·Table of Contents
·Materials Characterization and testing
New approach to optimise the ultrasonic testing of austenitic weldsDidier Flotté - Daniel Chauveau
Institut de Soudure - Zi les jonquières - 57365 Ennery - France
|Fig 1: Drawing of a 20mm thick Specimen|
Within a group sponsored project (10 partners - French, German and Canadian) led by INSTITUT DE SOUDURE with the collaboration of IZFP - Sarbrücken, the state of the art of the various techniques available for testing austenitic welds has been performed in order to determine the influence of the parent material and the type of welding process on the performances of ultrasonic testing.
To perform this programme, specimens in 316L, inconel and Duplex have been welded by several welding process. Artificial reference defects were machined in each specimen :
Several techniques have been compared: manual ultrasonic techniques, phased arrays, EMAT, TOFD and various signal processing applied. For conventional ultrasonic testing, the influence of the operating parameters has been studied: frequency, type of probes, propagation through the welded zones.
The aim of this conference is to explain how to prepare a conventional ultrasonic inspection taking account of these results by:
For each specimen described in figure 1 and for every probes, the following parameters have been measured:
The list of the main probes tested to elaborate the database is given here after :
A data base stores all the results obtained. This database shall be used carefully because it does not integrate the feeling of the operator.
The table below gives an example of values obtained from the database with the criteria :
|Probes||Waves||Freq.(MHz)||Angles(°)||Depth (mm)||Steel||Welding process||Gain (dB)||Attenuation(dB)||Signal/ noise(dB)|
In this case, two probes can be chosen, the KRAUTKRAMER MWB SW 4 MHz (C) and the KRAUTKRAMER 3XW composite dual element LW (I).
For each specimen, it is possible to select from the results stored in the data base the probes that give the best ability for testing austenitic steel. The choice must be however tempered by the feed back of the controller.
After these measurements, it was found that shear waves generated by composite probes could be a good compromise to test several types of austenitic welds.
To perform this frequency analysis, a specific procedure of measurement has been developed. The principle of this analysis is the measurement of the ultrasonic signal amplitude transmitted by reflection on the back wall of the material specimen at a given frequency like it is described in the drawing in figure 2.
|Fig 2: Configuration used for frequency analysis|
The generator is set as followed:
A reference transfer function between 1 MHz and 11 MHz is determined on a disc of silicon nitride. The same function is measured on the material under testing. The attenuation through the material versus the frequency is obtained by calculating the ratio between these two functions.
Figures 3 give as an example the curves obtained on a 316L specimen with 45° longitudinal waves.
|Fig 3: Curves Obtained with 45° Longitudinal waves on a 316L Specimen|
We can see that attenuation doesn't change in a regularly way with the frequency. Minima were observed. The frequency position and amplitude of these minima depends on steel composition, welding process and especially on the position along the weld. Amplitude variations are less important for the minima in the low frequency zone (around 2 MHz).
Using probes with operating frequencies around the attenuation minima observed in the high frequency zone is embarrassing. As a matter of fact, maximum of attenuation could be very important.
For shear waves, the increasing of attenuation according frequency is more regularly than with compression waves but is very important. It not depends on too much with the position along the weld. To use this type of wave, it is recommended to work at a frequency as low as possible.
An other way is to use an ultrasonic probe, piezoelectric or EMAT, as the receiver. After comparing the two methods, we have chosen a focused beam probe like the receiver for the beam drawing. As a matter of fact, this method offers a greater resolution than the EMAT technology and, in addition, the level of ferrite in the steel does not affect it. Contrary to EMAT, it is not possible to separate the contribution of shear waves to the contribution of longitudinal waves. But in the context of this programme, the two lobes are stored for two different spatial positions and can be separate easily.
The weld cap disturbs the acquisition made through the melted metal but a great part of the beam can be observed yet and, generally, the interesting parameters have been measured.
The principle of this analysis is the measurement of the amplitude of the ultrasonic signal transmitted by the probe in study and travelling through a block or a specimen made with the specified material. These analysis can be performed with the beam propagate through the parent metal or through the welded zone (figure 4).
|Fig 4: Configuration used for the beam drawing|
The first step of the procedure is to research the maximum of the amplitude by moving the receiver probe in the scanning plan (the emerging point of the ultrasonic beam and the focusing point of the receiver probe merge). The gain for the receiver channel is adjusts to obtain an amplitude of 90% of the full screen.
The origin of the scanning area is fixed at X and Y co-ordinates in relation to this position, for example -30 mm in X and Y direction. The acquisition windows are adjusted to perform the measurement of the signal amplitude during the complete cycle.
The example in figure 5 has been obtained through a weld. The definition of the various angles measured on the scan is also given. In this case, the beam of longitudinal waves is not very disturbed. The example in figure 6 gives the case of a very disturbed beam.
|Fig 5: Weld region drawing of beam explanation||Fig 6: An example with a very disturbed beam|
In all cases, the ultrasonic operator must not miss that testing welds with an austenitic structure already require a particular approach and result obtained with a sample must be used with caution.
|© AIPnD , created by NDT.net|||Home| |Top||