guest ::
TY - THES AU - Jäpel, Tom TI - Feasibility study on local elastic strain measurements with an ebsd pattern cross correlation method in elastic-plastically deforming materials : applications on high manganese TWIP steels PB - Aachen, Techn. Hochsch. VL - Dissertation CY - Aachen M1 - RWTH-2015-03440 SP - XII, 160 S. : Ill., graph. Darst. PY - 2015 N1 - Aachen, Techn. Hochsch., Diss., 2015 AB - This work explores the applicability of a relative local elastic strain imagingmethod in the SEM: HR-EBSD. The method is based on a Kikuchipattern post proceßing. By comparing sub-pixel shifts in a referenceKikuchi pattern to test patterns the elastic strain and rotation can becalculated. In the first part of the work, dynamically pattern simulationswith well-known geometry are used in order to reveal potentialweak points of the method. Parameters under observation are the precisionof the pattern shift measurement by croß-correlation, the differencebetween measured- and imposed strains, region of interest (ROI) sizeand position, ROI filter size, simulated pattern reflector number, patternnoise and binning size, pattern source point coordinate system accuracyand reference- to test pattern misorientation. In the simulations, the misorientationangle was the most significant influence factor. HR-EBSDshows an increasing error in streß measurement of up to 100 MPa up toabout 3 and can fail ultimately beyond that limit. The misorientationinduced error can be reduced by more than two order of magnitudesby limiting the croß-correlation to areas of small misorientation only.The second largest error influence is the pattern center accuracy. Witha careful pattern source point coordinate determination, errors of thatcomponent can be reduced to one-third in measured strain error. Inorder to validate HR-EBSD in the experimental case as well electronicgrade silicon material was measured for strain gradients. It is of utmostimportance to use a correct source point correction, when usinglarge scans (</td><td width="150"> AB - gt; 1 micron2 area) because measured elastic strain errors ofup to about 6 10-2 can occur (about 980 MPa elastic streß error). Forsmall scans (</td><td width="150"> AB - lt; 1 micron2 area) the poßible error is much smaller, butalso given, with about 2.4 10-4 in elastic strain measurement error. Ifthe scan size is small and the source point correction is applied properlyerrors in elastic streß measurement of smaller than about 36 MPa canbe expected. Two experiments were conducted on dominantly plastic deforming TWIP steels. The first experiment, a three point bending experiment,showed robust results of the HR-EBSD method for samplesregions of elastic deformation. All bending steps showed the expectedstrain component on the tension and compreßion side on the sample.The elastically and plastically deformed region could be distinguishedby occurring dislocation slip lines by the ECCI technique. By increasingthe three point bending displacement, the elastic zone was shrinking asexpected. For the first time the HR-EBSD- and ECCI technique wereapplied synergetically. The measured elastic strain component profilesof all bending steps showed the expected linearity in the elastic regionand a non-linearity in the plastically deformed regions. The second experimentcaptured the moment of twin formation in a polycrystallineTWIP steel by using standard Hough-based EBSD, ECCI and HR-EBSDin a tensile tested material. The HR-EBSD measurement supports thetheoretically calculated critical strain neceßary prior twinning of about2.5 10-3 by a measured value of about 2.6 10-3 at a grain boundarytriple point. The extending twin flank in the twinned grain shows anaverage elevation in elastic strain of about 1.25 10-3. LB - PUB:(DE-HGF)11 DO - DOI:10.18154/RWTH-2015-03440 UR - https://publications.rwth-aachen.de/record/480096 ER -