Method for determining electrical conductivities in samples by means of an eddy current sensor
10429349 · 2019-10-01
Assignee
Inventors
Cpc classification
International classification
Abstract
In the method for determining the electrical conductivity in samples by an eddy current sensor, an alternating electrical field is excited at a known measurement frequency, an alternating electromagnetic field which is directed against the alternating electrical field is thereby formed, detected by a suitable detector, and the complex impedance is determined, which procedure is repeated at different known measurement frequencies, once in air and once with the same measurement frequencies at a calibration body, differences of the real and imaginary portions and of the measured values in air and over the calibration body are then divided by the respective measurement frequency, wherein a product is associated with each value pair R/ and X/=L in accordance with the associated measurement frequency w and the known conductivity a of the calibration body and a locus is presented in a Nyquist diagram.
Claims
1. A method for determining the electrical conductivity in samples by means of an eddy current sensor in which an alternating electrical field is excited by an alternating electrical voltage applied to a transmitter coil at a known measurement frequency and an alternating electromagnetic field which is directed against the alternating electrical field is thereby formed which is detected by a receiver coil or by a suitable detector and the complex impedance Z=R+jX is determined wherein this procedure is carried out at different measurement frequencies once in air and once at the same measurement frequencies at a calibration body with a known electrical conductivity , and wherein a determination of the electrical conductivity of a sample is performed in accordance with one of the following: 1) differences of the real and imaginary portions R and X of the measured values in air and over the calibration body are divided by the respective measurement frequency ; wherein a product is associated with each value pair R/ and X/=L in accordance with the associated measurement frequency and the known conductivity of the calibration body and a locus shown in a Nyquist diagram with these value pairs; wherein a value pair R/ and X/ of the values in air is formed at the measured values over the sample at a measured impedance Z at a measurement frequency of a sample with an unknown conductivity and a point from the locus and thus a value is associated with this value pair R/ and X/=L and the unknown electrical conductivity of the sample material is determined from this using the known measurement frequency ; and 2) on the calibration at a calibration body and a measurement at a sample, the phase angle =tan(X/R) is calculated and the function of the product =f() is determined instead of the locus, with a product being associated with the phase angle on the measurement of an unknown conductivity of the sample material with this function and thus the electrical conductivity of the sample being determined at the known measurement frequency ; and 3) on the calibration at a calibration body and a measurement at a sample, the quotient q=X/R or its reciprocal 1/q is calculated and, instead of the locus, the function of the product =f(q) or =f(1/q) is determined and a product is associated with the quotient q or 1/q on the measurement of an unknown conductivity of a sample with this function f(q) or f(1/q) and thus the electrical conductivity of the sample material is determined.
2. A method in accordance with claim 1, characterized in that the respective measurement frequency is selected in dependence on the respective measurement range of the conductivity such that the product lies within the limits of the previously determined locus.
3. A method in accordance with claim 1, characterized in that only the difference of the real portion R / is determined in the calibration at a calibration body and on a measurement at a sample and, instead of the locus, the function of the product =f(R/)is determined, with a product being associated with the difference R/ on the measurement of an unknown conductivity of a sample with this function and with thus the electrical conductivity of the sample being determined.
4. A method in accordance with claim 1, characterized in that the difference of the imaginary portion X/32 L is determined on the calibration at a calibration body and on a measurement at a sample and, instead of the locus, the function of the product =f(L) is determined, with a product being associated with the difference L on the measurement of an unknown conductivity of a sample with this function and thus the electrical conductivity of the sample being determined.
5. A method in accordance with claim 1, characterized in that, instead of the circular frequency =2nf, the frequency f is utilized in the calculation of the values of the locus and of the derived functions, that is, work is carried out with f instead of .
6. A method in accordance with claim 1, characterized in that a calibration is carried out at a calibration body having a known electrical conductivity in air for all measurement frequencies f, .
Description
(1) The invention will first be presented very generally and explained in principle and with its backgrounds.
(2) An eddy current sensor is typically operated at a frequency f which in this respect represents the actual measurement frequency f. However, since the circular frequency , which results as =2*f, is typically used in the determination of impedance values, the circular frequency will generally be called the measurement frequency in the following.
(3) The generated electromagnetic field F(A) of an eddy current sensor can be functionally described by the following equation
(4)
(5) In this respect, A is the vector potential, S is the current density in the coil, B is the magnetic flux density, is the magnetic permeability of the sample material, is the electrical conductivity of the material, is the circular frequency and V is the volume over which integration is carried out.
(6) The first term in this respect describes the magnetic field, the second term the exciting electric field, that is the electromagnetic field generated by the electric coil current, and the third term the eddy current field in the material of a sample.
(7) On a closer observation of this expression, it can be recognized that the material properties are not contained in all three terms. The first term is only influenced by the permeability . With a non-magnetic material, the permeability is approximately as large as in air, that is, 1.0. The second term is fully independent of the material, and only in the third term is the electric conductivity contained as the product with the circular frequency .
(8) It is known that the imaginary portion X divided by the circular frequency produces the inductance L.
(9) If the impedance Z of the electric coil for different electric conductivities and circular frequencies is calculated by means of simulation, it can be recognized that with the same product values *, the inductance L and the phase angle between the real portion R and the imaginary part X adopt identical values. If the real portion R obtained in the simulation is likewise divided by the respective circular frequency , this value is also identical. Tables 1 to 3 show this.
(10) Table 1 contains the impedance values calculated using a simulation program for a circular frequency of 1 MHz and conductivities in the range from 16 to 1.610.sup.9 S/m.
(11) Table 2 indicates the impedance values for the fourfold circular frequency , that is, 4 MHz, with a simultaneous quartering of the electrical conductivities . The same is indicated in Table 3 for a tenfold circular frequency and a tenth of the electrical conductivity with respect to Table 1, that is, here in the range between 1.6 to 1.6*10.sup.8 S/m. This shows that the product * is always of an equal amount in the same table lines.
(12) (The value L is the difference of the inductance L in air from the inductance over the material, that is
L=L.sub.in airL.sub.over material. )
(13) TABLE-US-00001 TABLE 1 1 MHz R X Z L R/ L Air 3.0418 3.0418 484.12 1.6E+01 0.00022 3.0418 3.0418 89.996 484.12 0.035247 0.00 1.6E+02 0.00197 3.0417 3.0417 89.963 484.10 0.314293 0.02 1.6E+03 0.01654 3.0379 3.0380 89.698 483.50 2.632182 0.62 1.6E+04 0.08694 2.9873 2.9885 88.333 475.44 13.837180 8.68 1.6E+05 0.15915 2.7664 2.7710 86.708 440.29 25.330256 43.83 1.6E+06 0.09342 2.5581 2.6698 87.908 407.14 14.869330 76.98 1.6E+07 0.03526 2.4777 2.4779 89.185 394.33 5.612128 89.79 1.6E+08 0.01176 2.4517 2.4517 89.725 390.19 1.872195 93.93 1.6E+09 0.00378 2.4434 2.4434 89.911 388.88 0.601958 95.24
(14) TABLE-US-00002 TABLE 2 4 MHz R X Z L R/ L Air 12.1670 12.1670 484.12 4.0E+00 0.00089 12.1670 12.1670 89.996 484.12 0.035238 0.00 4.0E+01 0.00790 12.1670 12.1670 89.963 484.10 0.314293 0.02 4.0E+02 0.06615 12.1520 12.1520 89.688 483.50 2.632142 0.62 4.0E+03 0.34776 11.9490 11.9540 88.333 475.44 13.837339 8.68 4.0E+04 0.63659 11.0660 11.0840 86.708 440.29 25.329858 43.83 4.0E+05 0.37369 10.2320 10.2390 87.908 407.13 14.869091 76.99 4.0E+06 0.14104 9.9106 9.9116 89.185 394.33 5.611969 89.79 4.0E+07 0.04705 9.8066 9.8068 69.725 390.19 1.872115 93.93 4.0E+08 0.01513 9.7737 9.7737 89.911 388.88 0.601942 95.24
(15) TABLE-US-00003 TABLE 3 10 MHz R X Z L R/ L Air 30.4180 30.4180 484.12 1.6E+00 0.00221 30.4180 30.4180 89.936 484.12 0.035241 0.00 1.6E+01 0.01975 30.4170 30.4170 89.963 484.10 0.314293 0.02 1.6E+02 0.16538 30.3790 30.3800 89.688 483.50 2.632182 0.62 1.6E+03 0.86941 29.8730 29.8860 88.333 475.44 13.837498 8.68 1.6E+04 1.59150 27.6640 27.7100 86.708 440.29 25.330256 43.83 1.6E+05 0.93424 25.5810 25.5980 87.908 407.14 14.869330 76.98 1.6E+06 0.35261 24.7770 24.7790 89.185 394.33 5.612128 89.79 1.6E+07 0.11763 24.5170 24.5170 89.725 390.19 1.872195 93.93 1.6E+08 0.03782 24.4340 24.4340 89.911 388.88 0.601958 95.24
(16) As can immediately be recognized, the values for the product of the circular frequency and the electrical conductivity * are also of equal amounts in these lines. It results from this that the electrical conductivity and the circular frequency have an equal influence on the real and imaginary portions R and X, i.e. that the difficult and complex variation of the electrical conductivity can be replaced with a comparatively simple variation of the circular frequency in a calibration.
(17) When approaching the material of a sample, the inductance L of the electrical coil falls. This can be explained by the fact that the eddy current field generated in the material of the sample is directed against the excited field of the electrical coil. The inductance L is a measure for the energy of the magnetic field and can be calculated from the magnetic flux density B (see functional). It results from this that the difference of the magnetic field strength H without and with the material influence has to be equal to the magnetic field strength H of the eddy current field (B.sub.eddy current=B.sub.in airB.sub.over material).
(18) The increase in the real portion R on approaching the material of a sample is equally a measure for the electric field strength E of the eddy current field.
(19) The division of the real portion R by the circular frequency requires special note. In the simulation, the real portion R of an electrical coil in air is equal to zero since it is considered as an ideal electrical coil in this respect. The value can here immediately be divided by the circular frequency . In practice, the real portion R is composed of different proportions, The ohmic resistance of the windings of the electrical coil is a constant and its increase by the skin effect is not linearly dependent on the circular frequency . It would be wrong to divide these values by the circular frequency . Only the portion which is caused by the effect of the eddy current field generated in the material of a sample results in meaningful values after the division by the circular frequency . This proportion can be determined by forming the difference from the real portion R of the electrical coil in air with the same circular frequency .
(20) A locus of the product of the circular frequency and of the electrical conductivity is required for the measurement, said locus being determined from the difference of the coil impedance L in air and with respect to the electrical conductivity defined via a calibration body. This can be seen from the diagram shown in
(21) The difference of the real portions (divided by the circular frequency ) is entered on the abscissa and the difference of the inductances L is entered on the ordinate,
(22) A calibration according to the method in accordance with the invention has the following procedure:
(23) Step 1: The impedance Z of the electrical coil in air is measured at different circular frequencies .
(24) Step 2: The impedance Z of the electrical coil is measured at the same circular frequencies over a calibration body having a defined electrical conductivity .
(25) Step 3: The differences of real and imaginary portions R and X between the values in air and over the material are divided by the respective circular frequency and produce the locus.
(26) Step 4: The measurement range results from the desired measurement frequency and the range of the electrical conductivity to be measured. If e.g. electrical conductivities in the range between 25 S/m and 100 S/m are to be measured at a circular frequency of 1 MHz and if a calibration body having an electrical conductivity of 50 S/m is available, the circular frequency has to be varied from 500 kHz (for 100 S/m) up to 2 MHz (for 25 S/m). However, an electrical conductivity of 200 S/m at 250 kHz or 12.5 S/m at 4 MHz can equally be measured with this locus obtained therefrom, The product of the electrical conductivity and the circular frequency must always be within the calibration limits *.
(27) The differences of the real and imaginary portions R/ and X/=L divided by the circular frequency are again determined in the measurement over an unknown material. A point on the locus results relative to this. Since the respective measurement frequency is known, the associated electrical conductivity can now be determined.
(28) If the loci are calculated by simulation at different spacings of the electrical coil from the surface of a material (sample), the same points approximately lie on straight lines which correspond to the phase angle in a first approximation.
(29) A conclusion on a spacing tolerance (lift-off tolerance) of the values can be concluded from this.
(30) This is very logical: If the electrical coil is moved away from the material (surface of a sample), the magnetic field strength H in the material is reduced, which, however, equally produces a reduction of the magnetic field strength H of the eddy current field generated in the material of the sample. The ratio of the exciting field to the eddy current field is expressed by the phase angle and remains approximately the same. Only the repercussion on the electrical coil is smaller and the precision with which this phase angle can be determined thus falls.
(31) A function =f() can therefore also be determined from the locus. The phase angle is determined from the measured difference values R/ and X/=L; the value is determined with the phase angle using this function and from this, by means of the measurement frequency, the conductivity .
(32) Functions =f(R/) or =f(L) can also be determined from the difference values R/ and L and the conductivity can be determined by this.
(33) In the method in accordance with the invention, an alternating electrical field is excited within a sample by means of an eddy current sensor by an alternating electrical voltage applied to a transmitter coil at a known measurement frequency and an alternating electromagnetic field which is directed against the alternating electrical field is thereby formed which is detected by a receiver coil or by a suitable detector and the complex impedance Z=R+jX is determined,
(34) with this procedure being carried out at different measurement frequencies once in air and once with the same measurement frequencies at a calibration body with a known electrical conductivity .
(35) The electrical conductivity of a sample can then be determined in three alternatives named in the following.
(36) In the first alternative, the differences of the real and imaginary portions R and X of the measured values in air and over the calibration body are divided by the respective measurement frequency . A product is associated with each value pair R/ and X/=L in accordance with the associated measurement frequency and the known conductivity of the calibration body and
(37) a locus is shown in a Nyquist diagram with these value pairs. With a measured impedance Z and at a measurement frequency at a sample with unknown conductivity , a value pair R/ and X/ of the values in air is associated with those measured values over the sample and a point from the locus and thus a value is associated with this value pair R/ and X/=L. The unknown electrical conductivity of the sample material is determined from this at the known measurement frequency .
(38) In the second alternative in accordance with the invention, on the calibration at a calibration body and a measurement at a sample, the phase angle =tan(X/R) is calculated and the function of the product =f() is determined instead of the locus, with a product being associated with the phase angle on the measurement of an unknown conductivity of the sample material using this function and the electrical conductivity of the sample thus being determined at the known measurement frequency .
(39) In the third alternative in accordance with the invention, the quotient q=X/R or its reciprocal 1/q is calculated in the calibration at a calibration body and a measurement at a sample. Instead of the locus, the function of the product =f(q) or =f(1/q) is determined and a product is associated with the quotient q or 1/q on the measurement of an unknown conductivity of a sample with this function f(q) or f(1/q) and the electrical conductivity of the sample material is thus determined.
(40) The respective measurement frequency should be selected in dependence on the respective measurement range of the conductivity such that the product lies within the limits of the previously determined locus. The circular frequency should preferably be selected as the measurement frequency.
(41) In the calibration at a calibration body in air, all the measurement frequencies should be taken into account which are later taken into account in the determination of the electrical conductivity at samples.
(42) It is also possible to proceed in the invention such that only the difference of the real portion R/ is determined in the calibration at a calibration body and a measurement at a sample. The function of the product =f(R/) is determined instead of the locus; with a product being associated with the difference R/ on the measurement of an unknown conductivity of a sample using this function and with thus the electrical conductivity of the sample being determined.
(43) It is, however, also possible to determine the difference of the imaginary portion X/=L in the calibration at a calibration body and a measurement at a sample. The function of the product =f(L) is then determined instead of the locus, with a product being associated with the difference L on the measurement of an unknown conductivity of a sample using this function and the electrical conductivity of the sample thus being able to be determined.
(44) The selection of the evaluation with the previously named alternatives and options, namely whether the locus or one of the derived functions are used for the determination of the electrical conductivity , can be carried out in dependence on the respective measurement job. It can be observed in this respect that if the locus increases or drops very steeply in the desired range, the measurement value fluctuations are larger. If one of the derived functions delivers a more shallowly increasing or dropping curve extent in this range, it should be selected due to the smaller measurement value fluctuations.
(45) In the following Table 4, measurement values and calculated values are indicated with reference to simulation calculations in which an ideal coil for an eddy current sensor and a sample of copper are taken into account. The real portion in air is zero and L in air can be calculated from the imaginary portion. The L is the difference L L in air. R R in air is identical to Real in the simulation.
(46) The columns Frequency, Real (portion) and (Imaginary portion) are values obtained from the simulation. L and R/ are values divided by .
(47) The locus results from the columns in which the values are underlined. R/ for the abscissa, L for the ordinate. A value is associated for each point of the locus so that, with a known measurement frequency , a value for the corresponding electrical conductivity can be simply determined.
(48) In columns in which the values are shown in bold as well as the respective column L or R/, the derived functions =f(), =f(L/(R/)), =f(/R/) and =f(L) result together with the column.
(49) Real and Imag are the values for air; Cu Real and Cu Imag are the associated values for a sample of copper:
L(Imag)Cu Imag)/
R/=(Cu RealReal)/
(50) The specific locus results in an equivalent manner from this from the columns in which the values are underlined.
(51) TABLE-US-00004 TABLE 4 P9x5 Copper 58 *10e6 S/M L in air 484.13 nH Frequency Real[] Imag[] L[nH] R/[ns] L[nH] [] L/(R/) 0 0.5 0.013215 1.2338 392.74 4.207 91.388 29 1.52480 87.36710 21.724888 1 1 0.019005 2.4594 391.44 3.025 92.963 58 1.53818 88.13354 30.6439774 2 1.5 0.023449 3.6836 390.85 2.488 93.276 87 1.54413 88.47464 37.4891545 3 2 0.027196 4.9072 390.51 2.164 93.616 116 1.54768 88.67826 43.2555368 4 2.5 0.030498 6.1303 390.28 1.942 93.851 145 1.55011 88.81744 48.336877 5 3 0.033482 7.3531 390.11 1.776 94.024 174 1.55191 88.92031 52.9319148 6 3.5 0.036227 8.5756 389.97 1.647 94.162 203 1.55330 89.00032 57.157942 7 4 0.038782 9.798 389.86 1.543 94.268 232 1.55443 89.06480 61.0890403 8 4.5 0.041182 11.02 389.76 1.457 94.366 261 1.55536 89.11832 64.7867893 9 5 0.043452 12.242 389.69 1.383 94.444 290 1.55615 89.16357 68.2809525 10 5.5 0.04561 13.464 389.62 1.320 94.507 319 1.55683 89.20250 71.6035594 11 6 0.047673 14.686 389.57 1.265 94.560 348 1.55742 89.23643 74.7746678 12 6.5 0.049651 15.908 389.52 1.216 94.605 377 1.55795 89.26637 77.8156963 13 7 0.051555 17.13 389.49 1.172 94.644 406 1.55841 89.29303 80.7394342 14 7.5 0.053392 18.351 389.43 1.133 94.698 435 1.55883 89.31713 83.5783624 15 8 0.055168 19.573 389.40 1.098 94.726 464 1.55921 89.33879 86.305654 16 8.5 0.05689 20.794 389.36 1.065 94.769 493 1.55956 89.35863 88.9647515 17 9 0.058562 22.016 389.34 1.036 94.790 522 1.55987 89.37667 91.5286211 18 9.5 0.060188 23.237 389.30 1.008 94.826 551 1.56016 89.39338 94.0385709 19 10 0.061772 24.459 389.29 0.983 94.841 580 1.56043 89.40871 98.4858405 20 10.5 0.063317 25.68 389.26 0.960 94.871 609 1.56068 89.42302 98.8483708 21 11 0.064825 26.901 389.23 0.938 94.898 638 1.56091 89.43635 101.175113 22 11.5 0.066299 28.123 389.22 0.918 94.909 667 1.56113 89.44872 103.434005 23 12 0.067742 29.344 389.20 0.898 94.932 696 1.56133 89.46038 105.657723 24 12.5 0.069153 30.565 389.18 0.881 94.953 725 1.56152 89.47134 107.838559 25 13 0.07054 31.786 389.16 0.864 94.972 754 1.56170 89.48164 109.969581
(52) Diagrams produced using the values from Table 4 are shown in
(53)
(54)
(55)
(56)
(57)