Fraunhofer Institute for Mechanics of Materials IWM
Fraunhofer Institute for Mechanics of Materials IWM

Digital methods for lifetime assessment

© Fraunhofer IWM
Figure 1: Data set on fatigue behavior of high-strength steels, which covers a variety of material classes and influencing parameters (approx. 110 materials, 1100 fatigue test series, 22000 fatigue tests)

At Fraunhofer IWM, digital methods are developed which focus on a systematical use of material data for accelerated product development. Material data stored in structured data spaces and semantic knowledge graphs enable an interoperable connection to analyzation tools that cover the specific steps along the product development chain. In the frame of the internal Fraunhofer research project “UrWerk” such a data space has been developed for the “Fatigue assessment of high-strength steels” use case.

The life expectancy of components is influenced by a variety of parameters, including the metallurgical processing route, considerations concerning different hardness measurement methods, loading parameters in service and finally fracture surface characteristics under differing failure modes. Therefore, the change of parameters such as hardness and surface roughness as well as many others along each step of the process chain were tracked in a specific knowledge graph.

© Fraunhofer IWM
Figure 2: Snapshot of an ontology-based knowledge graph describing the process-microstructure-properties relationships with respect to the use case “Fatigue assessment of high-strength steels”.

To visualize correlations for technical applications as well as to be able to implement life expectancy predictions, data analyses by means of machine learning are coupled to the data space. In the project, a considerable data set (approx. 110 materials, 1100 fatigue test series, 22000 fatigue tests) was compiled, covering a large number of material classes and influencing parameters. This enabled a prognosis of the fatigue life related to these parameters using artificial intelligence methods (figure 1). Tailored ontology-based knowledge graphs support the data scientist to not only collect the data in tabular formats but to implement the data along the process chain. This increases the quality of the fatigue life expectancy prognosis for components as microstructure properties relationships can be taken into account.

to top