DE4342100A1 - Non-destructive testing of concrete by confocal magnetic tomography - Google Patents

Abstract

A method for the non-destructive testing of a reinforced / prestressed concrete building structure employs confocal magnetic tomography to obtain images of the magnetic distribution in planar fields at selected distances from the outer face of the structure. A magnetic sensor in a measurement plane (3) registers the combined effects of a magnetic source plane (2) incorporating the prestressed members (not shown) and a plane (1) nearer the concrete surface in which reinforcing rods (not shown) superimpose an unwanted distortion. Such distortion is suppressed by suitable filtration in order that faults/irregularities in the prestressing structure are reliably detected by a real-time Fourier transform programme.

Description

1. Problem

The method of magnetic leakage flux measurement enables non-destructive Material testing of magnetizable components or composite components with magnetizable Matrix. Here, the leakage flux locally exiting has a characteristic signature in the measured values for damage in the component. In principle, all open can Separation points (continuous breaks, cracks, notches, free ends) can be detected. In the In practice, however, the break signals are sometimes overlaid with large interference signals, which result from changes in permeability in the magnetizable material, e.g. B. by plastic Deformations, local heat influences, material inhomogeneities etc. result. Further Interference signals arise from the environment, e.g. B. magnetic interference fields from the operation of electrical machines, fields of magnetic components in the vicinity of the test object etc. The difficulty in interpreting the overall signals recorded is now from the superimposition of the different signals, the characteristic signals of e.g. B. Identify breaks clearly.

2. Retroactive accounting

Most interference signals can, however, be eliminated if one of the measured values two-dimensional distribution of the magnetic field strength in the measuring plane as Summation of all magnetic information from a three-dimensional object  represent, back to the existing magnetization in the object. The advantage is here that the magnetization in cracks and separation points becomes zero and thus this Make clearer of areas with permeability changes in which the magnetization in the is significantly different from zero. In addition, of course Recognize structure ends or periodic structure shapes.

Recalculation procedure, the three-dimensional magnetization distribution in the object require a large number of measuring points and a considerable one Computational effort, so that these procedures are very time consuming and not in real time can be carried out.

3. Model and proposed solution

In the following a simplified and faster procedure is to be presented, the so-called confocal magnetic tomography.

On the basis of / 1 /, in which the recalculation of measured field strengths to the generating current distribution in flat bodies is described, the presented here Process the magnetization distribution in bodies on planes parallel to the measuring plane in Calculate real time.

A simple model should be considered to describe the method. This is a flat body, see Fig. 1, which has a different magnetization distribution in a level 1 (interference level) at a distance -z₁ from the surface and in a further level 2 (source level) at a distance -z₂ from the surface . In a plane 3 (measuring plane) at a distance + z₃ from the surface, a suitable magnetic field sensor can be used to measure a two-dimensional distribution of the magnetic field strength component in a coordinate direction. This field strength component is a superimposition of the field strength resulting from the magnetization in level 1 and level 2. The aim of the method is now to calculate back as well as possible from the measurement of the field strength distribution in level 3 to the magnetization distribution (source distribution) in level 2 and thereby the magnetic Suppress information (interference signals) in level 1.

In the Fourier space, the relationship between the Fourier-transformed field strength component h _x (from the field strength component H _x ) and the Fourier-transformed magnetization component m _y (from the magnetization component M _y ) is described by

Here, k _x and k _{y are} the Fourier-transformed coordinates (frequencies) x and y and g (k _x , k _y , z) is the Fourier-transformed Green's function G (x, y, z), which can be solved analytically in this case and has the form of a low-pass filter, the filter strength being influenced by the distance of the measurement plane 3 from the body surface or the magnetization planes 1 and 3. In the problem presented here, the inverse problem must be solved, ie from equation (1) z. B. from the measured and transformed field h _x (k _x , k _y , z) the magnetization distribution m _y (k _x , k _y , z) can be calculated. This means that e.g. B. in the Fourier space, the y component of the magnetization can be determined simply by high-pass filtering the x component of the field strength. The magnetization in the x and y directions can also be clearly calculated from the field strength component in the z direction if one takes into account that in the quasi-static case the divergence of the magnetization is zero.

∇M = 0
-i _x x m _x (k _x , k _y ) -i k _{y.m y} (k _x , k _y ) = 0 (3)

The magnetization thus calculated in level 3 is, however, unsharp (Low frequency) mapping of the magnetization distribution in the body. A conversion of the Magnetization to other planes at a distance Δz is easily possible in the Fourier space by:

It is evident from equation (4) that when converting to levels closer to that Source planes are d. H. z is less than z₁, the high frequencies by the action of first factor to be amplified as a high-pass filter. It is figuratively spoken Magnetization shown sharper (higher frequency), z. B. it is then possible at a Conversion to level 1 the magnetization there as sharp as possible (in Dependence on the sensitivity and the signal-to-noise ratio of the Measuring system).

Here now there is the possibility of the high frequencies, which are essentially the include magnetic information from level 1, by a low pass filter and the magnetization filtered in this way is then further increased to level 2 using equation (4) to calculate back. This essentially has the blurred image of level 1 on the Level 2 removes and receives a good representation of the magnetization in Level 1, which is hardly is still influenced by the magnetization from level 1. With the above filtering in level 2, the filter strength must be selected so that the high frequency components of the Level 1 should be cut off as completely as possible, but without the frequency components of the Remove level 2, which is also included in level 1 as a blurred image. In the In practice, however, the filtering will not be carried out as ideally, since the upper one  Frequency range of level 1 and the lower frequency range of level 2 always overlap become. The cut-off frequency can only be selected so that the filtering is as possible little information is removed from level 2.

4. Influence of noise in the measurement

It is essential with this method that the measurement of the noise is as low as possible Field strength in level 3, since the noise as a high-frequency event in the case described here Process is disproportionately amplified and also with a filtering of the noise a large portion of the magnetic information would be removed at the same time.

5. Applications

One possible application of confocal magnetic tomography is as in Chapter 1 described in the detection of prestressing steel cracks in prestressed concrete buildings the one with a highly sensitive sensor near the component surface (level 3) Magnetic leakage flux from the breaking point is measured. This is usually short the sagging reinforcement (level 1) below the component surface and the deeper inside Tension bundle (level 2). With the described method, the interference signals from the Level 1 eliminated and the prestressing steel cracks can be recognized more clearly because the cracks in the Magnetization is zero.

Further applications can also be seen where there are wall anchors behind the reinforcement layer etc. must be recognized or the laying of reinforcement mats, e.g. B. on Liaison points to be checked.

A similar problem arises with the eddy current testing of structures riveted aluminum plates, the z. B. used in aircraft construction. If here from the outside cracks in the lower plates should be detected, then signals can be the actual rupture signal of the rivets and rivet holes from the plates above overlay. These interference signals can be eliminated by confocal magnetic tomography become.  

literature

B. J. Roth, N.G. Sepulveda and J.P. Wikswo, Using a Magnetometer to Image a Two-Dimensional Current Distribution, J. Appl. Phys. 65 (1), 1989.

Crack detection in prestressed concrete beams by recalculating the H _z (x, y) z = z₀ component:
For a level current or magnetization position, the sources can be back calculated after measuring the H _z component in a parallel plane.

In the case of two magnetization layers (sagging reinforcement), two Measurements at different z distances have one level removed because the sharpness of the image for the two levels in the measurement levels is different.

The magnetic signals of those under the prestressing steels Armouring disrupt the sensitive measurement of one level above the prestressing steel bundle whose cracks or tears should be measured.

The two-dimensional H _z (x, y) z = z _i distribution is calculated back by internal filtering (high-pass filtering). The measured image sharpness depends on the distance between the magnetization level and the measurement level and enables two measurements with different Z _i to differentiate between the two source levels.

Claims (1)

Process for the non-destructive testing of a component which has a flaccid magnetized reinforcement underneath its surface and also has a magnetized bundle of rods in a cladding tube located deeper relative to the component surface, in which the magnetic field is recorded with the aid of a magnetic field sensor at a defined distance from the component surface characterized in that with the help of the magnetic field values determined in this way, a calculation back to the magnetization distribution of the magnetized parts of the component in layers lying differently from the surface of the component is computationally determined and in this way a clearer representation of individual magnetized parts of the component containing levels are represented and thus damaged steel parts can be detected in the component.