POSITION DETECTION BY AN INDUCTIVE POSITION SENSOR

Abstract

For an easily implementable method for position determination using an inductive position sensor with increased precision of the position information, the position sensor generates a measurement signal from which a frequency functional dependent on the excitation frequency is formed, which represents a measure of the noise signal and the excitation frequency of the excitation signal is changed so that the frequency functional is minimized or maximized and the excitation frequency that minimizes or maximizes the frequency functional is used for the excitation signal.

Claims

1. A method for determining the position of a moving part relative to a stationary part by an inductive position sensor, wherein an excitation winding is arranged on one of the moving part or the stationary part and at least one secondary winding is arranged on the other of the moving part or the stationary part, the method comprising: feeding an electrical excitation signal with an excitation frequency and an excitation amplitude into the excitation winding, which generates an electromagnetic excitation field; detecting a measurement signal induced in the secondary winding by the generated electromagnetic excitation field, wherein the measurement signal includes a superimposed noise signal, forming a frequency functional which is dependent on the excitation frequency and represents a measure of the noise signal on the measurement signal, and changing the excitation frequency of the excitation signal so that the frequency functional is minimized or maximized and using the changed excitation frequency that minimizes or maximizes the frequency functional for the excitation signal, wherein the position of the moving part of the position sensor is determined from the measurement signal.

2. The method according to claim 1, wherein the frequency functional is used to examine an amplitude of the measurement signal with regard to a deviation from an expected signal curve of the measurement signal.

3. The method according to claim 1, wherein an amplitude information of the amplitude of the measurement signal is derived from the measurement signal and the frequency functional is a function of the amplitude information.

4. The method according to claim 3, wherein the measurement signal is demodulated in order to determine the amplitude information.

5. The method according to claim 3, wherein the measurement signal is demodulated with the excitation oscillation of the excitation signal in order to determine the amplitude information.

6. The method according to claim 5, wherein the measurement signal is demodulated with the excitation oscillation of the excitation signal in order to determine an I component of the measurement signal representing the amplitude.

7. The method according to claim 5, wherein the measurement signal is demodulated with the 90° out-of-phase excitation oscillation of the excitation signal, in order to determine a Q component of the measurement signal representing the phase.

8. The method according to claim 3, wherein a statistical variance of the amplitude information is used as the frequency functional.

9. The method according to claim 8, wherein the statistical variance is determined as the expected square deviation of the value of the amplitude information from the expected value of the amplitude information, wherein an arithmetic mean of the amplitude information is preferably used as the expected value.

10. The method according to claim 6, wherein a phase shift of the excitation signal is set so that the I components of the measurement signal are at a maximum.

11. The method according to claim 1, wherein the excitation amplitude is regulated to a predetermined amplitude setpoint.

12. An evaluation unit for an inductive position sensor having a moving part with one of an excitation winding or at least one secondary winding and a stationary part having the other of the one of an excitation winding or at least one secondary winding, in which an electrical excitation signal with an excitation frequency and an excitation amplitude is fed into the excitation winding, which generates an electromagnetic excitation field that induces a measurement signal in the secondary winding which is fed to the evaluation unit with a superimposed noise signal on the measurement signal, the evaluation unit, which determines a position of the moving part relative to the stationary part of the position sensor from the measurement signal, comprises: at least one memory and at least one processor configured to execute at least one set of instructions stored in the at least one memory to: form a frequency functional, which is dependent on the excitation frequency and represents a measure of the noise signal, from the measurement signal, and determine the excitation frequency of the excitation signal which minimizes or maximizes the frequency functional and the excitation frequency which minimizes or maximizes the frequency functional is used when using the evaluation unit for the excitation signal.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The present invention is described in greater detail in the following with reference to FIG. 1 to 7, which show advantageous embodiments of the invention by way of example, schematically, and in a non-limiting manner. In the drawings:

[0027] FIG. 1 and FIG. 2 show the measuring principle of a resolver as an inductive position sensor;

[0028] FIG. 3 shows a frequency spectrum of a noise signal;

[0029] FIG. 4 shows frequency tuning of the excitation frequency of the excitation signal according to the invention;

[0030] FIG. 5 shows phase tuning according to the invention;

[0031] FIG. 6 shows amplitude tuning according to the invention; and

[0032] FIG. 7 shows an advantageous implementation of an evaluation unit with frequency tuning, phase tuning and amplitude tuning.

DETAILED DESCRIPTION

[0033] The particulars shown herein are by way of example and for purposes of illustrative discussion of the embodiments of the present invention only and are presented in the cause of providing what is believed to be the most useful and readily understood description of the principles and conceptual aspects of the present invention. In this regard, no attempt is made to show structural details of the present invention in more detail than is necessary for the fundamental understanding of the present invention, the description taken with the drawings making apparent to those skilled in the art how the several forms of the present invention may be embodied in practice.

[0034] FIGS. 1 and 2 show the principle of an inductive position sensor 1 using the example of a conventional resolver. The resolver comprises an excitation winding EW, which is excited with an excitation signal ES. The excitation signal ES is generated with a generating unit 3, for example an electrical or electronic circuit. The excitation signal ES is an electrical alternating signal with a specific excitation amplitude R0 and excitation frequency ω, i.e., for example ES=R0.Math.cos(ωt), where t denotes the time. If the excitation winding EW moves relative to a secondary winding SW, an electrical voltage is induced in the secondary winding SW which is dependent on the position P (in this example the angle φ of the secondary winding SW relative to the excitation winding EW. In a resolver, for example, two secondary windings SW are arranged offset from one another by 90° (FIG. 2) and the excitation winding EW rotates in the resolver. However, only one secondary winding SW would be sufficient. In the case of a linear inductive position sensor 1, a linear relative movement would be provided between the excitation winding EW and the secondary winding SW, it also being possible for a plurality of secondary windings to be arranged one behind the other in the direction of movement. The voltage induced in the at least one secondary winding SW is output as a measurement signal MS of the inductive position sensor 1 and can be evaluated in an evaluation unit 2 in order to determine a position P of the moving part of the inductive position sensor 1, usually the excitation winding EW.

[0035] In the case of a resolver having two secondary windings SW, the measurement signal MS comprises two measurement signal tracks A, B. Due to the offset arrangement of the secondary windings SW, the measurement signal tracks A, B have a specific phase offset.

[0036] The excitation winding EW is arranged in an inductive position sensor 1 on a moving part, for example on a motor shaft of an electric motor, and the at least one secondary winding SW is on a stationary part, for example on a housing of the position sensor 1, which in turn can be arranged on a motor housing. However, this arrangement can also be reversed. The moving part and the stationary part are arranged to be movable relative to one another.

[0037] With a resolver as the inductive position sensor 1 and an excitation signal ES=R0.Math.cos(φ), with φ=ωt, and a 90° offset arrangement of the secondary windings SW, the measurement signal tracks A, B in the measurement signal MS result in, for example

A=R0.Math.u.Math.cos(ρ).Math.cos(φ−Δφ)

B=R0.Math.u.Math.sin(ρ).Math.cos(φ−Δφ)

[0038] Herein, u denotes a known transfer ratio of the inductive position sensor 1 and Δφ denotes a delay which results substantially from a propagation time of the inductive position sensor 1 and the processing of the measurement signal MS in the evaluation unit 2 (e.g. by filters and the like). The delay Δφ follows as Δφ=ωΔt+φ.sub.V, with a delay φ.sub.V resulting from the processing, and the propagation time Δt. ρ denotes the angular position of the excitation winding EW relative to the secondary windings SW (indicated in FIG. 2) and thus the position P of actual interest.

[0039] In other arrangements of an inductive position sensor 1, the measurement signal MS can of course also be different, also with more or fewer measurement signal tracks, but the measurement signal MS is always dependent on the position P, i.e. MS(P). The measurement signal MS comprises at least one measurement signal track A, B.

[0040] This principle of an inductive position sensor 1 is well known. It is also known that a noise signal RS is usually superimposed on the measurement signal MS, as indicated in FIG. 1. The noise signal RS is usually not known and can comprise different frequency bands. The noise signal RS can be white noise WR, for example, which also comprises narrow interference bands SB in specific frequency ranges. This is shown as an example in FIG. 3 by a frequency spectrum S(f) (frequency f) of the noise signal RS. These narrow-band interference bands SB around characteristic frequencies f.sub.P, 2f.sub.P can be coupled into the measurement signal MS from the environment (e.g. inductively). Of course, the noise signal RS can also have a different frequency spectrum S(f). The noise signal RS largely determines the signal-to-noise ratio of the measurement signal MS. In order to be able to extract the useful signal from the measurement signal MS (e.g. the measurement signal tracks A, B), the greatest possible signal-to-noise ratio is advantageous, i.e. a large distance between the actual useful signal in the measurement signal MS and the noise signal RS.

[0041] The basic idea of the invention is to use a frequency functional F.sub.f that is dependent on the excitation frequency ω as a measure of the noise signal RS in the measurement signal MS and to change the excitation frequency ω so that the frequency functional F.sub.f is optimized, which, depending on the formulation of the frequency functional F.sub.f, corresponds to a maximization or minimization. The excitation frequency ω, which minimizes or maximizes the frequency functional F.sub.f, is then used in the excitation signal ES for the operation of the inductive position sensor 1.

[0042] The frequency functional F.sub.f as a measure of the noise signal RS in the measurement signal MS thus makes it possible to obtain information on the noise signal RS in the measurement signal MS. The measure is, for example, a variable that compares the noise signal RS with the measurement signal MS or places the noise signal RS in relation to the measurement signal MS. The frequency functional F.sub.f can be formulated mathematically as a mathematical function of the excitation frequency ω.

[0043] The frequency functional F.sub.f can of course be formulated in a wide variety of ways, for example as a signal-to-noise ratio if the noise signal RS can be separated from the measurement signal MS, for example by filtering. A frequency functional F.sub.f can also be used to examine the amplitude of the measurement signal MS (specifically at least one measurement signal track) with regard to a deviation from the expected signal curve of the measurement signal MS. This deviation is a measure of the noise signal RS. In the case of the resolver, for example, this would be a modulated sine wave, with noise being noticeable as a deviation from this modulated sine wave. For this purpose, for example, a statistical variance of the amplitude of the measurement signal could be used as a frequency functional F.sub.f as a measure of this deviation, and thus as a measure of the noise signal RS. For this purpose, amplitude information FA of the amplitude of the measurement signal MS can be derived from the measurement signal MS, or from at least one measurement signal track of the measurement signal MS, which is then evaluated in the frequency functional F.sub.f. The amplitude information FA can be dependent on the excitation frequency ω. The frequency functional F.sub.f is therefore a function of the amplitude information FA, i.e. F.sub.f(FA(ω)), and is therefore dependent on the excitation frequency ω.

[0044] A particularly advantageous embodiment of the invention is explained with reference to FIG. 4. In this method, the at least one measurement signal track A, B is demodulated in order to obtain the useful signal, i.e. the variable that oscillates as a function of the position P, for example cos(ρ) or sin(ρ). The amplitude information FA of the measurement signal MS can be obtained from such a demodulation, from which a frequency functional F.sub.f can then be formed as described above. Since the excitation signal ES and the measurement signal MS usually have a phase shift due to a time delay Δφ, the well-known IQ method (in-phase & quadrature method) is suitable for demodulation, since it can be used to obtain both the amplitude information and the phase information. For demodulation, the measurement signal MS is multiplied by the excitation oscillation cos(ωt) in order to obtain the I component that represents the amplitude. If the measurement signal MS is optionally also multiplied by the excitation oscillation that is 90° out of phase, the phase information (Q component) is also obtained. In the case of a resolver, this can be expressed mathematically in the form of the periodic scalar products, with at least one measurement signal track A, B generally being present.

[00001]$A_I=\left(A|\cos \left(\phi \right)\right)=\frac{1}{2}R0.Math.u.Math.\cos \left(\rho \right).Math.\cos \left(\mathrm{\Delta \phi }\right)B_I=\left(B|\cos \left(\phi \right)\right)=\frac{1}{2}R0.Math.u.Math.\sin \left(\rho \right).Math.\cos \left(\mathrm{\Delta \phi }\right)\mathrm{And}\mathrm{optionally}\mathrm{or}\mathrm{as}\mathrm{needed}{A}_Q=\left(A|\sin \left(\phi \right)\right)=\frac{1}{2}R0.Math.u.Math.\cos \left(\rho \right).Math.\sin \left(\mathrm{\Delta \phi }\right)B_Q=\left(B|\sin \left(\phi \right)\right)=\frac{1}{2}R0.Math.u.Math.\sin \left(\rho \right).Math.\sin \left(\mathrm{\Delta \phi }\right)$

[0045] The I component A.sub.I, B.sub.I contains the amplitude information FA and can also be squared for further use FA=A.sub.I.sup.2, or alternatively the absolute value FA=|A.sub.I| could also be used. If there is a plurality of measurement signal tracks A, B, the sum of the squares FA=A.sub.I.sup.2+B.sub.I.sup.2 of the I components A.sub.I, B.sub.I or, alternatively, the sum of the absolute values FA=|A.sub.I|+|B.sub.I| can also be used. In principle, other arithmetic combinations of the I components are also conceivable.

[0046] A statistical variance V of the amplitude information FA, for example, which is dependent on the excitation frequency ω, can then be used as the frequency functional F.sub.f. The variance V can be specified as the expected quadratic deviation (i.e. the expected value E) of the value of the amplitude information FA from its expected value E(FA). The arithmetic mean can be used as the expected value E(FA). This can be expressed mathematically in the form

F.sub.f=V(FA)=E((FA−E(FA)).sup.2).

[0047] A plurality N of values of the amplitude information FA are preferably used to determine the variance V, which can be obtained by sampling the measurement signal MS with a predetermined sampling frequency (e.g. in the megahertz range). With a typical excitation frequency of 10 kHz, a sampling frequency of 1 MHz would result in a hundredfold oversampling. The variance V can thus be expressed as a short-term variance in the form

[00002]$F_f=V\left(FA\right)\approx \frac{1}{N-1}\underset{i=0}{\overset{N-1}{.Math.}}{\left({\mathrm{FA}}_i-\frac{1}{N}\underset{i=0}{\overset{N-1}{.Math.}}{\mathrm{FA}}_i\right)}^2.$

[0048] For example, 1,000 to 10,000 periods of the excitation signal ES are used for the determination, which would lead to an observation period for determining a value of the variance V of 100 ms to 1 s.

[0049] The frequency functional F.sub.f is now optimized with regard to the excitation frequency ω, which can be expressed mathematically by

[00003]$\omega =\underset{\omega }{\underbrace{\min }}\left(F_f\right)\mathrm{or}\omega =\underset{\omega }{\underbrace{\max }}\left(F_j\right).$

The excitation frequency ω is thus varied until the functional F.sub.f is optimized (in the sense of a minimum or maximum). The excitation frequency ω is varied within a specified frequency range.

[0050] In order to solve such optimization problems, there is a wealth of known solution algorithms, such as the gradient method, the Newton method, evolutionary methods or sequential quadratic programming, to name just a few. However, the choice of the solution algorithm is irrelevant for the invention, although it is of course advisable to choose a method that is advantageous in terms of computing effort and computing time. What the solution methods have in common is that possible solutions to the optimization problem are sought, usually iteratively, until a defined termination criterion is reached. The termination criterion can be, for example, a number of iterations, or the difference between the solutions of two successive iteration steps of the optimization problem falling below a limit value, or another termination criterion. The solution (i.e. excitation frequencies ω) in each iteration step is selected using the specified rules of the solution method, wherein a suitable choice of the solution can be specified as the starting value in the first iteration step. In the gradient method, for example, the gradient of the functional is determined (derivation of the functional with respect to the excitation frequency) and the manipulated variable for the next iteration step is selected along this gradient, with the increment from the current manipulated variable to the next manipulated variable being determined by the specified rules of the solution method.

[0051] In this way, the excitation frequency ω can be determined, with which the inductive position sensor 1 must be excited so that the influence of the noise signal RS is minimal. Narrow-band interference bands in particular can be avoided in this way. If the excitation frequency ω were selected, for example, in the range of the frequency f.sub.P (or an integral multiple thereof) (FIG. 3), then the signal-to-noise ratio of the measurement signal MS would drop compared to an excitation frequency ω away therefrom.

[0052] A possible implementation of a frequency tuning according to the invention as described above is shown in FIG. 4. The example again relates to a resolver with at least one measurement signal track A in the measurement signal MS.

[0053] The measurement signal track A is first demodulated and the I component A.sub.I is determined, which is squared in order to determine the amplitude information FA. This can be carried out in a frequency tuner FT. A second measuring signal track B, as is usual with resolvers, is indicated by dashed lines in FIG. 4. In this example, the amplitude information FA would then be the sum of the squared I components A.sub.I, B.sub.I of the measurement signal tracks A, B. The variance V(FA) of the amplitude information FA is used as the frequency functional F.sub.f, for example as a short-term variance as described above. The frequency functional F.sub.f is minimized with regard to the excitation frequency ω and the inductive position sensor 1 is then operated with this excitation frequency ω. This ensures the greatest possible signal-to-noise ratio in the respective application with the prevailing interference environment.

[0054] Low-pass filters LPF are also indicated in FIG. 4 after the determination of the I components A.sub.I, B.sub.I. Such a low-pass filter LPF could be used in order to suppress the double excitation frequencies ω that result during demodulation.

[0055] This frequency tuning could be carried out once when the inductive position sensor 1 is put into operation and the excitation frequency ω could then be left the same, assuming an unchanging interference environment. It would only have to be ensured that the sources of interference in the environment are active during the frequency tuning.

[0056] However, the frequency tuning could also be carried out continuously during operation of the inductive position sensor 1, for example at predetermined time intervals.

[0057] Ongoing frequency tuning could, for example, function in such a way that an ongoing spectral analysis of the at least one measurement signal track A, B is carried out, for example by means of an FFT (fast Fourier transformation) and the noise power density of the noise signal RS is quantified in the frequency spectrum as a frequency functional F.sub.f. The minimum of the noise power density can be identified and the excitation frequency W can then be set at the point with the minimum noise power density.

[0058] The frequency tuning according to the invention could also be combined with phase tuning. It can thus be achieved that the I components A.sub.I, B.sub.I of the demodulated measurement signal tracks A, B are at a maximum (which is advantageous for the frequency tuning) and that the Q components A.sub.Q, B.sub.Q disappear. This can be carried out, for example, with a phase controller R.sub.φ that controls a phase shift Δφ of the excitation signal ES, so that the Q components A.sub.Q, B.sub.Q disappear.

[0059] This can be implemented with a resolver, for example, by using a phase functional F.sub.φ=A.sub.IA.sub.Q+B.sub.IB.sub.Q, again using the scalar products.

[0060] This phase functional F.sub.φ is independent of the position ρ and results in

[00004]$F_{\phi }=\frac{1}{8}{R0}^2{u}^2\sin \left(2\mathrm{\Delta \phi }\right).$

This function has four zero points in the interval Δφ∈[0,2π], in two of which (at π/2, 3π/2) the I components A.sub.I, B.sub.I disappear and the Q components A.sub.Q, B.sub.Q are at a maximum and in the other two zeros (at 0, π) the Q components A.sub.Q, B.sub.Q disappear and the I components A.sub.I, B.sub.I are at a maximum.

[0061] A phase controller R.sub.φ can now control the phase functional F.sub.φ to zero, i.e. F.sub.φ=0, with the phase controller R.sub.φ ensuring that the phase shift Δφ locks at 0 or π. The phase controller R.sub.φ can be designed, for example, as a known I controller (integral controller).

[0062] The phase tuner PT can be implemented as shown in FIG. 5. The phase setpoint F.sub.φset is set to zero so that the phase shift Δφ is controlled by the phase controller R.sub.φ, so that the phase functional F.sub.φ becomes zero. The phase shift Δφ is then set in the excitation signal ES.

[0063] The frequency tuning according to the invention could also be combined with amplitude tuning, optionally also in combination with phase tuning. The idea behind the amplitude tuning is that the excitation amplitude R0 of the excitation signal ES is set to a maximum value. This maximum value depends on the implementation of the generating unit 3 for the excitation signal ES. For example, the amplitude R0 can be set to a value such that the full linearity range of an electronic circuit as generating unit 3 is used. This can be done, for example, with an amplitude controller R.sub.A that controls the excitation amplitude R0 to a predetermined setpoint.

[0064] This can be implemented with a resolver, for example, by using an amplitude functional F.sub.A=A.sub.I.sup.2+B.sub.I.sup.2+A.sub.Q.sup.2+B.sub.Q.sup.2. If the phase tuning described above is also used, the terms with the Q components disappear in the amplitude functional F.sub.A. This amplitude functional F.sub.A is independent both of the position p and of the phase shift Δφ and, with the above definitions of the I and Q components, results in

[00005]$F_A=\frac{1}{4}R0.Math.u.$

[0065] An amplitude controller R.sub.A can now control the amplitude functional F.sub.A to a predetermined amplitude setpoint F.sub.Aset. In this case, the amplitude setpoint F.sub.A, can result from an implementation of the generating unit 3. The amplitude controller R.sub.A can be designed, for example, as a known I controller (integral controller).

[0066] The amplitude tuner AT can be implemented as shown in FIG. 6. The excitation amplitude R0 is then set in the excitation signal ES.

[0067] FIG. 7 shows an advantageous position determination with a resolver as an inductive position sensor 1. The resolver is excited with an excitation signal ES with an excitation frequency ω, a phase shift Δφ and an excitation amplitude R0. The measurement signal tracks A, B of the measurement signal MS output by the resolver are evaluated in an evaluation unit 2. The measurement signal tracks A, B are demodulated by means of an IQ method, and the I components A.sub.I, B.sub.I as well as the Q components A.sub.Q, B.sub.Q are determined. The frequency functional F.sub.f, the phase functional F.sub.φ and the amplitude functional F.sub.A are determined from the I and Q components A.sub.I, B.sub.I, A.sub.Q, B.sub.Q. The excitation frequency ω is then set using a frequency tuner FT, in which the frequency tuning according to the invention described above is implemented. Likewise, the phase shift Δφ of the excitation signal ES is set using a phase tuner PT and the excitation amplitude R0 is set using an amplitude tuner AT.

[0068] Of course, the position P, in this case an angle ρ, can also be determined from the I components A.sub.I, B.sub.I (or optionally from the Q components A.sub.Q, B.sub.Q), for example with the a tan 2 function.

[0069] The evaluation unit 2 is preferably implemented digitally, as software on microprocessor-based hardware. Alternatively, evaluation unit 2 can be physically implemented by electronic (or optical) circuits such as logic circuits, discrete components, microprocessors, hard-wired circuits, memory elements, wiring connections, and the like, which may be formed using semiconductor-based fabrication techniques or other manufacturing technologies. Further, the blocks and/or modules depicted in the figures can be formed by analog instrumentation, e.g., analog electric/electronic circuits, analog computers, analog devices, etc., and/or can be formed as application specific integrated circuits (ASICs) or other programmable integrated circuits, and, in the case of the blocks and/or modules, which can be implemented by microprocessors or similar, they may be programmed using software (e.g., microcode) to perform various functions discussed herein and may optionally be driven by firmware and/or software. Alternatively, each block and/or module may be implemented by dedicated hardware, or as a combination of dedicated hardware to perform some functions and a processor (e.g., one or more programmed microprocessors and associated circuitry) to perform other functions.

[0070] In a digital implementation, the measurement signal tracks A, B output by the position sensor 1 are converted from analog to digital using suitable analog/digital converters (ADC). Likewise, the excitation signal ES can be generated digitally and applied to the position sensor 1 by a digital/analog converter (DAC). The position value P can also be output in analog or digital form.

[0071] At least one memory (not shown). e.g., a non-transitory computer readable medium or media, can be provided to store one or more sets of instructions to perform any of the methods or computer-based functions disclosed herein, either alone or in combination with the other described devices. The at least one memory, accessible by the processors, can be part of the resolver or remote from the resolver, e.g., a remotely located server, memory, system, or communication network or in a cloud environment.

[0072] It is noted that the foregoing examples have been provided merely for the purpose of explanation and are in no way to be construed as limiting of the present invention. While the present invention has been described with reference to an exemplary embodiment, it is understood that the words which have been used herein are words of description and illustration, rather than words of limitation. Changes may be made, within the purview of the appended claims, as presently stated and as amended, without departing from the scope and spirit of the present invention in its aspects. Although the present invention has been described herein with reference to particular means, materials and embodiments, the present invention is not intended to be limited to the particulars disclosed herein: rather, the present invention extends to all functionally equivalent structures, methods and uses, such as are within the scope of the appended claims.