P 1248 – Qualification of the instrumented indentation technique for the parameter identification of advanced high strength steels
The determination of mechanical properties of welded steel structures such as strength or ductility is a subject of high interest for the majority of companies in the area of metal processing. The material parameters can be obtained by performing the tensile test on the samples made from a part of a component. In some cases, it is highly expensive to produce the tensile specimens especially from the weld metal, which contains different type of microstructure such as weld seam or heat affected zone in an extremely small area.
Therefore, a novel method is described in this report to determine the material parameters of high strength steel structures and welded joints locally and without any additional effort to perform the tensile test. In this method, instrumented indentation technique (IIT), an indenter is pushed on the flat surface of a specimen in a certain period of time and simultaneously the applied force and the corresponding indentation path are measured. The data related to the force-indentation diagram is given as input to an artificial neural network (ANN) to obtain the material parameters. The ANN can be trained by generating the large qualitative data sets with numerical simulation of the IIT procedure. The simulation must be run several times with the different material model parameter sets to generate the numerous force-indentation diagrams as the inputs of ANN. Then, the trained ANN is validated by performing the IIT on the welded joints and comparing the obtained material parameters from ANN with the tensile test. Consequently, the mechanical properties of welded joints can be determined by performing the IIT and evaluating the resulting data by the ANN.
Prof. Dr.-Ing. M. Rethmeier