Diffusion bonded test blocks for ultrasonic inspection
TWI Bulletin, October 1985
by Peter Spanswick and Peter Bartle
Peter Bartle, CEng, MIM, MWeldl, is Section Head, Structural Monitoring Studies, in the NDT Department.
'Peter' Spanswick, TEng, (CEI), is in the Solid Phase Group of the Advanced Heavy Section Processes Department.
This update of an earlier article describes how diffusion bonding can be used to produce test blocks for ultrasonic inspection which can be used for equipment calibration (defect sizing), operator training, component replicas or comparison of responses to defects with different surface characteristics.
Production of diffusion bonded blocks
It is probable that diffusion bonding [1, 2] is the only process that allows the production of test blocks containing small or thin 'defects' of controlled size surrounded by homogeneous material.
In diffusion bonding, two clean, well machined surfaces are heated while held together (under a pressure below the bulk yield stress at the bonding temperature) in a protective environment so that a bond may form across the original interface ( Fig.1). The bond forms as a result of microde formation of surface asperities and the solution, by the parent metal, of the residual oxide and contaminant layers. In many bonds the original interface cannot be detected when sections are examined under an optical microscope.
Fig.1. Diffusion bond: a) Interface between two mild steel samples revealed only by change of grain structure etched in nital
Fig.1. Diffusion bond: b) Section through defect: etched in nital
Thus, for test blocks, 'defects' can be machined or indented into one or both of the surfaces to be bonded, the plane of which can be inclined at any suitable angle to the final external surfaces. The shape and surface finish of the defects can be tailored to suit requirements.
Additionally, the defects can be produced with special surface features such as those of a real crack by, for example, indenting a crack face into one or both of the surfaces to be bonded. An added attraction is that, before bonding, it is possible to take plastic replicas of the surface(s) to be bonded and the defects therein. This gives a permanent record of the defects within the block.
Calibration (sizing) blocks
For equipment calibration purposes simple rectangular blocks which contain suitable arrays of defects can be produced to supplement the normal BS 2704:1978 or IIW specification machined calibration blocks. The latter are basically adequate for calibrations relating to defect position (including time base calibration), but leave a good deal to be desired in respect of defect sizing (amplitude calibration). Defects can be 'ideally' positioned within diffusion bonded test blocks so that no problems are encountered as a result of back or side wall reflections, the detection of more than one defect at one time, or near field effects.
Alternatively, defects can be positioned to take account of these difficulties so that their effect on equipment response can be checked. Such blocks would be suitable either for calibrating signal amplitudes in relation to defect area, or for use as reference blocks to compare signal amplitudes with those from other testpieces which require examination.
For calibration, blocks containing five defects perpendicular to the main axis can be used ( Fig.2) the defects having diameters of 1, 2, 2.8, 4 and 5.7mm. A suitably designed single block of this type could be used for size calibration with normal probe techniques, angled probe transmission, and tandem probe work. The 1mm diameter defect serves to check size resolution, but the other four have area ratios of 1, 2, 4 and 8 for amplitude calibration. By off-setting the plane of the defects from the centre of a rectangular instead of square cross section block, calibrations for two surface-to-defect distances could be produced. Having used the above blocks to obtain a basic size calibration, blocks for single angle probe work, containing defects of two sizes (1 and 2mm diameter), could be used for check tests (assuming a repeat full calibration proves unnecessary).
Fig.2. Calibration block with 'defects' perpendicular to main axis
a) Arrangement of defects r, s, t, u, v;
b) Basic transmission and reflection, single and paired probe beam paths (edf, edg, hdj, hdk, ldl, ldm, mdm)
For single angled probe work both ideally and non-ideally positioned defects could be included in the same block ( Fig.3) but for defects perpendicular to the main axis, separate blocks would probably be advisable to avoid excessive block sizes. (Blocks of the order of 80 x 120mm section x 350mm long are currently envisaged).
Fig.3. Calibration block for single angled probe work
a) Arrangement of three 1mm diameter defects (a1, a2, a3) plus three 2mm diameter defects (b1, b2, b3);
b) Basic beam paths X1 to X6 (normal theta values 45°, 65° and 70°). Note: theta in a) is exceptionally shallow to show layout of defects
The same blocks could be used for reference purposes, or separate reference blocks could be produced with defect diameters (as opposed to areas) in fixed ratios. At present the defects used in reference blocks are flat bottomed drilled holes. Drilling narrow, relatively deep, flat bottomed holes is difficult, however, and the blocks are not suitable for transmission probe use.
Normally, for calibration purposes, simulated defects have flat surfaces to give strong reflections and well defined signals ( Fig.4) but, to establish equipment response more fully, defects with hemispherical, mottled or crack face surfaces could be produced. Additionally the production procedure allows blocks to be produced with defects either close to, or well below, the surface.
Fig.4. Oscillograph traces obtained during ultrasonic examination of diffusion bonded test block (normal probe)
a) Beam passing through bond clear of defect (back wall echo only);
b) Beam partially on defect (defect and back wall echoes)
Applications
Diffusion bonding has now been used at The Welding Institute to produce a variety of testpieces for ultrasonic inspection. Some of these are described below.
In order to reproduce defects in welded pipes, a pipe section with a material combination of ASTM A 1.82 1¼CrMo was welded to Paralloy H39W pipe using Inconel weld metal. This was sectioned as shown in Fig.5 and the mating faces machined and ground leaving a prep angle of 70°. Rectangular slots and flat bottom holes were spark eroded into each face of the Inconel weld metal as shown. The three components were then diffusion bonded together giving a pipe section with known buried defects which could be compared with reference standards. This enabled ultrasonic inspection to be assessed and proved for a situation where severe interpretational problems were being encountered.
Fig.5. Three sections of a pipe showing simulated defects in weld metal, prior to diffusion bonding
Figure 6 shows a section of thick walled pipe containing defects with simulated fatigue fracture surfaces. The layout of the 'defects', which were produced with a press tool, is depicted in the cast replica taken from the joint face. The press tools were manufactured by producing a fatigue fracture in a heat treatable material and then machining the fracture face to the required shape. The machined fracture faces were finally heat treated to produce a press tool.
Fig.6. Thick walled pipe containing defects with simulated fatigue fracture surfaces, as shown in plastic replica
Figure 7 shows a square block with an angled face illustrating a 'defect' which was produced on a shaping machine to give a knurled type finish. Many types of defects can be manufactured in differing shapes and sizes with a range of surface textures. One method of producing these is to use spark erosion which for example can be used in the production of a series of deep narrow slots in a block fabricated using only one bond interface. The Welding Institute's spark erosion machine, a Eurospark 740/50A, is shown in Fig.8, and this was used to produce the slots shown in Fig.9.
Fig.7. Simulated defect with diamond knurl type finish machined into a block with face angled at 70°, prior to being diffusion bonded to give a buried defect in a solid block
Fig.8. The Welding Institute's spark erosion machine - a Eurospark 740/50A
Fig.9. Slots produced using spark erosion
Training uses
Calibration and reference blocks of the type described above would be invaluable for initial operator training. They would need to be supplemented with similar blocks which contained larger defects and also with blocks of perhaps less regular shape with defects of size, orientation, and position not revealed to the operator until he had completed his inspection. For training purposes several forms of defect surface should be employed.
For instance, blocks with known artificial defects could be used in conjunction with testpieces containing real defects, e.g. weld defects, where the defect dimensions have been determined by ultrasonic sizing and radiography rather than premachining.
The use of blocks for training purposes will prove satisfactory only if they are employed as part of a well-founded course.
Component replicas
When applying ultrasonic inspection in practice it is important to know where geometric features of a structure or component cause difficulties: for example, parts of some tubular assemblies are difficult to inspect. If the validity of ultrasonic inspection for suspect joints can be demonstrated, both confidence in and usefulness of the technique are enhanced. In many cases it would be possible to diffusion bond replicas with appropriately sized defects suitably spaced and oriented to allow the integrity of ultrasonic inspection to be checked. Such replicas can also be valuable for planning and improving inspection procedures.
Response from defects
The amplitude of a reflection received from a defect depends heavily on the nature of the surface of the defect. These effects can be examined using blocks containing defects with surfaces of various types. In particular blocks canbe produced such that the same defect may be viewed from one side as a crack face and from the other as an identically shaped and sized machined face.
Three 'defects' produced with an end mill in a milling machine to an exact size and depth are shown in Fig.10a. These were then incorporated into a large block ( Fig 10b) to become buried defects of a known size and orientation.
Fig.10. Example test block
a) Detects produced by milling to exact size and depth;
b) Final test block produced containing detects shown
Summary
Diffusion bonding offers a method of producing calibration, reference, research, training, and replica blocks which contain sized defects of specified position, orientation and surface finish. These blocks can be produced in any material combination that can be diffusion bonded and it is envisaged that they would supplement calibration blocks to BS 2704:1978 or IIW specifications, and might replace current flat bottomed hole type reference blocks.
References
Members interested in this technique are invited to contact Peter Spanswick.
| N° | Author | Title | |
| 1 | Bartle P M: | 'Introduction to diffusion bonding'. Metal Construction 1969 10 (5), 241-4. | Return to text |
| 2 | Bartle P M: | 'Diffusion bonding - principles and applications'. Welding Institute Research Bulletin 1983 24 (3) 88-92. | |
Members interested in this technique are invited to contact Peter Spanswick.
