Cable material diameter detection method based on eddy current impedance effect

Abstract

The invention discloses a cable material diameter detection method based on an eddy current impedance effect, and particularly relates to the field of cable nondestructive testing, which comprises the following steps: S1, synchronously acquiring the resistance and inductance of a coil under a low-frequency signal and the resistance and inductance of the coil under a high-frequency signal by adopting a double-frequency excitation eddy current sensor; the environment temperature and the lift-off distance between the sensor and the cable are measured; s2, denoising is carried out through a combinatorial algorithm, and temperature compensation is carried out on the resistance and inductance of a sensor coil and the conductivity of a cable core wire; s3, calculating a material sensitive factor under low frequency, calculating a line diameter sensitive factor under high frequency, and correcting when the lift-off distance deviates; s4, calculating a material deviation rate to judge the material, fitting a linear regression model, and combining with conductivity to correct the skin depth to calculate the line diameter; and S5, according to the material and the wire diameter, comprehensively determining progressive output, and triggering a corresponding alarm in case of abnormality. The problems of parameter coupling and incomplete temperature compensation are solved, and the detection precision and the environmental adaptability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cable nondestructive testing, and more particularly to a cable material and wire diameter detection method based on eddy current impedance effect. BACKGROUND

[0002] In the field of cable production quality control and power grid operation and maintenance detection, the nondestructive testing technology based on eddy current impedance effect has become one of the mainstream technologies for cable material identification and wire diameter measurement due to its advantages of non-contact, no mechanical wear, and fast detection speed. The core principle is to generate an alternating magnetic field by passing an alternating current into an eddy current sensor coil, when a cable (conductive material) enters the magnetic field, an eddy current is induced on the surface, and the reverse magnetic field of the eddy current changes the coil impedance, and the impedance change is analyzed to realize the detection of cable characteristics.

[0003] At present, the existing eddy current cable detection scheme mostly adopts a single frequency excitation mode, that is, a single frequency alternating current is used to drive the sensor coil, and machine learning algorithms such as neural network and support vector machine are used to process the impedance signal to extract material and wire diameter information. This kind of scheme is widely used in the identification of aluminum instead of copper for copper / aluminum core cables, the online monitoring of wire diameter for high-speed wire drawing production lines, and the offline sampling inspection of overhead cables and cables in cable wells in the field of power grid operation and maintenance, trying to meet the needs of the industrial field for rapid verification of cable quality.

[0004] However, in actual use, there are still some shortcomings, such as under single-frequency excitation, the change of coil impedance is simultaneously affected by the cable material, wire diameter and sensor-cable lift-off distance, and the contribution of each parameter to the impedance change cannot be distinguished alone, especially when the cable wire diameter fluctuates or the lift-off distance changes slightly, material misjudgment and wire diameter out-of-tolerance missed detection problems are prone to occur, which is difficult to meet the requirements of GB / T3956-2008 on cable wire diameter A-level precision (error ≤ ±0.005mm), and the existing technology only compensates for the temperature drift of the sensor coil itself (such as the change of coil wire resistance and skeleton inductance with temperature), but ignores the key change of cable core conductivity with temperature, for example, the conductivity σ of copper core cable will decrease by about 3.9% when the temperature increases by 10℃, and the conductivity of aluminum core cable decreases by about 4.29%. When the temperature fluctuates by 10℃ in the industrial scene, the wire diameter measurement error will increase by more than 2 times, which cannot adapt to the detection environment of unstable temperature such as workshop and outdoor, and a constant temperature device needs to be additionally built, increasing the detection cost. Most schemes train a machine learning model by collecting ≥1000 groups of samples to realize the correlation analysis of material and wire diameter, but the black box characteristics of the algorithm make the detection error untraceable, and it is difficult to determine whether the abnormal result is caused by parameter setting, hardware failure or model deviation. At the same time, a large amount of time and sample resources are consumed for model training, and samples need to be collected and trained again for maintenance in industrial scenes, and the calibration time is usually ≥4 hours, which has high maintenance cost and cannot quickly adapt to the detection needs of different specifications of cables. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present application provide a cable material and wire diameter detection method based on the eddy current impedance effect, which solves the problems in the above background art by the following scheme.

[0006] In order to achieve the above-mentioned purpose, the present application provides the following technical scheme: a cable material and wire diameter detection method based on the eddy current impedance effect, comprising S1. Data acquisition: a double-frequency excitation eddy current sensor is used to synchronously collect the resistance R 10 , inductance L 10 of the coil under low-frequency excitation signal f1, the resistance R 20 , inductance L 20 of the coil under high-frequency excitation signal f2, and the environmental temperature T0, lift-off distance h0 between the sensor and the cable, wherein the f1 is set to 500kHz-1MHz according to the cable material, so that the eddy current skin depth δ1≤cable wire diameter d / 3; the f2 is set to 3MHz-5MHz according to the cable wire diameter, so that the eddy current skin depth δ2≥cable wire diameter d / 3;

[0007] S2. Data preprocessing:

[0008] S2.1. Noise suppression: the collected R 10 , L10 , R 20 , L 20 The denoised data R 12 , L 12 , R 22 , L 22 is obtained by using a combination algorithm of moving average filtering and wavelet threshold filtering.

[0009] S2.2. Full-link temperature compensation: resistance and inductance temperature compensation is performed on the sensor coil to obtain the compensated resistance R t1 , R t2 and inductance L t1 , L t2 ; at the same time, the temperature compensation is performed on the cable core conductivity to obtain the compensated conductivity σ T .

[0010] S3. Parameter decoupling extraction:

[0011] S3.1. Material sensitivity factor extraction: based on R t1 , L t1 at low frequency f1, the coil impedance change rate ΔZ1r is calculated, combined with the coil structure constant K1, the material sensitivity factor S1=μ r ・σ is extracted, where μ r is the relative magnetic permeability of the cable and σ is the conductivity of the cable.

[0012] S3.2. Wire diameter sensitivity factor extraction: based on R t2 , L t2 at high frequency f2, the coil impedance change amount ΔZ2 is calculated, and the wire diameter sensitivity factor S2=ΔZ2 is defined; if there is a deviation between h0 and the reference separation distance h ref =1mm, S2 is corrected to obtain S2'=S2×h ref / h0.

[0013] S4. Mathematical model analysis:

[0014] S4.1. Material identification: calculate the material deviation rate η Cu , η Al of the cable to be detected and the standard copper or aluminum test block; if η Cu ≤5% is determined as copper material, if η Al ≤5% is determined as aluminum material, otherwise it is determined as abnormal material.

[0015] S4.2. Wire diameter measurement: a linear regression model d ref =a・S2+b of the standard wire S2 and wire diameter d is fitted by using the least square method, combined with σ T to correct the skin depth, and the wire diameter d=d ref ×√(σ T / σref ), wherein σ ref is the standard conductivity of the cable at 20℃, and a and b are regression coefficients;

[0016] S5. Progressive output and abnormality processing: first output the material category and the deviation rate, then output the wire diameter value and the precision if the material is qualified, and finally output the comprehensive qualification determination, if the data is abnormal, the material is abnormal, or the wire diameter is out of tolerance, the corresponding alarm mechanism is triggered.

[0017] Technical effects and advantages of the present application:

[0018] 1. The present application designs a dual-frequency excitation scheme of low-frequency focusing material and high-frequency focusing wire diameter: the low frequency is set to make the eddy current skin depth ≤ cable wire diameter d / 3, to ensure that the impedance change is mainly dominated by the material; the high frequency is set to make the skin depth ≥ d / 3, to ensure that the impedance change is mainly dominated by the wire diameter, effectively avoiding the problem of misjudgment and missed detection;

[0019] 2. The present application constructs a full-link compensation mechanism of sensor coil, cable core wire and skin depth, which adopts copper wire and polytetrafluoroethylene skeleton temperature coefficient compensation for the sensor coil resistance / inductance, and adopts material-specific temperature coefficient compensation for the cable core wire conductivity, and combines the compensated conductivity to correct the high-frequency skin depth, further eliminating the influence of temperature on wire diameter measurement, without the need for additional constant temperature device, and being suitable for complex temperature scenes such as workshops and outdoors;

[0020] 3. All detection logics of the present application are based on the explicit mathematical model derived from electromagnetic theory: the material identification adopts the standard block deviation rate model, the wire diameter measurement adopts the linear regression and skin depth correction model, the model parameters can be directly calibrated by the standard block without sample training, reducing the calibration time and maintenance cost, at the same time, the error can be traced back through the formula, which is convenient for quickly locating the fault, and only needs to adjust the excitation frequency and standard parameters when adapting to different specifications of cables, the flexibility is significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 is a schematic diagram of the overall process of the present application;

[0022] Figure 2 is a step breakdown diagram of the present application. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0024] As shown in the drawings:Figure 1 The cable material diameter detection method based on the eddy current impedance effect comprises the following steps:

[0025] S1. Data acquisition: using a data acquisition device to synchronously acquire the resistance R 10 , inductance L 10 of the coil under a low-frequency excitation signal f1, the resistance R 20 , inductance L 20 of the coil under a high-frequency excitation signal f2, and the environmental temperature T0 and the lift-off distance h0 between the sensor and the cable, wherein the f1 is set to 500 kHz-1 MHz according to the cable material, so that the eddy current skin depth δ1≤cable diameter d / 3; and the f2 is set to 3 MHz-5 MHz according to the cable diameter, so that the eddy current skin depth δ2≥cable diameter d / 3.

[0026] The data acquisition device comprises a dual-frequency excitation eddy current sensor, a dual-frequency excitation module, an impedance acquisition module, an environmental parameter acquisition module, and a data processing unit.

[0027] It should be further explained that the dual-frequency excitation eddy current sensor adopts a solenoid structure, the coil is in a dense winding form, and the pitch is controlled to be 0.1 mm; the wire is selected to be oxygen-free copper enameled wire, the wire diameter is 0.15 mm, the insulating layer is made of polyimide material to ensure the high-temperature resistance and insulation performance, the number of turns of the coil is adjusted according to the diameter of the measured cable, the number of turns of the coil corresponding to the cable with a diameter ≤1 mm is 60 turns, and the number of turns of the coil corresponding to the cable with a diameter >1 mm is 30 turns; the inner diameter of the coil is 2 times the maximum diameter of the measured cable, and the length of the coil is 5 times the inner diameter of the coil. The sensor is built-in with a magnetic core, the magnetic core is made of Mn-Zn ferrite, the initial permeability μ i =1000, the diameter of the magnetic core is 0.8 times the inner diameter of the coil, the inductance range of the coil in the no-load state is 100 μH-500 μH, the Q value under the frequency of 1 MHz is not less than 50, and the lift-off distance measurement accuracy can reach ±0.01 mm.

[0028] The core chip of the dual-frequency excitation module adopts an AD9850 DDS signal generator, which has 2 independent output channels and is used for outputting low-frequency f1 and high-frequency f2 signals respectively; the module power supply is ±12 V DC, the power supply ripple is not more than 1%, the output frequency stability of the module can reach ±1 ppm / ℃, the output current is 50 mA±5 mA, and the distortion of the output sine wave is not more than 0.5%.

[0029] The core chip of the impedance acquisition module is AD5933 impedance converter, the measurement range covers resistance 0-100 Ω, inductance 0-1000 μH; the sampling rate is adjusted according to the detection scene, the sampling rate is 1 kHz in static detection, and the sampling rate is increased to 10 kHz in dynamic online detection. The measurement accuracy of the module is resistance ±0.05% FS, inductance ±0.1% FS, and the signal conversion time is not more than 1 ms.

[0030] The environmental parameter acquisition module includes a temperature sensor and a lift-off distance sensor. The temperature sensor is a DS18B20 digital sensor with a measurement accuracy of ±0.05°C. The lift-off distance sensor is a Keyence IL-300 laser displacement sensor with a measurement range of 0-10 mm and a measurement accuracy of ±0.01 mm.

[0031] The data processing unit uses an STM32H743 embedded CPU with a main frequency of 480 MHz to ensure real-time data processing. The storage unit is equipped with a ≥32GB SD card that supports real-time data storage and can cache 1000 groups of historical data for subsequent tracing and analysis.

[0032] The collection operation steps are as follows:

[0033] During the sensor installation and calibration process, the operation method needs to be adjusted according to the differences in the detection scene. In the online detection scene, the sensor is fixed on the production line detection station, and the coaxiality of the coil center and the cable conveying axis is calibrated by the laser displacement sensor to ensure that the deviation is not more than 0.1 mm.

[0034] At the same time, the lift-off distance between the sensor and the surface of the cable is set to 1 mm, which is achieved by fine adjustment with a micrometer, with an error of ±0.01 mm or less. In the static sampling scene, the cable is fixed on a V-shaped clamp made of polytetrafluoroethylene material to avoid introducing electromagnetic interference, and the sensor is moved to make the cable pass through the coil, keeping the lift-off distance constant at 1 mm. The excitation parameter setting needs to be combined with the cable characteristics. The low frequency f1 is determined according to the cable material. For copper / aluminum core cable, f1 is taken as 800 kHz ± 50 kHz. At this time, the skin depth δ1 of copper is about 0.2 mm, and the δ1 of aluminum is about 0.25 mm, which can ensure that δ1 ≤ d / 3, so that the material has a dominant influence on the impedance change.

[0035] The high frequency f2 is determined according to the cable diameter. When the wire diameter is ≤1 mm, f2 is taken as 3 MHz ± 100 kHz, and when the wire diameter is >1 mm, f2 is taken as 5 MHz ± 100 kHz. At this time, the δ2 of copper is about 0.08 mm, and the δ2 of aluminum is about 0.1 mm, which can ensure that δ2 ≥ d / 3, so that the wire diameter has a dominant influence on the impedance change.

[0036] During data acquisition, firstly, the excitation module is started, and the f1 and f2 signals are synchronously output; then, the impedance acquisition module continuously acquires 5 groups of data, each group of data containing the resistance R1 and the inductance L1 under the low frequency f1, and the resistance R2 and the inductance L2 under the high frequency f2; and the environmental parameter module synchronously acquires the environmental temperature T and the lift-off distance h;

[0037] During the acquisition process, the effectiveness of each group of data needs to be judged, if the lift-off distance h deviates more than 0.02 mm, it is determined as invalid data and reacquired; finally, the average value of the effective data is taken as the original data, denoted as [R 10 ,L 10 ,R 20 ,L 20 ,T0,h0].

[0038] S2. Data preprocessing:

[0039] S2.1. Noise suppression: for the common 50Hz power frequency interference, sensor thermal noise and cable vibration noise in the industrial scene, a combination filtering scheme of sliding average filtering and wavelet threshold filtering is adopted, and noise suppression is realized in two steps. The first step is sliding average filtering, for the original impedance data sequence {x n} (wherein n=1 to 5, corresponding to 5 groups of acquisition data), the window size k=2 (i.e. 3-point sliding average) is selected, and the calculation formula is: , wherein when n=1, x1=x2 is taken; when n=5, x5=x4 is taken, and the filtered data R 11 , L 11 , R 21 , L 21 are obtained through the calculation.

[0040] The second step is wavelet threshold filtering, the wavelet base db4 wavelet is selected, which takes into account the time domain locality and frequency domain resolution, and is suitable for non-stationary eddy current signal processing; the decomposition layer is set to 3 layers, after decomposition, the high frequency coefficients correspond to the noise signal, and the low frequency coefficients correspond to the effective signal; the threshold value is calculated by Stein unbiased risk estimation threshold, and the formula is (wherein N is the data length, and herein N=5); the soft threshold processing is adopted for the high frequency coefficients, and the processing formula is: w'=sign(w)⋅max(|w|-λ,0), and finally the denoised data R 12 , L 12 , R 22 , L 22 are obtained through wavelet reconstruction.

[0041] S2.2. Full-link temperature compensation: both the sensor coil and the cable core are considered, and two steps are executed to ensure the completeness of the compensation;

[0042] The first step is sensor coil temperature compensation, including resistance temperature compensation and inductance temperature compensation. The resistance temperature compensation is based on the temperature coefficient a of copper wire Cu =0.00393 / ℃ (according to GB / T3956-2008 standard), and the calculation formula is: The inductance temperature compensation is based on the thermal expansion coefficient a of the coil skeleton (polytetrafluoroethylene) PTFE =1.2×10 -4 / ℃, and the calculation formula is: Where T ref =20℃ is the reference temperature. Similarly, the temperature compensation is performed on R 22 and L 22 at high frequency to obtain the compensated R t2 and L t2 .

[0043] The second step is cable core temperature compensation (pre-compensation, preparing for subsequent wire diameter calculation), based on the electrical conductivity temperature coefficient g of the cable material, and the calculation formula is: s T = s ref ×(1−g(T0−T ref )), where s ref is the standard electrical conductivity of the cable material at 20℃ (according to GB / T351-2021, copper s Curef =58×10 6 S / m, and aluminum s Alref =38×10 6 S / m); g is the electrical conductivity temperature coefficient of the cable material (copper g Cu =0.00393 / ℃, and aluminum g Al =0.00429 / ℃).

[0044] S3. Parameter decoupling extraction:

[0045] S3.1. Material sensitivity factor extraction: at low frequency f1, the eddy current skin depth d1 is large, and the calculation formula is: (where m0=4p×10⁻ 7 H / m is the vacuum permeability), at this time the eddy current penetration depth can cover the cable core, and the impedance change is mainly determined by the relative permeability m of the cable material r and the electrical conductivity s, and the influence of the wire diameter on the impedance change can be ignored (experimental verification, when d≥3d1, the contribution of the wire diameter to the impedance change is not more than 2%).

[0046] First, calculate the coil impedance change. The coil impedance reference value (pre-calibrated) without cable is: Z 01 =√(R 01 ²+(2p f1L 01 )²); and the coil impedance with cable is: Zt1 =√(R t1 ²+(2πf1L t1 )²); and thus the rate of change of impedance is obtained: .

[0047] Subsequently, the material sensitivity factor S1 was derived. Through Maxwell's equations, it can be seen that at low frequencies, ΔZ1r and μ... r The relationship between σ and ΔZ1r is: ΔZ1r = −K1μ r •σ, where K1 is the coil structure constant, which is only related to the coil parameters and can be obtained through calibration using a standard test block. Its calculation formula is: (N is the number of coil turns, D is the inner diameter of the coil, and l is the length of the coil). Rearranging the above formula, we obtain the material sensitivity factor: S1 = μ r •σ=−K1ΔZ1r;

[0048] S3.2. Extraction of Line Diameter Sensitive Factor: At high frequency f2, the eddy current skin depth δ2 is relatively small, and the calculation formula is as follows: At this point, the eddy currents are concentrated only on the cable surface, and the impedance change is mainly determined by the cable diameter d (experiments show that when d ≤ 3δ², the material's contribution to the impedance change does not exceed 1.5%). First, define the cable diameter sensitivity factor S². The coil impedance reference value (pre-calibrated) without cable is: Z 02 =√(R 02 ²+(2πf2L 02 )²); When there is a cable, the coil impedance is: Z t2 =√(R t2 ²+(2πf2L t2 )²); The wire diameter sensitivity factor is the impedance change: S² = ΔZ² = Z t2 -Z 02 (Unit: Ω) If the actual lift-off distance h0 is different from the reference lift-off distance h ref If there is a deviation of 1mm, S2 needs to be corrected. The correction formula is: S2' = S2 × h ref / h0. Experimental verification shows that for every 0.1 mm change in the lift-off distance, S2 changes by approximately 5%. This correction can reduce the lift-off error to within 0.5%, ensuring the accuracy of the wire diameter sensitivity factor.

[0049] S4. Mathematical Model Analysis:

[0050] S4.1. Material Identification: First, standard material sensitivity factor calibration is performed. Prepare pure copper and pure aluminum test blocks, with the pure copper block having a purity of no less than 99.99% and the pure aluminum block no less than 99.95%. Both blocks have a diameter of 5mm and a length of 20mm. The test blocks are then tested at a reference temperature T. ref =20℃, reference lift-off distance h ref=1mm, the material sensitivity factor S1 of two standard test blocks is measured, and the standard sensitivity factor S 1Cu of copper and the standard sensitivity factor S 1Al of aluminum are obtained respectively. S1 To ensure the calibration accuracy, the standard deviation σ S1 =0.5×10 6 S·H / m) is calculated by repeating the measurement 20 times for each material.

[0051] Then the material deviation rate of the cable to be detected is calculated, and the copper material deviation rate is defined as: η Cu =|(S1-S 1Cu ) / S 1Cu |×100%, and the aluminum material deviation rate is defined as: η Al =|(S1-S 1Al ) / S 1Al |×100%.

[0052] Finally, the discrimination rule is established: if η Cu ≤2σ S1 / S 1Cu ×100% (usually the threshold value is not more than 5%), the cable is determined to be copper material; if η Al ≤2σ S1 / S 1Al ×100% (usually the threshold value is not more than 5%), the cable is determined to be aluminum material; if η Cu and η Al both exceed 5%, further calculation of the deviation of S1 and the standard value of impurity material (such as copper alloy, aluminum alloy) is performed, and if both still exceed the deviation, the cable is determined to be abnormal material (containing impurities or non-target material).

[0053] S4.2. Wire diameter measurement: five standard wires with known diameters are selected as calibration samples, and the diameters of the standard wires need to cover 10%, 30%, 50%, 70%, and 90% of the range of the wire diameter to be detected — for example, when the range of the wire diameter to be detected is 0.5mm to 2mm, the diameters of the standard wires are selected as 0.5mm, 0.9mm, 1.2mm, 1.6mm, and 2.0mm. Under the conditions of reference temperature T ref =20℃ and reference lifting distance h ref =1mm, the wire diameter sensitivity factor S2 of each standard wire is measured, and five sets of calibration data pairs (S 201 ,d 01 ) to (S 205 ,d 05 ) are obtained. Linear regression fitting is performed on the calibration data by using the least squares method, and the goodness of fit R²≥0.999 is required, and finally the wire diameter calculation formula d ref=a・S2+b, where a is the slope (unit: mm / Ω) and b is the intercept (unit: mm), for example, a=0.002 mm / Ω and b=0.05 mm can be obtained by calibration.

[0054] The second stage is the skin depth correction (temperature compensation). The high-frequency line diameter sensitivity factor S2 is related to the skin depth δ2, and δ2 is affected by the cable conductivity σ T , so the d ref obtained by linear regression needs to be corrected. The correction formula is: d=d ref ×δ ref / δ T , where δ ref is the skin depth at the reference temperature T ref , and the calculation formula is: ; δ T is the skin depth at the actual temperature T0, and the calculation formula is: Substitute the calculation formula of σ T into it, and after simplification, we get: Through this correction, the influence of temperature on the cable conductivity can be eliminated, and the line diameter measurement accuracy can be ensured.

[0055] S5. Progressive output and abnormality processing: first output the material category and the deviation rate, then output the line diameter value and the accuracy if the material is qualified, and finally output the comprehensive qualification determination. If the data is abnormal, the material is abnormal, or the line diameter is out of tolerance, the corresponding alarm mechanism will be triggered.

[0056] For abnormal situations that may occur during the detection process, a three-level abnormality processing mechanism is established.

[0057] The first is data abnormality processing. If the standard deviation of the collected data exceeds 0.5% (for example, the standard deviation of 5 groups of R 10 data exceeds 0.06Ω), the system will automatically re-collect data for 3 times. If the data is still abnormal after 3 times of re-collection, the system will output a prompt that the detection has failed (the data is unstable) and suggest checking the sensor connection state to eliminate hardware faults.

[0058] The second is material abnormality processing. When it is determined that the cable is abnormal in material, the system outputs a warning message that the material is unqualified (not copper / aluminum or contains impurities), and triggers a red audible and visual alarm with a frequency of 1 kHz and a duty cycle of 50%. This strong warning signal reminds the detection personnel to handle it in time.

[0059] The third is line diameter out-of-tolerance processing. When the line diameter value exceeds the pre-set qualified range, the system outputs a prompt that the line diameter is out of tolerance. In the online detection scene, the system outputs a 24V switching quantity signal at the same time, which links the production line to stop to avoid continuous production of unqualified products. In the offline sampling inspection scene, a yellow audible and visual alarm is triggered to prompt the detection personnel to mark the cable as unqualified.

[0060] Finally: the above only for the preferred embodiments of the present application, and not for limiting the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application, should be included in the scope of protection of the present application.

Claims

1. A method for detecting the wire diameter of cable material based on the eddy current impedance effect, characterized in that, include: S1. Data Acquisition: A dual-frequency excitation eddy current sensor is used to synchronously acquire the coil resistance R under the low-frequency excitation signal f1. 10 Inductor L 10 The resistance R of the coil under the high-frequency excitation signal f2 20 Inductor L 20 The ambient temperature T0 and the lift-off distance h0 between the sensor and the cable are also considered. The f1 is set to 500kHz~1MHz according to the cable material, so that the eddy current skin depth δ1 ≤ cable diameter d / 3; the f2 is set to 3MHz~5MHz according to the cable diameter, so that the eddy current skin depth δ2 ≥ cable diameter d / 3. S2. Data Preprocessing: S2.

1. Noise Suppression: For the acquired R 10 L 10 R 20 L 20 The denoised data R is obtained by using a combination of moving average filtering and wavelet threshold filtering algorithms. 12 L 12 R 22 L 22 ; S2.

2. End-to-end temperature compensation: Perform resistance temperature compensation and inductance temperature compensation on the sensor coil to obtain the compensated resistance R. t1 R t2 and inductor L t1 L t2 Simultaneously, conductivity temperature compensation is performed on the cable core wires to obtain the compensated conductivity σ. T ; S3. Parameter decoupling and extraction: S3.

1. Material Sensitive Factor Extraction: Based on R under low-frequency f1 t1 L t1 Calculate the coil impedance change rate ΔZ1r, and combine it with the coil structure constant K1 to extract the material sensitivity factor S1=μ. r ・σ, where μ r σ is the relative permeability of the cable, and σ is the electrical conductivity of the cable; S3.

2. Extraction of wire diameter sensitive factor: Based on R at high frequency f2 t2 L t2 Calculate the coil impedance change ΔZ2, and define the wire diameter sensitivity factor S2 = ΔZ2; if h0 is separated from the reference distance h ref =1mm has a deviation, so S2 is corrected to get S2'=S2×h ref / h0; S4. Mathematical Model Analysis: S4.

1. Material Identification: Calculate the material deviation rate η between the cable to be tested and the standard copper or aluminum test block. Cu η Al If η Cu ≤5% is considered copper material, if η Al If the content is ≤5%, it is determined to be aluminum; otherwise, it is determined to be a material abnormality. S4.

2. Wire Diameter Measurement: A linear regression model of S2 and wire diameter d for the standard conductor is fitted using the least squares method. ref =a・S2+b, combined with σ T By adjusting the skin depth, we obtain the wire diameter d = d ref ×√(σ T / σ ref ), where σ ref is the standard conductivity of the cable at 20℃, and a and b are regression coefficients; S5. Progressive output and anomaly handling: First, output the material type and deviation rate. If the material is qualified, output the wire diameter value and accuracy. Finally, output the comprehensive qualification judgment. If the data is abnormal, the material is abnormal, or the wire diameter is out of tolerance, trigger the corresponding alarm mechanism. A three-level anomaly handling mechanism was established to address any abnormalities that occurred during the testing process.

2. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The dual-frequency excitation eddy current sensor uses a solenoid-type tightly wound coil. For cables with a wire diameter ≤ 1mm, the coil has 60 turns; for cables with a wire diameter > 1mm, the coil has 30 turns. The inner diameter of the coil is twice the maximum diameter of the cable being measured, and the coil length is five times the inner diameter. The coil conductor is 0.15mm oxygen-free copper enameled wire, with an internal Mn-Zn ferrite core and an initial permeability μ. i =1000.

3. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The window size of the moving average filter is k=2, and the calculation formula is as follows: , where x n The original data sequence is n=1~5; the wavelet threshold filtering uses the db4 wavelet basis, with a decomposition level of 3. The threshold is calculated using Stein unbiased risk estimation, and the formula is as follows: N is the data length, and soft thresholding is used for high-frequency coefficients w′=sign(w)⋅max(∣w∣−λ,0).

4. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The formula for resistance temperature compensation is as follows: The formula for inductor temperature compensation is: , where α Cu =0.00393 / ℃ is the temperature coefficient of copper conductor, α PTFE =1.2×10 -4 / ℃ is the coefficient of thermal expansion of the polytetrafluoroethylene skeleton, T ref =20℃ is the reference temperature; the conductivity temperature compensation formula is σ T =σ ref ×(1−γ(T0−T ref Copper core cable γ Cu =0.00393 / ℃, γ of aluminum core cable Al =0.00429 / ℃.

5. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The formula for calculating the coil impedance change rate ΔZ1r is as follows: Z 01 =√(R 01 ²+(2πf1L 01 Z²) represents the reference value of the coil impedance when there is no cable. t1 =√(R t1 ²+(2πf1L t1 )²) represents the coil impedance when a cable is present, and the formula for calculating the coil structure constant K1 is as follows: N is the number of turns in the coil, D is the inner diameter of the coil, l is the length of the coil, and μ0 = 4π × 10⁻⁶. -7 H / m is the permeability of free space.

6. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The formula for calculating the wire diameter sensitivity factor S2 is S2=Z t2 -Z 02 Z 02 =√(R 02 ²+(2πf2L 02 Z²) represents the reference impedance value of the high-frequency coil without cables. t2 =√(R t2 ²+(2πf2L t2 )²) represents the high-frequency coil resistance when there is a cable.

7. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The material deviation rate η Cu =|(S1-S 1Cu ) / S 1Cu |×100%, η Al =|(S1-S 1Al ) / S 1Al |×100%, where S 1Cu S 1Al 20℃ and h respectively ref Material sensitivity factors of standard copper and aluminum test blocks under the condition of 1mm, with the purity of copper ≥99.99% and aluminum ≥99.95% respectively, and the size of the test blocks being 5mm in diameter and 20mm in length.

8. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The goodness of fit of the linear regression model is R² ≥ 0.

999. The calibration samples consist of five standard wires of known diameters, covering 10%, 30%, 50%, 70%, and 90% of the wire diameter range to be tested; the σ ref The standard conductivity of copper at 20℃ is σ. Curef =58×10 6 S / m, aluminum σ Alref =38×10 6 S / m.

9. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The progressive output includes three levels: the first level outputs the material type, deviation rate, and confidence level (confidence level = 1 - deviation rate); the second level outputs the wire diameter, measurement accuracy ±0.003mm, and temperature correction; and the third level outputs a comprehensive pass / fail judgment, where the wire diameter is within [d]. min ,d max If the range is within the acceptable range, it is considered acceptable; otherwise, it is considered out of tolerance.

10. The method for detecting cable material diameter based on eddy current impedance effect according to claim 1, characterized in that: The anomaly handling mechanism includes: automatically re-collecting data 3 times when data is abnormal, and outputting a detection failure if the data is still abnormal; triggering a red audible and visual alarm when the material is abnormal; and linking the production line to stop when the wire diameter is out of tolerance in online scenarios and triggering a yellow audible and visual alarm in offline scenarios.

Images