Contact

Dr. Eberhard Altstadt
Head Structural Materials
e.altstadtAthzdr.de
Phone: +49 351 260 - 2276
Fax: 12276, 2205

Mario Houska
Fracture Mechanic testing of Materials
m.houskaAthzdr.de
Phone: +49 351 260 - 2242

Determination of deformation and failure properties of radiated ductile materials by means of the small punch test and neural networks


Problem Outline

  • The ductile material behavior in structural components is changing due to
    • In service loading
    • Embrittlement by radiation (reactor pressure vessels)
    • Other environmental influences
  • In-situ monitoring of the material state is required
  • The material is assumed to be irradiated

Methods for Solving the Problem

  • The experimental method of the small punch test (SPT) is used to obtain the material response to external loading.
    • The small disk like punch specimen (Ø8 × 0.5 mm) allow a minimal invasive specimen sampling from larger structures, rather than typical tensile or fracture specimen.
    • The compact testing device can be put into a hot chamber, which is required for irradiated material.
    • The experimental result of the SPT is a load displacement curve, which contains information about the material strength and fracture thoughness
  • A finite element model of the small punch test is used predict the load displacement curve for different material parameters
    • The material behavior is modeled with a continuum damage model (GTN-model)
    • With a systematical parameter variation and FE-computation of the SPT a data base is build up
  • The computed data base is used to train a neural network
    • The task for the neural network is to determine the run of the load displacement curve depending on the material parameters of the GTN-model
    • The trained neural network now can be used as an approximation of the finite element computation
Strategie zur Identifikation von Materialparametern aus dem Small Punch Test
Abb.: Strategy for the identification of material parameters using the small punch test, a neural network and an optimization procedure.
Download
  • A identification routine is used to determine the material parameters
    • Starting with a arbitrary parameter set, the neural network computes a load displacement curve, which is compared to an experimental one
    • Since both curves does not match each other an error is computed
    • The identification routine tries to minimize this error by varying the material parameters
    • If the error becomes a minimum a material parameter set if found, which describes the material behavior properly
  • Using this parameter set, finite element computation of a tensile test is done, which delivers the true stress strain curve and the tensile strength of the material
  • Furthermore, computation of fracture tests can be performed to find values for the fracture thoughness of the material

Ductile Damage of Metallic Materials

  • Very pure metallic metallic material don't have any inclusions or voids. Then, the macroscopic failure of e.g. a tensile specimen is due to 100% diameter reduction.
  • Technical materials (steel) contain voids, inclusions and second phase particles, which are mostly wanted to obtain required material properties.
  • The ductile failure of such materials is a process, where
    • void nucleation
    • void growth
    • void coalescence
    due to mechanical deformation occur.
  •  
Bruchfläche einer GGG-40 SPT-Probe (100x). GGG-40 enthält kugelförmige Graphiteinschlüsse, die während der Verformung wie Hohlräume reagieren. Die Wabenstruktur ist typisch für duktile Bruchflächen.
Abb.: Bruchfläche einer GGG-40 SPT-Probe (100x). GGG-40 enthält kugelförmige Graphiteinschlüsse, die während der Verformung wie Hohlräume reagieren. Die Wabenstruktur ist typisch für duktile Bruchflächen.
Download
Bruchfläche einer 18Ch2MFA SPT-Probe (2000x). 18Ch2MFA enthält nichtmetallische Einschlüsse, die sehr viel kleiner als die kugelförmigen Graphiteinschlüsse im GGG-40 sind. Auch hier kann man die für duktilen Bruch typischen Wabenstrukturen erkennen.
Abb.: Bruchfläche einer 18Ch2MFA SPT-Probe (2000x). 18Ch2MFA enthält nichtmetallische Einschlüsse, die sehr viel kleiner als die kugelförmigen Graphiteinschlüsse im GGG-40 sind. Auch hier kann man die für duktilen Bruch typischen Wabenstrukturen erkennen.
Download

The Small Punch Test

In the small punch test a disk like specimen is deformed in a miniaturized deep drawing experiment. The specimen is clamped between a die (bore diameter d=4 mm, die edge radius r=0.5 mm) and a downholder and centrically deformed by the punch with a spherical head (radius R=1.25 mm). For irradiated material the device is put into a hot chamber.

Prinzip des Small Punch Test
Abb.: Prinzip des Small Punch Test
Download
Kraft-Verschiebungs-Kurven aus dem SPT für verschiedene Werkstoffe
Kraft-Verschiebungs-Kurven aus dem SPT für verschiedene Werkstoffe
Download

References

Linse, T.; Kuna, M.; Schuhknecht, J.; Viehrig, H.W.
Usage of the small-punch-test for the characterisation of reactor vessel steels in the brittle–ductile transition region
Engineering Fracture Mechanics 75 (2008) 3520–3533

Abendroth, M.; Kuna, M.
Identification of ductile damage and fracture parameters from the small punch test using neural networks
Engineering Fracture Mechanics 73(2006)6, 710-725

Abendroth, M.
Identifikation elastoplastischer und schädigungsmechanischer Materialparameter aus dem Small Punch Test
Dissertation, Freiberger Forschungshefte, Technische Universität Bergakademie Freiberg, 2005

Kuna, M.; Abendroth, M.
Identification and validation of ductile damage parameters by the small punch test
In Procceedings of the 15th European Conference of Fracture (ECF 15), 11.-13.08.2004, Stockholm, Sweden

Kuna, M.; Abendroth, M.
Prüfung duktiler Werkstoffe mit dem Small Punch Test
Materialprüfung 47(2005)1-2, 45-54

Abendroth, M.; Kuna, M.
Determination of ductile material properties by means of the small punch test and neural networks networks
Advanced Engineering Materials, Vol. 6(2004)7, 536-540


Contact

Dr. Eberhard Altstadt
Head Structural Materials
e.altstadtAthzdr.de
Phone: +49 351 260 - 2276
Fax: 12276, 2205

Mario Houska
Fracture Mechanic testing of Materials
m.houskaAthzdr.de
Phone: +49 351 260 - 2242