Method for Measuring Phase Currents of a Device Under Test, in Particular of an Inverter

Abstract

A method is for measuring phase currents of a device under test, in particular of an inverter, in which a sensor arrangement, which has a component including a crystal lattice with a defect, is arranged in a region of the device under test. The method includes using the sensor arrangement to detect a magnetic field formed by a vector of magnetic fields, the magnetic fields each in turn being brought about by one of the phase currents of the device under test, and calculating a vector of the phase currents from the vector of the magnetic fields based on a coefficient matrix.

Claims

1. A method for measuring phase currents of a device under test in which a sensor arrangement, which has a component comprising a crystal lattice with at least one defect, is arranged in a region of the device under test, the method comprising: using the sensor arrangement to detect a magnetic field formed by a vector of magnetic fields, which in turn are induced by one phase current of a plurality of phase currents of the device under test; and calculating a vector of the plurality of phase currents from the vector of the magnetic fields based on a coefficient matrix.

2. The method as claimed in claim 1, wherein the device under test has three phases and the method further comprises: positioning the sensor arrangement at approximately the same distance to the three phases of the device under test.

3. The method as claimed in claim 1, wherein the at least one defect has a fluorescent effect and the method further comprises: determining the magnetic field using an optically detected magnetic resonance (“ODMR”).

4. The method as claimed in claim 1 wherein the at least one defect induces a magnetic resonance and the method further comprises: determining the magnetic field using a photoelectric detection of the magnetic resonance (“PDMR”) measurement.

5. The method as claimed in claim 1, further comprising: determining the coefficient matrix using a calibration.

6. The method as claimed in claim 1, further comprising: determining the coefficient matrix using a trained neural network (550).

7. A sensor arrangement for measuring phase currents of a device under test, the sensor arrangement comprising: a component comprising a crystal lattice with at least one defect, the component arranged in a region of the device under test and the component is configured to (i) use the sensor arrangement to detect a magnetic field formed by a vector of magnetic fields, which in turn are induced by one phase current of a plurality of phase currents of the device under test, and (ii) calculate a vector of the plurality of phase currents from the vector of the magnetic fields based on a coefficient matrix.

8. The sensor arrangement as claimed in claim 7, wherein the crystal lattice is a diamond.

9. The sensor arrangement as claimed in claim 8, wherein the at least one defect is formed as a nitrogen defect or a nitrogen-vacancy center (“NV center”).

10. The sensor arrangement as claimed in claim 7, further comprising: at least one microwave source.

11. The sensor arrangement as claimed in claim 7, wherein the device under test includes an inverter.

12. The method as claimed in claim 1, wherein the device under test includes an inverter.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0024] FIG. 1 shows busbars of three phases at an output of an inverter.

[0025] FIG. 2 shows nitrogen defects (NV centers) in a diamond.

[0026] FIG. 3 shows an absorption and emission spectrum of the NV center.

[0027] FIG. 4 shows the Zeeman effect within the energy diagram of the negatively charged NV center.

[0028] FIG. 5 shows an optically detected magnetic resonance of a single NV center.

[0029] FIG. 6 shows a crystal unit cell.

[0030] FIG. 7 shows a schematic view of the components required for the sensor unit in one embodiment.

[0031] FIG. 8 shows busbars of three phases at an output of an inverter.

[0032] FIG. 9 shows a neural network for predicting currents.

EMBODIMENTS OF THE INVENTION

[0033] The invention is illustrated schematically in the drawings by reference to embodiments and is described in detail below with reference to the drawings.

[0034] FIG. 1 shows a perspective view of three busbars at the output of an inverter 310, namely a first busbar 312, a second busbar 314 and a third busbar 316. A first C-ring 322 is assigned to the first bus bar, a second C-ring 324 to the second bus bar 314, and a third C-ring 326 to the third bus bar 316. The C-rings 322, 324, 326 act as flux concentrators.

[0035] The method presented uses a magnetic field sensor in the following. An example of a magnetic field sensor is a magnetometer. This is a sensing device for measuring magnetic flux densities. Magnetic flux densities are measured in the Tesla (T) unit. Common magnetometers are, for example, Hall sensors, Förster probes, proton magnetometers, Kerr magnetometers, and Faraday magnetometers.

[0036] In addition to these types of magnetometer, it is also known to use diamonds, in the lattices of which defects or discontinuities are provided, which show a detectable behavior depending on an applied magnetic field. In these, a negatively charged color center consisting of nitrogen and a defect (nitrogen vacancy center, NV center) in a diamond is used for highly sensitive measurements of magnetic fields, electrical fields, mechanical stresses, and temperatures. Reference is made to FIG. 2 in this context.

[0037] The left-hand side of FIG. 2 shows a crystal lattice, in this case a diamond, the crystal lattice as a whole being designated by reference number 10. The crystal lattice 10 comprises a number of carbon atoms 12 and an NV center 14, which in turn has one nitrogen atom 16 and an adjacent defect or vacancy 18. The nitrogen defect 14 is aligned along one of the four possible bond directions in the diamond crystal.

[0038] On the right hand side, the energy level diagram 30 of the negatively charged NV center 14 is shown. A ground state .sup.3A.sub.2 32 is a spin triplet with total spin s = 1. The states 34 with magnetic spin quantum number m.sub.s = +- 1 are energetically shifted compared to state 36 with m.sub.s = 0. A state .sup.3E 38 and an intermediate state 40 are also shown. A microwave frequency of 2.87 GHz, corresponding to a splitting energy or zero-field splitting D.sub.gs, is shown in the bracket 42. The zero-field splitting is an intrinsic variable that is independent of the irradiated MW field or the MW frequency. It is approximately 2.87 GHz and is, in particular, temperature-dependent. The following relationship applies to the determination of the resonance frequency:

[00002]$\text{v}\text{}\pm \approx {\text{D}}_{\text{gs}}+\beta *\Delta \text{T}\pm {\text{y}}_{\text{NV}}*{\text{B}}_0;$

where ΔT indicates the deviation from room temperature, β the temperature-induced shift of the zero-field splitting with β approximately -74.2 kilohertz/Kelvin, y.sub.NV is the gyromagnetic ratio of the NV center, and B.sub.0 is the field strength of an external magnetic field.

[0039] The quantum technologies used in the magnetic field sensors presented offer crucial advantages over classical sensor principles, which underline the disruptive potential of quantum technology. With nitrogen defects, the following specific advantages are obtained: [0040] ultra-high sensitivities (1 pT/√Hz), [0041] vector magnetometry, i.e. a direction determination of the magnetic field is possible, [0042] high measurement range (> 1 Tesla), [0043] linearity (Zeeman effect), [0044] no degradation, as the measurement is based on quantum mechanical states, similar to those in the hydrogen atom, where the Rydberg constant is a fixed energy which is a constant for all atoms independent of location and time, [0045] it is possible to determine external magnetic fields vectorially at the same time based on the four possible spatial directions of the NV axes present in the diamond.

[0046] In order to read out a sensor based on NV centers, the magnetic resonance of the triplet of the ground state is optically detected, see .sup.3A state in FIG. 2 (ODMR, optically detected magnetic resonance). To do this, the NV center must be excited with green light. Reference is made to FIG. 3.

[0047] FIG. 3 shows a graph 50 of the absorption and emission spectrum of the NV center which is shown in FIG. 2. In the graph 50, the wavelength [nm] is plotted on an abscissa 52, the absorption coefficient [cm.sup.-1] is plotted on a first ordinate 54, and the fluorescence is plotted on a second ordinate 56. A first curve 60 shows the absorption spectrum, a second curve 62 shows the emission spectrum. A first arrow 70 indicates NV° ZPL, a second arrow 72 indicates NV.sup.- absorption, and a third arrow 74 indicates NV.sup.- fluorescence. Furthermore, NV.sup.- ZPL 76 at 637 nm is shown.

[0048] FIG. 4 shows the Zeeman effect in the ground state 150 of the NV center. The excited state 152 and the intermediate state 154 are also entered. A first arrow 160 shows a transition with a high probability or transition rate, a dashed arrow 162 shows a transition with a low probability or transition rate. A box 170 shows a transition 172 without a magnetic field and a transition 174 with a magnetic field.

[0049] FIG. 5 shows in a graph 100 the optically detectable magnetic resonance (ODMR) of a single NV center for various background magnetic fields. In the graph 100, the microwave frequency is plotted on an abscissa 102 and the fluorescence is plotted on an ordinate 106.

[0050] A first curve 110 shows the resonance at B = 0, a second curve 112 shows the resonance at B = 2.8 mT with the negative peaks ω.sub.1 114 and ω.sub.2 116, a third curve 120 shows the resonance at B = 5.8 mT and a fourth curve 122 shows the resonance at 8.3 mT.

[0051] FIG. 6 shows a crystal unit cell of diamond, which is labeled as a whole with the reference number 400. The diagram shows a nitrogen atom in the center and the four nearest lattice sites. A vacancy is located at one of these lattice sites, while the other three are occupied by C-atoms. The four possible positions of the vacancy span the space of the four possible directions of the NV center, which are called bonds in the following. There is a first bond a 410, a second bond b 412, a third bond c 414, and a fourth bond d 416. An arrow 450 depicts the magnetic field B.

[0052] On the right-hand side a graph 430 is shown in which the microwave frequency [GHz] is plotted on the abscissa 432 and the fluorescence [arb. units] is plotted on the ordinate 434. The graph 430 shows fluorescence dips, a first dip 440 for the first bond a 410, a second dip 442 for the second bond b 412, a third dip 444 for the third bond c 414, and a fourth dip 446 for the fourth bond d 416.

[0053] In this unit cell 400, the NV center 404 has four possible ways of arranging itself in the crystal. Since the angle between the magnetic field direction and crystal axis is different for each crystal axis, different degrees of frequency splits occur in the NV centers, depending on the respective crystal direction. This results in up to four associated pairs of fluorescence dips in the fluorescence spectrum, as can be seen on the right-hand side of graph 430. By evaluating the resonance frequency for each crystal direction, the magnetic field direction and strength can be extracted.

[0054] FIG. 7 shows a schematic diagram of the components required for the excitation and readout of a magnetometer with a diamond/NV-based sensor arrangement, which is designated as a whole with reference number 200. The illustration shows a laser source 202. Alternatively, an LED can also be used. Also shown are a lens 204, a microwave source 206, a diamond 208 with NV defects, a further lens 210, a long-pass filter (LP) 650 nm 212, a photodetector 214, an analog-to-digital converter 216, and a signal processor 218 which outputs an output 220 with magnetic field, temperature, pressure. For example, the long-pass filter 212 is a distributed Bragg reflector and ensures that the green excitation light is blocked and thus does not strike the photodetector 214. On the other hand, the long-pass filter 212 allows the fluorescent light emitted in the wavelength range > 650 nm, i.e. 650 - 800 nm, to pass.

[0055] The fluorescent light, red-shifted with respect to the excitation light, see FIG. 3, shows a characteristic dip in the energetic position of the electron spin resonance under additional irradiation by an alternating electromagnetic field (microwave), see FIG. 5. Due to the Zeeman effect, see FIG. 4, the position is linearly dependent on the magnetic field, see FIG. 5. This allows a highly sensitive magnetic field sensor to be constructed. FIG. 7 shows all the required components of a color-center-based magnetometer in schematic form.

[0056] Since the NV center in the single-crystalline diamond has four possible ways of arranging itself in the crystal lattice, see FIG. 6, the presence of a directed magnetic field leads causes the NV centers present in the crystal to react to the external magnetic field to different degrees depending on their position in the crystal, i.e. the projection of the magnetic field onto the NV axis is different. This means that in the maximum case four associated pairs of fluorescence minima can appear in the spectrum, from the position of which in the spectrum the magnitude and direction of the magnetic field can be uniquely determined.

[0057] In order to be able to obtain vector information even in the case of weak external magnetic fields, suitable technical measures should be taken to ensure that the contributions of all four possible NV orientations in the diamond crystal can be distinguished in the ODMR spectrum even without a magnetic field acting externally on the sensor. This can be effected by a static magnetic bias field, which is provided within the sensor by appropriate technical measures. However, this bias field should have a field strength that is as homogeneous as possible within the sensitive diamond volume, as inhomogeneities of the bias field affect the sensitivities of the sensor.

[0058] The sensitivity of a diamond magnetometer is calculated as:

[00003]$\eta =\delta \text{B}\sqrt{\Delta t}=\frac{\sqrt{I_0}}{\left(\partial I/\partial \text{B}\right)}\approx \frac{h}{g\text{}{\mu }_B}\frac{\Delta v}{\text{C}\sqrt{I_0}}$

[0059] The input parameters include a noise signal, which in optical sensors is equal to the photon shot noise as a physical limit, as well as the gyromagnetic ratio η.sub.NV, i.e. a material constant, the intensity I.sub.0 of the fluorescence, the contrast C of the ODMR (contrast) and the line width Δν, the FWHM of the optically detectable resonance.

[0060] Since in the sensitive diamond volume any deviation of the bias field from its target value would lead to a different position of the magnetic resonance, a bias-field gradient in the diamond volume leads to a broadening of the resonance and thus a deterioration of the sensitivity.

[0061] For effective splitting of the four NV orientations the strength of the bias field should be between 100 .Math.T and 10 mT, ideally around 1 mT. In order to enable a sensitivity of 1 pT/√Hz, the bias field must not deviate from the target value by more than 1 per mill in the region of the diamond volume, i.e. for 1 mT by a maximum of 1 .Math.T.

[0062] The method presented uses a vector magnetometer based on an ODMR measurement at NV centers in a diamond, which is typically positioned at approximately the same distance to the three phase conductors of an inverter. Reference is made to FIG. 8.

[0063] FIG. 8 shows busbars of the three phases at the output of an inverter with a diamond with NV centers for measuring the vector magnetic field that results from the current flows of the phases. The diagram shows a diamond 500, a first conductor 502 for phase 1, a second conductor 504 for phase 2, and a third conductor 506 for phase 3.

[0064] Reference number 520 shows a surface plot of the B.sub.y component of the magnetic flux density in Tesla, which is induced by the three phase currents, namely I.sub.1 = 1000 A, I.sub.2 = -500 A, I.sub.3 = -500 A.

[0065] From the linear relationship between the electric current and the magnetic flux density induced thereby on the one hand, and the linear superposition of the magnetic fields of individual current elements on the other, the linear relationship specified in (1) between the vector of the phase currents and the vector of the magnetic flux density is obtained, which is conveyed by the matrix of the coefficients b.sub.ij. If the coefficient matrix is known, the vector of the phase currents can be calculated by multiplying B by the inverse matrix, as indicated in (2).

[00004]$\underset{B}{.fwdarw.}=\left(\begin{array}{c}{B}_x\\ B_y\\ B_z\end{array}\right)=\left(\begin{array}{ccc}{b}_{11}& b_{12}& b_{13}\\ b_{21}& b_{22}& b_{23}\\ b_{31}& b_{32}& b_{33}\end{array}\right)\times \left(\begin{array}{c}{I}_1\\ I_2\\ I_3\end{array}\right)$

[00005]$\left(\begin{array}{c}{I}_1\\ I_2\\ I_3\end{array}\right)={\left(\begin{array}{ccc}{b}_{11}& b_{12}& b_{13}\\ b_{21}& b_{22}& b_{23}\\ b_{31}& b_{32}& b_{33}\end{array}\right)}^{\left(-1\right)}\times \left(\begin{array}{c}{B}_x\\ B_y\\ B_z\end{array}\right)$

[0066] The coefficients b.sub.ij are determined according to the method by calibration after the diamond with NV centers is placed at the output of the inverter at an approximately equal distance to the three busbars in the manner described above.

[00006]$\left(\begin{array}{c}{B}_x\\ B_y\\ B_z\end{array}\right)=\left(\begin{array}{c}{b}_{1j}\times I_j\\ b_{2j}\times I_j\\ b_{3j}\times I_j\end{array}\right)$

[00007]$\left(\begin{array}{c}{b}_{1\text{}j}\\ b_{2\text{}j}\\ b_{3\text{}j}\end{array}\right)=\left(\begin{array}{c}{B}_x/I_j\\ B_y/I_j\\ B_z/I_j\end{array}\right)$

[0067] For this purpose, only one phase is supplied with current at a time, while no current flows in the other two phases. With j as the index of the energized phase, equation (3) then applies, or (4) after solving for b.sub.1j, b.sub.2j and b.sub.3j. By means of three calibration measurements, j = 1, 2, 3, all nine coefficients of b.sub.ij are therefore known. These calibration coefficients are stored in the sensor and used to calculate the individual currents of the three phases of the inverter (I.sub.1, I.sub.2, I.sub.3) from the vector magnetic field B measured in the diamond by ODMR using equation (2).

[0068] As an alternative to the procedure described here, the coefficients b.sub.ij can also be determined as part of the calibration by determining the vector of the magnetic flux density B.sub.x, B.sub.y, B.sub.z for a plurality of arbitrary triples I.sub.1, I.sub.2, I.sub.3 and training a neural network on the basis of this data, which then determines the existing triple of the currents through the phases of the inverter from the measured vector of the magnetic flux density B.sub.x, B.sub.y, B.sub.z when the sensor is applied in the field. Reference is made to FIG. 9.

[0069] FIG. 9 shows a neural network 550 for predicting the currents through the three phases of the inverter 552, namely I.sub.1, I.sub.2, I.sub.3, from the vector components of the magnetic flux densities, namely B.sub.x, B.sub.y, B.sub.z, at the position of the diamond 554 with NV centers.

[0070] Another component of at least one embodiment of the presented method is the measurement of the magnetic field in the diamond with NV centers based on the photoelectric detection of the magnetic resonance (PDMR) of the NV centers in diamond. PDMR measures an electrical current, a photocurrent, instead of measuring the intensity of the fluorescent light. For this purpose, for example, an electrode structure consisting of an electrode and a counter electrode is applied on one side of the diamond, e.g. in the form of an interdigital structure made of a conductive material, e.g. metal. When an electrical voltage is applied, a photoelectric current flows when the diamond is excited with green light. If the microfield acting simultaneously on the diamond hits a magnetic resonance of the NV centers at approximately 2.87 GHz, plus or minus the frequency splitting by the magnetic field and the Zeeman effect, then the photocurrent collapses. The resonance can be determined from the measurement of the photocurrent and the magnetic field can be determined by means of the gyromagnetic ratio of the NV center.