Denoising method for measured data of eddy current angle sensor

Abstract

The invention discloses a denoising method for measured data of an eddy current angle sensor, which comprises the following steps of: carrying out differential processing on a signal output by the sensor to suppress common-mode noise, and dividing a rotation period into sectors according to fan blades; performing sector-level offset compensation and normalization compensation on the signal of the current large period based on the average value and extreme value of the signal of the sector corresponding to the previous large period so as to eliminate drift and gain difference; calculating a basic angle value of the current period by using the compensated signal; the method comprises the following steps of: calculating a difference value between angle values of sampling points corresponding to a previous large period and a previous large period, and performing Fourier analysis by taking a difference value sequence as a periodic function to obtain a direct current component and first four harmonic coefficients of the periodic function; and reconstructing a harmonic compensation value according to the coefficients, and superposing the harmonic compensation value to the basic angle value of the current period to obtain a final accurate angle value. According to the invention, periodic errors can be effectively compensated, and the measurement precision and the anti-interference capability of the eddy current angle sensor are significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data denoising, and in particular to a method for denoising measurement data of an eddy current angle sensor. Background Art

[0002] The eddy current angle sensor is a high-precision motor position angle sensor. The sensor is a magnet-free inductive position sensor that can be used for high-speed absolute position sensing in automotive, industrial, medical and consumer applications.

[0003] It uses the physical principles of eddy currents to detect the position of a metal target moving over a set of coils consisting of a transmitter coil and two receiver coils. These three coils are typically printed on a printed circuit board (PCB) in the form of copper traces. They are arranged so that the transmitter coil generates a secondary voltage in the two receiver coils, the voltage value of which depends on the position of the metal target above the coils. By demodulating and processing the secondary voltage from the receiver coils, a signal indicating the target's position on the coils is obtained. The target is made of a conductive material that allows eddy currents to flow. The target moves linearly relative to the transmitter and receiver coils. The inductive coupling between the transmitter coil and the two sets of receiver coils varies in a roughly sinusoidal or cosine curve depending on the position of the target. This allows a detector connected to the receiver coils to determine the target's position based on the induced signals from the receiver coils.

[0004] In a car, the eddy current sensor needs to be installed under the motor blades, without contact with the motor blades. A fixed gap should be maintained between the motor blades and the sensor coils, and they should be installed parallel to each other. The centers of the two need to be on the same axis. However, in actual assembly, it is impossible to completely guarantee that the installation position is on the same axis. In addition, when the car is actually driving, the motor blades may shake, vibrate, and bump, causing the motor blades and the sensor to be unable to be parallel and on the same axis. Therefore, a set of intelligent data processing algorithms is needed to eliminate the installation tilt and eccentricity of the sensor and the motor blades, as well as various noise interferences on the car.

[0005] Since the data is subject to interference from various external factors, such as electromagnetic interference between the internal electrical components of the vehicle and the influence of noise in the environment, it is necessary to denoise the data. Existing technologies usually use fixed filtering parameters to denoise it. However, since the interference from external factors is constantly changing, and the signal data of the sin signal and the cos signal also influence each other, the use of fixed filtering parameters cannot cope with this dynamic change well, resulting in low quality of the denoised data obtained by the eddy current sensor, which in turn affects the feedback angle accuracy. Summary of the Invention

[0006] The present invention mainly solves the technical problem that the fixed filtering parameters in the existing technology cannot cope with the constantly changing noise interference, which leads to low quality of denoised data obtained by the eddy current sensor, thereby affecting the system's calculation of the actual operating angle of the motor, and provides a high-quality eddy current angle sensor measurement data denoising method.

[0007] The present invention solves the above technical problems mainly through the following technical solutions: a method for denoising measurement data of an eddy current angle sensor, wherein the eddy current angle sensor is installed under the motor fan blade, comprising the following steps: S1: Perform differential processing on the signal received by the eddy current angle sensor to obtain sin signal and cos signal. Each group of signals received by the eddy current angle sensor includes sin+ signal, sin- signal, cos+ signal and cos- signal. Perform differential processing on the sin+ signal and sin- signal in each group (subtract sin- from sin+) to obtain sin signal. Perform differential processing on the cos+ signal and cos- signal in each group (subtract cos- from cos+) to obtain cos signal. Divide each fan blade of the motor into areas according to the angle. The angle range corresponding to a fan blade is regarded as a sector. Divide the sin signal and cos signal evenly into the corresponding sectors according to the angle. The sin signal corresponding to the i-th sector is X i_1 ,X i_2 ,……,X i_n , 1≤i≤m, m is the total number of sectors, similarly the cosine signal corresponding to the i-th sector is Y i_1 ,Y i_2 ,……,Y i_n ; Calculate the average value of the sin signal and the average value of the cos signal of the previous large cycle of each sector. The average value of the sin signal of the previous large cycle of the i-th sector is sin i_AVERAGE , the average value of the cosine signal of a large cycle in the i-th sector is cos i_AVERAGE ; The time it takes for each fan blade (i.e., each sector) to completely sweep across the eddy current angle sensor is considered a small cycle, and the time it takes for all sectors to completely sweep across the eddy current angle sensor (i.e., the motor's main shaft rotates 360 degrees) is considered a large cycle. The large cycle is equal to the sum of m small cycles. S2: Perform the first compensation on the sin signal and cos signal of the current large cycle of each sector. The j-th sin signal of the i-th sector after the first compensation is X i_j_1 =X i_j -sin i_AVERAGE , the jth cosine signal of the i-th sector is Y after the first compensation i_j_1 =Y i_j -cos i_AVERAGEThe time it takes for a sector to sweep across the sensor is reflected in the signal as one cycle, so the average value of all signals in the sector within the previous large cycle is subtracted from the entire sector. S3: Perform normalization compensation on the signal after the primary compensation; S4: Calculate the test angle value of each group of signals using the normalized and compensated sin and cos signals. The test angle value of the jth signal in the i-th sector is recorded as θ i_j ; S5: Calculate the harmonic compensation value by the difference between the test angle value of the previous large cycle and the test angle value θ i_j Superimpose to obtain the final test angle value.

[0008] As a preferred embodiment, step S5 is specifically as follows: calculating the test angle value θ of the jth signal in the i-th sector of the previous large cycle i_j_-1 and the test angle value θ of the jth signal in the i-th sector of the previous large cycle i_j_-2 The difference Δθ i_j =θ i_j_-1 -θ i_j_-2 ; The difference Δθ i_j Substitute F(θ) into the following formula to calculate the average value of the difference a0: ; The difference Δθ i_j Substitute F(θ) into the following formula to calculate the first-order coefficient a n and b n : ; ; The secondary coefficient Φ is calculated by the following formula n and C n : Φ n =ATAN2(b n ,a n ), n=1,2,3,4; ; The harmonic compensation value f(θ) is calculated according to the following formula: ; Compare the harmonic compensation value with the test angle value θ of the jth signal of the i-th sector of the current cycle i_j Superimpose to obtain the final test angle value.

[0009] Primary compensation is designed to compensate for possible sector-related DC offsets or slow drifts. The average value of the previous cycle is used to utilize historical information for compensation, avoiding the introduction of noise from the current cycle.

[0010] Normalization compensation aims to eliminate potential differences in signal amplitude (gain) between sectors, normalizing the signal amplitude of each sector to the range of [-1, 1]. This process also uses the extreme value information from the previous cycle.

[0011] Primary compensation and normalized compensation are based on the actual driving conditions of the vehicle and the noise level of the electrical signal in the previous circle. Each fan blade signal data is adaptively smoothed, and the sin and cos values of the next circle are compensated in real time using the sin and cos signals of the previous circle. The input signal is de-interfered, the input data is optimized, and the asymmetric signals on the entire vehicle are processed symmetrically.

[0012] Although primary compensation and normalization compensation can eliminate noise and signal glitches in the input signal, some sensors are affected by the installation environment. Due to the large motor blades, the eddy current sensor cannot fully cover the motor blades and can only be made into a semicircle or a quarter of a circle. Such a design will cause the eddy current coil to be symmetrical left and right and up and down, and the installation surface cannot be kept parallel, making it more susceptible to interference. Therefore, using the above-mentioned signal normalization processing alone cannot effectively remove interference and improve accuracy. Therefore, harmonic compensation is used to further optimize the algorithm.

[0013] Angle measurement contains periodic errors that vary with the angle, and these errors may have a certain degree of cyclical repeatability or slowly varying characteristics. By analyzing the differences in angle values from previous cycles, we can estimate the harmonic components of this error and use them to compensate for the angle value of the current cycle.

[0014] As a preference, step S4 obtains the test angle value θ of the jth signal of the i-th sector of the current cycle. i_j After that, calibration compensation is performed in the following ways: A1: The calibration theoretical angle value of the jth signal in the i-th sector is α i_j , the calibration test angle value is β i_j , calculate the error E between the calibration theoretical angle value and the calibration test angle value i_j =β i_j -α i_j ; A2: Set the error E i_j As F(θ), calculate the calibration harmonic compensation value f(x)' according to the method of claim 2; A3: Set the test angle value θ i_jSuperimposed with the calibrated harmonic compensation value f(x)' as the new test angle value θ i_j .

[0015] Before leaving the factory, the vehicle undergoes factory calibration. The calibration test angle values are obtained through the process described in steps S1 to S4. The corresponding calibration theoretical angle values are then measured and calculated. Calibration harmonic compensation values are then calculated. These are fixed parameters and are directly applied after shipment without further calculation. Calibration compensation automatically calculates a set of data based on actual vehicle installation test data, taking into account factors such as individual product installation errors, sensor errors, and vehicle glitches. This algorithm then accurately compensates the product's angle output, improving the energy efficiency of new energy vehicles. Calibration compensation can be performed offline, generating a compensation table or coefficients that can then be applied during actual operation. This can be combined with S5's dynamic harmonic compensation (for example, performing calibration compensation first, followed by S5 compensation), or it can replace S5 in certain scenarios where errors are primarily static.

[0016] As an example, the normalization compensation is specifically as follows: the jth sin signal of the current cycle of the i-th sector is normalized and compensated to X i_j_2 = X i_j_1 *2 / (max i_sin -min i_sin ), max i_sin is the maximum value of the sin signal in the last large cycle of the i-th sector, min i_sin is the minimum value of the sin signal of the last large cycle of the i-th sector; the j-th cos signal of the i-th sector after normalization compensation is Y i_j_2 = Y i_j_1 *2 / (max i_cos -min i_cos ), max i_cos is the maximum value of the cosine signal in the last large cycle of the i-th sector, min i_cos It is the minimum value of the cosine signal in the last large cycle of the i-th sector.

[0017] Preferably, the test angle value of the jth signal in the i-th sector is obtained by the following formula: θ i_j =ATAN2(Y i_j_2 , X i_j_2 ). The ATAN2 function can determine the quadrant of the angle based on the signs of the sin and cos signals, and output an unambiguous angle value covering [-π, +π] or [0, 2π]. i_j It is a preliminary angle calculation result of the current cycle.

[0018] The substantial effects brought about by the present invention are: 1. Flexibility: The algorithm can be used for vehicle bench testing and compensated according to the different working conditions of each vehicle, so that each new energy vehicle can be debugged to the optimal state; 2. Performance improvement: Without changing the hardware, the accuracy is improved from ±6° to ±1°, which is a 6% improvement in accuracy. It also saves hardware debugging costs and labor costs, and there is no need to debug the hardware of each vehicle. 3. Economical: The use of normalization and harmonic compensation improves the angle output accuracy of the eddy current sensor, making the sensor output accuracy closer to the actual angle value, and the feedback signal to the motor is more accurate, which reduces the energy consumption of the motor, increases the energy efficiency, and saves costs.

[0019] 4. Environmental protection: The motor has lower energy consumption, higher energy efficiency, and is more environmentally friendly, reducing the impact of heating of new energy motors on the local environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of a method for denoising measurement data of an eddy current angle sensor according to the present invention. DETAILED DESCRIPTION

[0021] The technical solution of the present invention will be further specifically described below through embodiments and in conjunction with the accompanying drawings.

[0022] Example: A method for denoising measurement data of an eddy current angle sensor, such as Figure 1 As shown, the following steps are included: S1: Perform differential processing on the signal received by the eddy current angle sensor to obtain sin signal and cos signal. Each group of signals received by the eddy current angle sensor includes sin+ signal, sin- signal, cos+ signal and cos- signal. Perform differential processing on the sin+ signal and sin- signal in each group (subtract sin- from sin+) to obtain sin signal. Perform differential processing on the cos+ signal and cos- signal in each group (subtract cos- from cos+) to obtain cos signal. Divide each fan blade of the motor into areas according to the angle. The angle range corresponding to a fan blade is regarded as a sector. Divide the sin signal and cos signal evenly into the corresponding sectors according to the angle. The sin signal corresponding to the i-th sector is X i_1 ,X i_2 ,……,X i_n , 1≤i≤m, m is the total number of sectors, similarly the cosine signal corresponding to the i-th sector is Y i_1 ,Y i_2 ,……,Y i_n ; Calculate the average value of the sin signal and the average value of the cos signal of the previous large cycle of each sector. The average value of the sin signal of the previous large cycle of the i-th sector is sini_AVERAGE , the average value of the cosine signal of a large cycle in the i-th sector is cos i_AVERAGE ; The time it takes for each fan blade (i.e., each sector) to completely sweep across the eddy current angle sensor is considered a small cycle, and the time it takes for all sectors to completely sweep across the eddy current angle sensor (i.e., the motor's main shaft rotates 360 degrees) is considered a large cycle. The large cycle is equal to the sum of m small cycles. S2: Perform the first compensation on the sin signal and cos signal of the current large cycle of each sector. The j-th sin signal of the i-th sector after the first compensation is X i_j_1 =X i_j -sin i_AVERAGE , the jth cosine signal of the i-th sector is Y after the first compensation i_j_1 =Y i_j -cos i_AVERAGE The time it takes for a sector to sweep across the sensor is reflected in the signal as one cycle, so the average value of all signals in the sector within the previous large cycle is subtracted from the entire sector. S3: Perform normalization compensation on the signal after the primary compensation; S4: Calculate the test angle value of each group of signals using the normalized and compensated sin and cos signals. The test angle value of the jth signal in the i-th sector is recorded as θ i_j ; S5: Calculate the harmonic compensation value by the difference between the test angle value of the previous large cycle and the test angle value θ i_j Superimpose to obtain the final test angle value.

[0023] The step S5 is specifically as follows: calculating the test angle value θ of the jth signal of the i-th sector of the previous large cycle i_j_-1 and the test angle value θ of the jth signal in the i-th sector of the previous large cycle i_j_-2 The difference Δθ i_j =θ i_j_-1 -θ i_j_-2 ; The difference Δθ i_j Substitute F(θ) into the following formula to calculate the average value of the difference a0: ; The difference Δθ i_j Substitute F(θ) into the following formula to calculate the first-order coefficient a n and b n : ; ; The secondary coefficient Φ is calculated by the following formula n and Cn : Φ n =ATAN2(b n ,a n ), n=1,2,3,4; ; The harmonic compensation value f(θ) is calculated according to the following formula: ; Compare the harmonic compensation value with the test angle value θ of the jth signal of the i-th sector of the current cycle i_j Superimpose to obtain the final test angle value.

[0024] Primary compensation is designed to compensate for possible sector-related DC offsets or slow drifts. The average value of the previous cycle is used to utilize historical information for compensation, avoiding the introduction of noise from the current cycle.

[0025] Normalization compensation aims to eliminate potential differences in signal amplitude (gain) between sectors, normalizing the signal amplitude of each sector to the range of [-1, 1]. This process also uses the extreme value information from the previous cycle.

[0026] Primary compensation and normalized compensation are based on the actual driving conditions of the vehicle and the noise level of the electrical signal in the previous circle. Each fan blade signal data is adaptively smoothed, and the sin and cos values of the next circle are compensated in real time using the sin and cos signals of the previous circle. The input signal is de-interfered, the input data is optimized, and the asymmetric signals on the entire vehicle are processed symmetrically.

[0027] Although primary compensation and normalization compensation can eliminate noise and signal glitches in the input signal, some sensors are affected by the installation environment. Due to the large motor blades, the eddy current sensor cannot fully cover the motor blades and can only be made into a semicircle or a quarter of a circle. Such a design will cause the eddy current coil to be symmetrical left and right and up and down, and the installation surface cannot be kept parallel, making it more susceptible to interference. Therefore, using the above-mentioned signal normalization processing alone cannot effectively remove interference and improve accuracy. Therefore, harmonic compensation is used to further optimize the algorithm.

[0028] Angle measurement contains periodic errors that vary with the angle, and these errors may have a certain degree of cyclical repeatability or slowly varying characteristics. By analyzing the differences in angle values from previous cycles, we can estimate the harmonic components of this error and use them to compensate for the angle value of the current cycle.

[0029] Step S4 obtains the test angle value θ of the jth signal of the i-th sector of the current cycle i_jAfter that, calibration compensation is performed in the following ways: A1: The calibration theoretical angle value of the jth signal in the i-th sector is α i_j , the calibration test angle value is β i_j , calculate the error E between the calibration theoretical angle value and the calibration test angle value i_j =β i_j -α i_j ; A2: Set the error E i_j As F(θ), calculate the calibration harmonic compensation value f(x)' according to the method of claim 2; A3: Set the test angle value θ i_j Superimposed with the calibrated harmonic compensation value f(x)' as the new test angle value θ i_j .

[0030] Before leaving the factory, the vehicle undergoes factory calibration. The calibration test angle value is obtained through the process of steps S1 to S4. The corresponding calibration theoretical angle value is then measured and calculated. The calibration harmonic compensation value is then calculated. The calibration harmonic compensation value is a fixed parameter and is directly applied after leaving the factory without further calculation. Calibration compensation is calculated based on actual vehicle installation test data to account for factors such as individual product installation errors, sensor errors, and vehicle glitches. An algorithm is then used to accurately compensate for the product's angle output, improving the energy efficiency of new energy vehicles. Calibration compensation can be performed offline, generating a compensation table or coefficients that can then be applied during actual operation. This can be combined with S5's dynamic harmonic compensation (for example, calibration compensation performed first, followed by S5 compensation), or it can replace S5 in certain application scenarios where errors are primarily static.

[0031] The normalization compensation is specifically as follows: the jth sin signal of the current cycle of the i-th sector is normalized and compensated to X i_j_2 = X i_j_1 *2 / (max i_sin -min i_sin ), max i_sin is the maximum value of the sin signal in the last large cycle of the i-th sector, min i_sin is the minimum value of the sin signal of the last large cycle of the i-th sector; the j-th cos signal of the i-th sector after normalization compensation is Y i_j_2 = Y i_j_1 *2 / (max i_cos -min i_cos ), max i_cos is the maximum value of the cosine signal in the last large cycle of the i-th sector, min i_cos It is the minimum value of the cosine signal in the last large cycle of the i-th sector.

[0032] The test angle value of the jth signal in the i-th sector is obtained by the following formula: i_j =ATAN2(Y i_j_2 ,X i_j_2 ). The ATAN2 function can determine the quadrant of the angle based on the signs of the sin and cos signals, and output an unambiguous angle value covering [-π, +π] or [0, 2π]. i_j It is a preliminary angle calculation result of the current cycle.

[0033] The specific embodiments described herein are merely illustrative of the spirit of the present invention. Persons skilled in the art may make various modifications, additions, or substitutions to the described specific embodiments without departing from the spirit of the present invention or exceeding the scope of the appended claims.

[0034] Although this document frequently uses terms such as differential processing, normalization compensation, and harmonic compensation, the use of other terms is not excluded. These terms are used solely to more conveniently describe and explain the essence of the present invention; interpreting them as any additional limitations would be contrary to the spirit of the present invention.

Claims

1. A method for denoising measurement data of an eddy current angle sensor, wherein the eddy current angle sensor is installed below a motor blade, and is characterized in that: The following steps are involved: S1: Perform differential processing on the signal received by the eddy current angle sensor to obtain the sin signal and cos signal. Divide each fan blade of the motor into areas according to the angle. The angle range corresponding to a fan blade is regarded as a sector. Divide the sin signal and cos signal evenly into the corresponding sectors according to the angle. The sin signal corresponding to the i-th sector is X i_1 ,X i_2 ,……,X i_n , 1≤i≤m, m is the total number of sectors, similarly the cosine signal corresponding to the i-th sector is Y i_1 ,Y i_2 ,……,Y i_n ; Calculate the average value of the sin signal and the average value of the cos signal of the previous large cycle of each sector. The average value of the sin signal of the previous large cycle of the i-th sector is sin i_AVERAGE , the average value of the cosine signal of a large cycle in the i-th sector is cos i_AVERAGE ; S2: Perform the first compensation on the sin signal and cos signal of the current large cycle of each sector. The j-th sin signal of the i-th sector after the first compensation is X i_j_1 =X i_j -sin i_AVERAGE , the jth cosine signal of the i-th sector is Y after the first compensation i_j_1 =Y i_j -cos i_AVERAGE ; S3: Perform normalization compensation on the signal after the primary compensation; S4: Calculate the test angle value of each group of signals using the normalized and compensated sin and cos signals. The test angle value of the jth signal in the i-th sector is recorded as θ i_j ; S5: Calculate the harmonic compensation value by the difference between the test angle value of the previous large cycle and the test angle value θ i_j Superimpose to obtain the final test angle value.

2. The method for denoising measurement data of an eddy current angle sensor according to claim 1, wherein: The step S5 is specifically as follows: calculating the test angle value θ of the jth signal of the i-th sector of the previous large cycle i_j_-1 and the test angle value θ of the jth signal in the i-th sector of the previous large cycle i_j_-2 The difference Δθ i_j =θ i_j_-1 -θ i_j_-2 ; The difference Δθ i_j Substitute F(θ) into the following formula to calculate the average value of the difference a0: ； The difference Δθ i_j Substitute F(θ) into the following formula to calculate the first-order coefficient a n and b n : ； ； The secondary coefficient Φ is calculated by the following formula n and C n : F n =ATAN2(b n ,a n ), n=1,2,3,4; ； The harmonic compensation value f(θ) is calculated according to the following formula: ； Compare the harmonic compensation value with the test angle value θ of the jth signal of the i-th sector of the current cycle i_j Superimpose to obtain the final test angle value.

3. The method for denoising measurement data of an eddy current angle sensor according to claim 2, wherein: Step S4 obtains the test angle value θ of the jth signal of the i-th sector of the current cycle i_j After that, calibration compensation is performed in the following ways: A1: The calibration theoretical angle value of the jth signal in the i-th sector is α i_j , the calibration test angle value is β i_j , calculate the error E between the calibration theoretical angle value and the calibration test angle value i_j =β i_j -α i_j ; A2: Set the error E i_j As F(θ), calculate the calibration harmonic compensation value f(x)' according to the method of claim 2; A3: Set the test angle value θ i_j Superimposed with the calibrated harmonic compensation value f(x)' as the new test angle value θ i_j .

4. The method for denoising measurement data of an eddy current angle sensor according to claim 1, wherein: The normalization compensation is specifically as follows: the jth sin signal of the current cycle of the i-th sector is normalized and compensated to X i_j_2 = X i_j_1 *2 / (max i_sin -min i_sin ), max i_sin is the maximum value of the sin signal in the last large cycle of the i-th sector, min i_sin is the minimum value of the sin signal of the last large cycle of the i-th sector; the j-th cos signal of the i-th sector after normalization compensation is Y i_j_2 = Y i_j_1 *2 / (max i_cos -min i_cos ), max i_cos is the maximum value of the cosine signal in the last large cycle of the i-th sector, min i_cos It is the minimum value of the cosine signal in the last large cycle of the i-th sector.

5. The method for denoising measurement data of an eddy current angle sensor according to claim 4, characterized in that: The test angle value of the jth signal in the i-th sector is obtained by the following formula: i_j =ATAN2(Y i_j_2 , X i_j_2 ).

Images