JOURNAL OF THEORETICAL AND APPLIED MECHANICS 44, 3, pp. 649-666, Warsaw 2006 NONDESTRUCTIVE DAMAGE CHARACTERIZATION WITH EXAMPLES OF THERMAL AGING, NEUTRON DEGRADATION AND FATIGUE Gerd Dobmann Iris Altpeter Klaus Szielasko Markus Kopp Institut für zerstörungsfreie Prüfverfahren IZFP Fraunhofer-Gesellschaft University, Saarbrücken, Germany; e-mail: gerd.dobmann@izfp.fraunhofer.de Nondestructive Testing (NDT) in the engineering community is normally associated with the objective to detect, to classify and to sizematerial non- conformities – for instance beginningwith nonmetallic inclusions of a size of some ten µm in steel or Aluminum alloys up to so-called ’material defects’ like macroscopic cracks of somemm size. This objective, however, is at the top of the list of activities concerning the number of applications in non- destructive material testing worldwide. Methodologies like UT (Ultrasonic Testing) andRT (Radiographic Testing) orMT (Magnetic Testing) are well introduced in awide field of product and component examination standards. In the last 15 to 20 years, the NDT technology was also developed for cha- racterizingmaterials, for instance in terms ofmicrostructure parameters, i.e. lattice defects, like distributions and densities of dislocations, precipitates, micro-voids, in order to describe strengthening and/or softening in mate- rials, mainly in metal alloys, but also to measure the applied and residual stresses (Dobmann et al., 1989). Key words: NDT, damage characterization,micromagnetic techniques 1. Introduction Starting in 1980, thefirst empirical approacheswere developed, to characterize materials in termsofmechanically and technologically definedparameters such like hardness, hardness depth, yield limit and ultimate strength (Dobmann et al., 1998), all of them standardized by a destructive procedure. In Borsutzki 650 G. Dobmann et al. (1998) itwas published that these parameters could bedetermined also on-line in the steel making process, i.e. in a hot-dip galvanizing line for car body steel sheet production. In this context, micromagnetic testing is especially suitable because mi- cromagnetic parameters obtained undermagnetic load showmany similarities to mechanical parameters derived under mechanical loads. This is why mi- crostructure parameters (lattice defects) impede dislocationmovement as well as Bloch wall movement and relationships between mechanical andmicroma- gnetic parameters are empirically derived by multiple parameter correlation, neural network and pattern recognition procedures (Altpeter et al., 2002). First publications to characterize thematerials damage in terms of micro- scopic damage parameters such as creep damage porosity and fatigue effects were published in Dobmann et al. (1993), Dobmann and Seibold (1992). In themeantime further research work was initiated (Dobmann, 2002; Dobmann et al., 2001a), some characteristic results of which are summarized in the fol- lowing contribution. They mainly document the maturity of micromagnetic properties to solve this inspection task. Metallurgists are familiar with the phenomenon of precipitation harde- ning which is usually associated with the elements carbon and nitrogen, the solubility of which is high in solid media, and which precipitate under heat treatment or specific service conditions as carbides or carbo-nitrides. Their contribution to the overall strength resides in the range of 10 to 12%, whe- re the strengthening effect occurs due to an impeded dislocation movement undermechanical loading. Mikheev andGorkunov (1979) have published abi- lities of micromagnetic parameters to characterize the strengthening effect of precipitation hardening. In iron alloys, copper behaves in a similar way, so that in combination with neutron irradiation, even small contents of Cu can act as a driving force for embrittlement. These conditions, for instance, are present in our western nuclear pressure vessel steels where the copper content ranges from 0.01 to 0.07wt%. In case of the German martensitic/bainitic structural steel WB 36 (15NiCuMoNb5, 1.6368) which is in service in fossil power plants but also in pipes and pressurizers of German nuclear power plants, the effects of preci- pitation hardening are much larger. WB 36 has an average composition with C– 0.15, Ni – 1.15, Cu – 0.65,Mo – 0.35 andNb– 0.025 (all amounts inwt%). The 2nd chapter of this contribution will present the results obtained at different heats ofWB 36 steel andwill introduce it in the applied 3MA-NDT- technology. Nondestructive damage characterization with examples... 651 In the 3rd chapter, the results obtained in GRETE (1999) and documen- ting the sensitivity of NDT-technology to characterize neutron degradation in pressure vessel materials are presented. However, in Germany the radiation conditions concerning nuclear pressure vessels are different compared with other countries and therefore the micromagnetic technology has to be valida- ted in an ongoing project. The 4th chapter of the contribution is devoted to the characterization of fatigue damage. In the low cycle as well as in the high cycle regime, austenitic and ferritic steel specimens were fatigued. A new NDT technique based on measurement of the transfer-impedance of an eddy current yoke transducer in combination with a Giant Magnetic Resistor (GMR) gradiometer (Grünberg, 1995) was applied to on-line monitoring of the fatiguing process. 2. ND-characterization of thermal aging of WB 36 steel 2.1. The material In the typical ”as delivered” state of WB 36, half of the contained Cu is already precipitated, while the other half remains in solid solution. After long term service exposure above 320◦C, damage was observed due to further pre- cipitation of Cu; an increase in yield strength ∆σy = ±150MPa and a shift of the fracture appearance transition temperature ∆FATT = +70 ◦C can be measured. Small angle neutron scattering revealed the fact that the mechani- cal property changes were caused byCu precipitates ranging from 1 to 1.5nm in size. The particles are coherent from the bcc structure and therefore arose a high level of compressive residual stress in their vicinity, balanced by ten- sile stresses in the environmental matrix. The precipitation hardening can be characterized by Vickers hardness measurements in the laboratory; the effect is in the range of 40HV 10 units. At a component in service, however, hard- ness measurements cannot be applied in an area-wide manner, which creates a demand for the development of a suitable NDT technique. Magnetic and, in particular, micro-magnetic techniques are suitable for the characterization of mechanical property changes. This relates to the fact that microstructure and lattice defects which impede the movement of dislo- cations and therefore contribute to themechanical strengthening, impede the movement of Bloch walls inmagnetizable materials in a similar way. This can be documented in micro-magnetic properties such as coercivity, Barkhausen noise, non-linearity, etc. Fraunhofer IZFP applies the so-called 3MA approach 652 G. Dobmann et al. (Altpeter et al., 2002) which combines several independent micro-magnetic properties in a regression or similar data fusionmodel, in order to predict the mechanical property changes at a component. This contribution discusses the results obtained with the 3MA approach in case of the steel WB 36. On a set of approximately 70 round samples (80mm in length, diameter 6mm) of the steel gradeWB36, thermal service exposurewas simulated in an acceleratedmanner through long-termannealing at 400◦C.The two examined heats E2 andE59 represent admissible variants ofWB36, whereE2 originates from a hot-rolled plate and E59 from a pressure vessel drum. The samples of the heat E59 were in service-exposed condition (57000h at 350◦C) and had to be recovery-annealed (3h at 600◦C) prior to service si- mulation. Half of the recovery-annealed samples underwent a stabilizing heat treatment at theMaterials Testing Institute (MPA) at theUniversity of Stut- tgart which also performed the subsequent service simulation and parts of the materials characterization. Although the samples of the heat E2 were not service-exposed in the be- ginning, 13 samples were recovery-annealed (3h at 600◦C) in order to relieve possible surface stresses and characterize their influence on the measured re- sults. In addition, 8 samples were plastically deformed to specific amounts (5%, 11% in longitudinal direction) for the simulation of practical disturban- ces. The service simulation andmechanical characterization of theE2material took place at IZFP. All heat-treatments were done using an electric furnace with internal air circulation. Samples of each initial condition were removed from the furnace piece bypiece in short time intervals in order to obtain a fixed set of differently aged material. All electro-magnetic tests were performed on this sample set. Additionally, theVickers hardness (HV5)was recordedwith a 95%confidence interval of less than ±5HV 5 by using a Krautkrämer TIV hardness tester. Fig.1 shows the obtained Vickers hardness values and the residual electrical resistance quotient (G – a measure for the amount of precipitated copper, obtained by MPA) for the heat E2. The maximal increase in hardness versus the initial material state was observed to be around 25HV 5 in any case, and it was reached after 600 to 1000 hours of service simulation. Due to a higher dislocation density, plastically deformed samples exhibit higher initial hardness and quicker ageing. Fig.2 shows the corresponding data for the heat E59, which has a higher initial hardness and slower ageing as compared to the heat E2. Themaximum hardness of E59 will probably be reached beyond 15000h of service simulation. Nondestructive damage characterization with examples... 653 Fig. 1. Vickers hardness (HV 5) and quotient of residual electrical resistance (G) as a function of service simulation time for the heat E2. Initially, the subsets A1 and A2 were as delivered, EGwas recovery-annealed, and PV5/PV11were plastically deformed by 5% and 11%, respectively Fig. 2. Vickers hardness (HV 5) and coefficient of residual electrical resistance (G) as a function of the service simulation time for the heat E59. The subset EGwas recovery-annealed from the service-exposed state B. The heat-treatment stabilized subset is denoted by SG 654 G. Dobmann et al. 2.2. The experiments AU-shaped electromagnet was used to excite an alternatingmagnetic field along the longitudinal axis of the sample. A disc-shaped pickup coil and a temperature-stabilized hall probewere used to recordBarkhausennoise events and magnetic field strength, respectively. The Barkhausen noise signal was amplified by 60dB and bandpass-filtered to a range of 5-200kHz. All signals were digitized using common data acquisition hardware. The Barkhausen no- ise signal was then digitally re-filtered for separate analysis of its different frequency components. Characteristic scalar quantities (Altpeter, 1990) we- re extracted from the envelope of the Barkhausen noise signal as a func- tion of the applied magnetic field strength. Moreover, an upper harmonics analysis (Pitsch, 1989) of the magnetic field strength signal was performed and characteristic quantities were derived. As changes in conductivity may be expected due to copper precipitation, a simplified eddy current analysis procedure was performed based on the relationship between magnetic field strength and exciting voltage of the electromagnetic coil. The scalar result quantities of all threemethods (Barkhausen noise, upper harmonics and eddy current analysis) are combined to a vector which characterizes the material condition. The total set of samples was split up into a large training set and a small test set. Using the above-mentioned electromagnetic methods, calibration da- ta were recorded for the training set and assigned the corresponding Vickers hardness values. For the replication of practical circumstances, an additional tensile load of 5 to 100MPa was applied to the sample in its longitudinal di- rection (a minimum load was required in order to prevent vibrations due to the alternating magnetic field). The amount of tensile load was varied both during calibration and the test, but remained unknown to the system, so as to constitute a realistic disturbing factor. The 3MA approach which is pur- sued at IZFP solves the inverse problem of target quantity prediction from a limited set of calibration data (Dobmann and Höller, 1990). In this case, a specialized pattern recognition algorithm (Tschuncky, 2004) based on a ne- arest neighbour search was used to obtain approximate values of the Vickers hardness during the repeatedmeasurement onboth the calibration set and the test set. Fig. 3 shows the obtained results in case of no external loading. TheRMS error (RMSE, residual standard deviation) of the predicted values is around 2HV 5, corresponding to a 95% confidence interval of ±4HV 5. The pre- diction error within the calibration set falls short of the reference confiden- ce (±5HV 5), which shows that the samples differ significantly from each Nondestructive damage characterization with examples... 655 other in terms of their electromagnetic properties. This enables the accompli- shment of additional disturbances like-superimposed tensile loads, as shown in Fig.4. The RMS error drops to around 4HV 5 in this case (equivalent to a 95% confidence interval of ±8HV 5), which is still within the expec- ted frame of accuracy, considering the disturbing influence of varying tensile load. Fig. 3. Predicted Vickers hardness forWB 36, heat E2, subsets A1, A2, EG, PV5 and PV11 versus actual Vickers hardness. Averaged statistics:R2 ≈ 0.98, RMSE≈ 2HV 5 Fig. 4. Same as above, but with superimposed tensile load from 5 to 100MPa during both calibration and test. Averaged statistics:R2 ≈ 0.91, RMSE≈ 4HV 5 In addition to the Vickers hardness, the 3MA approach predicted the qu- otient of residual electrical resistance (G) or the equivalent service simulation time with similar accuracy. These findings were also confirmed in case of the heat E59. However, as there has not been much change in hardness to da- te, an in-depth evaluation must follow as soon as the corresponding service- simulation is finished. 656 G. Dobmann et al. 3. ND-characterization of neutron degradation Depending on the specific design – which is different in different countries of the world – the pressure vessel material in nuclear power plants is exposed to a neutronfluxwhich is in the range between 5·1018n/cm2 (German design) in 32 years at 288◦C, and 8 ·1019n/cm2 at 254◦C in 14 years (French experien- ce). The energy input of the neutrons is directly producing lattice defects like vacancies and indirectly by stimulating the precipitation of Cu-rich precipita- tes. These are in the 3nmdiameter range and coherent in the bcc lattice, and can be detected by small angle neutron scattering. Both the vacancies and the precipitates reduce the toughness of the material, which can be characterized by a reduction of the Charpy energy and a shift in the fracture appearance transition temperature to higher temperatures. In practice the material degradation is characterized in surveillance pro- grammes by using a standardized Charpy V-notch specimen and the tensile test specimen made of the pressure vessel material and its weldments. The specimens are exposed in special radiation chambers near the NPP core at a higher fluence than that at the surface of the pressure vessel wall. These specimens from time to time are removed from the chambers and used for destructive tests in order to document the state of microstructure change. Under the headline of lifetime extension of theNPPs, longer than the∼ 30 years initially prospected, therewill be not enoughmaterial available and the- refore nondestructive tests should replace the destructive ones. Furthermore, to assure and to document a higher nuclear safety between two subsequent destructive tests, one would like to have many nondestructive tests and the ND-technology should also be developed to an in-service inspection method to be applied at the pressure vessel inner surface. That was the reason to investigate, in an inspection trial of the Euro- pean research programme EURATOM the ability of different ND-techniques to characterize the material degradation due to neutron irradiation. Different materials from surveillance programmes of different European countries we- re investigated by different teams in the hot cell of the research power plant in Petten, Netherlands. Table 1 gives an overview about the materials and the exposure conditions. There were two sets of specimens of France of the power plants Chinon and Dampierre and two sets coming fromGermany (re- search centre Rossendorf), a reference heat JRQ according to the Japanese standards and an optimized heat concerning the Cu-content, called JFL. In addition, SKODA from the Czech Republic has supplied with a set of speci- menaccording to theRussiandesign.Table 2 is documenting the experimental Nondestructive damage characterization with examples... 657 data. Whereas for the first four sets of specimens the transition temperature T09 is documented, the SKODA specimen are characterized by the transition temperature of the Charpy energy at 41joule. Table 1. Irradiation conditions and destructively determined mechanical properties Cu- content [%] Fluence in Irradiation Transition Supplier Heat 1MeV temp. temp. T09 1019n/cm2 [◦C] [◦C] France Chinon B1 0.07 0 −32.0 1.74 300 −10.0 3.03 300 7.0 4.65 300 13.0 6.64 300 21.0 France Dampierre M3 0.044 0 −26.0 1.74 300 −10.0 3.74 300 0 5.36 300 20.0 7.56 300 31.0 Germany JRQ 0.14 0 254 −8.6 0.72 254 103.4 5.61 254 171.9 9.58 254 240.4 Germany JFL 0.01 0 −45.9 0.62 254 −21.7 4.41 254 15.8 Table 2.Heats according to the Russian design Supplier Fluence Irradiation Transition SKODA in 0.5MeV temperature temperature WWER mat. 1019n/cm2 [◦C] T41 [◦C] 3 specimen 0 fresh −90 3 specimen 2.9 288 −72 3 specimen 9.7 288 −32 TheGermanheat JRQis theonewith thehighestCu-content andtherefore one can observe the greatest shift in the transition temperature. Generally the transition temperatures were calculated according to a th-curve fit from the Charpy energy data. 658 G. Dobmann et al. IZFP has applied 3MA-approaches (Altpeter et al., 2002) to calibrate re- gressionmodels. One part of each specimen set was used to calibrate and the other part – independently selected – was taken to test themodel. Because of the fact that all of the specimenswere only available as half-Charpy-specimens after performing the destructive test side-effects of residual stresses andplastic deformation were observed. In Figure 5 the results of the regression calibration and the tests are docu- mented according the specimens of Table 1. The correlation coefficient is 97.6 – very close to 100%and the residual standarddeviation is 15.6◦C(5.4%).Ho- wever, by testing the calibration with the independent test specimens, amuch larger standard error is obtained (54.1◦C, 18.9%) The reason for this discre- pancy was discussed before and it is the influence of the plastic deformation in these specimens. Fig. 5. The prediction of the T09-transition temperature by 3MA (correlation coefficient for calibration JFL, JRQ, Chinon, Dampierre together, cc=99.7, standard error is equal 15.6 (5.4%) for calibration samples and 54.1 (18.9%) for test samples) The same situation can be observed in Figure 6 when discussing the Rus- sian heat delivered by SKODA. In that case the regression calibration can be performed and an excellent correlation coefficient of 99.7 is obtained with a standard error of 1.8◦C (3.1%). Using the independently selected test speci- mens, however, the standard error of estimate is with 14.5◦C (24.9%) much larger. In order to overcome the side effects of the plastic deformation in further investigations, the 3MA-approach has to be calibrated at unbroken Charpy specimens. Nondestructive damage characterization with examples... 659 Fig. 6. The prediction of the T41-transition temperature at the Russian heat by 3MA (correlation coefficient cc=99.7, standard error is equal 1.8 (3.1%) for calibration samples and 14.5 (24.9%) for test samples) 4. ND-characterization of fatigue Austenitic stainless steels are inwidespread application in the chemical aswell as nuclear industry, mainly because of their high toughness and insensitivity against corrosion attack. However, under static aswell as fatigue load, thema- terial has the tendency to response with localized phase transformations from the non-magnetic γ- to themartensitic and ferromagnetic α′-phase. The pro- cess starts localized at positions of higher stress intensity, i.e. atmicrostructure inhomogeneities like non-metallic inclusions and carbo-nitride precipitates. The amount of martensite as well as its magnetic properties should pro- vide information about the fatigue damage. Fatigue experiments were car- ried out at different stress and strain levels at Room Temperature (RT) and at T =300◦C. The characterization methods included microscopic techniqu- es such as light microscopy, REM, TEM and Scanning Acoustic Microscopy (SAM) as well as magnetic methods, ultrasonic absorption, X-ray and neu- tron diffraction (Bassler, 1999; Lang, 2000). As the martensitic volume frac- tions are especially low, for in-service temperatures of about 300◦C highly sensitive measuring systems are necessary. Besides systems on the basis of HTC-SQUID (Kittel, 1971) (High Temperature Super Conducting Quantum Interference Devices), special emphasis was on the use of GMR-sensors [19] (Giant Magneto Resistors), which have the strong advantage to be sensitive also for DC-magnetic fieldswithout any need for cooling. In combination with an eddy-current transmitting coil and an universal eddy-current equipment, 660 G. Dobmann et al. as a receiver the GMR-sensors were used especially to on-line monitoring of the fatigue experiments in the servo- hydraulic fatigue machine. Pure ferritic steels do not showphase transformation as a function of cyclic or quasi-static loading. Therefore the strong magnetic effect observed in the austenitic material bymartensite formation is not detected. The plain carbon steel C15 (nominal 0.15% carbon) is a characteristic example for this case. The cyclic loading here only enhances the dislocation density (some orders of magnitude) and is influencing the development of characteristic dislocation networks and cell structures. The cyclic deformation behavior of this group of steels is well understood (Dobmann et al., 2001b). In stress-controlled fatigue tests the measured plastic strain amplitude is a sensitive quantity to measure the changes in thedislocation network andcell structure,whichare the reasons for the cyclic softening and cyclic hardening. 4.1. Materials Table 3documents the chemical compositionof the steel 1.4541 andTable 4 for the plain carbon steel C15. In the case of the austenitic stainless steel two heats were investigated in order to see also the effect of different charges, i.e. manufacturer influences of nominally the same quality. Table 3.Chemical composition of the stainless steel 1.4541 in mass [%] Elements C N Si Mn P S Cr Mo Ni Ti Heat 1 0.05 0.002 0.4 1.09 0.024 0.005 17.81 0.27 9.3 0.3 Heat 2 0.03 0.006 0.45 1.72 0.022 0.014 17.31 0.28 10.18 0.16 Table 4.Chemical composition of the plain carbon steel C15 inmass [%] Elements C Si Mn P S Cr Ni Mo Cu N Al 0.15 0.189 0.43 0.0134 0.025 0.132 0.037 0.011 0.014 0.007 0.037 In any case the experimentswere performed at specially designed (Bassler, 1999), ”hourglass”-shaped fatigue specimens as shown in Figure 7. 4.2. NDT-techniques applied Two types of magnetic sensors were utilized for characterizing the fatigue behavior, different SQUID-magnetometer (Kittel, 1971) and GMR [19]. For on-line measuring at the servo-hydraulic machine only the GMR sensors were Nondestructive damage characterization with examples... 661 Fig. 7. Hourglass-shaped fatigue specimen used in the experiments used. Figure 8 gives a view on the sensor together with the clip gage and the fatigue specimen in the machine. The GMR was controlled by standard eddy-current equipment and can be applied as magnetometer or canmeasure a transfer-impedance between a small electromagnetic yoke as transmitter coil and theGMRas a receiver. Themeasurement technique is described in detail in Lang (2000), Eifler andMacherauch (1990). Fig. 8. GMR-sensor for on-line fatigue monitoring 4.3. Fatigue tests and on-line monitoring results In the case of the steel 1.4541 the GMR-sensor was applied as magneto- meter and as an eddy-current receiver. Because of the increasing content of the ferromagneticmartensitic phase in the austeniticmatrix, a continuous en- hancement of the magnetic field strength is observed. Performing the on-line monitoring experiment some more effects can be documented. Figure 9 pre- sents the results for a strain-controlled LCF fatigue test (εa,p =0.2%, Rε =1, room temperature). 662 G. Dobmann et al. Fig. 9. Stress versusmagnetic field hysteresis curves In this case achangeof the stressamplitudeat the specimenwith increasing load cycles is observed, i.e. a cyclic strengthening. Similar to a cyclic hysteresis curve, cyclic stress versus magnetic field hysteresis curves can be measured. For different distinct load cycles these curves are presented. With increasing load cycle number, the area of the curves is increasing as well as the central point, which can be defined as a ”meanmagnetic field value”. Furthermore, a shearing of the curves is observed too. In stress-controlled fatigue tests the eddy current transfer-impedance me- asured by theGMR-sensor was found to be especially suitable to characterize the fatigue behavior. Figure 10 is an example for a multiple step fatigue test with a loadmix of different amplitudes and time dependences.The impedance ZGMRclearly shows in average a continuous increasing due to themartensite development (offset). However, this offset curve is modulated by a time func- tion, which exactly follows the plastic strain amplitude. Cyclic softening and cyclic hardening are visualized. The plain carbon steel C15 was fatigue-tested in single step stress- controlled HCF experiments. Also here theGMR-transfer-impedancewasme- asured to characterize the fatigue behavior. Figure 11 documents the total strain amplitude εa,t and the transfer-impedance ZGMR as a function of the load cycles in a cyclic deformation curve. The ferritic steel shows cyclic so- ftening followed by cyclic strengthening. The impedance as a function of the load cycles shows also an increase to a maximum followed by a decrease. Nondestructive damage characterization with examples... 663 Fig. 10. GMR transfer-impedance and plastic strain amplitude for a multiple step loadmix at the austenitic stainless steel 1.4541 However, here the dislocation multiplication, dislocation networking and cell sub-structuresprimarily influence the electrical conductivity and themagnetic permeability. Adislocationmultiplication –mainly in the first phase of fatigue – reduces the electrical conductivity. Therefore the impedance is in this first phase less sensitive than the strain amplitude, and the maximum is reached also later, with a time delay compared with the strain. However, the secon- dary strengthening is clearly visualized and is due to the permeability effects and can be discussed as an influence of compressive residual stresses of higher order. Fig. 11. Total strain amplitude and the GMR transfer-impedance at the plain carbon steel C15 664 G. Dobmann et al. 5. Conclusion • Micromagnetic techniques have a wide potential to characterize thema- terials damage. • Precipitation hardening by thermal aging can be described by use of a micromagnetic multiple parameter approach on a high confidence level. • Neutrondegradation is a complexmaterial damage and themicromagne- tic approach needs further investigations whenCharpy specimens should be used for calibration because of side-effects by plastic deformation and residual stresses in the broken specimens. Furthermore the sensitivity of the technology has to be validated for much lower fluences during lifetime. • The ability bymicromagnetic NDT to follow fatigue damage in the early stages, i.e. before surface cracking can be observed, and a special on-line monitoring technique was presented characterizing the cyclic softening and hardening effects. References 1. Altpeter I., 1990, Spannungsmessung und Zementitgehaltsbestimmung in Eisenwerkstoffen mittels dynamischer magnetischer und magnetoelastischer Messgrössen, PhD-Thesis, SaarlandUniversity (inGerman), Saarbrücken,Ger- many 2. Altpeter I., Becker R., Dobmann G., Kern R., Theiner W., Yashan A., 2002, Robust solutions of inverse problems in electromagnetic non- destructive evaluation, Inverse Problems, 18, 1-15, Institute of Physics Pu- blishing 3. Bassler H.-J., 1999,Wechselverformungsverhalten und verformungsinduzier- teMartensitbildung bei dem austenitischen Stahl X6 CrNiTi1810, PhD-Thesis, University Kaiserslautern (in German) 4. Borsutzki M., 1998, Process-Itegrated Dtermination of the Yield Strength and Deep Drawability Parameters r m und ∆r at Cold-Rolled and Hot-Dip- Galvanized Steel Sheets, Ph.D. Thesis, Saarland University (in German), Sa- arbrücken Nondestructive damage characterization with examples... 665 5. Dobmann G., 2002,On-linemonitoring of fatigue in the LCF andHCF range by using micro-magnetic NDT at plain carbon and austenitic stainless steel, 8th ECNDT, Barcelona, Spain, June 17-21,Technical Area Material Characte- rization, Conference Proceedings, Spanish Society for NDT 6. Dobmann G., et al., 1989, Progress in the micromagnetic multiparameter microstructure and stress analysis (3MA), In:Nondestructive Characterization of Materials, III, P. Höller, V. Hauk, G. Dobmann, C. Ruud, R. Green (edit.), Springer-Verlag, Berlin, p.516 7. Dobmann G., et al., 1998, Barkhausen noisemeasurements and relatedme- asurements in ferromagneticmaterials, In:Topics onNondestructive Evaluation Series, B.B. Djordjevic, H. DosReis (edit.), Vol. 1: Sensing for Materials Cha- racterization, Processing, and Manufacturing, G. Birnbaum, B. Auld (edit.), The American Society for Nondestructive Testing, Inc., ISBN 1-57117-067-7 8. DobmannG.,Debarberis L.,Coste J.-F., 2001,Agingmaterial evaluation and studies by non-destructive techniques (AMES-NDT), a European network project,Nuclear Engineering and Design, 26, 363-374 9. Dobmann G., Höller P., 1990, Non-destructive determination of material properties and stresses, 10th Int. Conf. NDE in the Nuclear and Pressure Ves- sels Industries, Glasgow, ASM Int. 10. Dobmann G., Kröning M., Theiner W., Willems H., 1993, Non- destructive characterizationofmaterialsbyultrasonicandmicro-magnetic tech- niques for strengthandtoughnesspredictionandthedetectionof earlycreepda- mage, Staatliche Materialprüfungsanstalt –MPA, Stuttgart: German-Japanese JointSeminar onStructural StrengthandNDEProblems,NuclearEngineering. Sec.3.4, Stuttgart 11. Dobmann G., Lang M., Eifler D., Bassler H.-J., 2001, On-line fatigue monitoring of austenitic stainless steel using aGMR-sensor, In: Studies in Ap- plied Electromagnetics and Mechanics, J. Pávo, G. Vértesy, T. Takagi, S.S. Udpa (edit.), Amsterdam, Washington, Tokyo: IOS Press, p.342, Proceedings of the InternationalWorkshop onElectromagnetic Non-Destructive Evaluation, June 2000, Budapest, ISSN: 1383-7281 12. Dobmann G., Seibold A., 1992, First attempts towards the early detection of fatigued substructures using cyclic-loaded 20MnMoNi 55 steel, Nuclear En- gineering and Design, 137, 363-369 13. Eifler D., Macherauch E., 1990, Microstructure and cyclic deformation behaviour of plain carbon and low-alloyed-steels, Int. Journal of Fatigue, 12, 3, 165-174 14. GRETE, 1999, Evaluation of NDT techniques for monitoring of material de- gradation, EURATOMprogramme, FIS5-1999-00280 15. Grünberg P., 1995,Riesenmagnetowiderstand inmagnetischen Schichtstruk- turen, Physikalische Blätter, 51, 1077-1081 666 G. Dobmann et al. 16. Kittel Ch., 1971,An Introduction to Solid State Physics, J.Wiley and Sons, NewYork 17. Lang M., 2000, Zerstörungsfreie Charakterisierung des Wechselverformung- sverhaltens und der verformungsinduzierten Martensitbildung bei dem austeni- tischen Stahl X6 CrNiTi 1810 mittels empfindlicher Magnetfeldsensoren, PhD- Thesis, SaarlandUniversity (in German), Saarbrücken 18. Mikheev M.N.,GorkunovE.S., 1979,Magneticmethods ofmonitoring qu- ality of heat treatment, Ninth World Conference on Non-Destructive Testing, 4A-10, Melbourne 19. NVEElectronics, Magnetizable Bead Detector, United States Patent Applica- tion, 20020060565, 2002 20. Pitsch H., 1989, Die Entwicklung und Erprobung der Oberwellenanalyse im Zeitsignal der magnetischen Tangentialfeldstärke als neues Modul des 3MA- Ansatzes (Mikromagnetische Multiparameter Mikrostruktur und Spannungsa- nalyse), PhD-Thesis, SaarlandUniversity (in German), Saarbrücken,Germany 21. Tschuncky R., 2004, Entwicklung eines Mustererkennungs- und Klassifika- tionsmoduls für die indirekteCharakterisierung vonWerkstoffeigenschaften, Di- ploma Thesis, SaarlandUniversity (in German), Germany Nieniszcząca charakteryzacja uszkodzenia materiału w odniesieniu do termicznego starzenia, degradacji neutronowej i zmęczenia Sreszczenie Badania nieniszczące (NDT) są zwykle, w społeczności inżynierskiej, wiązane ze zdolnością do wykrycia, klasyfikacji i wymiarowania niezgodności materiałowych – na przykład z początkiem niemetalicznych wtrąceń o rozmiarach kilku dziesiątek mikrometra dla stali lub stopów aluminium, aż do tak zwanych „defektów mate- riałowych” w rodzaju pęknięć makroskopowych o rozmiarach kilku milimetrów. Cel ten jest jednak na szczycie listy podejmowanych działań w szeroko rozumianych nie- niszczących badaniach materiałowych. Metodologie w rodzaju BU (badań ultradź- więkowych), BR (badań radiograficznych czy BM (badań magnetycznych) są dobrze wprowadzonew szerokiej dziedzinie standardowychbadańwyrobów i ich składników. W ostatnich 15-20 latach techniki nieniszczące rozwijano również w odniesieniu do charakteryzacji materiału na przykład w zakresie parametrów mikrostruktury, tzn. do defektów sieciowych typu rozkładów i gęstości dyslokacji, wtrąceń, mikropustek, aby opisać wzmocnienie i/lub osłabieniemateriałów, głównie stopówmetali, ale rów- nież domierzenia przyłożonych i resztkowych naprężeń (Dobmann et al., 1989). Manuscript received December 13, 2005; accepted for print March 15, 2006