Method and device for generating re-calibrated direct-current frequency signal

Abstract

The invention provides a method and a device for generating a re-calibrated direct-current frequency signal, and belongs to the technical field of direct-current frequency signal calibration. Comprising the following steps: S1, configuring at least one analog quantity output channel on primary frequency modulation test equipment, and respectively outputting direct current signals corresponding to expected power grid frequency; s2, setting an output relational expression between a direct-current voltage signal output by the analog quantity output channel and an expected power grid frequency; s3, when the primary frequency modulation test equipment is used for the first time, measuring point data of at least one analog quantity output channel is obtained, calibration is carried out, and the deviation degree of the fitting calibration curve and the output relational expression is judged until the deviation of the fitting calibration curve and the output relational expression does not exceed a set threshold value; and S4, after using for a period of time, re-calibrating at least one analog quantity output channel, and judging the deviation between the fitting re-calibration curve and the output relational expression until the deviation between the fitting re-calibration curve and the output relational expression does not exceed a set threshold value.

Description

Technical Field

[0001] The present invention relates to the technical field of direct current frequency signal calibration, and in particular to a method and device for generating a recalibrated direct current frequency signal. Background Art

[0002] In a primary frequency regulation test, it is necessary to generate a frequency signal that simulates the frequency change of the power grid in order to test whether the primary frequency regulation performance of the device under test meets the standard specification requirements.

[0003] For some specific test occasions, a DC frequency signal is required. Conventional primary frequency modulation test instruments do not provide such frequency signals. The current method for obtaining a DC frequency signal in the test is to generate the required DC frequency signal through another additional device. This method results in low integration, large equipment size, and no automatic calibration function for the accuracy of the output frequency. Long-term use may cause the signal accuracy to deteriorate, affecting the reliability of the test.

[0004] Therefore, it is very necessary to propose a recalibrated DC frequency signal generation method integrated in a primary frequency modulation device to improve the portability and usability of the device and to improve the disadvantage of unstable output frequency accuracy caused by aging of the device. Summary of the invention

[0005] In view of this, the present invention proposes a method and device for generating a direct current frequency signal with high integration and capable of automatically recalibrating the output frequency accuracy.

[0006] In one aspect, the present invention provides a method for generating a recalibrated DC frequency signal, comprising the following steps:

[0007] S1: On the primary frequency modulation test equipment, configure at least one analog output channel to output a DC signal corresponding to the expected grid frequency;

[0008] S2: By setting, the corresponding relationship between the DC voltage signal V output by at least one analog output channel and the expected grid frequency x is set to satisfy the output relationship: V = f(x), where f(x) is a polynomial;

[0009] S3: When the frequency modulation test equipment is used for the first time, obtain data of m measuring points of at least one analog output channel, perform calibration and fit the calibration curve; judge the degree of deviation between the fitted calibration curve and the output relationship formula, if the deviation between the fitted calibration curve and the curve corresponding to the output relationship formula does not exceed the set threshold, it means that the DC voltage signal output by the frequency modulation test equipment meets the requirements; if the deviation between the fitted calibration curve and the curve corresponding to the output relationship formula is greater than the set threshold, adjust the polynomial coefficient of the output relationship formula until the deviation between the fitted calibration curve and the output relationship formula does not exceed the set threshold;

[0010] S4: After the frequency modulation test equipment has been used for a period of time, at least one analog output channel is recalibrated, and n different measuring points are reacquired. Through m+n measuring points, the deviation between the recalibrated fitting recalibration curve and the output relationship is jointly determined. If the fitting recalibration curve and the output relationship do not exceed the set threshold, it means that the DC voltage signal output by the frequency modulation test equipment meets the requirements; if the fitting recalibration curve and the output relationship are greater than the set threshold, adjust the polynomial coefficients of the output relationship until the deviation between the fitting recalibration curve and the output relationship does not exceed the set threshold.

[0011] On the basis of the above technical solution, preferably, the corresponding relationship between the DC voltage signal V and the expected grid frequency x in step S2 satisfies the output relationship: V=a0+a1x+a2x 2 +…a K x K , the index value of the frequency-related term is k = 1, 2, 3, ..., K, K is an odd number greater than or equal to 5, a0, a1, a2, ..., a K are polynomial coefficients, and the initial values ​​of the polynomial coefficients are obtained through tuning.

[0012] Preferably, the specific content of step S3 is:

[0013] S31: Select m measuring points, one by one, to form (x m , V m ) data pair, x m is the measuring point, V m is the corresponding measurement point data; remove abnormal data pairs;

[0014] S32: Let (x m , V m The expression of the fitting calibration curve of the data pair is: where a i0 、a i1 、a i2 , …, a iK is the coefficient of the fitting rate curve, i = 1, 2, 3, ..., m, there are several combinations of coefficients of the fitting rate curve; set the first loss function E, Where λ is the first regularization parameter, λ∈[0,1], a iA 、a iB is the coefficient of any two different fitted calibration curves, a iA ≠a iB , a iA 、a iB ∈a i0 、a i1 、a i2 , ..., aiK , select the coefficient combination of the fitting rate curve to minimize the value of the first loss function;

[0015] S33: randomly extract M data pairs on the fitting calibration curves, M<m; if the deviations of the corresponding measuring point positions of the arbitrarily selected three or more non-adjacent measuring point data on the output relationship and the fitting calibration curve all exceed the set threshold, it is determined that the deviation of the coefficient combination of the current fitting calibration curve and the curve corresponding to the output relationship is greater than the set threshold, and does not meet the deviation requirement, and the coefficient combination of the current fitting calibration curve is eliminated from the combination of the coefficients of the fitting calibration curve, and the fitting that meets the minimum first loss function E is re-searched from the remaining coefficient combinations of the fitting calibration curve The coefficient combination of the calibration curve is repeated to compare the measurement point data deviation corresponding to the measurement point on the output relationship and the fitted calibration curve until a coefficient combination of the fitted calibration curve that meets the requirements is found; if the combination of the coefficients of the fitted calibration curve does not meet the measurement point data deviation requirements corresponding to the measurement point on the output relationship and the fitted calibration curve, then step S3 is repeated no more than K / 2 times until a coefficient combination of the fitted calibration curve that meets the requirements is found; if the number of times step S3 is repeated is greater than K / 2 times, it is determined that at least one analog output channel of the primary frequency modulation test equipment is faulty.

[0016] Further preferably, the step of removing abnormal data pairs in step S31 is to remove the data pairs (x m , V m ) m Substitute into the output relationship and compare the calculation result with x m If the deviation exceeds Δx, the data pair is directly eliminated and a new set of data pairs is added; if the deviation does not exceed Δx, the data pair is retained.

[0017] More preferably, the threshold value is set as follows: where x M and x Mi It is the output relationship with the same measuring point data and the corresponding measuring point on the fitting calibration curve.

[0018] Still further preferably, the n different measuring points reacquired in step S4 satisfy n<m, and the n measuring points are different from the m measuring points.

[0019] More preferably, the specific content of step S4 is:

[0020] S41: Combine m+n measurement points to construct m+n data pairs (x m+n , V m+n );

[0021] S42: According to m+n data pairs (xm+n , V m+n ), fitting the recalibration curve, where a j0 、a j1 、a j2 , …, a jK is the coefficient of fitting the recalibration curve, j = 1, 2, 3, ..., m + n; there are several types of coefficients for fitting the recalibration curve; set the second loss function L, Where λ0 is the second regularization parameter, λ0∈[0,1], a jA and a jB are the coefficients of the two different fitted recalibration curves, a jA 、a jB ∈a j0 、a j1 、a j2 , …, a jK , a jA ≠a jB ; Selecting a coefficient combination for fitting the recalibration curve so that the value of the first loss function is minimized;

[0022] S43: randomly extract P data pairs on the recalibration curves, M<P<m+n, and respectively search for the measurement point positions corresponding to the same measurement point data on the output relationship and the fitted recalibration curve. If the deviations of the corresponding measurement point positions of three or more non-adjacent different measurement point data selected arbitrarily on the output relationship and the fitted recalibration curve exceed the set threshold, it is determined that the deviation of the current coefficient combination of the fitted recalibration curve and the curve corresponding to the output relationship is greater than the set threshold and does not meet the deviation requirement, and the current coefficient combination of the fitted recalibration curve is eliminated from the combination of coefficients of the fitted recalibration curve, and the coefficient combination of the fitted recalibration curve is selected from the remaining combinations of coefficients of the fitted recalibration curve. Re-search for the coefficient combination of the fitted recalibration curve that minimizes the second loss function L, and repeat the process of comparing the measurement point data deviations corresponding to the measurement points on the output relationship and the fitted recalibration curve until a coefficient combination of the fitted recalibration curve that meets the requirements is found; if all combinations of the coefficients of the fitted recalibration curve do not meet the measurement point data deviation requirements corresponding to the measurement points on the output relationship and the fitted recalibration curve, repeat step S4 no more than K / 2 times until a coefficient combination of the fitted recalibration curve that meets the requirements is found; if the number of times step S4 is repeated is greater than K / 2 times, it is determined that at least one analog output channel of the primary frequency modulation test equipment is faulty.

[0023] Further preferably, the set threshold value of the recalibration process is the same as the set threshold value calibrated when the frequency modulation test equipment is used for the first time.

[0024] On the other hand, the present invention also provides a recalibrable DC frequency signal generating device, comprising a digital signal processor DSP, a digital-to-analog converter DAC, a filtering unit, an isolation unit and a driving unit. The digital signal processor DSP sends an analog-to-frequency signal output command to the digital-to-analog converter DAC in real time. The output signal of the digital-to-analog converter DAC is filtered out by the filtering unit to remove noise. The isolation unit transmits the signal to the side of the on-site device to be tested in an isolated manner. The signal output by the isolation unit is sent to the driving unit for amplification to increase the load capacity, and an analog frequency signal is output. The digital signal processor DSP adopts the above-mentioned recalibrable DC frequency signal generating method for calibration or recalibration.

[0025] The recalibrated DC frequency signal generation method and device provided by the present invention have the following beneficial effects compared with the prior art:

[0026] (1) By integrating a recalibrable DC frequency signal generating device into the existing primary frequency modulation test equipment, the calibration and recalibration process is performed at the first use and after a fixed interval, so as to realize self-calibration of the output parameters and maintain the stability of the DC frequency signal output, thereby improving the problem of signal accuracy degradation caused by equipment aging;

[0027] (2) The recalibration process not only uses the m measurement point data during calibration, but also further introduces the n current measurement point data as new information points. The output performance of the current analog output channel is evaluated by the combination of m+n measurement point data at different times. The relevant parameters of the original calibration state can be retained, which is of reference significance for subsequent recalibration.

[0028] (3) The present application integrates a DC frequency signal generating device directly into a primary frequency modulation test device, eliminating the need for additional DC signal generating equipment, thereby improving the portability and ease of use of the device. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 It is a calibration flow diagram of the recalibrable DC frequency signal generating method and device of the present invention;

[0031] Figure 2 A flow chart of recalibration of the recalibrable DC frequency signal generating method and device of the present invention;

[0032] Figure 3 The device structure block diagram of the recalibrated DC frequency signal generating method and device of the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] like Figure 1 and Figure 2 As shown, on the one hand, the present invention provides a method for generating a recalibrated DC frequency signal, comprising the following steps:

[0035] S1: On the primary frequency modulation test equipment, configure at least one analog output channel to output a DC signal corresponding to the expected grid frequency.

[0036] The signal output by at least one analog output channel may be a -10V to +10V analog voltage signal, or a 0-20mA analog current signal. For ease of explanation, the following embodiments take voltage signals as an example.

[0037] S2: Through tuning, the corresponding relationship between the DC voltage signal V output by at least one analog output channel and the expected grid frequency x is set to satisfy the output relationship: V=f(x), where f(x) is a polynomial.

[0038] The corresponding relationship between the DC voltage signal V and the expected grid frequency x satisfies the output relationship: V = a0 + a1x + a2x 2 +…a K x K , the index value of the frequency-related term is k = 1, 2, 3, ..., K, K is an odd number greater than or equal to 5, a0, a1, a2, ..., a K are polynomial coefficients, and the initial values ​​of the polynomial coefficients are obtained through tuning. The output relationship here is a simple theoretical value.

[0039] Through adjustment, the DC signal meets the requirements of the national standard "Technical Regulations and Test Guidelines for Primary Frequency Modulation of Grid-connected Power Supply" for testing purposes.

[0040] S3: When the frequency modulation test equipment is used for the first time, obtain data from m measuring points of at least one analog output channel, perform calibration and fit the calibration curve; determine the degree of deviation between the fitted calibration curve and the output relationship; if the deviation between the fitted calibration curve and the curve corresponding to the output relationship does not exceed the set threshold, it means that the DC voltage signal output by the frequency modulation test equipment meets the requirements; if the deviation between the fitted calibration curve and the curve corresponding to the output relationship is greater than the set threshold, adjust the polynomial coefficients of the output relationship until the deviation between the fitted calibration curve and the output relationship does not exceed the set threshold.

[0041] Step S3 specifically includes the following sub-steps:

[0042] S31: Select m measuring points, one by one, to form (x m , V m ) data pair, x m is the measuring point, corresponding to the frequency, V m is the corresponding measurement point data, is a DC signal; the abnormal data pairs are eliminated.

[0043] The removal of abnormal data pairs mentioned here is to remove the data pairs (x m , V m ) m Substitute into the output relationship and compare the calculation result with x m If the deviation exceeds Δx, the data pair is directly eliminated and a new set of data pairs is added; if the deviation does not exceed Δx, the data pair is retained. Eliminating abnormal data pairs is conducive to screening data with excessive noise and improving the accuracy and reliability of subsequent data comparison. As a preferred embodiment, Δx = 0.01.

[0044] S32: Let (x m , V m The expression of the fitting calibration curve of the data pair is: where a i0 、a i1 、a i2 , …, a iK is the coefficient of the fitting rate curve, i = 1, 2, 3, ..., m, there are several combinations of coefficients of the fitting rate curve; set the first loss function E, Where λ is the first regularization parameter, λ∈[0,1], a iA 、a iB is the coefficient of any two different fitted calibration curves, a iA ≠a iB , a iA 、a iB ∈a i0 、a i1 、a i2, …, a iK , select a coefficient combination for fitting the calibration curve to minimize the value of the first loss function.

[0045] The first loss function consists of two parts. The first summation term is the sum of the squares of the errors of each data pair, and the second summation term is a regularization term, which aims to introduce additional constraints by adding the sum of the squares of the differences between different coefficients in the current coefficient combination for fitting the calibration curve. The weight of the regularization term is controlled by the first regularization parameter λ. Such a combination of constraint terms improves the generalization ability, and the combination of the two summation terms can improve the accuracy of the fitted calibration curve.

[0046] S33: Randomly extract M data pairs on the fitted calibration curve, where M < m; if the deviation of the measurement point positions corresponding to any three or more non - adjacent different measurement point data in the output relation and the fitted calibration curve exceeds the set threshold, it is determined that the deviation between the current coefficient combination of the fitted calibration curve and the curve corresponding to the output relation is greater than the set threshold, and the deviation requirement is not met. Then, remove the current coefficient combination of the fitted calibration curve from the combination of the coefficients of the fitted calibration curve, and re - find the coefficient combination of the fitted calibration curve that minimizes the first loss function E from the remaining coefficient combinations of the fitted calibration curve. Repeat the process of comparing the deviation of the measurement point data corresponding to the output relation and the fitted calibration curve until a coefficient combination of the fitted calibration curve that meets the requirements is found; if none of the combinations of the coefficients of the fitted calibration curve meet the deviation requirements of the measurement point data corresponding to the output relation and the fitted calibration curve, then repeat step S3 no more than K / 2 times until a coefficient combination of the fitted calibration curve that meets the requirements is found; if the number of times of repeating step S3 is greater than K / 2 times, it is determined that there is a fault in at least one analog output channel of the primary frequency regulation test equipment. Although there are multiple combinations of the coefficients of the fitted calibration curve, as a whole, they still show a certain fluctuation trend around the curve corresponding to the output relation. Within a reasonable fluctuation trend range, there may be multiple combinations of the coefficients of the fitted calibration curve that minimize the deviation between the fitted calibration curve and the curve corresponding to the output relation. Generally speaking, to improve the efficiency of comparison, data can be sampled at fixed intervals to obtain measurement point data, and the corresponding measurement point values can be found on the curve corresponding to the output relation and the fitted calibration curve according to the measurement point data.

[0047] The determination content of the set threshold is where x M and x Mi are the corresponding measurement points on the output relation and the fitted calibration curve with the same measurement point data. Here, the frequency accuracy is set as ±0.006% of the initial value as the limit, and those exceeding this range are considered to have a large frequency deviation.

[0048] S4: After the frequency modulation test equipment has been used for a period of time, at least one analog output channel is recalibrated, and n different measuring points are reacquired. Through m+n measuring points, the deviation between the recalibrated fitting recalibration curve and the output relationship is jointly determined. If the fitting recalibration curve and the output relationship do not exceed the set threshold, it means that the DC voltage signal output by the frequency modulation test equipment meets the requirements; if the fitting recalibration curve and the output relationship are greater than the set threshold, adjust the polynomial coefficients of the output relationship until the deviation between the fitting recalibration curve and the output relationship does not exceed the set threshold.

[0049] The n different measurement points reacquired in step S4 satisfy n<m, and the n measurement points are different from the m measurement points. The acquisition time of the n measurement points and the m measurement points is different. Moreover, when acquiring the n measurement points, the external environment is as close as possible to the external environment of the calibration process.

[0050] The specific content of step S4 is similar to step S3, and specifically includes the following steps:

[0051] S41: Combine m+n measurement points to construct m+n data pairs (x m+n , V m+n );

[0052] S42: According to m+n data pairs (x m+n , V m+n ), fitting the recalibration curve, where a j0 、a j1 、a j2 , …, a jK is the coefficient of fitting the recalibration curve, j = 1, 2, 3, ..., m + n; there are several types of coefficients for fitting the recalibration curve; set the second loss function L, Where λ0 is the second regularization parameter, λ0∈[0,1], a jA and a jB are the coefficients of the two different fitted recalibration curves, a jA 、a jB ∈a j0 、a j1 、a j2 , …, a jK , a jA ≠a jB ; The coefficient combination of the fitting recalibration curve is selected to minimize the value of the first loss function; the values ​​of the first regularization parameter λ and the second regularization parameter λ0 are not completely equal, and the first loss function E and the second loss function L have the same value range. The value range of the first loss function E is associated with the value range of the first regularization parameter λ; the value range of the second loss function L is associated with the value range of the second regularization parameter λ0.

[0053] S43: Randomly extract P data pairs on the recalibration curve, where M < P < m + n. Respectively find the measuring point positions corresponding to the same measuring point data on the output relational expression and the fitted recalibration curve. If the deviations of the measuring point positions corresponding to any three or more non - adjacent different measuring point data on the output relational expression and the fitted recalibration curve exceed the set threshold, it is determined that the deviation between the coefficient combination of the current fitted recalibration curve and the curve corresponding to the output relational expression is greater than the set threshold, not meeting the deviation requirement. Then, eliminate the coefficient combination of the current fitted recalibration curve from the combination of the coefficients of the fitted recalibration curve, and re - find the coefficient combination of the fitted recalibration curve that meets the minimum of the second loss function L from the remaining combination of the coefficients of the fitted recalibration curve. Repeat the process of comparing the deviations of the measuring point data corresponding to the measuring points on the output relational expression and the fitted recalibration curve until a coefficient combination of the fitted recalibration curve that meets the requirements is found; if the combinations of the coefficients of the fitted recalibration curve do not meet the deviation requirements of the measuring point data corresponding to the measuring points on the output relational expression and the fitted recalibration curve, repeat step S4 no more than K / 2 times until a coefficient combination of the fitted recalibration curve that meets the requirements is found; if the number of times of repeating step S4 is greater than K / 2 times, it is determined that there is a fault in at least one analog quantity output channel of the primary frequency modulation test equipment. The structure of the second loss function L is basically the same as the composition of the first loss function E. It also takes the minimum value after summing the sum of the squared errors of each data pair and the sum of the squared differences of different coefficients in the coefficient combination of the current fitted recalibration curve as the screening index.

[0054] The set threshold in the recalibration process is the same as the set threshold calibrated during the initial use of the primary frequency modulation test equipment. Assume that the recalibration period is T. Then, after the period T, the number of measuring points obtained in step S4 is m + n, after the period 2T, the number of measuring points obtained in step S4 is m + 2n, and so on.

[0055] On the other hand, referring to the attached Figure 3 , the present invention also provides a re - calibratable DC - type frequency signal generating device, including a digital signal processor DSP, a digital - to - analog converter DAC, a filtering unit, an isolation unit, and a driving unit. The digital signal processor DSP sends a modulation frequency signal output command to the digital - to - analog converter DAC in real time. The output signal of the digital - to - analog converter DAC is filtered by the filtering unit to remove clutter, and the isolation unit is transmitted to the side of the on - site device to be measured in an isolation manner. The signal output by the isolation unit is sent to the driving unit for amplification to improve the load capacity and output an analog signal; the digital signal processor DSP performs calibration or recalibration using the above - mentioned re - calibratable DC - type frequency signal generating method. Figure 3 Shown is a structural block diagram of a device with two independent analog signals.

[0056] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating a recalibrated DC frequency signal, characterized in that: The steps include: S1: On the primary frequency modulation test equipment, configure at least one analog output channel to output a DC signal corresponding to the expected grid frequency; S2: By setting, the corresponding relationship between the DC voltage signal V output by at least one analog output channel and the expected grid frequency x is set to satisfy the output relationship: V = f(x), where f(x) is a polynomial; S3: When the frequency modulation test equipment is used for the first time, obtain data of m measuring points of at least one analog output channel, perform calibration and fit the calibration curve; judge the degree of deviation between the fitted calibration curve and the output relationship formula, if the deviation between the fitted calibration curve and the curve corresponding to the output relationship formula does not exceed the set threshold, it means that the DC voltage signal output by the frequency modulation test equipment meets the requirements; if the deviation between the fitted calibration curve and the curve corresponding to the output relationship formula is greater than the set threshold, adjust the polynomial coefficient of the output relationship formula until the deviation between the fitted calibration curve and the output relationship formula does not exceed the set threshold; S4: After the frequency modulation test equipment has been used for a period of time, at least one analog output channel is recalibrated, and n different measuring points are reacquired. Through m+n measuring points, the deviation between the recalibrated fitting recalibration curve and the output relationship is jointly determined. If the fitting recalibration curve and the output relationship do not exceed the set threshold, it means that the DC voltage signal output by the frequency modulation test equipment meets the requirements; if the fitting recalibration curve and the output relationship are greater than the set threshold, adjust the polynomial coefficients of the output relationship until the deviation between the fitting recalibration curve and the output relationship does not exceed the set threshold.

2. The method for generating a recalibrable DC frequency signal according to claim 1, characterized in that: The corresponding relationship between the DC voltage signal V and the expected grid frequency x in step S2 satisfies the output relationship: V = a0 + a1x + a2x 2 +...a K x K , the index value of the frequency-related term is k = 1, 2, 3, ..., K, K is an odd number greater than or equal to 5, a0, a1, a2, ..., a K are polynomial coefficients, and the initial values ​​of the polynomial coefficients are obtained through tuning.

3. The method for generating a recalibrable DC frequency signal according to claim 2, characterized in that: The specific content of step S3 is: S31: Select m measuring points, one by one, to form (x m , V m ) data pair, x m is the measuring point, V m is the corresponding measurement point data; remove abnormal data pairs; S32: Let (x m , V m The expression of the fitting calibration curve of the data pair is: where a i0 、a i1 、a i2 , ..., a iK is the coefficient of the fitting rate curve, i = 1, 2, 3, ..., m, there are several combinations of coefficients of the fitting rate curve; set the first loss function E, Where λ is the first regularization parameter, λ∈[0,1], a iA 、a iB is the coefficient of any two different fitted calibration curves, a iA ≠a iB , a iA 、a iB ∈a i0 、a i1 、a i2 , …, a iK , select the coefficient combination of the fitting rate curve to minimize the value of the first loss function; S33: randomly extract M data pairs on the fitting calibration curves, M<m; if the deviations of the corresponding measuring point positions of the arbitrarily selected three or more non-adjacent measuring point data on the output relationship and the fitting calibration curve all exceed the set threshold, it is determined that the deviation of the coefficient combination of the current fitting calibration curve and the curve corresponding to the output relationship is greater than the set threshold, and does not meet the deviation requirement, and the coefficient combination of the current fitting calibration curve is eliminated from the combination of the coefficients of the fitting calibration curve, and the fitting that meets the minimum first loss function E is re-searched from the remaining coefficient combinations of the fitting calibration curve The coefficient combination of the calibration curve is repeated to compare the measurement point data deviation corresponding to the measurement point on the output relationship and the fitted calibration curve until a coefficient combination of the fitted calibration curve that meets the requirements is found; if the combination of the coefficients of the fitted calibration curve does not meet the measurement point data deviation requirements corresponding to the measurement point on the output relationship and the fitted calibration curve, then step S3 is repeated no more than K / 2 times until a coefficient combination of the fitted calibration curve that meets the requirements is found; if the number of times step S3 is repeated is greater than K / 2 times, it is determined that at least one analog output channel of the primary frequency modulation test equipment is faulty.

4. The method for generating a recalibrated DC frequency signal according to claim 3, characterized in that: The step S31 of removing abnormal data pairs is to remove the data pairs (x m , V m ) m Substitute into the output relationship and compare the calculation result with x m If the deviation exceeds Δx, the data pair is directly eliminated and a new set of data pairs is added; if the deviation does not exceed Δx, the data pair is retained.

5. The method for generating a recalibrated DC frequency signal according to claim 4, characterized in that: The judgment content of setting the threshold is where x M and x Mi It is the output relationship with the same measuring point data and the corresponding measuring point on the fitting calibration curve.

6. The method for generating a recalibrable DC frequency signal according to claim 5, characterized in that: The n different measuring points reacquired in step S4 satisfy n<m, and the n measuring points are different from the m measuring points.

7. The method for generating a recalibrable DC frequency signal according to claim 6, characterized in that: The specific content of step S4 is: S41: Combine m+n measurement points to construct m+n data pairs (x m+n , V m+n ); S42: According to m+n data pairs (x m+n , V m+n ), fitting the recalibration curve, where a j0 、a j1 、a j2 , …, a jK is the coefficient of fitting the recalibration curve, j = 1, 2, 3, ..., m + n; there are several types of coefficients for fitting the recalibration curve; set the second loss function L, Where λ0 is the second regularization parameter, λ0∈[0,1], a jA and a jB are the coefficients of the two different fitted recalibration curves, a jA 、a jB ∈a j0 、a j1 、a j2 , …, a jK , a jA ≠a jB ; Selecting a coefficient combination for fitting the recalibration curve so that the value of the first loss function is minimized; S43: randomly extract P data pairs on the recalibration curves, M<P<m+n, and respectively search for the measurement point positions corresponding to the same measurement point data on the output relationship and the fitted recalibration curve. If the deviations of the corresponding measurement point positions of three or more non-adjacent different measurement point data selected arbitrarily on the output relationship and the fitted recalibration curve exceed the set threshold, it is determined that the deviation of the current coefficient combination of the fitted recalibration curve and the curve corresponding to the output relationship is greater than the set threshold and does not meet the deviation requirement, and the current coefficient combination of the fitted recalibration curve is eliminated from the combination of coefficients of the fitted recalibration curve, and the coefficient combination of the fitted recalibration curve is selected from the remaining combinations of coefficients of the fitted recalibration curve. Re-search for the coefficient combination of the fitted recalibration curve that minimizes the second loss function L, and repeat the process of comparing the measurement point data deviations corresponding to the measurement points on the output relationship and the fitted recalibration curve until a coefficient combination of the fitted recalibration curve that meets the requirements is found; if all combinations of the coefficients of the fitted recalibration curve do not meet the measurement point data deviation requirements corresponding to the measurement points on the output relationship and the fitted recalibration curve, repeat step S4 no more than K / 2 times until a coefficient combination of the fitted recalibration curve that meets the requirements is found; if the number of times step S4 is repeated is greater than K / 2 times, it is determined that at least one analog output channel of the primary frequency modulation test equipment is faulty.

8. The method for generating a recalibrable DC frequency signal according to claim 7, characterized in that: The set thresholds during the recalibration process are the same as the set thresholds used when the frequency modulation test equipment is calibrated for the first time.

9. A recalibrable DC frequency signal generating device, characterized in that: The invention comprises a digital signal processor DSP, a digital-to-analog converter DAC, a filtering unit, an isolation unit and a driving unit. The digital signal processor DSP sends an analog-to-frequency signal output command to the digital-to-analog converter DAC in real time. The output signal of the digital-to-analog converter DAC is filtered out by the filtering unit to remove clutter. The isolation unit transmits the signal to the side of the on-site device to be tested in an isolated manner. The signal output by the isolation unit is sent to the driving unit for amplification to improve the load capacity, and an analog frequency signal is output. The digital signal processor DSP adopts the recalibrable DC frequency signal generation method described in any one of claims 1 to 8 for calibration or recalibration.

Images