Method and device for measuring the mechanical angular position of a rotor

Abstract

The invention relates to a method and a device for measuring the mechanical angular position of a rotor, comprising, during a calibration phase and then during a setting phase: - acquisition of Ns measurement signals at measurement positions offset by an angle of 2Pi / Nc radians from the corresponding mechanical angle, the measurement positions being offset by Pi / (2xNc) radians for Ns=2 and by 2Pi / 3Nc radians for Ns=3; - calculation of an instantaneous electrical angular position value, the instantaneous electrical angular position value taking the inverse tangent of the ratio of the values of the two measurement signals for the time considered, or the inverse tangent of the ratio of the values of the two transforms obtained by applying a Clarke transform to the three measurement signals; - determination of an electrical calibration signature (SIGcb) and an electrical setting signature (SIGi); determination of an angular measurement offset value by a reset signature operation.

Description

Technical Field

[0001] This invention relates to the field of measuring the mechanical angular position of a rotor relative to a stator when the rotor's angular travel relative to the stator is greater than 360 degrees (i.e., 2Pi radians) around the axis of rotation of the rotor relative to the stator.

[0002] This problem arises in many applications. For example, one might want to know the mechanical angular position of the rotor of an electric motor relative to its stator. One might also want to know the mechanical angular position of the crankshaft or camshaft of an internal combustion engine, which is the rotor, or any rotating accessory of such a system, or the drive screw of a screw-nut system. Background Technology

[0003] To measure the mechanical angular position of a rotor relative to a stator, many known sensor systems include a detector and a target, the detector being able to transmit a signal representing the target's position. According to a first type of technology, the target is a magnetic target, and the detector includes one or more measuring units capable of measuring the magnetic field or changes in the magnetic field generated by the target near the measuring unit. According to another type of technology, the target is a metallic target, and the detector includes a sensing element and means for measuring the induced changes caused by the target or its relative displacement with respect to the detector.

[0004] In all cases, such a sensor is used by attaching a target to one or the other of the rotor or stator, and by attaching a detector to the other of the rotor and stator. It should be noted that in the remainder of this document, for the purpose of describing exemplary embodiments, the rotor is considered a component movable relative to a general reference frame, and the stator is a component fixed relative to the same general reference frame. However, since it involves relative position measurement, it is not important which one is movable and which one is fixed. In this way, from the perspective of the scope of the invention, the rotor may be fixed relative to a given reference frame, and the stator may be movable relative to that given reference frame. Alternatively, both the rotor and the stator may be movable relative to a given reference frame.

[0005] To measure angular position, a known device simultaneously and continuously acquires an integer Ns electrical measurement signals, each representing the intensity of an electrical or magnetic variable at one of the Ns measurement positions. The Ns measurement positions are separate and fixed relative to the stator. They are offset by a given mechanical angle around the axis of rotation of the rotor relative to the stator. Typically, Ns is an integer equal to 2 or 3.

[0006] Typically, each measurement at each measurement location is performed by one or more measurement units. Therefore, to measure electrical variables at two separate locations, at least two measurement units are required, each capable of including one or more measurement elements. In cases where a measurement unit includes multiple measurement elements, each unit is associated with a specific measurement location, and the two measurement units corresponding to the two specific measurement locations can share one or more measurement elements. However, for reasons of compactness, ease of assembly, and integration of sensors into systems comprising rotors and stators, it is advantageous to reduce the number of measurement locations, and especially to contain the measurement locations as finitely as possible within an angular sector around the rotor's axis of rotation.

[0007] In these devices, the change in the electrical measurement signal at the measurement position is caused by the rotation of a target mechanically connected to the rotor in front of the measurement position under consideration. To provide sufficient accuracy over 360-degree angles (i.e., 2Pi radians) and larger angular travels, and also to allow for the geographical inclusion of individual measurement positions within a limited angular sector (which benefits sensor compactness), the target has an integer Nc of greater than or equal to 2 contrast regions. Within each contrast region, the target exhibits a contrast in conductivity, permeability, and / or magnetization. The term "contrast" should be understood as referring to the change in the electromagnetic properties of the target's conductivity, permeability, and / or magnetization according to the angular position of the target in question about the axis of rotation. This change can be a binary change, a continuous change, a step change, etc. It is the presence of this contrast that produces the change in the measurement signal according to the change in the angular position of the target relative to the measurement position as the contrast region passes in front of the detector.

[0008] The contrast regions Zk are distributed angularly in a periodic pattern around the axis of rotation on the target, with an integer Nc pattern period. The pattern is considered periodic in the sense that all contrast regions have an equal angular range around the axis of rotation and that each contrast region extends around the axis of rotation at a mechanical angle of 2Pi / Nc radians. Each contrast region Zk can be identified by a class k (an integer contained in the range from 1 to Nc, corresponding to the order of the mechanical angular positions of the contrast regions relative to other contrast regions around the axis of rotation). Therefore, class k corresponds to the physical arrangement of the contrast regions on the target. Thus, as the target passes in front of the considered position, one sees the contrast region of class 1 passing through, then the contrast region of class 2, then the contrast region of class 3, and so on until the contrast region of class Nc, which, after a mechanical angular position of 2Pi radians of the rotor, will again be a contrast region of class 1. Of course, if the rotor rotates in the opposite direction, the order of passage will also be reversed. Finally, as will be understood below, the first contrast region seen by the detector at the start of a measurement session does not necessarily have to be a contrast region of class 1.

[0009] In some devices, each contrast region Zk causes the passage of the contrast region of the target in front of each measurement position during rotation about the axis of rotation to result in a change in the electrical measurement signal acquired at that position. This change is a sine or quasi-sine function of the electrical angular position θe of the contrast region relative to the measurement position. The electrical angular position θe can be calculated from the acquired signal, for example, expressed in radians. Therefore, for an angular change of 2Pi / Nc radians in the mechanical angular position θm of the target about the axis of rotation, the electrical angular position θe changes by 2Pi radians. A number of Nc contrast regions Zk generate a sine or quasi-sine period of measurement signals S1(θe), S2(θe) equal to the number of contrast regions Nc over one mechanical revolution of the stator.

[0010] Based on such a device, and by utilizing the sinusoidal or quasi-sinusoidal properties of the signal according to the relative angular position of the rotor with respect to the stator, the electrical angular position θe of the rotor relative to the stator can be easily determined with sufficient accuracy within the angular range of the contrast region and thus within the period or quasi-period of the signal. However, for a device with two or more periodic contrast regions for the target, it is impossible to know the mechanical angular position θm of the rotor at a mechanical angle of 2Pi radians relative to the stator. In fact, due to the periodicity of the contrast regions, the number Nc of mechanical angular positions θm of the rotor is the same as the number Nc of contrast regions, where the electrical angular positions θe that can be calculated based on the acquired signal are equal. In fact, the periodicity of the contrast regions implies that there are Nc positions of the target that produce the same electromagnetic conditions in front of the detector.

[0011] Document US 2019 / 056251 describes a device in which the measured angular position is corrected to compensate for a non-sinusoidal signal caused by geometric errors. The correction is based on a scaling factor and uses a lookup table with predetermined data from tests and simulations.

[0012] Document WO 2020 / 006659 describes a method for obtaining accurate absolute position measurements by using an absolute but imprecise initial measurement and an accurate but not absolute measurement.

[0013] If one is only interested in changes in the rotor's angular position (e.g., the rotor's angular velocity), or when the system associated with the rotor and stator is itself entirely periodic, with the number of cycles being the same as the number of contrast regions Nc of the target, such a sensor can be perfectly satisfactory.

[0014] However, sometimes it is necessary to know the mechanical angular position of the rotor relative to the stator, that is, to obtain a bijective relationship between the value representing the position transmitted by the sensor and the actual angular position of the rotor relative to the stator at a mechanical angle exceeding 2Pi radians. By comparison, the electrical angular position θe defined above only makes it possible to establish a bijective relationship between the value representing the position transmitted by the sensor and the actual angular position in an angular sector with an angular size equal to the angular range of a comparison region.

[0015] To achieve this, document EP-2.385.353 describes a sensor that implements some of the features described above. This sensor belongs to the category of magnetic sensors, where the target comprises magnetic elements, and in this case, the detector comprises two units for measuring the magnetic field, such as Hall effect units. Therefore, the sensor has a target with contrast regions. In the example described in this document, the contrast region is formed by two juxtaposed basic magnets, one of which presents its north pole in the direction of the measuring unit, while the other presents its south pole. Thus, the orientation difference of the basic magnets in the same contrast region produces the contrast within the sense of the invention. All contrast regions formed by the two juxtaposed basic magnets have the same angular range around the rotor, which is 20° in the example of this document. However, the document specifies that the difference between each contrast region and its adjacent contrast region lies in the different ratio of the corresponding angular ranges of the two basic magnets. For some contrast regions, this ratio appears to be approximately one to one, while for others, the ratio appears to be approximately one to four or even one to five. However, it should be understood that within the same contrast region, when the first fundamental magnet passes in front of the measuring unit, it will generate a signal that can be described as predominantly positive, while when the second fundamental magnet passes in front of the same measuring unit, it will generate a signal that can be described as predominantly negative. Therefore, the measured physical variable (here, the magnetic field) will exhibit a first positive half-alternation and a second negative half-alternation. In the case of the apparatus described in this document, the characteristic is that even though the sum of the durations of the two half-alternations within the same contrast region is constant for all contrast regions, the width of each half-alternation varies from one contrast region to another. In other words, the signal generated by a given contrast region exhibits two half-alternations, and for some contrast regions, these two half-alternations have individual durations that can be as high as a ratio of one to four or one to five.

[0016] In document EP-2.385.353, the detector comprises two measuring units spaced 10 degrees apart by a mechanical angle, which corresponds to half the angular range of a contrast region. The document appears to state that the sum of the signals recorded by the two measuring units varies according to a law that, if the ratio of the angular range of the basic magnet in the continuous contrast region is appropriately chosen, varies as a sine function over a mechanical angle of 2Pi radians around the rotor axis. This is based on the document's... Figure 2 The curve shown is marked with reference numeral 22 in the figure.

[0017] However, the device described in the document is particularly complex and expensive to manufacture because it requires creating basic magnets of different sizes for each contrast region and assembling them in a strict order.

[0018] Therefore, the object of the present invention is to create a measurement method that is particularly easy to implement with conventionally constructed devices to measure the mechanical angular position of the rotor relative to the stator at a 2Pi radian angle. Summary of the Invention

[0019] This invention relates to a method for measuring the mechanical angular position of a rotor, which can rotate multiple times relative to the stator about a rotation axis, wherein:

[0020] In this method, a number of Ns electrical measurement signals are acquired, each electrical measurement signal representing the intensity of an electrical or magnetic variable at one of a number of Ns measurement positions, wherein the Ns measurement positions are individual, fixed relative to the stator and offset about a rotation axis by a given mechanical angle, and Ns is an integer equal to 2 or 3.

[0021] The change in the electrical measurement signal at the measurement position is caused by the rotation of the target in front of the measurement position under consideration. The target is mechanically connected to the rotor and has a number of Nc individual contrast regions, where the number Nc is greater than or equal to 2. The target includes contrasts of conductivity, permeability and / or magnetization.

[0022] The contrast regions are distributed at an angle on the target in a periodic pattern around the rotation axis. The pattern has Nc quasi-periods, and each contrast region extends around the rotation axis at a mechanical angle of 2Pi / Nc.

[0023] Each contrast region causes a change in the electrical measurement signal acquired at each measurement position as the contrast region passes through it in front of the measurement position during rotation about the axis of rotation. The change in the electrical measurement signal is a quasi-sine function of the electrical angular position of the contrast region relative to the measurement position, which changes by 2Pi radians relative to the mechanical angular position of the target about the axis of rotation.

[0024] At least two of the comparison regions have physical differences from each other, which produce differences in the magnitude of the intensity of the physical variables measured in the at least two individual quasi-cycles of the measurement signal at the same number of revolutions on the target.

[0025] The mechanical angle corresponding to the modulus offset is the angle of the measurement position around the rotation axis in 2Pi / Nc radians. For Ns=2, the measurement position is offset in Pi / (2×Nc) radians, and for Ns=3, the measurement position is offset in 2Pi / 3Nc radians.

[0026] The method includes, during the calibration phase, and then again during the setup phase of the measurement session:

[0027] a) Acquire the Ns electrical measurement signals over one mechanical revolution of the target around the rotation axis;

[0028] b) For different mechanical angular positions at one mechanical revolution around the target axis, the instantaneous value of the electrical angle position of the considered mechanical angular position is calculated using the calculation rule for electrical angle position. The calculation rule for electrical angle position considers:

[0029] ●For Ns=2, the arctangent of the ratio of the values ​​of the two measured signals for the time under consideration, or

[0030] ●For Ns=3, the arctangent of the ratio of the values ​​of the two transforms obtained by applying the Clark transform to the three measured signals of the time under consideration;

[0031] c) Calculate the instantaneous increment electrical angle position value, which is obtained by incrementing the counter X by one unit for any 2PI radian change in electrical angle position, the change occurring in the same rotational direction;

[0032] d) Determine at least one electronic signature of the target by computer, said at least one electronic signature including an electronic calibration signature determined during the calibration phase and an electronic setup signature determined during the setup phase, each electronic signature being determined by a signature value pair or a series of signature value pairs, the signature value pairs including:

[0033] ● An amplitude signature value derived from at least one value of at least one measurement signal from at least one measurement signal at at least one angular position relative to the target;

[0034] ●And the angular position signature value corresponding to one or more angular positions of the target that are considered to determine the amplitude signature value.

[0035] During the calibration phase, the electrical calibration signature of the target is recorded by computer.

[0036] During the setup phase of the measurement session, an angle offset measurement value is determined by a reset operation that includes calculating the angle measurement offset value. This angle measurement offset value is applied to the angle position signature value of the electrical setup signature or electrical calibration signature, thereby minimizing the difference between the electrical setup signature and the electrical calibration signature.

[0037] During a measurement session, the mechanical angular position of the rotor at a given time is determined by increasing the electrical angular position correction by an amount equal to the angular measurement offset value.

[0038] Other features of this approach are indicated below; these features are optional and can be implemented individually or in combination.

[0039] The instantaneous increasing electrical angle position value can be obtained according to the following relationship:

[0040] θe_inc=mod((θe+X×2Pi) / Nc,2Pi)

[0041] Furthermore, in this case, the mechanical angular position of the rotor at a given time can be determined as an incremental electrical angular position value modulo 2Pi, corrected by an amount equal to the angular measurement offset value.

[0042] In one variation, the instantaneously increasing electrical angle position value can be obtained according to the following relationship:

[0043] θe_inc=mod((θe+X×2Pi),Nc×2Pi)

[0044] Furthermore, in this case, the mechanical angular position of the rotor at a given time can be determined as an incremental electrical angular position modulo 2Pi, divided by the number of comparison areas and corrected by an amount equal to the angular measurement offset value.

[0045] Determining the signature value pair of an electronic signature can include:

[0046] i) At each angular position of the rotor over one mechanical revolution around the axis of rotation:

[0047] *The instantaneous amplitude value is calculated by considering the calculation rule of the instantaneous amplitude value of at least one of the two measurement signals for the time under consideration. The instantaneous amplitude value represents the intensity of the physical variable with respect to the angular position.

[0048] *The instantaneous electrical angle position value is calculated using the calculation rules for instantaneous electrical angle position values;

[0049] ii) The signature value pair is recorded by computer. According to the rules for determining the signature value, the signature value pair includes:

[0050] * An amplitude signature value derived from at least one of the instantaneous amplitude values;

[0051] * and an incremental electrical angle position signature value derived from instantaneous electrical angle position values ​​at one or more angular locations, taking into account one or more instantaneous amplitude values ​​to derive an amplitude signature value.

[0052] The target's electronic signature may include at least one signature value pair, the at least one signature value pair having an amplitude signature value, the value of which is unique over a mechanical revolution.

[0053] The electrical signature of the target may include at least one signature value pair having an amplitude signature value that exhibits the greatest difference from all other instantaneous amplitude values, which correspond to the same instantaneous value of an electrical angle over a mechanical revolution.

[0054] The target's electronic signature can include an ordered series of signature value pairs, with each signature value pair corresponding to a comparison region.

[0055] The series of signature value pairs are sorted according to the order in which the corresponding comparison region Zk passes before each measurement position.

[0056] For a given angular position, the rule for determining the signature value can consider at least one of the following values:

[0057] - The sum of squares of the measured signal values ​​at this angular position, where each measured signal value can reduce the offset of the measured signal or the square root of the sum of squares;

[0058] - A linear combination of the squares of the measured signal values ​​at that angular position, where each measured signal value can reduce the offset of the measured signal or the square root of the linear combination;

[0059] - A linear combination of the absolute measurement signal values ​​at this angular position, where each measurement signal value can reduce the offset value of the measurement signal;

[0060] - A sum or linear combination of the measured signal values ​​at this angular position, where each measured signal value can reduce the offset value of the measured signal.

[0061] For a given angular range, such as a sinusoidal quasi-period of a measured signal, the rule for determining the signature value can consider at least one of the following values:

[0062] - The maximum absolute value of one or more measurement signals within this angular range, where each measurement signal value can reduce the offset value of the measurement signal;

[0063] - The average value of one or more measurement signals over this angular range, where each measurement signal value is able to reduce the offset value of the measurement signal.

[0064] For a given angular range, such as a sinusoidal quasi-period of a measured signal, the rule for determining the signature value can consider at least one of the following values:

[0065] - The average value of the square root of the sum of the squares of the measured signal values ​​over the angular range, where each measured signal value can reduce the offset of the measured signal;

[0066] - The average value of the square root of a linear combination of the squares of the measured signal values ​​over the angular range, where each measured signal value can reduce the offset of the measured signal;

[0067] - The average value of a linear combination of absolute measurement signal values ​​over the angular range, where each measurement signal value reduces the offset of the measurement signal.

[0068] The rules for determining signature values ​​can identify at least one signature value pair, which corresponds to an identifiable value in a value pair formed by an instantaneous amplitude value and a corresponding incremental electrical angle position.

[0069] The identifiable value can be an instantaneous amplitude value, which takes one of the following values: local maximum value, local maximum value over a quasi-cycle of the electrical angle position, local minimum value, local minimum value over a quasi-cycle of the electrical angle position, absolute maximum value over a mechanical revolution of the target, absolute maximum value over a quasi-cycle of the electrical angle position, absolute minimum value over a mechanical revolution of the target, absolute minimum value over a quasi-cycle of the electrical angle position, previously determined value, average value over a mechanical revolution of the target, average value over a quasi-cycle of the electrical angle position, and average value over a half-cycle of the electrical angle position.

[0070] The identifiable value can be an electrical angle position or an incremental electrical angle position.

[0071] The rules for determining signature values ​​can determine at least one signature value pair, which corresponds to one or more predetermined values ​​of electrical angle position and / or corresponds to a predetermined series of incremental electrical angle positions.

[0072] Starting with an electrical calibration signature having M pairs of reference values, and capable of being written as:

[0073] SIGcb={(Asigcb1; θe_inc_sigcb1); (Asigcb2; θe_inc_sigcb2),

[0074] …,

[0075] (Asigcb M ;θe_inc_sigcb M )}

[0076] And the electrical signature setting was written as

[0077] SIGi={(Asigi1; θe_inc_sigi1); (Asigi2; θe_inc_sigi2-),…,

[0078] (Asigi M ;θe_inc_sigi M )},

[0079] An ordered vector Asigi = {Asigi1; Asigi2; ...; Asigi...} can be computed to be applied to the amplitude signature value extracted from the electrical setting signature SIGi. M The value of the number of cyclic permutations of} is "n". min The difference between the ordered vectors in a cyclic arrangement is:

[0080] Asigi[n]={Asigi 1+n,在[1,M]的范围内 ;

[0081] Asigi 2+n,在[1,M]的范围内 ...

[0083] Asigi M+n,在[1,M]的范围内}

[0084] And an ordered vector Asigcb{Asigcb1; Asigcb2; ...; Asigcb} of amplitude signature values ​​extracted from the electrical calibration signature SIGcb. M} is the smallest, and the angle measurement offset value can be calculated from the difference for at least one value of j:

[0085] delta_i_θm0=θe_inc_sigi j+nmin,在[1,M]的范围内 -θe_inc_sigcb j .

[0086] The physical difference between two contrast regions in the contrast region can be a spontaneous difference in the design of the sensor, which includes the target and the device for acquiring the Ns electrical measurement signals.

[0087] The physical difference between two contrast regions in the contrast region can be a non-spontaneous difference related to the manufacturing or installation dispersion of the sensor, which includes a target and means for acquiring the Ns electrical measurement signals.

[0088] The present invention also relates to a device for measuring the mechanical angular position of a rotor, the rotor being capable of rotating multiple revolutions relative to the stator around a rotation axis, the device comprising:

[0089] - Detector, acquires a number of Ns electrical measurement signals, each electrical measurement signal representing the intensity of an electrical or magnetic variable at one of a number of Ns measurement positions, the Ns measurement positions being individual, fixed relative to the stator and offset about a rotation axis by a given mechanical angle, Ns being an integer equal to 2 or 3;

[0090] - Target (12), mechanically connected to the rotor and having a number of targets Nc individual contrast regions (Zk), the number of targets Nc being greater than or equal to 2, wherein the targets include contrasts of electrical conductivity, magnetic permeability and / or magnetization.

[0091] The contrast regions are distributed at an angle on the target in a periodic pattern around the rotation axis. The pattern has Nc quasi-periods, and each contrast region extends around the rotation axis at a mechanical angle of 2Pi / Nc.

[0092] Each contrast region causes a change in the electrical measurement signal acquired at each measurement position as the contrast region passes through it in front of the measurement position during rotation about the axis of rotation. The change in the electrical measurement signal is a quasi-sine function of the electrical angular position of the contrast region relative to the measurement position, which changes by 2Pi radians relative to the mechanical angular position of the target about the axis of rotation.

[0093] At least two of the comparison regions have physical differences from each other, which produce differences in the magnitude of the intensity of the physical variables measured in the at least two individual quasi-cycles of the measurement signal at the same number of revolutions on the target.

[0094] In this device, the mechanical angle α corresponding to the modulus offset is the angle of 2Pi / Nc radians around the rotation axis. For Ns=2, the measurement position is offset by Pi / (2×Nc) radians, and for Ns=3, the measurement position is offset by 2Pi / 3Nc radians.

[0095] Additionally, the device includes an electronic control unit (28), which is programmed to:

[0096] a) Acquire the Ns electrical measurement signals (S1, S2) over one mechanical revolution of the target around the rotation axis.

[0097] b) For different mechanical angular positions at one mechanical revolution around the rotation axis of the target, the instantaneous value of the electrical angle position θe of the considered mechanical angular position is calculated using the calculation rule for electrical angle position. The calculation rule for electrical angle position considers:

[0098] ●For Ns=2, the arctangent of the ratio of the values ​​of the two measured signals (S1, S2) for the time under consideration, or

[0099] ●For Ns=3, the arctangent of the ratio of the values ​​of the two transforms obtained by applying the Clark transform to the three measured signals of the time under consideration;

[0100] c) Calculate the instantaneous increment electrical angle position value, which is obtained by incrementing the counter X by one unit for any 2PI radian change in electrical angle position, the change occurring in the same rotational direction;

[0101] d) Determine at least one electronic signature of the target by computer, each electronic signature being determined by a signature-value pair or a series of signature-value pairs, the signature-value pairs including:

[0102] ● An amplitude signature value derived from at least one value of at least one measurement signal from at least one measurement signal at at least one angular position relative to the target;

[0103] ●And the angular position signature value corresponding to one or more angular positions of the target that are considered to determine the amplitude signature value.

[0104] The device includes an electronic memory in which the electrical calibration signature of the target, determined during the calibration phase, is recorded by a computer.

[0105] The electronic control unit is programmed to determine the electrical setup signature and the angle measurement offset value during the setup phase of the measurement session by including a reset operation that calculates the angle measurement offset value, which is applied to the angle position signature value of the electrical setup signature or the electrical calibration signature, thereby minimizing the difference between the electrical setup signature and the electrical calibration signature.

[0106] Additionally, during the measurement session, the electronic control unit determines the rotor's mechanical angular position at a given time by increasing the electrical angular position correction by an amount equal to the angular measurement offset value.

[0107] Other features of such a device are indicated below, which are optional and can be implemented individually or in combination.

[0108] The electronic control unit can be programmed to:

[0109] a) Acquire the Ns electrical measurement signals over one mechanical revolution of the target around the rotation axis.

[0110] b) For different mechanical angular positions at one mechanical revolution around the target axis, the instantaneous value of the electrical angle position of the considered mechanical angular position is calculated using the calculation rule for electrical angle position. The calculation rule for electrical angle position considers:

[0111] ●For Ns=2, the arctangent of the ratio of the values ​​of the two measured signals (S1, S2) for the time under consideration, or

[0112] ●For Ns=3, the arctangent of the ratio of the values ​​of the two transforms obtained by applying the Clark transform to the three measured signals of the time under consideration;

[0113] c) Calculate the instantaneous increment electrical angle position value, which is obtained by incrementing the counter X by one unit for any 2PI radian change in electrical angle position, the change occurring in the same rotational direction;

[0114] d) Determine at least one electrical calibration signature of the target by computer, said at least one electrical calibration signature being determined by a signature value pair or a series of signature value pairs, the signature value pair including:

[0115] ● An amplitude signature value derived from at least one value of at least one measurement signal from at least one measurement signal at at least one angular position relative to the target;

[0116] ●And the angular position signature value corresponding to one or more angular positions of the target that are considered to determine the amplitude signature value.

[0117] The physical difference between two contrast regions in the contrast region can be a spontaneous difference in the design of the target.

[0118] The physical differences between two contrast regions in the contrast region can be non-spontaneous differences related to the fabrication or installation dispersion of sensors, including targets and detectors.

[0119] The electronic control unit can be programmed to implement the methods described above.

[0120] The detector can be manufactured in the form of a detector box, which includes a preferably sealed box containing a measurement unit, an electronic control unit, and a computerized communication interface. Attached Figure Description

[0121] [ Figure 1 ] Figure 1 This is a schematic diagram illustrating an example of a sensor that enables the implementation of the present invention.

[0122] [ Figure 2 ] Figure 2 yes Figure 1 A schematic diagram of the sensor's detector.

[0123] [ Figure 3 ] Figure 3 It shows Figure 1 A schematic diagram illustrating the operating principle of this type of sensor.

[0124] [ Figure 4 ] Figure 4 This is shown in a variant embodiment. Figures 1 to 3 A schematic diagram illustrating the operating principle of this type of sensor.

[0125] [ Figure 5 ] Figure 5 It is a graph showing the measurement signal acquired using a sensor with four contrast regions during one mechanical revolution.

[0126] [ Figure 6 ] Figure 6 It shows that based on Figure 5 The graph shows an example of the calculation rules for the electrical angular position and instantaneous amplitude value of the measured signal, where the instantaneous amplitude value represents the intensity of the physical variable of the angular position under consideration.

[0127] [ Figure 7 ] Figure 7 It is a graph showing the measurement signal acquired using a sensor with 6 contrast regions during one mechanical revolution.

[0128] [ Figure 8 ] Figure 8 It shows that based on Figure 7 The graph shows an example of the calculation rules for the electrical angular position and instantaneous amplitude value of the measured signal, where the instantaneous amplitude value represents the intensity of the physical variable of the angular position under consideration.

[0129] [ Figure 9 ] Figure 9 It is a graph showing the variation of the function used as the calculation rule for the instantaneous amplitude value over a complete mechanical revolution of the target, and the average value of the instantaneous amplitude value over each quasi-cycle.

[0130] [ Figure 10 ] Figure 10 It is a graph showing the variation of the function used as the calculation rule for the instantaneous amplitude value over a full mechanical revolution of the target, and the average value of that instantaneous amplitude value over each quasi-cycle, with a noise function added to the measurement signal.

[0131] [ Figure 11 ] Figure 11 The data is presented in the form of six radar charts. Figures 7 to 10 The figures show the electrical calibration signature and electrical setup signature of sensors of the same type as the sensors used to measure the signals and their derived values. In each figure, the electrical calibration signature is shown as a closed dashed curve, and the electrical setup signature is shown as a closed solid curve. The vertices of the closed curves represent paired signature values ​​in polar coordinates. Figure A shows two figures generated by the measurement. Figures B through F each correspond to the value of the number of cyclic increments of the ordered vector of paired signature values ​​of the electrical setup signature.

[0132] [ Figure 12 ] Figure 12 It is a graph showing each of the "n" values ​​representing the number of cyclic permutations of the ordered vectors of paired signature values ​​for the electrically set signature, indicating the "Manhattan distance" between the ordered vectors of the electrically calibrated signature and the electrically set signature.

[0133] [ Figure 13 ] Figure 13 It is similar to Figure 12 The graph shows each of the "n" values ​​representing the number of cyclic increments of the ordered vectors of paired signature values ​​for the electrical setting signature, indicating the "Euclidean distance" between the ordered vectors of the electrical calibration signature and the electrical setting signature. Figure 13 ). Detailed Implementation

[0134] Figure 1 An exemplary embodiment of sensor 10 is schematically illustrated, enabling the measurement method according to the invention to be implemented. The sensor 10 includes a target 12 and a detector 14. The detector is located in... Figure 2 It is shown separately in the text.

[0135] In this example, sensor 10 is an inductive sensor, wherein detector 14 is capable of generating an electromagnetic field and measuring a value representing the electric field. Target 12 has electromagnetic properties that allow it to modify the characteristics of the electromagnetic field as it passes in front of detector 14, such modification being detected by detector 14. For example, the target comprises a metallic element, wherein the electromagnetic field generated by detector 14 induces eddy currents, which then perturb the electromagnetic field, which can be detected by detector 14. However, note that the invention can also be implemented using magnetic sensors.

[0136] In all cases, sensor 10 is intended for measuring the mechanical angular position of two components that are rotatably movable relative to each other about a rotational axis A1. Sensor 10 is capable of measuring and indicating angular positions exceeding 360 degrees of mechanical angle for angular travel of the rotor relative to the stator exceeding one mechanical revolution, and therefore for angular travel exceeding 360 degrees of mechanical angle and thus exceeding 2Pi radians of mechanical angle. For angular travel exceeding one mechanical revolution, sensor 10 does not necessarily give an indication of the number of revolutions traveled, but is able to give the relative angular position of the rotor and stator in 360 degrees of mechanical angle even after several mechanical revolutions. Target 12 is intended to be attached to one of these components, which will be arbitrarily referred to herein as the rotor. Detector 14 is intended to be attached to the other of these components, which will be arbitrarily referred to herein as the stator. In practice, it is generally easier to attach detector 14 to a stationary component, thus the name stator is reasonable, because detector 14 is made to exchange information (especially electrical measurement signals) with a wider range of systems into which sensors can be integrated (e.g., electronic control systems of motors including rotors and stators). Conversely, target 12 is typically a passive component that does not require any electrical connection, so its mounting on a rotating component does not cause any specific problems. However, nothing prevents detector 14 from being mounted on a rotating component and target 12 from being mounted on a stationary component, or both detector 14 and target 12 from being mounted on a single component, wherein both components rotate relative to each other and both rotate relative to the stationary environment.

[0137] The target 12 has Nc individual contrast regions, where Nc is greater than or equal to 2. These contrast regions Zk (where k = 1, 2, ..., Nc) are distributed angularly on the target in a periodic pattern around the rotation axis A1. Typically, the number of individual contrast regions Nc is greater than or equal to 4. The present invention is particularly advantageous for targets having an integer number of Nc individual contrast regions less than or equal to 20.

[0138] As a first approximation, the contrast regions are identical to each other and can be considered to follow each other. There is no discontinuity between the contrast regions in a 360-degree mechanical angle around the rotation axis of the target 12, forming a continuous periodic pattern whose repeating pattern elements are contrast regions that follow each other. However, in detail, it will be seen that the present invention utilizes the fact that a real sensor is not an ideal sensor and will require at least two contrast regions Zk that are physically different from each other.

[0139] In the example shown, target 12 includes radial teeth 18 carried by support member 16. Figure 1 In the example, the support 16 takes the form of an angular ring around axis A1, and radial teeth 18 extend radially inward from the annular ring 16. However, referring to... Figure 4As will be seen, a target may be provided comprising radial teeth 18 extending radially outward from a central support 17 (which may be annular or may not be annular).

[0140] exist Figure 1 In the example, the radial tooth 18 is defined at an angle by side edges 20, which are straight and each oriented in a radial direction originating from and passing through the axis A1. However, the tooth may have a different geometry, and in particular, may not include any side edges 20 that are straight or oriented in a radial direction from the axis A1.

[0141] exist Figure 1 In the example, the radial teeth 18 are defined radially inward by a radial end edge 22, which is a semi-circular shape centered on axis A1 and defined by two side edges 20. The radial teeth 18 are evenly spaced at an angle around axis A1.

[0142] In theory, the radial teeth 18 are considered identical, and any differences are due to manufacturing tolerances or uncontrolled deformation. In this case, the radial teeth 18 are all considered to have the same angular dimension around axis A1.

[0143] Similarly, the radial teeth 18 are assumed to be arranged at the same radial distance from the axis A1, and they also have the same radial dimension between their radial end edges 22 and the annular ring 16.

[0144] The radial teeth 18 are made of a material that has specific characteristics for at least one particular electromagnetic property (e.g., conductivity, permeability, and / or magnetization). In this example, the radial teeth are made of or include conductive materials, such as metallic materials.

[0145] Two consecutive radial teeth 18 around axis A1 are separated by an inter-tooth space 24, which has characteristics different from those of the radial teeth 18 for the same electromagnetic properties. In this example, the inter-tooth space 24 contains no conductive material. Note that the inter-tooth space 24 is shown as an empty space. However, note that with respect to the electromagnetic properties measured by the sensor, the target 20 may include a body made of a neutral material, for example, to ensure its attachment to the rotor. In the example of the inductive sensor, the target 20 may therefore include an attachment made of a plastic material, which may be adjacent to or even coated with the radial teeth 18. Similarly, the annular ring 16 may be made partially or entirely of a non-conductive material, such as a plastic material.

[0146] Therefore, in this example, each inter-tooth space 24 is defined at an angle by the side edges 20 of two adjacent consecutive radial teeth 18 on one side of that inter-tooth space 24. Thus, each inter-tooth space 24 has an angular dimension defined by two adjacent radial teeth. In this embodiment, all inter-tooth spaces 24 have the same angular dimension. Similarly, in the illustrated embodiment, the inter-tooth spaces 24 have the same angular dimension as the radial teeth 18.

[0147] The target 12 includes a plurality of radial teeth 18 and an inter-tooth space 24 of the same number Nc as the number of radial teeth 18. Figure 1 In the example, the target includes eight radial teeth 18, and therefore the same number of inter-tooth spaces 24. In the example, the radial teeth 18 and the adjacent inter-tooth spaces 24 form a contrast region Zk within the meaning of this invention, and thus the contrast region has a mechanical angular dimension about axis A1, which has an angle value of 2Pi / Nc radians.

[0148] As seen above, each contrast region Zk can be identified by an integer level k contained in the range 1 to Nc. Level k corresponds to the order of the mechanical angular positions of the contrast regions relative to other contrast regions around the axis of rotation. Therefore, level k corresponds to the physical arrangement of the contrast regions on the target. Thus, as the target passes in front of the considered position, one sees the contrast region of level 1 passing through, then the contrast region of level 2, then the contrast region of level 3, and so on until the contrast region of level Nc, which, after a mechanical angular rotation of 2Pi radians by the rotor, will again be the contrast region of level 1. Of course, if the rotor rotates in the opposite direction, the order of passage will also be reversed.

[0149] Detector 14 typically includes multiple measurement units (here, two measurement units) arranged at separate measurement positions (here, P1 and P2), and each measurement unit is capable of measuring the intensity of the same electrical or magnetic variable at its corresponding measurement position. Detector 14 may include three measurement units arranged at three separate measurement positions. Preferably, these units are identical in the sense that they produce one and the same electrical signal value if the same electromagnetic variable is present at their respective measurement positions. However, simulations have shown that the invention operates satisfactorily, including slight inhomogeneities between the units and / or their signal processing channels. In the context of inductive sensors, each measurement unit may, for example, include measurement windings B1, B2. Figure 2 In the example, the measuring windings B1 and B2 of the two units are identical, but they are offset at an angle around axis A1 by a known mechanical angle value "α". Figure 3In the example, the measurement windings B1 and B2 of the two units are different. A measurement position can be defined for each measurement unit. For each unit, this measurement position can be arbitrarily defined. This measurement position is, of course, related to the position and geometry of the measurement winding of the measurement unit under consideration. Regarding the invention, the measurement position can, for example, be arbitrarily defined as a position relative to the center of the measurement winding, which is... Figure 2 The diagram illustrates this point. Such a central position is defined, for example, in a manner similar to the determination of the center of gravity. Preferably, the measurement position will be defined identically for all measurement units.

[0150] In this example, the measuring windings B1 and B2 are manufactured as printed circuits on the same printed circuit board 26. The measuring windings B1 and B2 are connected to an electronic control unit 28, which is also supported by the same printed circuit board 26. The electronic control unit 28 is connected to a computer communication interface (e.g., connector 30), which may also be carried by a printed circuit board. The electronic control unit 28 of the sensor 10 includes, for example, a microprocessor with electronic memory and electronic input / output interfaces. Connector 30 enables a computer link with an external electronic system via a computer communication network, such as a serial link of the CAN bus type or a multiplexed network, wherein the external electronic system is electrically connected to the sensor 10, and the external system itself can include one or more electronic control units using the angular position information transmitted by the sensor. Of course, as a computer communication interface, connector 30 can be a wireless electronic communication unit (e.g., (Or other types) replacement or completion. Sensor 10 may be specified to include: a detector 14 manufactured in the form of a detector housing, including a preferably sealed housing in which measurement units B1, B2 are arranged; a computing unit 28, integrated into the housing, which is the computing unit within the housing; and a computer communication interface. Such a detector can be described as an "intelligent" detector, which can be calibrated and tested before its integration into an external system.

[0151] According to the invention, the measurement positions P1 and P2 of the units are offset by a given mechanical angle "α" around axis A1. In the illustrated example where the measurement units are mounted on the same printed circuit board, the mechanical angle "α" between the measurement positions of the units is determined during detector manufacturing. However, it is also conceivable to have a detector made of multiple components, with the measurement units arranged on different components of the detector. In this case, the mechanical angle "α" between the measurement units, and therefore the measurement positions, will be determined at a later stage of detector assembly or during the assembly of the different components of the detector on associated components. In operation, the mechanical angle "α" between the measurement positions P1 and P2 is fixed.

[0152] In the example shown, note that the measuring unit corresponds to a separate measuring position, but the corresponding measuring windings B1 and B2 overlap at an angle around the axis. The windings can be offset at a complete angle relative to each other.

[0153] Regarding the sensing type sensor, detector 14 includes a primary winding B0, which is designed to generate a magnetic field in space relative to the measuring unit. In the example, the primary winding B0 extends in a plane perpendicular to axis A1 and defines the primary winding profile containing two measuring windings B1 and B2 in this plane.

[0154] Examples of inductive sensors and their operating principles are known, for example, from documents US2019017845, EP-0.182.085-A2 or FR-3.023.611-A1.

[0155] Figure 3 The operation of an exemplary embodiment of the sensing sensor is illustrated very schematically. A voltage source 32 causes current to flow through the primary winding B0, which generates a primary magnetic field in front of the primary winding B0. For each measuring winding B1, B2, the detector includes elements M1, M2 for measuring the intensity of the voltage and / or current in the measuring winding B1, B2. These measuring elements M1, M2 transmit raw electrical signals SB1, SB2 representing the intensity of the voltage and / or current in the considered measuring winding B1, B2 (and thus the magnetic field in front of the considered measuring winding). Without any perturbing elements, each measuring winding B1, B2 is energized by current or generates a voltage whose value depends on the primary electromagnetic field generated by the primary winding B0. Figure 3 A radial tooth 18 is shown, which loops in front of the measuring windings B1, B2, and therefore also in front of the primary winding B0. When the radial tooth 18 faces the primary winding B0, it is exposed to the primary magnetic field generated by the primary winding B0, causing eddy currents to be induced in the radial tooth 18. The eddy currents themselves generate an anti-electromagnetic field, which then locally perturbs the primary electromagnetic field generated by the primary winding B0 relative to the radial tooth 18. In this way, when the radial tooth 18 faces one or the other of the measuring windings B1, B2, the magnetic field observed by the considered measuring windings B1, B2 consists of the superposition of the primary magnetic field and the back electromotive force field generated by the radial tooth 18, which manifests as changes in the current and / or voltage in the measuring windings B1, B2, and therefore as changes in the electrical signals SB1, SB2 transmitted by the corresponding measuring elements M1, M2.

[0156] The electrical signals SB1 and SB2 transmitted by the measuring elements M1 and M2 are transmitted to the electronic control unit 28. The electronic control unit 28 can transmit the corresponding electrical signals S1 and S2 through appropriate electronic processing. Each of the electrical signals S1 and S2 represents the intensity of the electric or magnetic variable at the corresponding measurement position, which here represents the magnetic field at the corresponding measurement position.

[0157] In a manner known per se, the sensor is designed such that as each comparison region rotates past the measurement position about the axis of rotation, each comparison region causes a change in the electrical measurement signals S1, S2 acquired at that measurement position, which is a quasi-sine function of the electrical angular position θe of the comparison region relative to the measurement position.

[0158] The electrical angular position θe is associated with the mechanical angular position θm of the target relative to the measurement position, but it is only related to the angular range of the contrast region under consideration. Locally, for a given contrast region in front of the measurement position, and if the electrical angular position θe and the mechanical angular position θm are expressed in the same units, this gives the following relationship:

[0159] θe = θm × Nc, modulo 2Pi.

[0160] Note that for a mechanical angular position θm of the target around the axis of rotation that changes by 2Pi / Nc radians, the electrical angular position θe changes by 2Pi radians.

[0161] Over one mechanical revolution of the rotor relative to the stator, Nc comparison regions Zk generate sinusoidal quasi-period measurement signals S1 and S2 with the same number of comparison regions Nc as the number of comparison regions Nc.

[0162] In an ideal sensor comprising Nc strictly identical contrast regions, the sensor will cause each contrast region to produce a perfect sinusoidal change in the electrical measurement signals S1 and S2 based on the electrical angular position θe as it rotates about the axis of rotation in front of the measurement position. Each contrast region will correspond to one electrical cycle of the signal. In one mechanical revolution of the target, the number of electrical cycles of the signal will equal the number of contrast regions. Using such a sensor, at a given electrical angular position θe, it is impossible to identify which contrast region is facing or close to the sensor positioning region by simple, direct analysis of the signal.

[0163] Unlike ideal sensors, this invention aims to utilize the non-ideal nature of sensors to determine, for a given angular position, which contrast region faces or is closer to the sensor-positioned contrast region, thereby determining the rotor's position relative to the stator, which is a mechanical angle of 2Pi radians. This non-ideal property is characterized by the fact that at least two of the contrast regions Zk exhibit physical differences from each other. In practice, there are typically more than two contrast regions that exhibit physical differences from each other in pairs.

[0164] Because the contrast regions continuously follow the axis of rotation as the target rotates relative to the detector, the number Nc contrast regions Zk produce a quasi-sinusoidal period of the same number Nc as the number of contrast regions Nc over one mechanical revolution of the rotor, referred to below as a quasi-period for each measurement signal S1, S2. Specifically, the physical differences between the contrast regions of the target produce differences in the measurement signal over two quasi-periods of the measurement signal at the same revolution of the target. At least two contrast regions Zk exhibit physical differences from each other, which produce differences in the amplitude of the intensity of the physical variable measured in the at least two individual quasi-periods of the measurement signal, respectively, over the same signal cycle. This difference in the signal may specifically relate to the maximum and / or minimum values ​​of the intensity of the physical variable measured in the at least two individual quasi-periods of the measurement signal. These differences justify the name "quasi-period".

[0165] It should be understood that the physical difference between at least two contrasting regions must be sufficient to allow them to be repeatedly measured in the measurement signal.

[0166] The physical difference between two contrasting regions can arise from the differences between the two contrasting regions, which may involve, for example:

[0167] -Inherent geometric differences, such as dimensional differences, flatness differences, orientation differences, etc.;

[0168] - and / or differences in electromagnetic characteristics, such as those related to component materials or material inhomogeneities in the comparison region;

[0169] - and / or other parameters.

[0170] The physical difference between the two contrast regions can also be caused by the difference in the relative positions of the contrast regions of the target and the detector 14 with respect to the two contrast regions. For example, in the above embodiment, the radial teeth 18 can be made to extend in a plane perpendicular to the rotation axis A1. However, it is possible that one or more radial teeth 18 have an inclination relative to this plane, such that the distance between the radial teeth of one contrast region and the detector is different from the distance between the radial teeth of the other contrast region and the detector along the direction of the rotation axis A1. In this case, even if all other parameters are equal between the two contrast regions, the amplitude of the quasi-sinusoidal signal measured at the measurement position given by the detector 14 will be different for the two contrast regions.

[0171] The physical difference between two contrasting regions can be spontaneous or non-spontaneous. It is spontaneous if the difference corresponds to the design of the sensor or its installation, and is intentionally chosen in the design to have such a difference, which can be known or unknown.

[0172] The physical difference between two contrast regions can be non-spontaneous, as it may be related to variations in the manufacture or installation of the sensor. Specifically, for sensors comprising a detector mounted on one component and a target mounted on another component movable relative to the aforementioned component, it is virtually unavoidable that the attachment of the detector and target to their respective components will be performed independently. In this way, the relative position of the target along the rotation axis A1 relative to the sensor depends on how one component and the other are attached. However, if the target is mounted on the corresponding component (e.g., a rotor) where there is a slight misalignment between the target's axis and the actual rotation axis of the rotor, the motion of the target relative to the detector as it rotates will not be a perfect circular and planar motion, but rather a quasi-circular motion, exhibiting, for example, eccentricity and / or a rhythmic pattern corresponding to the change in distance between the contrast region and the detector 14 along the rotation axis A1 from one contrast region to another, as described above.

[0173] This phenomenon is, for example, in Figure 5The diagram illustrates signals S1 and S2 corresponding to measurements of physical variables at a first measurement position P1 and a second measurement position P2 using a sensor comprising four contrast regions. The signals are represented here as variations over one full mechanical revolution of the target, i.e., variations over a mechanical rotation angle of 2Pi radians of the rotor relative to the stator. Note that within the 2Pi radian mechanical angle range, each signal comprises four quasi-cycles, each quasi-cycle having a positive half-cycle and a negative half-cycle relative to the offset values ​​S1m and S2m of the signal under consideration. The offset value of the signal is, for example, the average value of the signal over a defined time period corresponding to an integer number of quasi-cycles (preferably over Nc quasi-cycles, i.e., one mechanical revolution), such as an arithmetic mean.

[0174] Within the meaning of this invention, an example of a quasi-sine function for the electrical angular position θe is defined, which can be written as:

[0175] S1(θe,θm)=Anom1×(1+dA(θm+phi1))×sin(θe+phi1)+S1m,

[0176] S2(θe,θm)=Anom2×(1+dA(θm+phi2))×sin(θe+phi2)+S2m,

[0177] S3(θe,θm)=Anom3×(1+dA(θm+phi3))×sin(θe+phi3)+S3m,

[0178] in

[0179] The amplitude values ​​of Anom1, Anom2, and Anom3 can be considered as constants.

[0180] dA(θm) is the amplitude modulation function.

[0181] phi1, phi2, and, where applicable, phi3, depending on the phase shift value at the selected measurement location, are also considered constants.

[0182] The offset values ​​S1m, S2m, and S3m are considered constant and may, for example, correspond to slight non-uniformities between cells and / or their signal processing channels.

[0183] And recalling that within the same comparison area:

[0184] θe = θm × Nc, modulo 2Pi.

[0185] Therefore, this quasi-sine function has a sinusoidal basis sin(θe+phi) with a defined quasi-period, and its angular range has a value of 2PI / Nc, which corresponds to the angular range of the contrast region. The amplitude modulation function dA(θm+phi) is a function (preferably continuous) that is periodic over one mechanical revolution of the target, and its value varies at most between +0.5 and -0.5, preferably at most between +0.25 and -0.25. As its name suggests, the amplitude modulation function exhibits the irregularity observed with respect to the local maxima of the absolute value of the signal at each point in its half-period.

[0186] exist Figure 5 In the example shown, the offset values ​​S1m and S2m are non-zero. This offset is not desirable, but it is experienced. It can be, for example, particularly due to the inhomogeneity of the generated magnetic field, which can be greater near the edges of the primary winding B0 than at its center. In some embodiments, this offset may be due to the fact that windings B1 and B2 cannot both be centered relative to the primary winding B0. It is quite clear that the maximum value of the absolute value of the signal is not the same relative to the offset value from one quasi-cycle to another. In this way, the signal thus has a quasi-sinusoidal characteristic in a mechanical angle of 2Pi radians, as it comprises consecutive positive and negative half-cycles. The signal exhibits at least a variation in signal amplitude between quasi-cycles. Within the same quasi-cycle, or even within the same half-cycle, the signal does not perfectly follow a sine curve. However, through construction, the deviation from the sine curve is sufficiently reduced so that the electrical angular position θe of the contrast region relative to the measurement position can be determined with acceptable accuracy. This is accomplished by considering the calculation rule of the arctangent of the ratio of the values ​​of the two measured signals over the considered time period. Conventionally, if the electrical angular position θe of the comparison area relative to the measurement position is determined by considering the arctangent of the ratio of the values ​​of two measurement signals within the considered time period, and the error is less than 10% of one leveling period, and the angular range has a value of 2PI / Nc, preferably less than 5% of one leveling period, and more preferably less than 1% of one leveling period, then acceptable accuracy will be achieved. Specifically, this level of accuracy is necessary because, as will be seen further, the calculated value of the electrical angular position involves the mechanical angular position information to be transmitted by the sensor.

[0187] In the case of achieving two measurement signals at two separate measurement positions, the method specifies that the measurement positions are offset around the rotation axis by a corresponding mechanical angle "α", which is modulo the angular range of the comparison area, and therefore modulo 2Pi / Nc radians, resulting in an angle of [Pi / (2×Nc)] radians, which is one-quarter of a leveling cycle. Preferably, the mechanical angle "α" between the two measurement positions has an angle value of [Pi / (2×Nc)] radians, such that the two measurement positions are as close as possible to make the detector as compact as possible. Therefore, the compactness of the measurement positions ensures that the two signals undergo similar amplitude modulation. In other words, the modulation function dA(θm) does not include any significant changes in the interval [Pi / (2×Nc)]. Therefore, the measurement error caused by this modulation will be small.

[0188] However, the mechanical angle "α" between the two measurement positions can be specified to have an angle value of [Pi / (2×Nc)]+(ke×2Pi / Nc) radians, where ke is a non-zero integer interval exponent. In this case, the two measurement positions are far apart from each other, which facilitates the manufacture and / or installation of the detector. In this case, it must be ensured that the modulation function dA(θm) does not include any significant changes within the interval of [Pi / (2×Nc)]+(ke×2Pi / Nc).

[0189] In both cases, due to this configuration, the changes of the two signals acquired at two separate locations can be written as two functions, one as the sine of the same variable and the other as the cosine of the same variable.

[0190] Therefore, under these conditions, it is possible to write at least within a given quasi-period Tk corresponding to a given contrast region Zk:

[0191] S1(θe,θm)=Anom1×(1+dA(θm))×sin(θe+phi1)+S1m,

[0192] S2(θe,θm)=Anom2×(1+dA(θm))×cos(θe+phi1)+S2m,

[0193] And θe=Nc×θm

[0194] And the difference between dA(θm+α) and dA(θm) is very small.

[0195] By properly positioning the two sensors, phi1 can be made equal to 0.

[0196] Under these conditions, the electrical angular position θe of the comparison area relative to the measurement position can be determined very easily by taking into account the arctangent of the ratio of the values ​​of the two measurement signals over the considered time period. This calculation is performed by a computer, for example by the electronic control unit 28, whether it is integrated into the detector box as described above, or by a remote electronic control unit.

[0197] To determine the value of θe, the function atan2(S2,S1), present in many programming languages, can be used. The two-variable atan2 function is based on the arctangent function. In mathematical terms, atan2 returns the principal value in the interval [-Pi,Pi] of the function applied to the complex number (x+iy). In the example, it is preferable to perform the function in the interval [0; 2PI], thereby allowing the atan2 function to take the return value within this interval. In other words, to determine an estimate of the electrical angular position θe, a function defined as follows can be used:

[0198] If S1≥0 and S2>0, then θe=arctan(S1 / S2)

[0199] If S2 < 0, then θe = arctan(S1 / S2) + PI

[0200] If S1 < 0 and S2 > 0, then θe = arctan(S1 / S2) + 2 * PI

[0201] If S1 > 0 and S2 = 0, then θe = PI / 2

[0202] If S1 < 0 and S2 = 0, then θe = 3 * PI / 2

[0203] If S1 = 0 and S2 = 0, then θe is undefined.

[0204] For the case of using three measurement signals acquired at three separate locations, the present invention specifies that the measurement positions are offset by a mechanical angle "α" around the rotation axis A1, which corresponds to an angle of [2Pi / (3×Nc)] radians, or one-third of the quasi-period of the electrical angle. Next, a Clarke transform is applied to the three measurement signals within the considered time period, resulting in two transforms. A calculation rule is then applied to these transforms, taking into account the arctangent of the ratio of the values ​​of the two obtained transforms, to determine the electrical angle position θe, as described above.

[0205] Note that this invention enables the realization of a target without any "holes," i.e., without missing contrast regions. In other words, the target used is one in which, as a first approximation, the contrast regions are sufficiently identical to each other to determine the electrical angular position θe of the contrast region relative to the measurement position, and this is done by taking into account the calculation rule of the arctangent of the ratio of the values ​​of two measurement signals over the considered time period, giving correct values ​​for all contrast regions in the 360-degree mechanical angle around the rotation axis of the target 12. Similarly, the contrast regions can be considered to follow each other, with no discontinuity between them in the 360-degree mechanical angle around the rotation axis of the target 12, and the estimated value of the electrical angular position θe of the contrast region relative to the measurement position is determined by taking into account the calculation rule of the arctangent of the ratio of the values ​​of two measurement signals over the considered time period, which gives correct values ​​in the 360-degree mechanical angle around the rotation axis of the target 12.

[0206] The present invention provides a measurement method comprising a calibration phase, followed by at least one measurement session, the at least one measurement session including a setup phase. Between the calibration phase and the measurement session, there may be a relative rotation of the target 12 of sensor 10 relative to the detector 14 of sensor 10, even if sensor 10 cannot “see” this rotation, for example, because it is no longer powered. In other words, sensor 10 can be “turned off” between the calibration phase and the measurement session.

[0207] During the calibration phase, and then again during the setup phase of the measurement session, the method provides the acquisition of Ns electrical measurement signals S1, S2. Preferably, the acquisition is performed at at least one target revolution. For the setup phase of a given measurement session, signal acquisition can be completed at a single mechanical revolution of the target around the rotation axis. In a variation, for the setup phase of a given measurement session, signals can be acquired at multiple mechanical revolutions of the target around the rotation axis, and for each mechanical angular position in a 360-degree angle, the average value of the different measurements performed at different revolutions is acquired. Therefore, averaging can be taken and thus measurement noise can be filtered. Preferably, acquisition is performed continuously at one mechanical revolution of the target. Preferably, acquisition is performed with an angular resolution less than a given electrical angle (and therefore less than 5 / Nc mechanical angle), preferably less than or equal to one electrical angle (and therefore less than or equal to 1 / Nc mechanical angle). Preferably, acquisition is performed for each mechanical revolution of the target at one mechanical revolution of the target around the rotation axis. Each angular position of the target is understood here to be defined by the maximum angular resolution of the sensor.

[0208] Note that during the calibration phase, when the relative rotational speed between the rotor and stator is constant, it is possible to acquire a signal. This allows for the acquisition of a time-based signal, which can be converted into a signal based on the mechanical angular position, without the need for an additional absolute reference position sensor on the calibration bench. In this case, a rotational marker defining the origin of the mechanical position will be sufficient. In certain special applications, such as when the mechanical angular position transmitted by the sensor is only used to define the linearization law of the sensor over one mechanical revolution, the origin of the mechanical position can also be arbitrarily defined during the calibration phase without any external reference. If the calibration bench has an absolute position sensor, the signal based on the mechanical angular position can be obtained based on the acquired time-based signal, even if the relative rotational speed between the rotor and stator is not constant during the acquisition of the target over one mechanical revolution.

[0209] Based on this, the instantaneous value of the electrical angle position θe for the considered mechanical angle position θm can then be calculated. This calculated instantaneous value of the electrical angle position θe is, of course, an estimate of the instantaneous angle position θe value. This calculation can be performed for each mechanical angle position over one mechanical revolution of the target around the rotation axis, where measurement signals S1 and S2 have already been acquired. This calculation is performed using the calculation rule for the electrical angle position, as described above, which considers:

[0210] For Ns = 2, the arctangent of the ratio of the values ​​of the two measured signals with respect to the time under consideration, or

[0211] For Ns=3, the arctangent of the ratio of the transform values ​​obtained by applying the Clark transform to the three measured signals within the considered time period.

[0212] The electrical angle position θe determined in this way is an angle position that can be defined in the range of 2PI values. This range corresponds to a contrast region, and therefore corresponds to the mechanical angle sector, the range of which takes the value of a contrast region, that is, a mechanical angle of 2PI / Nc radians.

[0213] In the example below, starting from the electrical angle position θe, the instantaneous value θe_inc of the increasing electrical angle position is calculated, for example in electronic control unit 28, including when it is integrated into the detector box as described above, or in a remote electronic control unit. The instantaneous value of the increasing electrical angle position corresponds to the electrical angle position θe, and is therefore evaluated based on the quasi-period of the signal corresponding to a contrast region, to which an angular range corresponding to the quasi-period that has elapsed since the start of the increment is added. This value is calculated modulo one revolution of the target.

[0214] For example, for any change of 2Pi in the electrical angle position θe, the instantaneous value of the increasing electrical angle position can be obtained by incrementing the counter X by one unit, the change occurring in the same rotational direction.

[0215] Therefore, in the first example, the instantaneous value of the increasing electrical angle position can be obtained, for example, according to the following relationship:

[0216] θe_inc=mod((θe+X×2Pi) / Nc,2Pi)

[0217] In this example, it can be seen that for one mechanical revolution of the target, the instantaneous value of the increasing electrical angle position changes by an angle of 2Pi radians.

[0218] However, the instantaneous values ​​of the increasing electrical angle positions can be obtained, for example, from the following relationship:

[0219] θ'e_inc=mod(θe+X×2Pi,Nc×2Pi)

[0220] In this example, it can be seen that the instantaneous value of the increasing electrical angle position changes by an angle of Nc×2Pi radians with one mechanical revolution of the target.

[0221] In both cases, regardless of whether the instantaneous value of the increasing electrical angle position is obtained directly from the value of the electrical angle position θe, it is increased as a bijective function of the mechanical angle position θm of the target. In this respect, the instantaneous value of the increasing electrical angle position describes a complete mechanical revolution of the target.

[0222] In the remainder of this paper, the first example above will be developed in more detail, where the instantaneous values ​​of the increasing electrical angle positions can be obtained, for example, according to the following relationship:

[0223] θe_inc=mod(θe+X×2Pi) / Nc,2Pi)

[0224] Therefore, it should be understood that, due to its bijective nature, the instantaneous value of the incremental electrical angle position corresponds to a value outside the 2Pi radian mechanical angle. However, at this stage, the instantaneous value θe_inc of the incremental electrical angle position cannot determine the unique relative position of the rotor with respect to the stator in the 2Pi radian mechanical angle. Specifically, since the initial relative position of the rotor with respect to the stator in the 2Pi radian mechanical angle is unknown at the start of the increment (i.e., when calculating the incremental electrical angle position θe_inc), the instantaneous position relative to the stator in the 2Pi radian mechanical angle is also unknown. In effect, the instantaneous value θe_inc of the incremental electrical angle position allows us to know the relative angular position of the contrast region with respect to the stator, but it is impossible to know which contrast region is in front of the detector. In other words, the incremental electrical angle position θe_inc can be considered equivalent to the mechanical angle position θm, but with an unknown initial angular phase shift. The original position for calculating the incremental electrical angular position θe_inc is the same as the mechanical angular position θm0 of the rotor relative to the stator.

[0225] Therefore, during the measurement, θm = θe_inc + θm0 is given.

[0226] In practice, it is advantageous to use the reference value (e.g., 0) of the electrical angle position θe as the origin for calculating the incremental electrical angle position θe_inc. In the formula above, θe_inc = mod((θe + X × 2Pi) / Nc, 2Pi), therefore, the original position for calculating the incremental electrical angle position θe_inc is the position where θe = 0 and X = 0.

[0227] It should be understood that, without any specific precautions, the mechanical angular position θm0 of the rotor relative to the stator is different for the original position calculated for the incremental electrical angular position θe_inc, for the calibration phase and for each setup phase of each measurement session. This gives the original angular phase shift value θm0_cb for the calibration phase and the original angular phase shift value θm0_i for the setup phase of the "i"th measurement session.

[0228] Therefore, during the calibration phase, the mechanical angular position of the rotor relative to the stator is given by the following formula:

[0229] θm=θe_inc+θm0_cb.

[0230] Therefore, during the i-th subsequent measurement session, the mechanical angular position of the rotor relative to the stator is given by the following formula:

[0231] θm=θe_inc+θm0_i

[0232] Therefore, the challenge lies in determining the angular measurement offset between the mechanical angular positions of the rotor and the stator, delta_i_θm0=(θm0_i-θm0_cb), the original position for calculating the incremental electrical angular position θe_inc, which is used on the one hand for the setup phase of the measurement session under consideration, and on the other hand for the calibration phase.

[0233] If precautions are taken, and the reference value (e.g., 0) of the electrical angle position θe is used as the starting point for calculating the incremental electrical angle position θe_inc, then this angle measurement offset corresponds to an integer number of angle ranges within a comparison region, which gives:

[0234] delta_i_θm0=(θm0_i-θm0_cb)=ki×2Pi / Nc

[0235] Where ki is an integer specific to each measurement session. Note that it is possible to adopt the convention that the integer ki associated with a measurement session is contained in the range from 0 to (Nc-1). The number of levels of the offset for a given measurement session is then called ki.

[0236] However, the method according to the invention does not impose any such precautions.

[0237] In all cases, the angle measurement offset delta_i_θm0 is the difference between the mechanical angular position of the rotor relative to the stator at the original measurement position, which is used on the one hand for the setup phase of the measurement session under consideration, and on the other hand for the calibration phase.

[0238] The method according to the present invention specifies that at least one electronic signature of the target is determined by computer.

[0239] During the calibration phase, the electrical calibration signature is determined by a computer.

[0240] In the setup phase of the subsequent measurement session, the electrical setup signature is determined.

[0241] Each electronic signature is determined by a pair of signature values ​​or a series of signature value pairs, where the "j"th pair of signature values ​​includes:

[0242] - An amplitude signature value Asig derived from at least one of the values ​​of at least one of the measurement signals S1 and S2 at at least one angular position relative to the target. j ;

[0243] - and an incremental electrical angle position signature value θe_inc_sig for one or more angular positions of a target corresponding to one or more instantaneous amplitude values. j One or more instantaneous amplitude values ​​are considered to determine the amplitude signature value Asig. j .

[0244] Further examples of rules for determining electronic signatures will be given, particularly for determining the amplitude signature value Asig. j and the incrementing electrical angle position signature value θe_inc_sig j Examples of rules. Electrical signatures are important for paired features including targets and detectors respectively arranged on the rotor and stator. Changes to the target and / or detector, and / or removal / reassembly of the target and / or detector will preferably result in a new determination of the electrical calibration signature. Simultaneously, an electrical setup signature is determined during the setup phase at the beginning of each measurement session. The measurement session can be triggered, for example, at each startup of the rotor / stator system where the angular position is to be measured, and thus, for example, each time the sensor is re-energized.

[0245] Therefore, it should be understood that for both the calibration and setup phases, the electronic signature is determined as a function of the incremental electrical angular position, and thus as a function of the original angular phase shift θm0, which corresponds to the mechanical angular position of the rotor relative to the stator at the original position calculated for the incremental electrical angular position θe_inc. Since this original angular phase shift θm0 is different for the calibration phase and the setup phase of each measurement session, the electronic signature is essentially associated with this original angular phase shift θm0.

[0246] Typically, the amplitude signature value Asig can be used. j The incremental electrical angle position signature value θe_inc_sig j Any pair of related values ​​(Asig) j ;θe_inc_sig j Preferably, each pair of values ​​corresponds to a single occurrence of that pair of values ​​at one mechanical revolution of the target, corresponding to the closest measurement tolerance.

[0247] Therefore, based on general examples, the electronic signature SIG can take the following form:

[0248] SIG={(Asig1; θe_inc_sig1); (Asig2; θe_inc_sig2),…}

[0249] In this general form, for the same sensor, the electrical calibration signature can therefore be determined, which can be written as

[0250] SIGcb = {(Asigcb1; θe_inc_sigcb1); (Asigcb2; θe_inc_sigcb2), ...} and determines the electrical setup signature for the corresponding "i"th measurement session, which can then be written as

[0251] SIGi={(Asigi1;θe_inc_sigi1);

[0252] (Asigi2;θe_inc_sigi2-),…},

[0253] In general, this form is therefore a two-row matrix, with one row for the magnitude signature value Asig of a pair of signature values. j One row represents the incrementing electrical angle position value for a pair of reference values. The number of columns in the matrix depends on the number of pairs of reference values. In such a matrix, the magnitude signature value Asig for the pair of reference values ​​is... j The row can be considered as an ordered vector of magnitude signature values ​​Asig = (Asig 1, Asig2,…), and the incremental electrical angle position signature value θe_inc_sig of paired reference values. j The rows can be considered as an ordered vector of increasing electrical angle position signature values ​​θe_inc_sig=(θe_inc_sig1,θe_inc_sig2,…).

[0254] Of course, the signature can also be written in transpose, and therefore in the form of a two-column matrix, one column for the magnitude signature value Asig of the paired reference values. j And a column for increasing electrical angle position values ​​for paired reference values. Then, the number of rows in the matrix depends on the number of logarithms of the reference values. In such a matrix, the magnitude signature value Asig of the paired reference values. j The columns can be considered as an ordered vector Asig of amplitude signature values, and the electrical angle position signature values ​​θe_inc_sig are paired reference values ​​in an increasing manner. j The columns can be considered as an ordered vector θe_inc_sig of increasing electrical angle position signature values.

[0255] In some cases, as an electronic signature, it may be sufficient to retain (especially by computer recording) either an ordered vector Asig containing only the amplitude signature value or an ordered vector θe_inc_sig containing only the incrementing electrical angle position signature value. For example, in some cases, the rules for determining the electronic signature directly determine either the ordered vector Asig containing the amplitude signature value or the ordered vector θe_inc_sig containing the incrementing electrical angle position signature value, making it unnecessary to record them by computer. Therefore, in such cases, especially when recorded by computer, it is sufficient to retain only the other of the ordered vector Asig containing the amplitude signature value or the ordered vector θe_inc_sig containing the incrementing electrical angle position signature value (i.e., the one not directly determined by the rules for determining the electronic signature value).

[0256] To increase the robustness of the system, it is advantageous, for example, to use at least one pair of signature values ​​(Asig). j ;θe_inc_sigj ) as a signature value pair, wherein the at least one pair of signature values ​​(Asig) j ;θe_inc_sig j The instantaneous amplitude value Asig j It is most clearly distinguished from other instantaneous amplitude values ​​that correspond to the same (non-incremental) electrical angle position value.

[0257] To determine a pair of signature values, there may be reasons to determine the magnitude signature value Asig. j According to the rules for determining the amplitude signature value, this value must be derived from at least one of the values ​​of at least one of the measurement signals S1 and S2 for at least one angular position of the target. According to an embodiment, the amplitude signature value is derived from at least one instantaneous amplitude value representing the intensity of the physical variable at the considered angular position. The instantaneous amplitude value is determined, for example, by a calculation rule for instantaneous amplitude values. This calculation rule uses at least one of the values ​​of at least one of the measurement signals S1 and S2 for at least one angular position of the target. Furthermore, the rules for determining the amplitude signature value use one or more instantaneous amplitude values.

[0258] Now refer to Figure 6 The first example describing the determination of a digital signature is based on, for example, Figure 5 The measurement signals S1 and S2 are shown.

[0259] Figure 6 Measurement signals S1 and S2 are shown again. The horizontal axis corresponds to the axis of the increasing electrical angle position. Recall that the increasing electrical angle position corresponds to the mechanical angle position, which corresponds to the most recent original angular phase shift, which was unknown at this stage.

[0260] Figure 6 The example also shows the electrical angle position θe calculated using the calculation rule for electrical angle position as defined above. This example uses the calculation rule derived from the atan2 function, but it returns θe values ​​between 0 and 2Pi:

[0261] If S1≥0 and S2>0, then θe=arctan(S1 / S2)

[0262] If S2 < 0, then θe = arctan(S1 / S2) + PI

[0263] If S1 < 0 and S2 > 0, then θe = arctan(S1 / S2) + 2 * PI

[0264] If S1 > 0 and S2 = 0, then θe = PI / 2

[0265] If S1 < 0 and S2 = 0, then θe = 3 * PI / 2

[0266] Therefore, the electrical angle position here is a discontinuous function over one mechanical revolution of target 12, but has quasi-periods T1, T2, ..., T4, with an angular range of 2Pi / Nc, and in each quasi-period corresponding to a contrast region, the electrical angle position θe is a linear function of the increasing electrical angle position. For each contrast region passed in front of the detector, the electrical angle position varies from 0 to 2Pi over the angular range of the contrast region (i.e., the mechanical angle range with values ​​of 2Pi / Nc radians). This recalls that in the context of... Figure 5 and Figure 6 In the example, the target includes four contrast regions.

[0267] Figure 6 The variation of the function A = square_root(S1^2 + S2^2) is also shown. This function is an example of a rule for calculating the instantaneous amplitude value of the intensity of a physical variable at a considered angular position, which can be implemented in the rule for determining the amplitude signature value. Therefore, the function A = square_root(S1^2 + S2^2) is equivalent to calculating the square root of the sum of squares of the measured signal values ​​at the target's angular position. Thus, it truly represents the intensity of the physical variable measured at the considered angular position. The function A is calculated for different angular positions over one mechanical revolution of the target.

[0268] However, other calculation functions can be implemented as rules for calculating instantaneous amplitude values, which represent the intensity of a physical variable with respect to the considered angular position. For example, the function A = (S1^2 + S2^2) can be used. According to another example, a function of the type involving a linear combination of the measured signal values ​​at that angular position can be used, which can be written, for example, as A = a S1 + b S2 + c, where a, b, and c are parameters; or A = square_root(a S1^2 + b S2^2 + c), where a, b, and c are parameters; or A = (a S1^2 + b S2^2 + c), where a, b, and c are parameters. Note that in these last three examples, one or the other of parameters a and b can be zero, but not both. Similarly, parameter c can be zero.

[0269] According to another example, a function of the type that is a linear combination of the values ​​of the absolute measurement signals S1 and S2 relating to that angular position can be used. In such an example, the linear combination of absolute values ​​can simply be the sum of the absolute measurement signal values ​​for that angular position.

[0270] Since the values ​​of signals S1 and S2 are used to determine the absolute position, it is important to consider all parameters that may affect these signals. It may be advantageous to operate with relative values ​​(i.e., minimum and maximum values ​​marked at one mechanical revolution) associated with the peak-to-peak values ​​at one mechanical revolution to compensate for slow effects, such as, for example, temperature changes. If the sensor includes an amplification stage with variable gain, the gain value can be considered, or the peak-to-peak measurement can be repeated at one mechanical revolution after each gain change.

[0271] In all cases, in the rules for calculating the instantaneous amplitude value of the intensity of a physical variable for the angular position under consideration, each measured signal value can be corrected (e.g., reduced) by the offset value of that signal.

[0272] exist Figure 6 In the example shown, it can be seen that the instantaneous amplitude value given by the function A = square_root(S1^2 + S2^2) is not constant over the entire range of one complete mechanical revolution of the target. Furthermore, note that the instantaneous amplitude value given by the function A = square_root(S1^2 + S2^2) is not constant over the angular range corresponding to a contrast region. However, it should be noted that if processing is performed using an ideal sensor transmitting perfect sinusoidal signals S1 and S2, the instantaneous amplitude value given by the function A = square_root(S1^2 + S2^2) is constant over one complete mechanical revolution of the target.

[0273] Note that, over a complete number of mechanical revolutions, the set of paired values ​​that associate the instantaneous amplitude value with the corresponding incremental electrical angular position forms an example of the sensor's electrical signature. In this case, the rule for determining the amplitude signature value includes all calculated values ​​considering the instantaneous amplitude value. However, depending on the number of measurement points over a complete number of revolutions on the target, the amount of data required to represent this electrical signature, and therefore the amount of memory required for storage, will be considerable. This will also affect the computational load required to process the electrical signature, which itself will become quite large.

[0274] In addition, another set of rules for defining signature values ​​can be used to define the electronic signatures of sensors that require a smaller amount of data.

[0275] For example, instead of retaining a pair of values ​​for all measurement points, one can retain only pairs of values ​​corresponding to identifiable amplitude values ​​by applying a deterministic rule that includes selecting amplitude signature values ​​for only a few instantaneous amplitude values. Therefore, it is possible to select and retain only two pairs of values ​​corresponding to the local maximum values ​​of the instantaneous amplitude values ​​given by function A, which is used as the rule for calculating the instantaneous amplitude values. Figure 6In the example, there are two local maxima, Amax1 and Amax2, corresponding to the increasing electrical angle positions θe_inc_max1 and θe_inc_max2, respectively. In this case, the electrical signature SIG can be defined as a set of two pairs of signature values:

[0276] SIG={(Amax1; θe_inc_max1); (Amax2; θe_inc_max2)}

[0277] Of course, only the pairs of values ​​corresponding to the local minimum of the instantaneous amplitude value given by function A can be retained, where the signature will have the form {…;(Amin j θe_inc_min j The signature values ​​are in the form of a set of values. In another variation, pairs of values ​​can be retained that correspond to the local maximum values ​​of the absolute values ​​of the instantaneous amplitude values ​​given by function A.

[0278] In cases where there are many local maxima or local minima, the rules for determining the amplitude signature value can select some of them, such as retaining only a predefined number of local maxima or minima, and / or a predefined number of local maxima and / or minima for each quasi-period Tk of the electrical angle position θe, etc.

[0279] In this case, the rule for determining the amplitude signature value can retain only one pair of signature values, for example, corresponding to the absolute maximum value Amax or absolute minimum value Amin of the instantaneous amplitude value given by the calculation function A representing the intensity of the physical variable. For the case where the rule for determining the amplitude signature value retains only a single pair of signature values ​​corresponding to the absolute maximum value Amax or absolute minimum value Amin of the instantaneous amplitude value given by the calculation function A representing the intensity of the physical variable, this pair of signature values ​​can be specified to include:

[0280] -The absolute maximum value Amax and absolute minimum value Amin of the instantaneous amplitude value given by function A are obtained by using the amplitude signature value.

[0281] - The electrical angular positions corresponding to the absolute maximum value Amax and the absolute minimum value Amin, respectively, through the angular position signature values, implicitly form an increasing electrical angular position associated with the extrema of function A, because it can be uniquely determined over a mechanical revolution by its association with the extrema.

[0282] Typically, a pair of signature values ​​corresponds to a pair of identifiable values ​​derived from an instantaneous amplitude value and a pair of values ​​formed by the corresponding incremental electrical angle position.

[0283] As seen in the previous example, the identifiable value can be a specific instantaneous amplitude value. Therefore, an instantaneous amplitude value can be defined as an identifiable value, for example, taking one of the following values: a local maximum, a local maximum within a quasi-period of the electrical angle position, a local minimum, a local minimum within a quasi-period of the electrical angle position, an absolute maximum over a mechanical revolution of the target, an absolute maximum over a quasi-period of the electrical angle position, an absolute minimum over a mechanical revolution of the target, an absolute minimum over a quasi-period of the electrical angle position, etc. The identifiable value is not necessarily based on a maximum or minimum value. An identifiable value can be defined, for example, as an instantaneous amplitude value taking a given value, such as an amplitude reference value, which can be zero, an average over a mechanical revolution of the target (which, for example, corresponds to an offset value), an average over a quasi-period of the electrical angle position, an average over a half-period of the electrical angle position, etc. In the latter case, the average is, for example, an arithmetic mean. Typically, in this case, only the ordered vector θe_inc_sig of the increasing electrical angle position signature values ​​corresponding to these identifiable instantaneous amplitude values ​​can be retained as the signature. In this ordered vector, the amplitude reference value forms an amplitude signature value associated with the incremental electrical angle position signature value.

[0284] In a variant, the identifiable value can be a specific electrical angle position θe. Therefore, a rule for determining the amplitude signature value can be implemented that retains, for example, one or more pairs of signature values ​​corresponding to specific conditions of the electrical angle position θe. Thus, the rule for determining the amplitude signature value can be designed to retain one or more pairs of signature values ​​corresponding to one or more predetermined values ​​of the electrical angle position θe. In this case, the signature can include as many signature value pairs as the number of comparison areas, because over one mechanical revolution of the target, there are as many quasi-periods of electrical angle positions θe as there are comparison areas. In a variant, a rule for determining the amplitude signature value can be implemented that is designed to retain one or more pairs of signature values ​​corresponding to a predefined series of increasing electrical angle positions θe_inc. Typically, in this case, only an ordered vector Asig of amplitude signature values ​​corresponding to these identifiable values ​​of the increasing electrical angle positions can be retained as the signature. In such an ordered vector, the order rank of each amplitude signature value in the vector forms an increasing electrical angle position signature value associated with the amplitude signature value.

[0285] Figure 7 , Figure 8 and Figure 9 An example of another rule for determining the signature value is shown here, specifically the magnitude signature value. In this example, Figure 7 With Figure 5Signals S1 and S2, corresponding to the measurements of physical variables at the first measurement position P1 and the second measurement position P2, are shown in the same manner, with the sensor comprising six contrast regions in this example. The variation of the signal over one full mechanical revolution of the target 12 is also shown here, i.e., the variation over a mechanical rotation angle of 2Pi radians relative to the stator. Note that within the 2Pi radian mechanical angle range, each signal comprises six quasi-cycles, each quasi-cycle having a positive half-cycle and a negative half-cycle relative to the offset value of the signal under consideration. For the previous example, the maximum absolute value of the signal is not the same from one quasi-cycle to another. In this way, the signal thus has a quasi-sinusoidal characteristic over a 2Pi radian mechanical angle because it comprises consecutive positive and negative half-cycles.

[0286] Figure 8 The calculation of the electrical angle position θe, as defined above, is also shown, where the horizontal axis is equivalent to the axis of the increasing electrical angle position. Therefore, the electrical angle position here is a discontinuous function over one mechanical revolution of the target 12, but has half-cycles T1, T2, ..., T6, with an angular range of 2Pi / Nc radians, and over each quasi-cycle T1, T2, ..., T6, the electrical angle position θe is a linear function of the increasing electrical angle position.

[0287] Figure 8 The variation of the function A = square_root[(S1 - S1m)^2 + (S2 - S2m)^2] is also shown, which serves as the formula for calculating the instantaneous amplitude value representing the intensity of the physical variable at the considered angular position. In other words, for this example, each measured signal value is reduced by the offset value of the signal, but uncorrected eigenvalues ​​can be used. The variations in the electrical angular position θe and the instantaneous amplitude value are also shown here, for their variation over one full mechanical revolution of target 12, i.e., the variation in the mechanical rotation angle of the rotor relative to the stator by 2Pi radians.

[0288] Figure 9 Again, the variation of the function A = square_root[(S1-S1m)^2+(S2-S2m)^2] used as the calculation rule for the instantaneous amplitude value over one full mechanical revolution of target 12 is shown at a larger scale on the vertical axis. It also shows the average values ​​for each quasi-cycle T1, T2, ..., T6 for the electrical angle position θe, for example, the arithmetic mean Am1, Am2, ..., Am6 of the instantaneous amplitude values ​​within the quasi-cycle of the considered electrical angle position. It can be seen that the average value of the instantaneous amplitude value within the considered half-cycle is different for all quasi-cycles. In this example, each quasi-cycle has different values ​​Am1, Am2, ..., Am6 of the average instantaneous amplitude value over the considered quasi-cycle.

[0289] Therefore, in the rules for determining the amplitude signature value, one, more, or all of the values ​​Am1, Am2, ..., Am6 of the average value of the instantaneous values ​​can be selected to form one, more, or Nc signature value pairs.

[0290] In this example, the average of the instantaneous amplitude values ​​within the considered quasi-period is an example that allows a value representing the intensity of a physical variable within a given angular range to be attributed to that range over an angular range of 2 Pi radians at increasing electrical angular positions. However, another deterministic rule for calculating such a representative value can be used.

[0291] For example, the integral of the instantaneous amplitude value over the considered angular range can be calculated as a representative value for that range, such as for a quasi-period. Alternatively, for a quasi-period, the following can be chosen as representative values ​​for the considered angular range:

[0292] -A linear combination of the squares of the measured signal values ​​or the average value of the square root of such a linear combination within the quasi-periodic angle range;

[0293] - The average value of a linear combination of absolute measurement signal values ​​within this angular range;

[0294] - The minimum or maximum instantaneous amplitude value within the range of angles considered.

[0295] Then, for any given angular range, this final variation is a summary of what was described above in the example of the maximum or minimum value over a quasi-period of the electrical angular position.

[0296] In all cases, each measured signal value can reduce the offset of that signal.

[0297] In each pair of signature values, the representative value used as the amplitude signature value (here, the average values ​​Am1, Am2, ..., Am6) is associated with an incremental electrical angular position signature value corresponding to one or more angular positions of the target considered to determine one or more instantaneous amplitude values ​​of the amplitude signature value.

[0298] Figure 10 The advantages of using the average of the instantaneous amplitude values ​​within the considered quasi-period as the amplitude signature value are shown. Specifically, the figure illustrates the case where signals S1 and S2 are particularly noisy. This inevitably leads to a particularly noisy curve showing the instantaneous amplitude values ​​over one mechanical revolution of the target itself. On the other hand, it can be seen that the noise has a weak effect on the average of the instantaneous amplitude values ​​within the considered quasi-period. Therefore, a determination rule for a deterministic signature that is particularly robust to noise and measurement uncertainty is obtained.

[0299] According to one possibility, for a pair of signature values, the incremental electrical angle position signature values ​​can, for example, correspond to the electrical angle positions θe_inc_T1, θe_inc_T2, ..., θe_inc_T6... at the beginning, end, or middle of the corresponding quasi-periods T1, T2, ..., T6, where the average of the instantaneous amplitude values ​​within the considered quasi-period is calculated. In this case, the electrical signature can be written in the following form:

[0300] SIG={(Am1; θe_inc_T1); (Am2; θe_inc_T2); (…)}

[0301] According to another possibility, for a pair of signature values, the increasing electrical angle position signature value can correspond to the rank of the corresponding quasi-period of the electrical angle position for which the amplitude signature value has already been calculated. In this case, the rank corresponds to the order of the considered quasi-periods of the electrical angle position relative to other quasi-periods of the electrical angle position within one quasi-period of the increasing electrical angle position. In this case, the signature can simply be written in the form of an ordered sequence of amplitude signature values, hereinafter referred to as an ordered vector, because the order of the sequence corresponds to the association with the rank, and the rank corresponds to the increasing electrical angle position signature value of one or more angular positions of the target, which correspond to one or more instantaneous amplitude values ​​considered to determine the amplitude signature value. The signature can then be written in the form of an ordered vector of amplitude signature values:

[0302] SIG={Am1;Am2;…;AmNc}

[0303] In the example above, the amplitude signature value is calculated for each quasi-cycle at the electrical angle position. However, for example, it is likely that a finer resolution could be chosen by calculating the amplitude signature value for each half-cycle or each quarter-cycle.

[0304] In all cases, the number of signature value pairs reserved for a signature can be increased by implementing a rule for determining the signature value (e.g., a rule for determining the amplitude signature value), retaining signature value pairs corresponding to different types of identifiable values ​​as defined above, for example, by retaining some value pairs based on identifiable values ​​related to amplitude, and for others based on identifiable values ​​related to electrical angle position.

[0305] Therefore, a signature constructed in this way is one that allows indexing of a specific value representing the intensity of a physical variable measured at a specific location, expressed in terms of an increasing electrical angular position term.

[0306] As described above, the determination of the sensor's electrical signature is completed for the first time during the sensor's calibration phase. During this calibration phase, the signature determined by the computer is the electrical calibration signature. Therefore, the incremental electrical angle position signature value is indexed relative to the original angle calibration phase shift θm0_cb generated by the original selection of the incremental electrical angle position calculated during the calibration phase.

[0307] Note that the electrical calibration signature can be determined by the detector itself via a computer, and then the calculations required to determine the signature are performed, for example, within the electronic control unit 28 (including when it is integrated into the detector housing as described above). However, if the calibration phase is completed on a calibration bench, the determination of all or part of the sensor's electrical calibration signature can be performed by another electronic control unit belonging to the calibration bench. It is also conceivable that the electrical calibration signature is determined by the electronic control unit of an external system to which the detector is connected.

[0308] Once the electrical calibration signature is determined, it is recorded by a computer. This recording can be made in an electronic memory integrated into the sensor, for example, forming part of the electronic control unit 28 (including when it is integrated into the detector cartridge as described above). However, alternatively or additionally, the recording can be made in an electronic memory located away from the sensor, for example, forming part of an external system to which the detector is connected.

[0309] The electrical calibration signature is recorded in a way that allows it to be invoked at every stage of a large number of subsequent measurement sessions.

[0310] It should be understood that the electrical calibration signature is essentially associated with the original angle calibration phase shift value θm0_cb. This relationship is generated by determining one or more incremental electrical angle position signature values.

[0311] In use, when it is desired to implement the sensor to measure the mechanical angular position θm of the rotor by implementing the i-th measurement session, the i-th measurement session begins from the setup phase for the measurement session.

[0312] During this setup phase, the signature determined by the computer is the electrical setup signature. Therefore, the incremental electrical angle position characteristic value is indexed relative to the original angle setup phase shift θm0_i generated by the original selection of the incremental electrical angle position calculated during this setup phase, and is thus valid for the i-th measurement session.

[0313] Note that the electrical setup signature can be determined by the detector 14 itself via a computer, and then the calculations required to determine the signature are performed within the electronic control unit 28 (especially when it is integrated into the detector housing as described above). It is also conceivable that the electrical setup signature is determined by the electronic control unit of an external system to which the detector is connected.

[0314] Once the electrical setup signature is determined, the electrical setup signature of the target can be selectively recorded by computer, for example, in an electronic memory integrated into the sensor that forms part of the electronic control unit 28 integrated into the detector housing as described above, and / or in an electronic memory remote from the sensor that forms part of an external system to which the detector is connected.

[0315] During the setup phase, the electrical setup signature is reset relative to the electrical calibration signature.

[0316] Specifically, since this involves the same sensors mounted on the rotor and stator in the same manner, it is assumed that the amplitude signature values ​​are identical, achieving the closest possible measurement uncertainty. On the other hand, each incremental electrical angular position signature value associated with these amplitude signature values ​​is offset angularly by the same angular measurement offset delta_i_θm0=(θm0_i-θm0_cb).

[0317] If precautions are taken for both the calibration and setup phases, using the reference value (e.g., 0) of the electrical angle position θe as the starting point for calculating the incremental electrical angle position θe_inc, then the angle measurement offset corresponds to an integer number of angle ranges in the comparison area. This gives delta_i_θm0=(θm0_i-θm0_cb)=ki×2Pi / Nc, making the reset operation equivalent to the search offset level number ki.

[0318] Therefore, the reset operation is equivalent to searching for the angle measurement offset delta_i_θm0 = (θm0_i - θm0_cb), preferably by searching for the offset's level number ki. This allows the electrical calibration signature to be found again by "rotating" the electrical setting signature, which consists of the value of the angle measurement offset. Thus, the reset operation is a search for the angle measurement offset, which makes it possible to set the electrical calibration signature at an angle using the electrical setting signature. This is accomplished, for example, by calculating the angle measurement offset, which is applied to the incremental electrical angle position signature value of either the electrical setting signature or the electrical calibration signature, so that the difference between the electrical setting signature and the electrical calibration signature can be minimized.

[0319] Graphical reasoning is performed using the curve of the change of a function relative to the instantaneous amplitude value of the intensity, which is used to represent the physical variable of the considered angular position. Figure 6 and Figure 7 In the example, this is equivalent to finding the ring offset, which is the number of electrical angle alignment periods. For this number of electrical angle alignment periods, the curve acquired during the setup phase needs to be translated to superimpose it as well as possible onto the curve acquired during the calibration phase.

[0320] When a signature can be written as an ordered vector of signature values, such as an ordered vector of magnitude signature values, e.g., SIG = Asig = (Asig1; Asig2; ...; AsigNc}, the reset operation corresponds to determining the value of the number of cyclic permutation increments that must be applied to the signature values ​​of the ordered vector of electrically set signatures in order to find the ordered vector of electrically calibrated signatures again. The number of cyclic permutation increments gives the offset level number ki.

[0321] Note that if precautions have been taken to use the reference value (e.g., 0) of the electrical angle position θe as the starting point for calculating the incremental electrical angle position θe_inc, then the determination of the angle measurement offset can be performed through simple and fast calculations, with the possibility of accepting high measurement uncertainties. Specifically, the number of possible solutions is limited to the number of comparison regions NC.

[0322] In practice, Figure 11 The six radar charts show the relationship with Figures 7 to 10 The diagram shows the electrical calibration signature SIGcb and electrical setup signature SIGi of the same type of sensor as the sensor used to measure the signal and its derived values. Therefore, a sensor with six contrast regions is used. In each diagram, the electrical calibration signature SIGcb is shown as a closed dashed curve, and the electrical setup signature SIGi is shown as a closed solid curve. The vertices of the closed curves represent paired signature values ​​in polar coordinates. In these diagrams, the angle between two adjacent vertices of the closed curve relative to the center of the diagram is directly an image of the angular range of a contrast region. Figure A shows two diagrams generated by the measurement. Figures B through F each correspond to the value of the number "n" of the cyclic permutation increments applied to the electrical setup signature SIGi, which is also designated as SIGi[n] in Figures B through F, where n is the number of cyclic permutation increments. Note that Figure E shows two diagrams with the number of cyclic permutation increments that ensure signature reset. It corresponds to four cyclic permutation increments in the clockwise direction, and therefore two cyclic permutation increments in the counterclockwise direction. It can be easily deduced that in order to reset the electrical setting signature corresponding to the electrical calibration signature, the electrical setting signature needs to be offset clockwise by 4 times the angle range of the comparison area, and therefore counterclockwise by 2 times the angle range of the comparison area.

[0323] Mathematically, the distance between ordered vectors representing signatures can be determined for each value of the number of cyclic increments "n", such as the "Manhattan distance". Figure 12 ) or "Euclidean distance" Figure 13 This reset is performed. Other distances, such as Minkowski distance or Chebyshev distance, can be used. Therefore, for an electrically calibrated signature, it can be written as an ordered vector.

[0324] SIGcb=Asigcb{Asigcb1;Asigcb2;…;Asigcb Nc}

[0325] And the electrical setup signature SIGi obtained for the "i"th measurement session, which can also be written in the form of an ordered vector as

[0326] SIGi=Asigi={Asigi1;Asigi2;...;Asigi Nc},

[0327] Then, the electrical calibration signature SIGcb and the electrical setting signature SIGi are represented by ordered vectors formed by Nc amplitude reference values ​​for calibration and setting, respectively.

[0328] For each value of the number "n" of the cyclic permutation increment, n varying from 1 to Nc, the electrical setting signature is transformed into an offset electrical setting signature:

[0329] SIGi[n]={Asigi 1+n,在[1,Nc]的范围内 Asigi 2+n,在[1,Nc]的范围内 ...

[0330] Asigi Nc+n,在[1,Nc]的范围内}

[0331] Therefore, for each value of the number of cyclic permutation increments "n", where n varies from 1 to Nc, the Euclidean distance DE(n) can be calculated:

[0332] [Mathematical Expression 1]

[0333]

[0334] Of course, since the square root function is always increasing, the same reasoning can be applied to the square of the distance, so it is not necessary to calculate the square root.

[0335] Whether it's Manhattan distance, Euclidean distance, Minkian distance, or Chebyshev distance, the value n of the increment of the cyclic permutation with the minimum distance is... min Its electrical setting signature must be offset to be as close as possible to the value of the electrical calibration signature. This value n min The value ki is given, which is the rank number for measuring the offset of session i, where

[0336] delta_i_θm0=(θm0_i-θm0_cb)=ki×2Pi / Nc

[0337] And in the direction of target rotation, ki = n min Or ki = Nc - n min .

[0338] In this way, the mechanical angular position during the i-th measurement session is given by the following relationship:

[0339] θm = mod(θe_inc + n) min ×2Pi / Nc+θm0_cb,2Pi).

[0340] To explain the principle of resetting by applying a cyclic permutation to the electrical calibration signature, we can consider a sensor for which an electrical calibration signature has been acquired, having: a quantity M of reference values, M>=Nc, and for this sensor, for this electrical calibration signature, different incremental electrical angular position signature values ​​θe_inc_sigcb. j They are equidistant. In this case, after determining the electrical setting signature of the same sensor, the following method can be used: apply a cyclic arrangement to the M setting amplitude reference values ​​Asigi. j The ordered vector Asigi (Asigi1, Asigi2, ...) is formed, and each cyclic permutation of the ordered vector Asigi is combined with M calibration amplitude reference values ​​Asigcb. j The resulting ordered vectors Asigcb{(Asigcb1;θe_inc_sigcb1);(Asigcb2;θe_inc_sigcb2),…} are compared.

[0341] This method is equivalent to, for example, calculating ki in the following way:

[0342] ki=round{(Nc / M)×arg min[1<=n<=M]D(Asigcb,C[n](Asigi))}

[0343] in:

[0344] - "D" is an operator that calculates any distance between an ordered vector Asigcb and C[n](Asigi), such as the distance from one of the distance examples given above;

[0345] - "round" is a function that rounds to the nearest integer;

[0346] The operator "arg min[1<=n<=M]" is an operator that returns the value that is the smallest increment from D(Asigcb,C[n](Asigi)) among the integer values ​​of the number "n" that vary from 1 to M.

[0347] -C[n] is an operator that performs a cyclic permutation of "n" elements, thus performing a cyclic permutation on the components of an ordered vector (Acgi), and it can be mathematically defined by the following relation:

[0348] C[n](X) = {X_mod(j + n, M)} where j ranges from 0 to M - 1.

[0349] If M is very large, for example much larger than Nc, it may also be advisable not to apply the permutation method directly to the amplitude reference values. For example, if M is very large, pre - compressing the data may be beneficial. This compression includes transforming any signature SIG = {(Asig j ; θsig j )(where j ranges from 1 to M) into a signature of smaller size to be represented as SIGcomp = {(Asigcomp j ; θsigcomp j ), where j ranges from 1 to P}, where P is the size of the compressed signature (where P < M and for example P >= Nc). This transformation can include, for example, taking the average of the amplitude reference values Asigj over P angular segments. The same compression must be applied to the ordered vector Asigi j formed by M set amplitude reference values Asigi to obtain the compressed ordered vector Asigicomp, and applied to the ordered vector Asigcb j formed by M calibrated amplitude reference values Asigcb to obtain the compressed ordered vector Asigcbcomp. Next, in the same way as previously described, it is sufficient to apply the permutation method to the P values of the vectors thus formed. Using these new definitions, this will give:

[0350] ki = round{(Nc / P) × arg min[1 <= n <= P]D(Asigcbcomp, C[n](Asigicomp))}

[0351] Another way to calculate the angular measurement offset delta_i_θm0 = (θm0_i - θm0_cb) is to use the Fourier transform. Based on the electrical calibration signature SIGcb = {(Asigcb j ; θe_inc_sigcb j ), where j ranges from 1 to M}, calculate the following two projections:

[0352] T_sin_cb = Sum(Asigcb j x sin(θe_inc_sigcb j )), where j ranges from 1 to M;

[0353] T_cos_cb = Sum(Asigcb j × cos(θe_inc_sigcb j )), where j ranges from 1 to M.

[0354] Next, calculate the equivalent phase beta_cb of the electrical calibration signature: beta_cb = atan2(Tsin_cb, Tcos_cb).

[0355] The electrical setup signature for the setup phase of the "i"th measurement session is performed in the same manner, providing the electrical setup signature SIGi = {(Asigi j ;θe_inc_sigi j ),…}, where j is in the range from 1 to M, calculate the following two projections:

[0356] Tsin_cgi=Sum(Asigi j ×sin(θe_inc_sigi j ), where j is in the range from 1 to M;

[0357] Tcos_cgi = Sum(Asigi) j ×cos(θe_inc_sigi j )), where j is in the range from 1 to M.

[0358] Next, calculate the equivalent phase beta_cgi for the electrical signature: beta_cgi = atan2(Tsin_cgi, Tcos_cgi).

[0359] This is ultimately given by performing the following calculation:

[0360] ki = round{(Nc / 2Pi)(beta_cgi-beta_cbi)}, the rounding operator is the operator that returns to the nearest integer.

[0361] To put it very simply, it can also be shown that the electrical calibration signature is a signature that can be within the range of an ordered vector of signature values ​​at its electrical angular position:

[0362] SIGcb=θe_inc_sigcb={θe_inc_sigcb j ), where j is in the range from 1 to M.

[0363] For example, SIGcb = {17°, 88°, 112°, 130°, 310°}. Assuming the angle measurement offset value delta_i_θm0 = 172°, this gives the electrical calibration signature that will be written as SIGi = {122°, 189°, 260°, 184°, 302°}. Knowing SIGcb and SIGi, it is easy to find the angle measurement offset value again, since finding the value "n" of the number of cyclic permutations is sufficient, where the difference (θe_inc_sigcb) is... j+n -θe_inc_sigcb j If is a constant, then this constant is the angle measurement offset value. In this example, measurement noise is ignored. To account for measurement noise, the value "n" of the number of cyclic permutations needs to be chosen to obtain the minimum dispersion.

[0364] According to another example, starting from the general case of an electrical calibration signature with M pairs of reference values, it can be written as:

[0365] SIGcb={(Asigcb1; θe_inc_sigcb1); (Asigcb2; θe_inc_sigcb2),…,(Asigcb M ;θe_inc_sigcb M )}

[0366] And the electrical setting signature, used for the corresponding "i"th measurement session, can then be written as

[0367] SIGi={(Asigi1; θe_inc_sigi1); (Asigi2; θe_inc_sigi2-),…,(Asigi M ;θe_inc_sigi M )},

[0368] First, we can determine that this will be applied to the ordered vector Asigi = {Asigi1; Asigi2; ...; Asigi...} M The value of the number of cyclic permutations of} is "n". min This value can be extracted from the electrical setting signature SIGi, where the difference between ordered vectors is:

[0369] Asigi[n]={Asigi 1+n,在[1,M]的范围内 ;

[0370] Asigi 2+n,在[1,M]的范围内 ...

[0372] Asigi M+n,在[1,M]的范围内}

[0373] And the ordered vector Asigcb{Asigcb1; Asigcb2; ...; Asigcb} extracted from SIGcb. M} is the minimum value. Whether it's Manhattan distance, Euclidean distance, Minkian distance, or Chebyshev distance, this can be determined, for example as described above, by finding the value n of the number of increments of the cyclic permutation that minimizes the distance. min This completes the process. This value makes it very easy to find the angle measurement offset again, because it gives the value regardless of the value of j:

[0374] delta_i_θm0=θe_inc_sigi j+nmin,在[1,M]的范围内 -θe_inc_sigcb j

[0375] In addition, to compensate for unavoidable measurement errors, an average value (e.g., the average of the angle offset calculated according to the formula for multiple values ​​of j, or even for all values ​​of j, or even for all values ​​of j varying from 1 to M-1) can be used as the angle offset value.

[0376] It can be seen that the absolute position of the rotor relative to the stator in terms of angle depends on the original calibration angle phase shift value θm0_cb during the calibration phase. Recall that the original calibration angle phase shift value θm0_cb is the original position of the rotor relative to the stator in terms of mechanical angle position calculated for the incremental electrical angle position θe_inc during the calibration phase.

[0377] In some applications, it is not necessary to know the original calibration phase shift value θm0_cb. In this case, an arbitrary value of 0 can then be given, such that the mechanical angular position of the rotor relative to the stator is then given by the following relationship:

[0378] θm=mod(θe_inc+delta_i_θm0,2Pi)

[0379] In some of the examples above, it is written as:

[0380] θm = mod(θe_inc + n) min ×2Pi / Nc,2Pi)

[0381] In all cases, the mechanical angular position given by the sensor during the measurement session is a relative mechanical angular position, unique over a 360-degree rotation of the rotor relative to the stator. However, this position is given relative to an unknown original calibration phase shift value θm0_cb, which remains the same as long as the calibration is valid. If the relative mechanical angular position given by the sensor is used, for example, as a drive compensation rule, which itself is established as a function of the same original calibration phase shift value θm0_cb, then not knowing the original calibration phase shift value θm0_cb is not an obstacle.

[0382] Furthermore, note that if the original calibration phase shift value θm0_cb is unknown, but during calibration, care has been taken to use the reference value (e.g., 0) of the electrical angle position θe as the origin for calculating the incremental electrical angle position θe_inc, then the relative mechanical angle position given by the sensor can be used, for example, to drive a compensation law that has itself been established as a function of the periodic properties of the sensor. For example, if the angle position value transmitted by the sensor is used to drive a motor, it can be placed in this situation. In this case, it is advantageous to specify that the sensor has the same number of contrast regions as the number of pole pairs of the motor.

[0383] In other applications, it may be desirable to know this initial angular phase shift value relative to a given geometric position (e.g., relative to a reference frame associated with the stator), and therefore, the absolute mechanical angular position. This can be accomplished during calibration using a calibration bench equipped with an absolute position sensor. In this case, the initial absolute mechanical position of the rotor relative to the stator can be recorded on the calibration bench as the initial position for calculating the incremental electrical angular position θe_inc during the calibration phase, in the form of the relative position between the rotor's reference frame and the stator's reference frame. This initial absolute mechanical position is then used as the initial angular phase shift value for calibration. Therefore, for any subsequent measurement session where the same calibration remains valid, there will be an absolute mechanical angular position given by the following relationship:

[0384] θm=mod(θe_inc+delta_i_θm0+θm0_cb,2Pi)

[0385] In some of the examples above, it is written as:

[0386] θm = mod(θe_inc + n) min ×2Pi / Nc+θm0_cb,2*pi)

[0387] The calculations required to determine the mechanical angular position are performed, for example, within the electronic control unit 28 (particularly integrated into the detector housing as described above). It is also conceivable that the mechanical angular position is determined by the electronic control unit of an external system to which the detector is connected.

[0388] Of course, the mechanical angular position transmitted by the sensor can be corrected and compensated, for example, based on the mechanical angular position given by one of the methods described above, to improve its accuracy. In particular, during calibration, correction parameters to be applied to the mechanical angular position given by the methods described above can be defined, for example, based on data from calibration sensors available on the calibration bench. These corrections or compensations can be applied to either the absolute mechanical angular position or the relative mechanical angular position given by the methods described above.

Claims

1. A method for measuring the mechanical angular position θm of a rotor, characterized in that, The rotor can move relative to the stator around the axis of rotation (A1) multiple times, wherein: - Acquire a number of Ns electrical measurement signals (S1, S2), each electrical measurement signal representing the intensity of an electrical or magnetic variable at one of a number of Ns measurement positions (P1, P2), the Ns measurement positions being individual, fixed relative to the stator and offset about the axis of rotation by a given mechanical angle α, where Ns is an integer equal to 2 or 3; - The change in the electrical measurement signal at the measurement position is caused by the rotation of the target (12) in front of the measurement position under consideration. The target is mechanically connected to the rotor and has a number of Nc individual contrast regions (Zk) of the target, the number of which Nc is greater than or equal to 2, wherein the target includes contrasts of conductivity, permeability and / or magnetization. - The contrast regions (Zk) are distributed at an angle on the target in a periodic pattern around the axis of rotation. The pattern has Nc quasi-periods, and each contrast region extends around the axis of rotation at a mechanical angle of 2Pi / Nc. - Each contrast region (Zk) causes a change in the electrical measurement signal (S1, S2) acquired at each measurement position when the contrast region passes in front of each measurement position as it rotates around the axis of rotation. The change in the electrical measurement signal is a quasi-sine function of the electrical angular position θe of the contrast region relative to the measurement position, which changes by 2Pi radians with respect to the mechanical angular position θm of the target around the axis of rotation. - At least two of the contrast regions (Zk) have physical differences from each other, which produce differences in the magnitude of the intensity of the physical variables measured in the at least two individual quasi-cycles of the measurement signal at the same number of revolutions on the target; The mechanical angle α corresponding to the modulus offset is the angle of the measurement position around the rotation axis in 2Pi / Nc radians. For Ns=2, the measurement position offsets in Pi / (2×Nc) radians, and for Ns=3, the measurement position offsets in 2Pi / 3Nc radians. The method includes performing the following steps during the calibration phase, and then performing the following steps again during the setup phase of the measurement session: a) Acquire the Ns electrical measurement signals (S1, S2) over one mechanical revolution of the target around the rotation axis. b) For different mechanical angular positions at one mechanical revolution around the rotation axis of the target, the instantaneous value of the electrical angle position θe of the considered mechanical angular position is calculated using the calculation rule for electrical angle position. The calculation rule for electrical angle position considers: For Ns=2, the arctangent of the ratio of the values ​​of the two measured signals (S1, S2) for the time under consideration, or For Ns=3, the arctangent of the ratio of the values ​​of the two transforms obtained by applying the Clark transform to the three measured signals of the time under consideration; c) Calculate the instantaneous increment of the electrical angle position value θe_inc, which is obtained by incrementing the counter X by one unit for any 2PI radian change in the electrical angle position θe, which occurs in the same direction of rotation. d) Determine at least one electronic signature of the target by computer, said at least one electronic signature including an electronic calibration signature determined during the calibration phase and an electronic setup signature determined during the setup phase, each electronic signature being determined by a signature value pair or a series of signature value pairs, the signature value pairs including: An amplitude signature value derived from at least one value of at least one measurement signal from at least one measurement signal at at least one angular position relative to a target; And the angular position signature value corresponding to one or more angular positions of the target that are considered to determine the amplitude signature value; During the calibration phase, the electrical calibration signature of the target is recorded by computer. During the setup phase of the measurement session, the angle offset measurement value delta_i_θm0 is determined by a reset operation that includes calculating the angle measurement offset value. The angle measurement offset value is applied to the angle position signature value of the electrical setup signature or electrical calibration signature, so as to minimize the difference between the electrical setup signature and the electrical calibration signature. During the measurement session, the mechanical angular position θm of the rotor at a given time is determined by increasing the electrical angular position correction by an amount equal to the angular measurement offset value.

2. The method according to claim 1, characterized in that, The instantaneous increasing electrical angle position value θe_inc is obtained according to the following relationship: θe_inc= mod((θe + X × 2Pi) / Nc, 2Pi) The mechanical angular position θm of the rotor at a given time is determined as an incremental electrical angular position θe_inc, modulo 2Pi, corrected by an amount equal to the angular measurement offset value.

3. The method according to claim 1, characterized in that, The instantaneous increasing electrical angle position value θe_inc is obtained according to the following relationship: θe_inc= mod((θe + X × 2Pi), Nc × 2Pi) The mechanical angular position θm of the rotor at a given time is determined as an incremental electrical angular position θe_inc modulo 2Pi, divided by the number of comparison areas and corrected by an amount equal to the angular measurement offset value.

4. The method according to claim 1, characterized in that, The signature value pair that determines the electronic signature includes: i) At each angular position of the rotor over one mechanical revolution around the axis of rotation: The instantaneous amplitude value is calculated by considering the calculation rule of the instantaneous amplitude value of at least one of the two measurement signals for the time under consideration. The instantaneous amplitude value represents the intensity of the physical variable with respect to the angular position. The instantaneous electrical angle position value is calculated using the calculation rules for instantaneous electrical angle position values; ii) The signature value pair is recorded by computer. According to the rules for determining the signature value, the signature value pair includes: An amplitude signature value derived from at least one of the instantaneous amplitude values; And an incremental electrical angle position signature value derived from instantaneous electrical angle position values ​​at one or more angular positions, taking into account one or more instantaneous amplitude values ​​to derive an amplitude signature value.

5. The method according to claim 1, characterized in that, The target's electronic signature includes at least one signature value pair, which has an amplitude signature value that is unique over a mechanical revolution.

6. The method according to claim 1, characterized in that, The electrical signature of the target includes at least one signature value pair having an amplitude signature value that exhibits the greatest difference from all other instantaneous amplitude values, which correspond to the same instantaneous value of an electrical angle over a mechanical revolution.

7. The method according to claim 1, characterized in that, The target's electronic signature consists of an ordered series of signature value pairs, with each signature value pair corresponding to a comparison region.

8. The method according to claim 7, characterized in that, The series of signature value pairs are sorted according to the order in which the corresponding comparison region (Zk) passes before each measurement position.

9. The method according to claim 4, characterized in that, For a given angular position, the rule for determining the signature value considers at least one of the following: - The sum of squares of the measured signal values ​​at this angular position, where each measured signal value can reduce the offset of the measured signal or the square root of the sum of squares; - A linear combination of the squares of the measured signal values ​​at that angular position, where each measured signal value can reduce the offset of the measured signal or the square root of the linear combination; - A linear combination of the absolute measurement signal values ​​at this angular position, where each measurement signal value can reduce the offset value of the measurement signal; - A sum or linear combination of the measured signal values ​​at this angular position, where each measured signal value can reduce the offset value of the measured signal.

10. The method according to claim 4, characterized in that, For a given angle range, the rule for determining the signature value considers at least one of the following: - The maximum absolute value of one or more measurement signals within this angular range, where each measurement signal value can reduce the offset value of the measurement signal; - The average value of one or more measurement signals over this angular range, where each measurement signal value is able to reduce the offset value of the measurement signal.

11. The method according to claim 10, characterized in that, The angle range is one quasi-sine period of the measurement signal.

12. The method according to claim 4, characterized in that, For a given angle range, the rule for determining the signature value considers at least one of the following: - The average value of the square root of the sum of the squares of the measured signal values ​​over the angular range, where each measured signal value can reduce the offset of the measured signal; - The average value of the square root of a linear combination of the squares of the measured signal values ​​over the angular range, where each measured signal value can reduce the offset of the measured signal; - The average value of a linear combination of absolute measurement signal values ​​over the angular range, where each measurement signal value reduces the offset of the measurement signal.

13. The method according to claim 12, characterized in that, The angle range is one quasi-sine period of the measurement signal.

14. The method according to claim 4, characterized in that, The rules for determining signature values ​​identify at least one signature value pair, which corresponds to an identifiable value in a value pair formed by an instantaneous amplitude value and a corresponding incremental electrical angle position.

15. The method according to claim 14, characterized in that, The identifiable value is the instantaneous amplitude value, which takes one of the following values: local maximum value, local maximum value over a quasi-cycle of the electrical angle position, local minimum value, local minimum value over a quasi-cycle of the electrical angle position, absolute maximum value over a mechanical revolution of the target, absolute maximum value over a quasi-cycle of the electrical angle position, absolute minimum value over a mechanical revolution of the target, absolute minimum value over a quasi-cycle of the electrical angle position, previously determined value, average value over a mechanical revolution of the target, average value over a quasi-cycle of the electrical angle position, and average value over a half-cycle of the electrical angle position.

16. The method according to claim 14, characterized in that, The identifiable value is the electrical angle position or the incremental electrical angle position.

17. The method according to claim 4, characterized in that, The rules for determining the signature value determine at least one signature value pair, which corresponds to one or more predetermined values ​​of electrical angle position θe and / or corresponds to a predetermined series of incremental electrical angle positions θe_inc.

18. The method according to claim 1, characterized in that, Starting with an electrical calibration signature having M pairs of reference values, and capable of being written as: SIGcb = {(Asigcb1; θe_inc_sigcb1); (Asigcb2; θe_inc_sigcb2), …, (Asigcb M ; θe_inc_sigcb M )} And the electrical signature setting was written as SIGi = {(Asigi1; θe_inc_sigi1) ; (Asigi2 ; θe_inc_sigi2 -),…, (Asigi M ; θe_inc_sigi M )}, The calculation is to be applied to the ordered vector Asigi = {Asigi1 ;Asigi2 ; ... ; Asigi} extracted from the electrical setting signature SIGi. M The value n of the number of cyclic permutations of} min The difference between the ordered vectors in a circular permutation is: Asigi[n] = {Asigi 1+n, 在[1, M]的范围内 ; Asigi 2+n, 在[1, M]的范围内 ; ... ; Asigi M+n, 在[1, M]的范围内 } And an ordered vector Asigcb {Asigcb1 ; Asigcb2 ;... ; Asigcb M } is the smallest, and the angle measurement offset value can be calculated from the difference for at least one value of j: delta_i_θm0 = θe_inc_sigi j+nmin, 在[1, M]的范围内 - θe_inc_sigcb j 。 19. The method according to claim 1, characterized in that, The physical difference between the two contrast regions in the contrast region (Zk) is a spontaneous difference in the design of the sensor, which includes the target and the device for acquiring the Ns electrical measurement signals (S1, S2).

20. The method according to claim 1, characterized in that, The physical difference between the two contrast regions in the contrast region (Zk) is a non-spontaneous difference related to the manufacturing or installation dispersion of the sensor, which includes a target and means for acquiring the Ns electrical measurement signals (S1, S2).

21. A device for measuring the mechanical angular position θm of a rotor, characterized in that, The rotor is capable of rotating multiple times relative to the stator around a rotation axis (A1), and the device includes: - Detector (14) acquires a number of Ns electrical measurement signals (S1, S2), each electrical measurement signal representing the intensity of an electrical or magnetic variable at one of a number of Ns measurement positions (P1, P2), the Ns measurement positions being individual, fixed relative to the stator and offset around the rotation axis by a given mechanical angle α, Ns being an integer equal to 2 or 3; - Target (12), mechanically connected to the rotor and having a number of targets Nc individual contrast regions (Zk), the number of targets Nc being greater than or equal to 2, wherein the targets include contrasts of conductivity, permeability and / or magnetization; - The contrast regions (Zk) are distributed at an angle on the target in a periodic pattern around the axis of rotation. The pattern has Nc quasi-periods, and each contrast region extends around the axis of rotation at a mechanical angle of 2Pi / Nc. - Each contrast region (Zk) causes a change in the electrical measurement signal (S1, S2) acquired at each measurement position when the contrast region passes in front of each measurement position as it rotates around the axis of rotation. The change in the electrical measurement signal is a quasi-sine function of the electrical angular position θe of the contrast region relative to the measurement position, which changes by 2Pi radians with respect to the mechanical angular position θm of the target around the axis of rotation. - At least two of the contrast regions (Zk) have physical differences from each other, which produce differences in the magnitude of the intensity of the physical variables measured in the at least two individual quasi-cycles of the measurement signal at the same number of revolutions on the target; The mechanical angle α corresponding to the modulus offset is the angle of 2Pi / Nc radians around the rotation axis at the measurement position. For Ns=2, the measurement position offset is Pi / (2×Nc) radians, and for Ns=3, the measurement position offset is 2Pi / 3Nc radians. The device includes an electronic control unit (28), which is programmed to: a) Acquire the Ns electrical measurement signals (S1, S2) over one mechanical revolution of the target around the rotation axis. b) For different mechanical angular positions at one mechanical revolution around the rotation axis of the target, the instantaneous value of the electrical angle position θe of the considered mechanical angular position is calculated using the calculation rule for electrical angle position. The calculation rule for electrical angle position considers: For Ns=2, the arctangent of the ratio of the values ​​of the two measured signals (S1, S2) for the time under consideration, or For Ns=3, the arctangent of the ratio of the values ​​of the two transforms obtained by applying the Clark transform to the three measured signals of the time under consideration; c) Calculate the instantaneous increment of the electrical angle position value θe_inc, which is obtained by incrementing the counter X by one unit for any 2PI radian change in the electrical angle position θe, which occurs in the same direction of rotation. d) Determine at least one electronic signature of the target by computer, each electronic signature being determined by a signature-value pair or a series of signature-value pairs, the signature-value pairs including: An amplitude signature value derived from at least one value of at least one measurement signal from at least one measurement signal at at least one angular position relative to a target; And the angular position signature value corresponding to one or more angular positions of the target that are considered to determine the amplitude signature value; The device includes an electronic memory in which an electrical calibration signature of the target, determined during the calibration phase, is recorded by a computer. The electronic control unit (28) is programmed to determine the electrical setup signature and the angle measurement offset value delta_i_θm0 during the setup phase of the measurement session by a reset operation including calculating the angle measurement offset value, the angle measurement offset value being applied to the angle position signature value of the electrical setup signature or the electrical calibration signature, thereby minimizing the difference between the electrical setup signature and the electrical calibration signature. During the measurement session, the electronic control unit determines the mechanical angular position θm of the rotor at a given time by increasing the electrical angular position correction by an amount equal to the angular measurement offset value.

22. The apparatus according to claim 21, characterized in that, The electronic control unit (28) is programmed to, during the calibration phase; a) Acquire the Ns electrical measurement signals (S1, S2) over one mechanical revolution of the target around the rotation axis. b) For different mechanical angular positions at one mechanical revolution around the rotation axis of the target, the instantaneous value of the electrical angle position θe of the considered mechanical angular position is calculated using the calculation rule for electrical angle position. The calculation rule for electrical angle position considers: For Ns=2, the arctangent of the ratio of the values ​​of the two measured signals (S1, S2) for the time under consideration, or For Ns=3, the arctangent of the ratio of the values ​​of the two transforms obtained by applying the Clark transform to the three measured signals of the time under consideration; c) Calculate the instantaneous increment of the electrical angle position value θe_inc, which is obtained by incrementing the counter X by one unit for any 2PI radian change in the electrical angle position θe, which occurs in the same direction of rotation. d) Determine at least one electrical calibration signature of the target by computer, said at least one electrical calibration signature being determined by a signature value pair or a series of signature value pairs, the signature value pair including: An amplitude signature value derived from at least one value of at least one measurement signal from at least one measurement signal at at least one angular position relative to a target; And angular position signature values ​​corresponding to one or more angular positions of the target that are considered to determine the amplitude signature value.

23. The apparatus according to claim 21, characterized in that, The physical difference between the two contrast regions in the contrast region (Zk) is a spontaneous difference in the design of the target (12).

24. The apparatus according to claim 21, characterized in that, The physical difference between two contrast regions in the contrast region (Zk) is a non-spontaneous difference related to the fabrication or installation dispersion of the sensor, which includes the target and the detector (14).

25. The apparatus according to claim 21, characterized in that, The electronic control unit (28) is programmed to implement the method according to any one of claims 1 to 18.

26. The apparatus according to claim 21, characterized in that, The detector (14) is manufactured in the form of a detector box, which includes a box containing a measurement unit (B1, B2), an electronic control unit (28), and a computerized communication interface.

27. The apparatus according to claim 26, characterized in that, The box is sealed.

Images