阅读说明：本技术 一种旋转变压器信号包络检测方法 (Rotary transformer signal envelope detection method ) 是由 易吉良 周雪纯 李中启 黄晓峰 刘建华 李军军 谷志茹 张晋 贺力克 于 2020-04-26 设计创作，主要内容包括：本发明公开了一种旋转变压器信号包络检测方法,该方法采用自适应单向量S变换检测旋转变压器信号包络；首先根据归一化频率<I>f</I><Sub><I>r</I></Sub>确定单向量S变换核的高斯窗函数的窗宽比系数<I>w</I><Sub><I>k</I></Sub>,然后根据<I>w</I><Sub><I>k</I></Sub>和高斯窗非零阈值<I>z</I>确定高斯窗的总长度<I>N</I>和非零长度<I>N</I><Sub><I>z</I></Sub>,进而确定高斯窗函数并对待求旋转变压器信号做单向量S变换,最后根据非零长度<I>N</I><Sub><I>z</I></Sub>去掉变换结果的端部后求模即得到对应信号段的包络。本发明能够快速精确地求出旋转变压器信号的包络,为旋转变压器信号解码提供基础。(The invention discloses a method for detecting signal envelopes of a rotary transformer, which adopts self-adaptive single vector S transformation to detect the signal envelopes of the rotary transformer; first according to the normalized frequency f r Determining window-width ratio coefficients for a gaussian window function of a single vector S transform kernel w k Then according to w k And a non-zero threshold value of the Gaussian window z Determining the total length of a Gaussian window N And a non-zero length N z Further determining a Gaussian window function, performing one-way quantity S transformation on the resolver signal to be solved, and finally performing non-zero length transformation N z And after the end part of the conversion result is removed, the module is solved to obtain the envelope of the corresponding signal segment. The method can quickly and accurately calculate the envelope of the resolver signal and provide a basis for resolver signal decoding.)

1. A resolver signal envelope detection method comprises an adaptive Gaussian window sequence generation module and an envelope detection module, and is characterized in that the adaptive Gaussian window sequence generation module generates a Gaussian window sequence according to normalized frequency and outputs the sequence to the envelope detection module.

2. The resolver signal envelope detection method according to claim 1, wherein the adaptive gaussian window sequence generation module comprises the steps of:

s1: the parameters are given as follows: sampling frequency f_sFrequency f of the excitation signal_eNon-zero threshold value z, go to step S2;

s2: the normalized frequency f is calculated using the formula_r：

Proceeding to step S3;

s3: calculating the window width coefficient w of the Gaussian window by using the following formula_k：

Proceeding to step S4;

s4: calculating the non-zero number N of the Gaussian window sequence by adopting the following formula_z：

In the formula, ceil () is an rounding-up function, and the process proceeds to step S5;

s5: the total number of points N of the gaussian window sequence is calculated using the following formula:

N＝2^nextpow2(Nz)

wherein nextpow2() is an exponentiation function to the power of 2 satisfying N or more_zProceeding to step S6;

s6: the normalized gaussian window sequence is calculated using the following formula:

proceeding to step S7;

s7: g'_F(N) carrying out period continuation by the period N, and taking a 0-N-1 main value interval to obtain a self-adaptive Gaussian window sequence g output to the envelope detection module_F(n)。

3. The resolver signal envelope detection method according to claim 1, wherein the envelope detection module comprises the steps of:

t1: input adaptive Gaussian window sequence generationModule determined parameters Nz, N, f_rAnd an adaptive Gaussian sequence g_F(n), go to step T2;

t2: calculating the number of single-sided end points N by using the following formula_end：

Entering step T3;

t3: calculating the number of updating segment points N by adopting the following formula_ce：

N_ce＝N-2×N_end

Entering step T4;

t4: calculating the spectral peak sequence number I of the data segment to be analyzed by adopting the following formula_p：

I_p＝ceil(f_r×N)

Entering step T5;

t5: judging whether the unprocessed data volume of the data buffer is more than or equal to N_ceIf yes, go to step T6, otherwise return to wait;

t6: extracting N from a data buffer_ceThe data replaces the oldest sample in the original analysis data to form a data sequence h to be analyzed with the length of N_N(n), go to step T7;

t7: with g_F(n) for h_N(n) frequency spectrum I_PPoint-solving one-way quantity S transformation to obtain sequence S_N(n), go to step T8;

t8: to S_N(n) obtaining a modulo envelope sequence | S by modulo_N(n) |, proceed to step T9;

t9: from the serial number N_endInitially, extract | S_NN of (N) |_cePoint data yielding a sequence of modulo envelopesEntering step T10;

t10: will be provided withMultiplying by polarityValue-derived envelope sequenceReturning to T5.

4. The envelope detection module of claim 3, wherein the envelope sequence in step T10 isThe method specifically comprises the following steps:

TA 1: by corresponding toMaximum value of signal to be analyzed of segment and its sequence number I_maxTo obtain a polarity value P_s：

P_s＝sign(v_e(I_max)×v_s(I_max))

Where sign () is a sign function, v_e(I_max) And v_s(I_max) Respectively representing the resolver excitation signal and the sine/cosine signal at sequence number I_maxTo step TA 2;

TA 2: judgment ofWhether a minimum value exists or not, if not, entering a step TA3, otherwise, entering a step TA 4;

TA 3: envelope sequence is calculated using the following equation

Entering TA 8;

TA 4: judgment ofCorresponding to the sequence number I_minWhether or not less than I_maxIf yes, go to step TA5, otherwise, go to step TA 6;

TA5：at sequence number I_minMultiplying the previous part by-P_s，I_minAfter part is multiplied by P_sGo to step TA 7;

TA6：at sequence number I_minMultiplying the previous part by P_s，I_minAfter part is multiplied by-P_sGo to step TA 7;

TA7：in sequence number I_minThe value of (A) is set to 0, and an envelope sequence is formed by combining the steps TA 4-TA 6Entering step TA 8;

TA 8: returning to step T5 of claim 3, waiting for the next calculation.

Technical Field

The invention relates to the field of signal processing of rotary transformers, in particular to a rotary transformer signal envelope detection method.

Background

The rotary transformer is a motor rotation angle and speed sensor commonly used in the industrial fields of electric automobiles, rail transit, aerospace and the like, and output signals of the rotary transformer are sine and cosine signals for modulating excitation signals. The output signal of the resolver is an analog signal, and the motor rotation angle and speed are obtained by resolving the analog signal.

At present, the resolver output signal is generally obtained by a dedicated chip, but this increases the cost and complicates the hardware circuit. Therefore, there is also a scheme of solving the signal of the resolver using a software algorithm using a general-purpose chip. This is advantageous for reducing costs and enhancing system flexibility, but it is not easy to achieve the desired resolution accuracy and real-time. Envelope detection of the output signal of the rotary transformer is the basis for realizing accurate calculation, but the accurate detection of the envelope is difficult to realize in a noise environment, and the patent provides an effective method aiming at the problem.

Disclosure of Invention

The invention provides a resolver signal envelope detection method, which is used for extracting the envelope of a resolver output signal and providing a basis for resolving the rotating position and speed of a motor.

The invention relates to a resolver signal envelope detection method which comprises a self-adaptive Gaussian window sequence generation module and an envelope detection module, wherein the self-adaptive Gaussian window sequence generation module generates a Gaussian window sequence according to normalized frequency and outputs the sequence to the envelope detection module.

The invention relates to a method for detecting signal envelope of a rotary transformer, wherein a self-adaptive Gaussian window sequence generating module comprises the following steps:

s1: the parameters are given as follows: sampling frequency f_sFrequency f of the excitation signal_eNon-zero threshold value z, go to step S2;

s2: the normalized frequency f is calculated using the formula_r：

Proceeding to step S3;

s3: calculating the window width coefficient w of the Gaussian window by using the following formula_k：

Proceeding to step S4;

s4: calculating the non-zero number N of the Gaussian window sequence by adopting the following formula_z：

In the formula, ceil () is an rounding-up function, and the process proceeds to step S5;

s5: the total number of points N of the gaussian window sequence is calculated using the following formula:

N＝2^nextpow2(Nz)

wherein nextpow2() is an exponentiation function to the power of 2 satisfying N or more_zProceeding to step S6;

s6: the normalized gaussian window sequence is calculated using the following formula:

proceeding to step S7;

s7: g'_F(N) performing period extension by the period N, and taking a main value range of 0-N-1 to obtain an adaptive Gaussian window sequence g for an envelope detection module_F(n)。

The invention relates to a method for detecting signal envelopes of a rotary transformer, wherein an envelope detection module of the method comprises the following steps:

t1: parameters Nz, N, f determined by an input adaptive Gaussian window sequence generation module_rAnd an adaptive Gaussian sequence g_F(n), go to step T2;

t2: calculating the number of single-sided end points N by using the following formula_end：

Entering step T3;

t3: calculating the number of updating segment points N by adopting the following formula_ce：

N_ce＝N-2×N_end

Entering step T4;

t4: calculating the spectral peak sequence number I of the data segment to be analyzed by adopting the following formula_p：

I_p＝ceil(f_r×N)

Entering step T5;

t5: judging whether the unprocessed data volume of the data buffer is more than or equal to N_ceIf yes, go to step T6, otherwise return to wait;

t6: extracting N from a data buffer_ceThe data replaces the oldest sample in the original analysis data to form a data sequence h to be analyzed with the length of N_N(n), go to step T7;

t7: with g_F(n) for h_N(n) frequency spectrum I_PPoint-solving one-way quantity S transformation to obtain sequence S_N(n), go to step T8;

t8: to S_N(n) obtaining a modulo envelope sequence | S by modulo_N(n) |, proceed to step T9;

t9: from the serial number N_endStart ofExtracting | S_NN of (N) |_cePoint data yielding a sequence of modulo envelopesEntering step T10;

t10: will be provided withMultiplying by the polarity value to obtain the envelope sequenceReturning to T5.

The envelope detection module of the invention is the envelope sequence in the step T10The method specifically comprises the following steps:

TA 1: by corresponding toMaximum value of signal to be analyzed of segment and its sequence number I_maxTo obtain a polarity value P_s：

P_s＝sign(v_e(I_max)×v_s(I_max))

Where sign () is a sign function, v_e(I_max) And v_s(I_max) Respectively representing the resolver excitation signal and the sine/cosine signal at sequence number I_maxTo step TA 2;

TA 2: judgment ofWhether a minimum value exists or not, if not, entering a step TA3, otherwise, entering a step TA 4;

TA 3: envelope sequence is calculated using the following equation

Entering TA 8;

TA 4: judgment ofCorresponding to the sequence number I_minWhether or not less than I_maxIf yes, go to step TA5, otherwise, go to step TA 6;

TA5：at sequence number I_minMultiplying the previous part by-P_s，I_minAfter part is multiplied by P_sGo to step TA 7;

TA6：at sequence number I_minMultiplying the previous part by P_s，I_minAfter part is multiplied by-P_sGo to step TA 7;

TA7：in sequence number I_minThe value of (A) is set to 0, and an envelope sequence is formed by combining the steps TA 4-TA 6Entering step TA 8;

TA 8: returning to step T5 of claim 3, waiting for the next calculation.

The method has the advantages that a good foundation can be provided for resolver signal calculation by providing the resolver signal envelope detection method, and the defects of high cost, complex circuit and the like of a hardware method are overcome.

Drawings

FIG. 1 is a flow chart of an adaptive Gaussian window sequence generation module according to the present invention.

Fig. 2 is a flow chart of the envelope detection module of the present invention.

Fig. 3 is a flow chart of envelope sequence calculation of the present invention.

Fig. 4 is an illustration of an embodiment of the present invention.

Detailed Description

The following description of the preferred embodiments of the present invention with reference to the accompanying drawings is provided for further illustration of the present invention and is not intended to limit the scope of the present invention.

The invention relates to a resolver signal envelope detection method which comprises a self-adaptive Gaussian window sequence generation module and an envelope detection module, wherein the self-adaptive Gaussian window sequence generation module generates a Gaussian window sequence according to normalized frequency and outputs the sequence to the envelope detection module.

FIG. 1 shows the steps of the adaptive Gaussian window sequence generation module according to the present invention:

s1: the parameters are given as follows: sampling frequency f_sFrequency f of the excitation signal_eNon-zero threshold value z, go to step S2;

s2: the normalized frequency f is calculated using the formula_r：

Proceeding to step S3;

s3: calculating the window width coefficient w of the Gaussian window by using the following formula_k：

Proceeding to step S4;

s4: calculating the non-zero number N of the Gaussian window sequence by adopting the following formula_z：

In the formula, ceil () is an rounding-up function, and the process proceeds to step S5;

s5: the total number of points N of the gaussian window sequence is calculated using the following formula:

N＝2^nextpow2(Nz)

wherein nextpow2() is an exponentiation function to the power of 2 satisfying N or more_zProceeding to step S6;

s6: the normalized gaussian window sequence is calculated using the following formula:

proceeding to step S7;

s7: g'_F(N) performing period extension by the period N, and taking a main value range of 0-N-1 to obtain an adaptive Gaussian window sequence g for an envelope detection module_F(n)。

Fig. 2 is a diagram illustrating the steps adopted by the envelope detection module according to the present invention:

t1: parameters Nz, N, f determined by an input adaptive Gaussian window sequence generation module_rAnd an adaptive Gaussian sequence g_F(n), go to step T2;

t2: calculating the number of single-sided end points N by using the following formula_end：

Entering step T3;

t3: calculating the number of updating segment points N by adopting the following formula_ce：

N_ce＝N-2×N_end

Entering step T4;

t4: calculating the spectral peak sequence number I of the data segment to be analyzed by adopting the following formula_p：

I_p＝ceil(f_r×N)

Entering step T5;

t5: judging whether the unprocessed data volume of the data buffer is more than or equal to N_ceIf yes, go to step T6, otherwise return to wait;

t6: extracting N from a data buffer_ceData, instead of original scoreAnalyzing the oldest sample in the data to form a data sequence h to be analyzed with the length of N_N(n), go to step T7;

t7: with g_F(n) for h_N(n) frequency spectrum I_PPoint-solving one-way quantity S transformation to obtain sequence S_N(n), go to step T8;

t8: to S_N(n) obtaining a modulo envelope sequence | S by modulo_N(n) |, proceed to step T9;

t9: from the serial number N_endInitially, extract | S_NN of (N) |_cePoint data yielding a sequence of modulo envelopesEntering step T10;

t10: will be provided withMultiplying by the polarity value to obtain the envelope sequenceReturning to T5.

The envelope detection module calculates that the single vector S transformation in step T7 is only for sequence number I_pThe spectral peak point of (A) is calculated by firstly intercepting a main value sequence by spectrum cyclic shift, and then using the main value sequence and g_F(n) multiplying, and finally solving IFFT for the product to obtain S_N(n)。

FIG. 3 is a diagram illustrating the envelope sequence of the envelope detection module in step T10 according to the present inventionThe adopted calculation steps are as follows:

TA 1: by corresponding toMaximum value of signal to be analyzed of segment and its sequence number I_maxTo obtain a polarity value P_s：

P_s＝sign(v_e(I_max)×v_s(I_max))

Where sign () is a sign function, v_e(I_max) And v_s(I_max) Respectively representing the resolver excitation signal and the sine/cosine signal at sequence number I_maxTo step TA 2;

TA 2: judgment ofWhether a minimum value exists or not, if not, entering a step TA3, otherwise, entering a step TA 4;

TA 3: envelope sequence is calculated using the following equation

Entering TA 8;

TA 4: judgment ofCorresponding to the sequence number I_minWhether or not less than I_maxIf yes, go to step TA5, otherwise, go to step TA 6;

TA5：at sequence number I_minMultiplying the previous part by-P_s，I_minAfter part is multiplied by P_sGo to step TA 7;

TA6：at sequence number I_minMultiplying the previous part by P_s，I_minAfter part is multiplied by-P_sGo to step TA 7;

TA7：in sequence number I_minSet the value at (1) to 0, and combine the steps TA 4-TA 6Envelope sequenceEntering step TA 8;

TA 8: returning to step T5 of claim 3, waiting for the next calculation.

Fig. 4 is an application example of the resolver signal envelope detection method of the present invention, and fig. 4(a) is a resolver signal to be analyzed, in which a bold section is a cut-out section for a certain analysis; FIG. 4(b) shows the intermediate state of the thick segment of FIG. 4(a) after analysis by the method of the present invention, where the total length of FIG. 4(b) is N-256 points and the thick segment is N_cePoint 152, is a valid result retained; FIG. 4(c) is a complete die envelope obtained by multiple segmentation and truncation analysis of FIG. 4(a) by the method of the present invention, wherein the bold sections are the bold sections in FIG. 4(b), and N is removed from the beginning and the end of each bold section_zThe number of end points of/2; fig. 4(d) is the complete envelope obtained by multiplying fig. 4(c) by the polarity value, i.e. the final result obtained by the method of the present invention.

The above-described embodiments of the present invention are not intended to limit the scope of the present invention, and various modifications and changes may be made to the embodiments of the present invention without departing from the spirit and scope of the present invention.