The investigation of these input and output variables or their consideration with uncertainties and their optimization by means of fuzzy arithmetic is thus an important component of process-oriented tolerance management in order to optimize the interface between process simulation and tolerance simulation.A new goal of the overall project in the second phase is to take the wear behavior during operation into account during component tolerance. For the TP, the new goal is to create a virtual level of this situation. The modelling of the wear behaviour during operation leads to an additional temporal component in the modelling. In addition to the measurement uncertainty and the surface structure, operating parameters and wear over time have to be taken into account in the metamodel.The time-dependent components represent a challenge. The focus of wear is undoubtedly the gear pairing. The Archard model will be used to make a virtual prediction. Real wear tests will be compared with calculations and characterized taking into account measurement uncertainties, surface topography and operating parameters. The interaction of the components, their prediction and their effects will be analysed exemplarily at the demonstrator from the first phase.