Re: TOFD Milos:
A couple weeks ago you asked about smallest detectable flaw using TOFD. You stated:
I am looking for answer to the following problem.
It is known that the size of the smalest defect detectable by reflection of UT waves is about the hallf of the wave length. What is the theoretical limit of detection when diffraction (TOFD) instead of reflection is used ? Is it the same for PE technique and TOFD? Is it possible to estimate the limits of detection (size of the smalest defect) by TOFD method without experimenting on artificial defects (EDM notches)?
This discussion on smallest detectable flaw goes back a long way on this Forum, to its earliest days in 1996 (http://www.ndt.net/wshop/wshop_tr/messages/87.htm).
I think it appropriate to add a word about the much used Rule of thumb being invoked in this discussion. The idea that the smallest detectable flaw is half the wavelength is just a guideline and not a specific target or absolute limiting factor.
Long ago the limits todetection were considered and half a wavelength was NOT a limit. Lord Rayleigh (Theory of Sound, London, 1926) provided a relationship between a spherical reflector and the wavelength for the condition where the wavelength was much greater than the sphere diameter (amplitude is proportional to D^3/Lamda^2). Ermolov derived similar relationships for several of the more common basic shapes.
In clean low carbon steel I have detected pores on the order of 0.2mm diameter using a 5MHz TOFD. That is on the order of 5-6 times smaller than the wavelength! But when the same TOFD technique is applied to austenitic stainless steel with grain size on the order of 50-100 microns, the scatter makes it virtually impossible to detect anything but the largest of flaws. Grain size in chrome stainless steels are typical of this order of magnitude.
Based on my readings of Krautkramer and Ermolov, the detection being discussed by the Rule of Thumbis based on the traditional pulse-echo principles, whereby the beam axis is centred on the flaw and the concern is for the reflected wavefront. Finite element modelling or visualisation can be used to illustrate that a slight planar nature of a reflected wavefront can be maintained when the flaw is flat and its dimensions are on the order of a wavelength (perhaps down to approximately a half wavelength). Such a condition would produce a peak in the reflection directivity plot and could be used to maximise the echo response by placing the receiver along the maximum reflection axis.
But when we are not concerned with Reflections and it is a backscattered or forward scattered Diffraction that we are using to detect imperfections, the ratio of flaw size to wavelength (and the on-axis prerequisite) can be relaxed to something less than the half-wavelength guideline. TOFD is a forward scatter technique. The goal is to detect diffracted wavefronts from flaws in the pressure envelope of the transmitted pulse. We try to optimise detection by placing a receiver in the region where the expected maximum pressure from the forward scattered diffraction will occur from a crack-tip (see Charlesworth and Temple, Engineering Applications of Ultrasonic TOFD, Research Studies Press, 2001).
Properties of the materials tested (not just the alloy, since an alloy may be made in many forms of different grain sizes) and the purpose of the test will be critically important factors when considering the answer to your question. We can discuss theoretical limits to detections; but in UT (including and especially in TOFD) the response of the indication of concern over the background scatter noise from grain structure will be crucial. If you are expecting to RELIABLY detect flaws smaller than the grain size you will probably not be successful. The lateral wave will cause you problems at the lower end of thicknesses due to the dead zone (but I have seen shear wave TOFD used on 4mm wall in fine grained zirconium tubing). Thick sections (200-300mm) can also be tested by TOFD but accumulation of scatter increases with increasing soundpaths. In all cases the signal to noise ratio you can achieve will be the limiting factor for practical use of TOFD.
So even the use of EDM notches will be an approximation for the specific material tested. For surface connected conditions they may be adequate. But as a tool to investigate embedded flaws they cannot be made without access to the ends of the material so you will always have the effect of length to consider as well as the vertical extent (and shape and surface roughness). Ultimately the grain noise will be your limiting factor and the best you can do is increase receiver gain so there is a maximum noise level of about 10-15% (as illustrated in BS-7706).
no magic bullets for this. It will be empirical.
----------- Start Original Message -----------
: I am looking for answer to the following problem.
: It is known that the size of the smalest defect detectable by reflection of UT waves is about the hallf of the wave length. What is the theoretical limit of detection when diffraction (TOFD) instead of reflection is used ? Is it the same for PE technique and TOFD? Is it possible to estimate the limits of detection (size of the smalest defect) by TOFD method without experimenting on artificial defects (EDM notches)?
------------ End Original Message ------------