Eddy current sensor detection error elimination method based on roundness compensation
Through the eddy current sensor detection error elimination method based on roundness compensation, the problem of eddy current sensor detection error in magnetic levitation rotary joints is solved, and the relative posture detection accuracy of the stator rotor and the reliability of the sensor system are improved. It is suitable for super-precision and super-precision Stable and ultra-quiet point to control.
Patent Information
- Application Number
- CN202411402788.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-09
- Publication Date
- 2025-02-14
- Estimated Expiration
- 2044-10-09
AI Technical Summary
The prior art is difficult to completely eliminate the detection error of eddy current sensors in magnetic levitation rotary joints, affecting the relative posture detection accuracy of the stator of the magnetic levitation rotary joints.
The eddy current sensor detection error elimination method based on roundness compensation is adopted. Through differential measurement solution and data fusion, a compensation curve is established to eliminate the roundness error of the stator shaft at the detection cross-section, and the least squares estimation of the magnetic suspension bearing stator Instant axis.
It improves the relative posture detection accuracy of the stator rotor of the magnetic levitation rotary joint, enhances the reliability of the sensor system, and maintains high accuracy when some sensors fail, and is suitable for ultra-precision, ultra-stable, and ultra-static pointing control.
Smart Images

Figure CN119437016A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of signal processing and relates to a method for eliminating detection errors of an eddy current sensor inside a magnetic suspension rotating joint. Background Art
[0002] Magnetic bearings are high-performance bearings that use magnetic field force to suspend the rotor. Because magnetic bearings have many advantages such as no friction, no wear, no lubrication, low cost, low loss, and long life, they can be used in both high-speed motion and low-speed clean situations, and have a wide range of applications.
[0003] Applying magnetic bearings to rotating scanning satellites constitutes a magnetically suspended rotating scanning imaging remote sensing satellite. The rotating scanning satellite realizes the intersection of the scanning trajectory and the flight trajectory through the swing scanning and circular scanning (also known as cone scanning) of the payload (such as a camera, etc.) rotating at a steady speed. In the fast scanning imaging or splicing imaging mode of the camera and other payloads, it can achieve fast, large-scale, high-resolution and ultra-wide imaging of thousands of kilometers, which is used for detailed ground inspections, as well as rapid search and large-scale tracking of moving targets on the sea, land and air.
[0004] The magnetic levitation rotary scanning imaging remote sensing satellite is mainly composed of a platform cabin that provides energy, attitude and orbit control, thermal control and other system services, and a payload cabin that mainly consists of a rotary scanning camera and star sensors. The two cabins are softly connected by a magnetic levitation rotary joint. The magnetic levitation rotary joint can not only isolate the broadband disturbance of the platform cabin, but also perform secondary fine adjustment on the position and posture of the payload cabin, so that the pointing of the camera and other payloads has the "super" potential of ultra-precision, ultra-stable, ultra-quiet and ultra-wide, creating a good working environment for payload imaging; at the same time, it can also ensure fast and effective data transmission of wireless communication by adjusting the relative position between the stator and the rotor. In addition, the magnetic levitation rotary joint provides precise support and rotation control for the payload cabin. Through the magnetic levitation rotary joint angle measurement component (rotary transformer) and displacement measurement component (eddy current sensor), the platform cabin attitude measurement component, and the payload cabin attitude measurement component, the control accuracy of the load cabin shaft pointing is guaranteed, and high-precision pointing measurement data is provided for the camera optical axis, etc. Therefore, measurement feedback is performed through eddy current sensors, signal processing and control signal generation are performed through the controller, and the attitude maneuvering and maintenance of the payload cabin can be achieved through the actuation of magnetic bearings. The pointing accuracy of the magnetic levitation rotary joint is the result of the combined action of the controller and the controlled object.
[0005] It can be seen that the premise of ultra-wide-width and high-resolution imaging of the magnetic levitation rotating scanning satellite is that the magnetic levitation rotating joint achieves ultra-precise, ultra-stable and ultra-static pointing performance. On the one hand, system optimization design is carried out in the design stage, including structural design indicators, process parameter indicators and control performance indicators, to improve the design performance indicators of the magnetic levitation rotating joint; on the other hand, in the product production process, it is necessary to strictly implement quality control to ensure the quality of processing, assembly and adjustment of each component in the product at each link, and directly calibrate and correct the performance indicators of the product in the final link.
[0006] From the perspective of parts, traditional magnetic levitation shaft systems almost all improve feedback accuracy by increasing the machining accuracy of the detection pitch circle and improving the assembly accuracy of eddy current sensors, etc., while the machining accuracy and installation errors can only be gradually reduced but cannot be completely eliminated. Therefore, relying solely on machining and assembly to improve the accuracy of suspension pointing control will undoubtedly increase the product development cost. Summary of the invention
[0007] The technical problem solved by the present invention is: to overcome the shortcomings of the prior art and provide a method for eliminating detection errors of eddy current sensors based on roundness compensation. The eddy current sensors provided by the magnetic levitation rotating joint are used to eliminate the roundness and other morphological errors of the stator shaft at the detection section through differential measurement solution and data fusion, thereby improving the detection accuracy of the relative posture of the stator and rotor of the magnetic levitation rotating joint, and providing a reference for the design of control laws for high-precision attitude control of subsequent satellites.
[0008] The technical solution of the present invention is: a method for eliminating detection errors of eddy current sensors based on roundness compensation, specifically:
[0009] The magnetic bearing is supported by a mechanical shaft system, and under the mechanical shaft system support state, the rotor of the magnetic bearing is driven by a rotary motor to slowly rotate for several cycles, and the displacement value sensed by the eddy current sensor of the magnetic bearing itself and the phase value measured by the rotary transformer are collected;
[0010] Using the displacement value and the phase value, a compensation curve for compensating for the roundness error of the stator shaft of the magnetic bearing is established through phase alignment;
[0011] The magnetic suspension bearing is supported by a magnetic suspension shaft system, and the compensation curve is used to compensate the detection data of the eddy current sensor of the magnetic suspension bearing itself according to the phase, so as to eliminate the shape error of the stator shaft at the detection section.
[0012] Furthermore, the method of the present invention also includes estimating the instantaneous axis center of the stator of the magnetic bearing by least squares using the measured value after compensation of the eddy current sensor of the magnetic bearing itself.
[0013] Furthermore, the eddy current sensor comprises 8 sensor probes, each sensor probe is installed at an interval of 45°, and the installation orientation is (q-1)π / 4,q∈{1,2,…,8}.
[0014] Furthermore, the compensation curve for compensating the roundness error of the stator shaft of the magnetic bearing is established by phase alignment, specifically:
[0015] (41) collecting and recording the displacement values sensed by each group of eddy current sensors and the phase values measured by the rotary transformer;
[0016] (42) interpolating at a given phase point using the measured data, and replacing randomly collected data with fixed-point data;
[0017] (43) In combination with the installation phase of each eddy current sensor, the fixed-point data of each group of eddy current sensors are phase-shifted and aligned according to the phase angle;
[0018] (44) Calculate the average value of the displacement measured by the same eddy current sensor in each cycle to obtain a single-cycle phase-average displacement curve;
[0019] (45) The root mean square curve of the single-turn phase-average displacement curve measured by each eddy current sensor is calculated as the morphological characteristic curve of the detection pitch circle on the stator, which is also the phase compensation curve of each sensor.
[0020] Furthermore, in step (43), the fixed-point data of each group of eddy current sensors are phase-shifted and aligned according to the phase angle, specifically: the measurement signals of any pair of eddy current sensors installed on the same straight line are periodically transformed:
[0021] δ k (α)=-[σ k (α)-σ k+4 (α+π)],k∈{1,2,3,4}
[0022] Among them, σ q (α) is the roundness of the stator shaft measured by each eddy current sensor in the eddy current sensor installation plane, q∈{1,2,…,8}, α is the angle of the rotor rotation, δ k (α) is the differential displacement measurement value of a pair of eddy current sensors installed on the same straight line.
[0023] Furthermore, the compensation curve is specifically:
[0024]
[0025] Furthermore, the compensation curve is used to compensate the eddy current sensor detection data of the magnetic bearing itself according to the phase, specifically:
[0026] (71) Under the support state of the magnetic bearing, determining the current phase according to the rotary transformer;
[0027] (72) At any acquisition time, the phase values of each group of eddy current sensors are used to interpolate the compensation curve to obtain the compensation value of the group of measurement data at that time;
[0028] (73) Subtract the compensation value at that moment from the measurement value of each group of eddy current sensors to obtain the corrected displacement value.
[0029] Furthermore, the relative distance value between the stator and rotor axes after correction is:
[0030]
[0031] Among them, χ k (α) is the relative displacement of the stator and rotor axes measured by differential measurement of any pair of sensor probes, ρ k0 Provides installation deviation for the sensor probe.
[0032] Furthermore, the instantaneous axis center of the stator of the magnetic bearing is estimated by least squares using the measured value compensated by the eddy current sensor of the magnetic bearing itself:
[0033]
[0034] Wherein, e is the relative position deviation between the stator axis and the rotor axis of the magnetic bearing, and β is the angle between the straight line where e is located and the horizontal direction.
[0035] The advantages of the present invention compared with the prior art are:
[0036] (1) The present invention aims at the periodic rotation of the parts errors of the magnetic suspension rotating joint during the rotation process, fully utilizes the high-precision measurement capability of the differential sensor, eliminates the installation eccentricity of each sensor, and performs real-time compensation for the roundness error in the measurement value of each sensor, thereby eliminating the roundness and other morphological errors of the stator shaft at the detection pitch circle; because the measurement information of all sensors is fully utilized, the drift effect of any sensor can be reduced, the accuracy of the shaft center position detection is improved, and when some sensors fail, the relative position of the shaft center can still be detected with high accuracy, thereby improving the reliability of the sensor system;
[0037] (2) The present invention uses the relative position of the stator and rotor under the mechanical shaft system as a reference to measure the periodic error model of the roundness and other morphological errors at the detection pitch circle, which can be applied to the detection data compensation of the corresponding magnetic suspension rotating joint, has direct engineering application value, and provides a guarantee for the ultra-precision, ultra-stable and ultra-static pointing control of the magnetic suspension joint;
[0038] (3) Based on the periodic error model, the present invention is expected to obtain a higher-precision relative position of the stator and rotor of the magnetic suspension rotating joint through real-time monitoring and compensation, which is more conducive to giving full play to the linear advantages of the Lorentz force magnetic bearing and providing a better guarantee for the realization of ultra-precision, ultra-stable and ultra-static pointing control indicators;
[0039] (4) The present invention can be extended to high-precision monitoring and real-time compensation of roundness errors and installation errors of other magnetic levitation rotary actuators. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 A simplified model schematic diagram of the magnetic levitation rotary joint of the present invention;
[0041] Figure 2 It is a schematic diagram of the layout of the eddy current sensor probe of the present invention;
[0042] Figure 3 This is a flow chart of error elimination in eddy current sensor detection according to the present invention. DETAILED DESCRIPTION
[0043] like Figure 1 The figure shows a simplified model schematic diagram of the magnetic suspension rotary joint of the present invention, which mainly includes: ① measurement feedback sensors: upper eddy current sensor, lower eddy current sensor and rotary transformer; ② rotary motor drives the stator and rotor to rotate relative to each other; ③ mechanical shaft system supporting the stator and rotor: upper mechanical bearing and lower mechanical bearing; ④ magnetic suspension shaft system supporting the stator and rotor: upper radial magnetic bearing, lower radial magnetic bearing, axial magnetic bearing. The mechanical shaft system and the magnetic suspension shaft system are backup for each other, and the support mode between the stator and the rotor can be switched arbitrarily between the two working modes of the mechanical shaft system and the magnetic suspension shaft system through the switching mechanism in the magnetic suspension rotary joint, which can not only meet the different performance requirements in different modes, but also improve the reliability of the satellite.
[0044] like Figure 2 As shown in the figure, the eddy current sensor is installed on the rotor, and reflects the relative position relationship between the stator and the rotor by sensing the relative displacement change of the stator and the rotor axis. In view of the fact that the pointing accuracy and stability of the magnetic levitation joint are required to be at the arc second level, the requirements for the actuators and sensors in the magnetic levitation joint must also be improved, such as requiring the measurement accuracy of the eddy current displacement sensor to be at the micron level. Due to the inevitable combined effects of part errors such as the roundness of the stator shaft detection surface, assembly errors, and measurement errors such as sensor temperature drift, constant drift, and random drift, the sensor system in the magnetic levitation joint has measurement errors. In order to better meet the high-precision requirements of the magnetic levitation rotary joint, it is necessary to improve the measurement accuracy of the sensor system, and compensating for the sensor measurement error is a simple and effective way.
[0045] To this end, the present invention proposes a method for phase compensation of the roundness error of the stator shaft detection pitch circle detected by the eddy current sensor. First, under the mechanical shaft support state, the eddy current sensor of the magnetic suspension rotating joint is used to detect the roundness error and other morphological errors of the detected stator shaft, and the curve model to be compensated is established through phase alignment; then, under the magnetic suspension shaft support state, the eddy current sensor detection data is compensated in phase by the optimization method, thereby eliminating the roundness and other morphological errors of the stator shaft at the detection section, and improving the relative posture detection accuracy of the stator and rotor of the magnetic suspension rotating joint.
[0046] 1. Probe layout
[0047] like Figure 2 As shown in the figure, the magnetic levitation rotary joint adopts the outer rotor mode. Assuming that the outer rotor rotates in the negative direction of the ox axis, it can be considered that the stator rotates in the positive direction of the ox axis relative to the rotor. The radial eddy current sensor installed on the rotor in the figure includes two sets, upper and lower; each set has 8 sensor probes, and each sensor probe is installed at an interval of 45°. Assuming that the installation orientation is: (k-1)π / 4, k∈{1,2,…,8}, the position relationship between the sensor probe and the stator axis is as follows Figure 2 As shown. Component processing errors and assembly errors will make it difficult for the probe to be accurately arranged on a standard detection pitch circle, and there will be certain differences in their distances from the axis. The rectangle in the figure represents the sensor probe, the dotted circle represents the position when the shaft is fully aligned, that is, the position when the mechanical shaft system is supported, which is the reference position of the magnetic levitation joint; and the solid circle represents the position of the stator and rotor axes after relative offset, that is, the position when the magnetic levitation shaft system is supported.
[0048] The two opposite eddy current sensor probes on the same detection surface are used in pairs to differentially measure the relative displacement of the stator and rotor at the detection surface along the axis direction of the sensor probe. Each set of 8 sensor probes constitutes 4 pairs. The installation orientations of these 4 pairs of eddy current sensors are: (k-1)π / 4, k∈{1,2,3,4}. Generally, the two pairs of sensor probes installed along the coordinate axis of the 4 pairs of sensor probes are set as the main pair, that is, Figure 2 The black probe in the middle; the other two pairs of sensor probes rotated 45° are set as backup, that is, Figure 2 Medium grey probe. The primary and backup can be used independently. In fact, any two of the four pairs of eddy current sensors can calculate the radial relative distance change between the stator and the rotor on the detection surface. If four sets of sensors are used at the same time, the accuracy and redundancy of the measurement system can be further improved by data fusion.
[0049] 2. Principle of roundness error compensation
[0050] Differential measurement can highlight the relative position change of the stator and rotor, and the midpoint of the sensor probe can be used as the zero point of the measurement. However, if the midpoint does not coincide with the zero point specified by the system, the system error of the sensor installation eccentricity still exists, and manual compensation is required, and the shape errors such as the roundness of the stator shaft at the detection section cannot be eliminated. In fact, the measurement value of any sensor, similar to the measurement of the radial runout of the stator shaft, contains geometric features such as the roundness of the stator shaft, and these characteristic errors are periodic and can also be eliminated from the detection value.
[0051] In view of the large support stiffness of the mechanical bearing, the relative shaking between the stator and rotor is small during the slow rotation of the rotor, and the assembly, large system integration and reference transfer processes are almost all completed under the support state of the mechanical bearing. Therefore, it is assumed that the relative position relationship of the stator and rotor under the support state of the mechanical shaft system is the ideal state, and the mechanical shaft system support state is used as the reference to detect the relative displacement fluctuation of the stator and rotor. Specifically:
[0052] The magnetic suspension rotary joint is switched to a mechanical bearing support state through a switching mechanism in the magnetic suspension rotary joint, and the rotor shaft (coaxial with the stator) in the mechanical shaft support state is regarded as a geometric axis;
[0053] The magnetic suspension rotary joint rotor is driven by a rotary motor to slowly rotate for several cycles, and the displacement values sensed by each group of eddy current sensors and the phase values measured by the rotary transformer are collected and recorded at the same time;
[0054] Using the measured data to interpolate at a given phase point, using fixed-point data instead of randomly collected data;
[0055] Combined with the installation phase of each sensor, the fixed-point data of each group of sensors are phase-shifted and aligned according to the phase angle;
[0056] Calculate the average displacement of the same sensor in each cycle to obtain the single-cycle phase-average displacement curve;
[0057] The root mean square curve of the single-turn phase-average displacement curve measured by each sensor is calculated as the morphological characteristic curve of the detection pitch circle on the stator, which is also the phase compensation curve of each sensor.
[0058] The measurement value of any single sensor is similar to measuring the radial runout of the stator shaft, which includes geometric characteristics such as the roundness of the shaft; while the symmetrically arranged differential sensors reflect the radial runout difference in the diameter direction of the shaft, so the topographic characteristic curve includes geometric information such as the roundness of the stator shaft, so repeated measurements can be used and eliminated.
[0059] If the sensor measurement value is subtracted from the compensation curve in phase under the mechanical shaft support state, the sensor measurement value is corrected to zero, indicating that the morphological features of the detection pitch circle on the stator are completely eliminated.
[0060] If the sensor measurement value is subtracted from the above compensation in phase under magnetic levitation, the compensated displacement value will no longer contain the geometric information of the stator shaft, and the test result will have higher measurement accuracy.
[0061] 3. Modeling of morphological features under mechanical bearing support
[0062] Figure 2 In the equation, it is assumed that the distance between the kth eddy current sensor probe (installed on the rotor) and the rotor axis is ρ k ((k-1)π / 4), k∈{1,2,…,8} is a constant value ρ k ; Assume that the angle of the rotor rotation at a certain moment is α, and the distance from the sensor probe to the stator shaft surface is λk(α+(k-1)π / 4). In the sensor detection surface, assume that the theoretical radius of the stator shaft is R; the roundness of the stator shaft on the detection surface is σk(α+(k-1)π / 4), k∈{1,2,…,8}. And assume that the offset distance between the stator and the rotor is e, and the projection of e in the direction of the kth pair of sensor axes is ek(α+(k-1)π / 4). Then the following relationship exists:
[0063]
[0064] Since the rotation of the rotor is periodic, the roundness information and the eddy current sensor measurement signal are also periodic, with a period of 2π. Therefore, the measurement signals of any pair of sensors can be transformed periodically:
[0065]
[0066] This achieves the phase alignment of any sensor measurement signal. The first formula is transformed into:
[0067] λ k (α) = ρ k -R-σ k (α)-e k (α),k∈{1,2,…,8} (3)
[0068] For the magnetic levitation rotary joint supported by mechanical bearings, based on the compensation principle in the second part above, the flexibility of the stator and rotor is not considered because the support stiffness of the mechanical bearing is relatively large. Since the inner and outer rings of the mechanical bearing are coaxial, there is e k (α) = 0, equation (2) becomes:
[0069]
[0070] Since the distances from different sensor installation positions to the axis centerline may be different, the measurement signal of a single sensor includes error information such as the installation error of each sensor relative to the stator shaft and the roundness of the detection pitch circle. If differential measurement with symmetrically installed sensors is used, the differential displacement is:
[0071] δ k (α) = λ k (α)-λ k+4 (α+π)=ρ k -ρ k+4 -[σ k (α)-σ k+4 (α+π)],k∈{1,2,3,4} (5)
[0072] By integrating the above formula over one cycle, we can get the average value of the measurement data of any pair of sensors over one cycle:
[0073]
[0074] It can be seen that under the mechanical shaft support state, the average value of the measurement data of any pair of sensors in one cycle is equal to the probe installation deviation ρ k0 =ρ k -ρ k+4 ; It can be compensated from the subsequent measurement value through the data acquisition software, that is, the installation deviation is adjusted to 0, ρ k -ρ k+4 =0, then formula (5) can be simplified to:
[0075] δ k (α)=-[σ k (α)-σ k+4 (α+π)],k∈{1,2,3,4} (7)
[0076] It can be seen that in the state of mechanical shaft support, after stripping the installation position deviation of the differential measurement sensor, the differential measurement value of the sensor is only related to the fluctuation information such as the roundness of the stator shaft detection pitch circle.
[0077] For each sensor, its data fluctuation represents the radial runout of the stator shaft at the pitch circle. Therefore, the differential displacement fluctuation in the above formula is the superposition of the geometric errors on the opposite side of the detection pitch circle. In traditional measurement systems, this part of the fluctuation is often ignored. In fact, the geometric errors of parts are often in the order of several microns, which cannot be tolerated by high-precision control. For this reason, the present invention further strips it away.
[0078] Since each pair of sensors in the same set (upper or lower) tests the same stator shaft section, the roundness and other topographic errors measured at the same phase are almost the same. Therefore, the measured data can be averaged at the same phase to obtain:
[0079]
[0080] The above formula is the shape error information at the detection pitch circle measured by all differential sensors. It is used as the phase compensation curve of the sensor and the relative displacement of the stator and rotor axes can be obtained by compensating formula (7).
[0081] That is, the roundness of the pitch circle is further tested after further stripping the stator shaft. The differential measured displacement after φ is:
[0082]
[0083] After compensating the sensor probe installation deviation ρ k0 =ρ k -ρ k+4 Check the roundness of the pitch circle of the stator shaft Finally, because the stator and rotor are coaxial under the mechanical shaft support state, the measurement value of the sensor probe is theoretically 0.
[0084] 4. Compensation of measured values under magnetic bearing support
[0085] In the magnetic suspension state, the roundness is compensated using the compensation curve and the measured values of each sensor.
[0086] In the state of magnetic bearing support, the current phase is determined according to the rotary transformer;
[0087] At any acquisition time, the compensation curve is interpolated using the phase values of each group of sensors to obtain the compensation value of the group of measurement data at that time;
[0088] The compensation value at that moment is subtracted from the measured value of each group of sensors to obtain the corrected displacement value.
[0089] In the state of magnetic suspension shaft support, according to formula (3), the differential displacement measured by any pair of sensor probes is:
[0090] δ k (α) = λ k (α)-λ k+4 (α+π)
[0091] =ρ k -ρ k+4 -[σ k (α)-σk+4 (α+π)]-[e k (α)-e k (α+π)],k∈{1,2,3,4} (10)
[0092] That is, interpolation is performed according to the compensation curve of formula (8) to calculate the compensation value of the current phase; the measured relative displacement of the stator and rotor axes is:
[0093]
[0094] From this, the relative distance between the stator and rotor axes can be solved as:
[0095]
[0096] (5) Optimal determination of instantaneous axis
[0097] In the magnetic suspension state, the instantaneous axis center is identified by least squares using the compensated measurement values.
[0098] The axis of the rotating shaft is identified by using multiple sets of corrected displacement values through the least square method.
[0099] After a certain period of time, the magnetic suspension support state is switched to the mechanical shaft support state again, and the above steps are repeated to calibrate the sensor and eliminate the constant drift of the sensor.
[0100] Given that the magnetic levitation joint is subject to strict thermal control, the drift caused by temperature is small and can be ignored.
[0101] like Figure 2 As shown in the figure, assuming that the angle between the relative position deviation e of the stator and rotor axes and the oy axis at any time is β, then the angle between e and the kth pair of sensors is β-(k-1)π / 4. Obviously, the projection component of e on the axial direction of the kth pair of sensors is ecos(β-(k-1)π / 4), which is the theoretical measurement value of the differential sensor. It is different from the measured value e k There is a certain error between (α) and ecos(β-(k-1)π / 4)-e k (α), in order to accurately estimate the measurement results, assuming that the weighted sum of squares of the projection error is the smallest, then combined with its distribution position in the structure β-(k-1)π / 4, k∈{1,2,3,4}, the closer the azimuth of the deviation e is to the sensor installation azimuth, the greater its weight; when the two are perpendicular, the weight is 0. Then the model is:
[0102]
[0103] The least squares fitting of the compensated data can obtain the relative position error e and angle β of the stator and rotor axes. Then it can be estimated that the relative position of the stator and rotor axes is (e, α), and the projection in the direction of the sensor axis is for:
[0104]
[0105] The whole process is like Figure 3 shown.
[0106] From the perspective of magnetic levitation rotary joint signal processing, the accuracy of measurement feedback is improved to meet the model mission's requirements for arc-second-level attitude control accuracy of the magnetic levitation rotary joint, improve the pointing control performance of the magnetic levitation rotary joint, and ensure the high-precision pointing requirements of the rotor.
[0107] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.
Claims
1. A method for eliminating detection errors of eddy current sensors based on roundness compensation, characterized in that: The magnetic bearing is supported by a mechanical shaft system, and under the mechanical shaft system support state, the rotor of the magnetic bearing is driven by a rotary motor to slowly rotate for several cycles, and the displacement value sensed by the eddy current sensor of the magnetic bearing itself and the phase value measured by the rotary transformer are collected; Using the displacement value and the phase value, a compensation curve for compensating for the roundness error of the stator shaft of the magnetic bearing is established through phase alignment; The magnetic suspension bearing is supported by a magnetic suspension shaft system, and the compensation curve is used to compensate the detection data of the eddy current sensor of the magnetic suspension bearing itself according to the phase, so as to eliminate the shape error of the stator shaft at the detection section.
2. The method for eliminating detection errors of eddy current sensors based on roundness compensation according to claim 1 is characterized in that: The method also includes estimating the instantaneous axis center of the stator of the magnetic bearing by using the compensated measurement value of the eddy current sensor of the magnetic bearing itself through least squares.
3. The method for eliminating detection errors of eddy current sensors based on roundness compensation according to claim 1 or 2, characterized in that: The eddy current sensor comprises 8 sensor probes, each of which is installed at an interval of 45°, and the installation orientation is (q-1)π / 4, q∈{1,2,…,8}.
4. The method for eliminating detection errors of eddy current sensors based on roundness compensation according to claim 3 is characterized in that: The compensation curve for compensating the roundness error of the stator shaft of the magnetic bearing established by phase alignment is specifically: (41) collecting and recording the displacement values sensed by each group of eddy current sensors and the phase values measured by the rotary transformer; (42) interpolating at a given phase point using the measured data, and replacing randomly collected data with fixed-point data; (43) In combination with the installation phase of each eddy current sensor, the fixed-point data of each group of eddy current sensors are phase-shifted and aligned according to the phase angle; (44) Calculate the average value of the displacement measured by the same eddy current sensor in each cycle to obtain a single-cycle phase-average displacement curve; (45) The root mean square curve of the single-turn phase-average displacement curve measured by each eddy current sensor is calculated as the morphological characteristic curve of the detection pitch circle on the stator, which is also the phase compensation curve of each sensor.
5. The method for eliminating detection errors of eddy current sensors based on roundness compensation according to claim 4 is characterized in that: In the step (43), the fixed-point data of each group of eddy current sensors are phase-shifted and aligned according to the phase angle, specifically: the measurement signals of any pair of eddy current sensors installed on the same straight line are periodically transformed: d k (a)=-[s k (a)-s k+4 (a+π)],k∈{1,2,3,4} Among them, σ q (α) is the roundness of the stator shaft measured by each eddy current sensor in the eddy current sensor installation plane, q∈{1,2,…,8}, α is the angle of the rotor rotation, δ k (α) is the differential displacement measurement value of a pair of eddy current sensors installed on the same straight line.
6. The method for eliminating detection errors of eddy current sensors based on roundness compensation according to claim 5 is characterized in that: The compensation curve is specifically:
7. The method for eliminating detection errors of eddy current sensors based on roundness compensation according to claim 4 is characterized in that: The method of using the compensation curve to compensate the eddy current sensor detection data of the magnetic bearing itself according to the phase is specifically as follows: (71) Under the support state of the magnetic bearing, determining the current phase according to the rotary transformer; (72) At any acquisition time, the phase values of each group of eddy current sensors are used to interpolate the compensation curve to obtain the compensation value of the group of measurement data at that time; (73) Subtract the compensation value at that moment from the measurement value of each group of eddy current sensors to obtain the corrected displacement value.
8. The method for eliminating detection errors of eddy current sensors based on roundness compensation according to claim 7 is characterized in that: The relative distance value of the stator and rotor axes after correction is: Among them, χ k (α) is the relative displacement of the stator and rotor axes measured by differential measurement of any pair of sensor probes, ρ k0 Provides installation deviation for the sensor probe.
9. The method for eliminating detection errors of eddy current sensors based on roundness compensation according to claim 4, characterized in that: The instantaneous axis center of the stator of the magnetic bearing is estimated by the least square method using the measured value compensated by the eddy current sensor of the magnetic bearing itself: Wherein, e is the relative position deviation between the stator axis and the rotor axis of the magnetic bearing, and β is the angle between the straight line where e is located and the horizontal direction.
Citation Information
Patent Citations
Radial magnetic bearing electrical vortex sensor integrated structure for magnetic levitation high-speed electric machine
CN102829709A
Unbalance identification and vibration suppression control system for magnetic suspension rotating machinery
CN103425051A
Unbalanced compensation algorithm of polygonal iterative search with variable step size for rotor imbalance coefficient
CN107133387A
Eddy current displacement sensor for high temperature magnetic bearing in axial direction
CN108489371A
Displacement detection circuit of maglev rotor system and displacement self-sensing system thereof
US20230117718A1