·Home ·Table of Contents ·General

Potentialities of ultrasonics for evaluating residual stresses: Influence of microstructure H. WALASZEK, Y. ABDALLAHOUI, H.P. LIEURADE Centre Technique des Industries Mécaniques 52, avenue Félix Louat -B.P. 80067 60304 SENLIS Contact

ABSTRACT

The good mastership of residual stresses level in mechanical components is an important factor, particularly for a good fatigue-strenght of these components. This paper presents advances obtained at CETIM in the field of development of an ultrasonic method for stress measurements. This method is potentialy advandageous by its non destructivity, its good portability on site, and its relative easiness for operation.

In the paper are discussed the results obtained with ultrasonics on steel welded plate, and a comparison is made with stress measurement obtained by incremental hole drilling method and X Ray diffraction. These results are also validated by thermal relaxation of the plates.

The paper discusses also the microstructure influence on ultrasonic measurements and the precision obtained.

In conclusion is outlined the interest for studying the ability of the ultrasonic residual stress measurement method in different industrial cases.

Keywords: Non destructive measurement - ultrasonic measurement - subsurface wave - acoustoelastic coefficient - residual stresses - microstructure - welded plate.

Introduction

Some technical means enabling evaluation of the stresses already exist but are generaly destructive or semi-destructive, costly and complexe to use. The currenty known methodology are following :

• X Ray diffraction method
• Incremental hole drilling-method
• Ultrasonic method
• Barkhausen (or ferromagnetic) noise method
Among these methods, X Ray diffraction and hole drilling method are the most commonly used. X Ray method is non destructive but allows only measurement of stresses in a subsurface zone of ten micrometers depth. Hole drilling methods is semi destructive but allows to measure the stress gradient in the depth of the material.
Both Barkhausen noise and ultrasonic method are already in a semi-study state.
Meanwhile, our aim is to develop a methodology dealing with mechanical parts, which needs measurements on length less than 10 mm.
Therefore CETIM works on the theme of non destructive evaluation of residual stresses, by the way of non linear effects of these stresses on ultrasonic speed. So, an equipment designed by CETIM to measure stresses in bolt assembled mechanical components is commercialy spread in industry under license since several years .
In the residual stresses domain, CETIM uses now the speed variation of ultrasonic waves to describe the stress-state in subsurface zones .
The aim of this paper is to present one of the aspect of CETIM knowledge in the field of residual stresses evaluation on welded components. The results obtained by using ultrasonic measurements are compared with those obtained by hole drilling method and X Ray diffraction. Good accordance of ultrasonic results and those of the two methods taken as reference is discussed. Factors influencing measurement accuracy are also discussed.
In conclusion, the paper introduces the potentialities of ultrasonic method to check the quality of various mechanical and thermomecanical processes.

2. Sensitivity of ultrasonic waves to stress : Acoustoelastic coefficients

To evaluate experimentally residual stresses, we need the value of acoustoelastic coefficient. This can be done through a calibration (tensile test for example). These coefficients are determinated for a given material, by measuring propagation speed (or propagation time) of ultrasonic waves versus stress to which the material is submitted. The value of this ratio can be obtained experimentaly by calibrating.
In this study, we define an acoustoelastic coefficient K_i which represents the value of the curve of the relative speed variation versus stress. As the distance between ultrasonic transmitter and ultrasonic receiver is maintained constant, measurements are not carried out in speed itself, but on ultrasonic travel-time which is directly measurable :


Where t₀ is the ultrasonic travel time in non stressed condition
t is the ultrasonic travel time in stressed condition
s is the applied stress

3. Microstructure produced by welding - Effect of thermal relaxation process

The welding operation is characterized by a great and localized heat flux on welded component. It appears that, if we move our point of view on the weld from the fusion line to the parent-metal, we meet typical microstructures which nature depends on the maximum temperature reached and on the cooling rate at this point. We may distinguish, from parent metal to weld axis, a microstructure evolution from ferrito-perlitic structure, in parent metal, to a bainitic or a martensitic-bainitic microstructures, respectively, in the heat affected zone (HAZ) and the melted zone.
The micrographs obtained in the different zones after heat treatment show that structures remain stable before and after heat treatment. X Ray diffraction were made in these three zones. These measurements showed than stresses are close to zero state after heat treatment. This confirms that this heat treatment leads to relaxation of stresses, without affecting microstructure.
The values of acoustoelastic constant estimated in the three zones are given in table 1.

4. Experimental procedure methodology

4.1 Sample used
The involved assemblies are butt welded plates structures (S355 steel grade : 8 and 30 mm thick), with a X chamfer.(figure1)
The parent metal yield strength, measured in the longitudinal or transverse direction is 400 MPa. It is the same for the deposit metal.

Fig 1: Butt welded joint-steel S355 (thickness 8 or 30 mm)

Fig 2: refracted wave at the surface of the propagation medium

Zone	Acoustoelastic coefficient K1 x 10-5 [MPa]-1
Parent Metal (P.M.)	K1PM (measured) = - 1.25 ± 5 10-2
Heat Affected Zone H.A.Z	K1HAZ (estimated) # 90% K1PM
Melted Metal (M.M)	K1MM (estimated) # 80% K1PM
Table1 : Acoustoelastic coefficient measured in parent metal and estimated in H.A.Z and in deposit metal

4.2 Measurement method
Measurements are made with a transmitter-receiver transducer generating longitudinal subsurface waves (figure 2). The transducer axis is parallel to weld axis.
Ultrasonic travel time is taken on 2 mm spaced points, on melted zone and heat affected zone, and
5 mm spaced points, in parent metal zone.
The residual stresses measured by this ultrasonic method are localized in near surface zone. Measurement is integrated on a 10 mm length. Integration depth of measurement is estimated at 1,5 mm for 10 MHz and 2,5 mm for 5 MHz. The metallurgical texture is taken into account through a double calibration i.e in longitudinal and transverse direction, when acoustoelastic constants are determinated.
So, the influence of texture was estimated to 10 %. This effect can be substracted from the measurement when the rolling direction is known.

5. Results and discussion

5.1 Evaluation of welding stresses gradient
We give in figure 3 the evolution of longitudinal stress (in zone C, parallel to weld axis), evaluated by ultrasonic method and using the acoustoelastic constant determined in parent metal. This constant is given in table 1.
The stress is evaluated by measurement of ultrasonic travel time on each point. One of this point defined by the mean value of ultrasonic travel time measured from the plate end is considered as the reference point, and « zero stressed ». Therefore the given values of stresses are relative. Moreover, the stresses obtained do not take into account the effect of the microstructure variations induced by welding, on the measurement.
If we consider the instrumental error and the error due to non taking into account of the biaxiality of stresses, and the reproductibility error, the precision on ultrasonic stress evaluation method is estimated to ± 50 MPa.

5.2 Comparaison of measurements obtained with different methods, before and after microstructure correction
For first validation of ultrasonic measurements, we performed stress measurement by incremental hole drilling. The error on hole drilling is also here estimated to ± 50 MPa. Ultrasonic and hole drilling measurements are integrated on a 2 mm depth.
Figure 3 gives the residual stress profile obtained by the ultrasonic method, and results of stress measurements by incremental hole drilling obtained in parent metal (2 points) and weld (3 points).


Fig 3: 5 MHz longitudinal stress profile obtained by ultrasonic and hole drilling method

Comparison between ultrasonic and hole drilling stress measurements give the following results :

• the evolution of stress profiles obtained by using the two measurement method is the same, whatever is the considered measurement point.
• Comparison of stress profiles obtained by these methods especially in parent metal zone
  (at ± 120 mm) shows that ultrasonic measurement needs to be corrected by an offset adjustment of around - 100 MPa due to arbitrary choice of the reference point.
• Values obtained by ultrasonics on welded joints are higher than those obtained by hole drilling method, even if we take into account the offset correction we discussed higher. In fact, we used in weld zone the same ultrasonic calibration than in parent metal. As we will see in further paragraph, using the right calibration allows a good coïncidence between ultrasonic and hole drilling stress measurements.

Fig 4: effect of progressive thermal stress relief on ultrasonic measurements - materiel S355 steel - 8 mm thick

5.3 Influence of thermal relaxation


• 8mm thick S355 welded plate.

On figure 4 are given stress profiles obtained by ultrasonics after incremental temperature steps of thermal stress relaxation. At each step the stress relaxation makes the ultrasonic stress profile flatter. Finaly we present on figure 5 the comparison between stress profiles obtained either by ultrasonic, or by hole drilling method after complete relaxation. The difference in results obtained by these methods is very low. The only zone in which is noticeable a difference is the welded joint zone, for which the maximum deviation between the two profiles is 100 MPa.

Fig 5: comparison of stress profiles obtained by ultrasonics and hole drilling method after complete thermal stress relief processing - Material S355 steel - 8 mm thick

• 30 mm thick S355 - welded plate
On figures 6 and 7 are given comparative stress measurement results obtained by ultrasonic or X Ray diffraction method.


Fig 6: Comparison of stress profiles obtained by ultrasonic and X Ray diffraction, before stress relief - Material S355 stell - 30 mm thick

The measured stresses are integrated on a thickness of 2.5 mm for ultrasonic and of 0,5 mm for
X Ray measurements. In fact, X Ray measurement integrate stresses on only 10 micrometers. Measurement of stresses at 0,5 mm depth bellow the surface was enabled after chemical dissolution. Reproductibility of X Ray measurements are in our case estimated by ± 10 MPa.
Adding a - 100 MPa offset to ultrasonic measurement in figure 6, and reminding that we consider the ultrasonic measurement as relative, ultrasonic and X Ray measurements meet correctly together, except in melted zone.
After stress relaxation, (figure 7) the comparison of ultrasonic and X Ray measurement stress profiles show a good accordance, without offset correction. Nevertheless, we observe on weld a 200 MPa stress value by ultrasonics, compared to 100 MPa are measured by X Ray. As we explain further, this is an effet of microstructure variation.

Fig 7: Same as fig. 6 but after thermal stress relief

5.4 Microstructure effect
The ultrasonic stress measurement results presented in the paper give a relative stress level, because we are not able to define precisely the reference zero-stress points on the sample which is investigated.
Apparently these results display stress level by ultrasonics higher than the yield point of the parent metal wich is surprising at a first glance.
To improve the validity of the comparison between ultrasonic method and others, we have to take into account on the microstructure effect from figure 6 (measurement on 30 mm thick plate). So, figure 8 is derived from figure 6 by the following process applied to the ultrasonic measurement:

• application of a - 100 MPa offset.
• in H.A.Z, the stress value is reduced by 50 MPa (10 %) due to microstructure offset, following table 1
• in melted zone, the stress value is reduced by 150 MPa, due to microstructure effect, following table1
After these three correction steps, we finally found a good coïncidence of ultrasonic and X Ray profiles. Ultrasonic stress measurements appear now lower that the yield strength of parent metal.

Fig 8: Idem fig. 6 after offset and "microstructure" correction

6. Conclusion

The ultrasonic stress measurement method described in this paper enabled in laboratory experimental conditions, the evaluation of residual stresses on welded plates.
The measurements have been validated by two reference methods (hole drilling and X Ray diffraction) and confirmed by thermal stress relief, in the following cases :

• parent metal
• heat affected zone (HAZ)
• melted zone, with a correction of microstructure effects
Studies on the ability of applying the ultrasonic methodology are planned in the case of some insolved industrial problems, such as :
• evaluation of residual stresses on plates (static stresses)
• checking of a thermal stress relief quality on welded joints and plates
• checking of cold hammering quality
• optimization of rolling process quality

© AIPnD , created by NDT.net

|Home|    |Top|