A method for automatic identification of power inductor saturation parameters
By applying high-frequency pulse voltage to the power inductor and performing wavelet high-frequency decomposition, the problem of being unable to accurately measure saturation parameters in the prior art is solved, the recognition accuracy and anti-interference ability are improved, and the comprehensive saturation parameters of the power inductor are achieved. Evaluate.
Patent Information
- Application Number
- CN202210604824.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-05-31
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-05-31
AI Technical Summary
The existing power inductor saturation parameter measurement methods cannot accurately measure the saturation current, saturation inductor value and volt-second capacity, and there are interference problems, resulting in a low recognition success rate.
By applying high-frequency pulse voltage to the power inductor, performing high-speed ADC sampling, performing wavelet high-frequency decomposition and reconstruction, removing the low-frequency part, calculating the variance threshold, judging the saturation state, and accurately evaluating the saturation parameters.
It improves the accuracy of current sampling and anti-interference ability, can more accurately identify the saturation parameters of the power inductor, and reduces the influence of human factors.
Smart Images

Figure CN115065276B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of motor control, and more specifically, to a method for automatically identifying a saturation parameter of a power inductor. Background Art
[0002] The saturation parameters of power inductors mainly include: saturation current, saturation inductance value, saturation volt-second capacity, etc. The measurement of saturation parameters is very important for the quality evaluation of power inductors. If the capacity of the power inductor does not meet the standard, it is very easy to burn out in the power electronic application circuit. The existing instruments for measuring power inductors are mainly LCR bridges. The measured inductance value is only the initial inductance value of the power inductor, and it is impossible to accurately measure parameters such as saturation current, saturation inductance value and volt-second capacity. Moreover, the measurement signal emitted by the LCR bridge is usually a low-voltage, low-frequency sinusoidal signal, which is very different from the actual working conditions of the power inductor (usually a high-frequency pulse signal). Therefore, the measurement result cannot correctly characterize the actual value of the power inductor. At present, the engineering mainly uses an oscilloscope to view the current waveform of the power inductor to determine whether the inductor is saturated. The judgment result is greatly affected by human factors, especially when the inductor is in a critical saturation state and the oscilloscope waveform cannot clearly see the saturation, then the inductor will have a greater probability of failure when used in the circuit.
[0003] However, some patents use sampling of the current data of the power inductor to determine whether the power inductor is saturated based on waveform recognition. However, the recognition method used is mainly to calculate the slope of the sampled current waveform and determine whether the power inductor is saturated by whether the slope changes too much. This method is theoretically feasible, but in practice, the collection of current waveforms will introduce more interference, especially when the inductor is charging or discharging, it will introduce a lot of spike interference. When calculating the slope of the curve, it will also introduce more interference, making the recognition success rate low.
[0004] Therefore, the present invention applies a high-frequency pulse voltage to the power inductor, samples the current of the power inductor at high speed, then performs wavelet high-frequency decomposition and reconstruction on the current waveform, removes the low-frequency part that is not of much concern, then takes the variance of the reconstructed signal segment by segment, sets a variance threshold, and judges whether the power inductor is saturated by comparing with the threshold. This method simulates the actual working environment of the power inductor, has high current sampling accuracy, and has strong anti-interference ability, and can evaluate the saturation parameters of the power inductor more comprehensively. Summary of the invention
[0005] In order to overcome the above-mentioned defects existing in the prior art, the present invention provides a method for automatically identifying the saturation parameters of a power inductor, which simulates the actual working environment of the power inductor, has high current sampling accuracy, and strong anti-interference ability, and can comprehensively evaluate the saturation parameters of the power inductor.
[0006] The above technical objectives of the present invention are achieved through the following technical solutions: A method for automatically identifying saturation parameters of a power inductor, comprising the following steps:
[0007] S1. Perform high-speed ADC sampling on the charge and discharge current of the power inductor to be tested in the saturated state. The sequence of ADC current sampling numbers is expressed as:
[0008] A(n)=[a1,a2a3......a n ]
[0009] n=512, that is, the number of samples each time is 512;
[0010] S2, the beginning and the end of the sequence A(n) are periodically extended by 9 data points, and low-pass filtered through a low-pass filter to obtain the output sequence:
[0011] B(n)=[b1,b2......b n ]
[0012] Among them, the number of sequence points of B(n) is the same as the number of sequence points of A(n);
[0013] S3, perform wavelet high frequency decomposition on the data obtained after step S2, and obtain the sequence E(n)=[d2, d4, d6...d 516 ,d 518 ], the number of points in the sequence E(n) is 256+1;
[0014] S4, reconstruct the sequence E(n) described in step S203 into the time domain, and obtain the reconstructed sequence K(n)=[k1,k2...k 512 ]=[j8,j9......j 519 ], the number of data points is 512;
[0015] S5. Divide every 8 consecutive points of the sequence K(n) starting from the first element into a group, that is, K(n) = [kk1(n), kk2(n)......kk 64 (n)] = {[k1, k2...k8], [k9, k 10 ...k 16 ]......[k 505 ,k 506 ...k 512 ]}, where kk1(n)=[k1,k2......k8], kk2(n)=[k9,k 10 ......k 16 ]......kk 64 =[k 505 ,k 506 ......k512 ], then for the sequence kk m (n) calculates the variance and obtains the sequence δ(n) = [δ1,δ2.......δ 64 ], by setting a threshold δ th , compare the elements in δ(n) with δ th The size of δ n Greater than δ th And δ n The previous data are all less than δ th When δ n The element at saturation, then the accurate saturation moment is obtained according to the data label identification in the sequence K(n);
[0016] S6. Calculate saturation parameters, where the saturation parameters include saturation current, saturation inductance, saturation volt-second capacity, and initial inductance.
[0017] Furthermore, the low-pass filter in step S2 adopts a 10th-order FIR filter with a cutoff frequency of 1 MHz, and the sequence is expressed as:
[0018] h fir (k)=[h1,h2......h 10 ]
[0019] Furthermore, the specific steps of step S3 are as follows:
[0020] The DB4 high frequency decomposition filter structure is used, and the filter coefficients are Hd(n) = [-0.2304, 0.7148, -0.6309, -0.028, 0.187, 0.0308, -0.0329, -0.0106]; for the sequence B(n) = [b1, b2......b n ] to extend the cycle and obtain the sequence C(n)=[b8,b7,...b1,b2......b 512 ,b 511 ...b 505 ], the first and last parts of the sequence are the data of periodic extension, and the number of data points is 512+14. The sequence C(n) is convolved with the sequence Hd(n) to obtain the result sequence D(n), that is Downsample the sequence D(n), take data from the second data at intervals of 1, and get the sequence E(n) = [d2, d4, d6...d 516 ,d 518 ], the number of points in the sequence E(n) is 256+1.
[0021] Furthermore, the reconstruction filter parameters used in step S4 are: Hr(n)=[-0.0106,-0.0329,0.0308,0.1870,-0.0280,-0.6309,0.7148,-0.2304], and the sequence E(n) is upsampled, that is, 0 is inserted at every other point, to obtain the sequence F(n)=[f1,f2...f 519 ]=[0,d2,0,......d 518 ,0], the number of data points of the sequence F(n) is 512+7, the sequence F(n) is extended from the beginning to the end, and the sequence G(n) = [g1,g2......g 533 ]=[f8,f7...f1,f2......f 519 ,f 518 ...f 512 ], the number of data points of sequence G(n) is 533, and then sequence G(n) is convolved with Hr(n) to obtain sequence The number of data points of the sequence J(n) is 533-7=526. Remove the extended part of the sequence J(n), that is, remove 7 points at the beginning and the end, and get the reconstructed sequence K(n)=[k1,k2......k 512 ]=[j8,j9......j 519 ], the number of data points is 512.
[0022] Furthermore, the saturation current described in step S6 is the current value at the saturation moment on the current curve; the saturation inductance value is the ratio of the excitation voltage at the saturation moment to the slope of the current curve; the saturation volt-second capacity is the product of the excitation voltage at the saturation moment and the saturation time; the initial inductance value is the ratio of the excitation voltage at the initial moment of the current curve to the slope of the current curve.
[0023] In summary, the present invention has the following beneficial effects:
[0024] 1. Compared with the existing method of manually observing the power inductor current waveform through an oscilloscope to determine whether it is saturated, the present invention achieves greatly improved accuracy by accurately sampling the power inductor current data and then performing algorithmic recognition on the current waveform.
[0025] 2. Compared with the method of directly calculating the slope of the current data and judging whether it is saturated according to the slope, the present invention adopts the method of wavelet high-frequency decomposition to perform waveform recognition in the transform domain, which greatly improves the recognition accuracy and anti-interference ability.
[0026] 3. Compared with the existing method of judging whether saturation is achieved by judging the slope, the method of comparing with the variance threshold adopted in the present invention greatly improves the accuracy and anti-interference ability. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 is a structural diagram of a current waveform sampling circuit in an embodiment of the present invention;
[0028] Figure 2 is a schematic diagram of power inductor excitation and current curves in an embodiment of the present invention;
[0029] Figure 3 It is a flow chart of automatic identification of saturation parameters in an embodiment of the present invention. DETAILED DESCRIPTION
[0030] The following is combined with Figure 1-3 The present invention is described in further detail.
[0031] Embodiment: A method for automatically identifying saturation parameters of a power inductor, such as Figure 1 As shown, the following steps are included:
[0032] S1. Perform high-speed ADC sampling on the charge and discharge current of the power inductor to be tested in the saturated state. The sequence of ADC current sampling numbers is expressed as:
[0033] A(n)=[a1,a2a3......a n ]
[0034] n=512, that is, the number of samples each time is 512;
[0035] S2, the beginning and the end of the sequence A(n) are periodically extended by 9 data points, and low-pass filtered. Here, a 10th-order FIR filter with a cutoff frequency of 1 MHz is used:
[0036] h fir (k)=[h1,h2......h 10 ]
[0037] The output sequence is:
[0038] B(n)=[b1,b2......b n ]
[0039] Before performing the FIR filtering operation, the purpose of periodically extending the beginning and end of the original sequence A(n) by 9 data points is to achieve the same number of sequence points of B(n) as that of A(n);
[0040] S3, perform wavelet high frequency decomposition processing on the data obtained after step S2, adopt DB4 high frequency decomposition filter structure, the filter coefficient is Hd(n)=[-0.2304,0.7148,-0.6309,-0.028,0.187,0.0308,-0.0329,-0.0106]; for the sequence B(n)=[b1,b2......b n ] to extend the cycle and obtain the sequence C(n)=[b8,b7,...b1,b2......b 512 ,b 511 ...b 505 ], the first and last parts of the sequence are the data of periodic extension, and the number of data points is 512+14. The sequence C(n) is convolved with the sequence Hd(n) to obtain the result sequence D(n), that is Downsample the sequence D(n), take data from the second data at intervals of 1, and get the sequence E(n) = [d2, d4, d6...d 516 ,d 518 ], the number of points in the sequence E(n) is 256+1;
[0041] S4, reconstruct the sequence E(n) described in step S203 into the time domain, and use the reconstruction filter parameters: Hr(n) = [-0.0106, -0.0329, 0.0308, 0.1870, -0.0280, -0.6309, 0.7148, -0.2304], upsample the sequence E(n), that is, insert 0 at every other point, and obtain the sequence F(n) = [f1, f2...f 519 ]=[0,d2,0,......d 518 ,0], the number of data points of the sequence F(n) is 512+7, the sequence F(n) is extended from the beginning to the end, and the sequence G(n) = [g1,g2......g 533 ]=[f8,f7...f1,f2......f 519 ,f 518 ...f 512 ], the number of data points of sequence G(n) is 533, and then sequence G(n) is convolved with Hr(n) to obtain sequence The number of data points of the sequence J(n) is 533-7=526. Remove the extended part of the sequence J(n), that is, remove 7 points at the beginning and the end, and get the reconstructed sequence K(n)=[k1,k2......k 512 ]=[j8,j9......j 519 ], the number of data points is 512;
[0042] S5. Divide every 8 consecutive points of the sequence K(n) starting from the first element into a group, that is, K(n) = [kk1(n), kk2(n)......kk 64 (n)] = {[k1, k2...k8], [k9, k 10 ...k 16 ]......[k 505 ,k 506 ...k 512 ]}, where kk1(n)=[k1,k2......k8], kk2(n)=[k9,k 10 ......k 16 ]......kk 64 =[k 505 ,k 506 ......k 512 ], then for the sequence kk m (n) calculates the variance and obtains the sequence δ(n) = [δ1,δ2.......δ 64 ], by setting a threshold δ th , compare the elements in δ(n) with δ th The size of δ n Greater than δ th And δ n The previous data are all less than δ th When δ n The element at saturation, then the accurate saturation moment is obtained according to the data label identification in the sequence K(n);
[0043] S6. Calculate saturation parameters, which include saturation current, saturation inductance, saturation volt-second capacity and initial inductance, wherein the saturation current is the current value at the saturation moment on the current curve; the saturation inductance is the ratio of the excitation voltage at the saturation moment to the slope of the current curve; the saturation volt-second capacity is the product of the excitation voltage at the saturation moment and the saturation time; the initial inductance is the ratio of the excitation voltage at the initial moment of the current curve to the slope of the current curve.
[0044] like Figure 2 As shown, the current waveform sampling circuit structure includes an H-bridge voltage excitation circuit, an H-bridge driving circuit, a driving isolation circuit, a control processor, a serial port screen, a high-speed ADC sampling circuit and an isolation sampling circuit; the H-bridge voltage excitation circuit includes MOS1 switch tube, MOS2 switch tube, MOS3 switch tube and MOS4 switch tube that together with the inductor to be measured Lx form an H-bridge, and a power supply is provided between the MOS1 switch tube and the MOS3 switch tube; a sampling resistor Rx is provided between the MOS2 switch tube and the MOS4 switch tube.
[0045] Among them, the opening and closing signals of the MOS1 switch tube, the MOS2 switch tube, the MOS3 switch tube and the MOS4 switch tube are sent by the control processor, and are driven by the isolation circuit to drive the H-bridge drive circuit; the voltage of the sampling resistor Rx is collected to the control processor through the isolation sampling circuit and the high-speed ADC sampling circuit; the saturation identification algorithm is performed in the control processor, and the result is output to the serial port screen for display after the calculation is completed.
[0046] When the MOS1 switch tube and the MOS4 switch tube are turned on, and the MOS2 switch tube and the MOS4 switch tube are turned off, a positive excitation voltage is applied to the inductor Lx to be measured. When the inductor Lx to be measured is in a saturated state, the voltage of the sampling resistor Rx is detected, and the charging current curve of the inductor Lx to be measured in the saturated state is calculated; when the MOS2 switch tube and the MOS4 switch tube are turned on, and the MOS1 switch tube and the MOS4 switch tube are turned off, a negative excitation voltage is applied to the resistor Lx to be measured. When the inductor Lx to be measured is in a saturated state, the voltage of the sampling resistor Rx is detected, and the discharge current curve of the inductor Lx to be measured in the saturated state is calculated. The discharge current curve of the inductor Lx to be measured is combined with the charging current curve of the inductor Lx to be measured described in step S102 to obtain a complete current curve of the inductor Lx to be measured in the saturated state.
[0047] like Figure 3 As shown in the figure, it is a schematic diagram of the current curve when a pulse voltage excitation is applied to the power inductor. In the initial state, when a positive excitation voltage is applied to the power inductor, the power inductor is charged, and the current increases linearly at first; when the inductor is saturated, the inductor energy is full, and the inductor loses its inductance function and becomes a resistor with a small resistance value. The inductance value of the power inductor will decrease sharply, and the current will increase sharply, which is manifested as an "upward" state on the current curve. When a reverse pulse excitation voltage is applied to the inductor, the power inductor discharges and the current continues to decrease. By identifying the saturation point of the power inductor current waveform curve, combined with the two parameters of the applied pulse excitation voltage and application time, the saturation parameters of the power inductor can be accurately calculated.
[0048] This specific embodiment is merely an explanation of the present invention and is not a limitation of the present invention. After reading this specification, those skilled in the art may make non-creative modifications to the present embodiment as needed. However, as long as they are within the scope of the claims of the present invention, they are protected by the patent law.
Claims
1. A method for automatically identifying saturation parameters of a power inductor, characterized in that: The following steps are involved: S1. Perform high-speed ADC sampling on the charge and discharge current of the power inductor to be tested in the saturated state. The sequence of ADC current sampling numbers is expressed as: ; n=512, that is, the number of samples each time is 512; S2, sequence The beginning and the end of are extended periodically by 9 data points, and low-pass filtered by a low-pass filter to obtain the output sequence: ; in, The number of sequence points and The sequence points of are the same; S3, perform wavelet high frequency decomposition on the data obtained after step S2 to obtain the sequence ,sequence The number of points is 256+1; S4, the sequence described in step S203 Reconstructed into the time domain, the reconstruction filter parameters are: , for the sequence Perform upsampling, that is, insert 0 at every other point, and get the sequence ,sequence The number of data points is 512+7, and the sequence Extend the first and last cycles to get the sequence ,sequence The number of data points is 533, and then the sequence and Perform convolution operation to obtain the sequence ,sequence The number of data points is 533-7=526. The first and last extension parts are removed, that is, 7 points are removed from the first and last parts respectively, and the reconstructed sequence is obtained. , the number of data points is 512; S5. Sequence Every 8 consecutive points starting from the first element are grouped into a group, that is ,in , then the sequence Calculate the variance and get the sequence , by setting a threshold ,Compare Internal elements and The size of Greater than and The previous data are all less than When is the element when saturated, then according to the sequence Internal data label recognition to obtain accurate saturation time; S6. Calculate saturation parameters, where the saturation parameters include saturation current, saturation inductance, saturation volt-second capacity, and initial inductance.
2. The method for automatically identifying the saturation parameters of a power inductor according to claim 1, characterized in that: The low-pass filter described in step S2 adopts a 10th-order FIR filter with a cutoff frequency of 1 MHz, and the sequence is expressed as: 。 3. The method for automatically identifying the saturation parameters of a power inductor according to claim 1, characterized in that: The specific steps of step S3 are as follows: Using DB4 high frequency decomposition filter structure, the filter coefficient is ; for sequence Extend the cycle and get the sequence , the first and last parts of the sequence are the data of periodic extension, and the number of data points obtained is 512+14; with sequence Perform convolution operation to obtain the result sequence ,Right now , for the sequence Perform downsampling, take data from the second data at intervals of 1 point, and get the sequence ,sequence The number of points is 256+1.
4. The method for automatically identifying the saturation parameters of a power inductor according to claim 1, characterized in that: The saturation current described in step S6 is the current value at the saturation moment on the current curve; the saturation inductance value is the ratio of the excitation voltage at the saturation moment to the slope of the current curve; the saturation volt-second capacity is the product of the excitation voltage at the saturation moment and the saturation time; the initial inductance value is the ratio of the excitation voltage at the initial moment of the current curve to the slope of the current curve.
