Self-adaptive pulsed eddy current corrosion defect detection method and detection device

Abstract

The invention discloses a self-adaptive pulsed eddy current corrosion defect detection method and a detection device, and belongs to the technical field of electromagnetic nondestructive detection. The method comprises the following steps: respectively collecting pulse eddy current signals of an air domain and a measured piece, and constructing a differential signal after filtering and logarithmic transformation; based on the peak moment of the differential signal, the slope ratio of the ascending section to the descending section and the initial change time difference, adaptively determining the optimal fitting interval of the later slope; performing linear fitting on the signal of the measured piece in the interval, and extracting a later slope characteristic value; and calculating the residual wall thickness of the measured position through a formula by utilizing the known thickness and the corresponding slope of the calibrated position, so as to realize quantitative evaluation of the corrosion defect. According to the method, the problems that in the prior art, a fitting interval depends on a fixed threshold value and the adaptability is poor are solved, different detection conditions can be automatically adapted, the stability of feature extraction and the reliability of a detection result are remarkably improved, and the method is particularly suitable for on-line corrosion detection of a ferromagnetic component with a coating layer.

Description

Technical Field

[0001] This invention relates to the field of electromagnetic nondestructive testing technology, and in particular to an adaptive pulsed eddy current corrosion defect detection method and device. Background Technology

[0002] Pulsed eddy current (PEC) testing is a typical time-domain electromagnetic nondestructive testing method. By analyzing the attenuation process of eddy currents in the test piece after rapid changes in excitation current, it enables the detection of changes in component wall thickness and corrosion defects. Compared with traditional sinusoidal excitation eddy current testing methods, PEC testing has advantages such as a wide excitation spectrum, strong penetration capability, and insensitivity to surface conditions. It can perform online testing of magnetic pressure-bearing equipment without removing the cladding or performing surface treatment, thus showing broad application prospects in the fields of petrochemical, energy, and special equipment safety testing.

[0003] Pulsed eddy current testing signals typically exhibit a nonlinear voltage response curve that decays over time. Numerous studies have shown that information closely related to material wall thickness and corrosion state is concentrated in the later stages of signal decay, particularly the near-linear decay characteristics observed in logarithmic coordinates. The mechanism lies in the fact that as material thinning occurs, its equivalent wall thickness decreases, accelerating the eddy current decay process and leading to a larger slope in the later stages. Therefore, defects are manifested as differences in electromagnetic response caused by changes in remaining wall thickness, rather than directly presenting geometric morphology. Consequently, the slope of the later-stage signal has gradually become one of the commonly used characteristic quantities in pulsed eddy current testing, capable of characterizing changes in wall thickness or the degree of corrosion thinning in ferromagnetic components.

[0004] In existing technologies, some invention patents have attempted to extract the slope as a detection feature by linearly fitting the later attenuation segment of the pulsed eddy current detection signal. For example, invention patent CN104266579B proposes a method for least-squares fitting of the signal segment of the induced voltage measurement curve after 0.1 times the characteristic attenuation time in a semi-logarithmic coordinate system, and uses the slope of the fitted line as a detection feature for measuring the wall thickness of ferromagnetic components. This method, to some extent, utilizes the linear attenuation characteristics of the later signal, providing a new approach for feature extraction of pulsed eddy current signals.

[0005] However, in practical engineering applications, this method still has certain limitations: on the one hand, the “feature decay time” it uses lacks a unified and clear basis for determination, and the selection of the fitting start point depends on empirical judgment, making it difficult to maintain consistency under different detection objects and detection conditions; on the other hand, this method does not clearly define the fitting termination interval. When the detection signal further decays and tends to flatten or is affected by noise or system drift in the later stage, continuing to participate in the fitting is prone to introducing errors, thereby affecting the stability and reliability of the slope feature.

[0006] Furthermore, existing technologies have proposed determining late-signal intervals through time-window analysis. For example, invention patent CN112946065B discloses a pulse eddy current detection method based on the slope of late-signal signals. This method performs logarithmic transformation and normalization on the detected signal, selects a fixed number of late-signal time windows based on the difference between adjacent time windows, and performs linear fitting on the signal within this interval to obtain slope characteristic values. This method, to some extent, avoids directly using the characteristic decay time as the fitting starting point.

[0007] However, the determination of the late signal interval in this type of method still relies on a pre-set fixed time window rule, and the length and position of the fitting interval are not dynamically adjusted with the overall changes in signal characteristics. When the material, wall thickness range, coating thickness, or signal noise level of the detected object changes, the fixed time window rule cannot guarantee that the fitting interval is always in the effective linear decay range of the signal, which may still lead to unstable slope feature extraction results, and to some extent depends on the selection of manual empirical parameters.

[0008] In summary, existing feature extraction methods based on the slope of the later signal in pulsed eddy current detection generally suffer from problems such as fixed rules for selecting the fitting interval, insufficient adaptability to signal changes, and strong reliance on human experience. There is a lack of a pulsed eddy current detection method that can automatically and continuously determine the fitting interval of the later slope based on the signal's own variation characteristics, while balancing fitting accuracy and stability. Therefore, it is necessary to propose a new adaptive fitting method for the later signal slope to improve the reliability and engineering applicability of feature extraction in pulsed eddy current detection. Summary of the Invention

[0009] Therefore, the purpose of this invention is to provide an adaptive pulsed eddy current corrosion defect detection method. By analyzing the variation characteristics of the pulsed eddy current detection signal and its differential signal, the starting point and ending point of the later slope fitting interval are automatically determined. Without the need to manually set the endpoints of the fitting interval, the method can achieve stable and reliable extraction of the slope feature values ​​of the later signal, thereby improving the accuracy and engineering applicability of wall thickness and corrosion defect assessment in pulsed eddy current detection.

[0010] To achieve the above objectives, the present invention provides an adaptive pulsed eddy current corrosion defect detection method, comprising the following steps: S1. Acquire data; including acquiring the detection signal of the pulse eddy current sensor when it is placed in air without load, as the reference signal in the air domain; and acquiring the detection signal of the test device when the pulse eddy current sensor is placed above the test device. S2. After preprocessing the air domain reference signal and the test device detection signal, differential calculation is performed. The preprocessing includes filtering and logarithmic transformation. S3. Plot the differential signal as a curve, and mark the coordinates of the plotted curve when it reaches the maximum peak value as the starting fitting point index1. S4. Starting from the initial fitting point index1, adaptively determine the fitting endpoint index2 based on the variation characteristics of the differential signal, and use the initial fitting point index1 and the fitting endpoint index2 as the boundary of the subsequent slope fitting interval. S5. In the preprocessed test component detection signal, extract the signal segment within the slope fitting interval and perform linear fitting to obtain the subsequent slope feature value. S6. Calculate the corrosion thinning amount of the tested part based on the later slope characteristic value, and determine the defect location.

[0011] More preferably, in S2, the logarithmic transformation adopts the following formula: in, The original signal, This is the transformed signal.

[0012] More preferably, in S2, the data of the air domain reference signal after preprocessing is represented as follows: The data after preprocessing the detection signal of the test piece is represented as The difference calculation is expressed by the following formula: ; in, This represents the differential signal between the air domain reference signal and the detection signal of the device under test.

[0013] More preferably, in S4, the initial fitting point index1 and the fitting endpoint index2 are used as the boundaries of the later slope fitting interval, including the following steps: S401. Using the initial fitting point index1 as the boundary, divide the differential signal into rising segments. and falling segment signal ; S402. Using the sliding slope method to analyze the rising segment of the differential signal. and falling segment signal Perform local least squares linear fitting separately; S403, Calculate the rising segment signal maximum slope With falling segment signal minimum slope ; S404, Calculate the maximum slope With minimum slope The absolute value of the ratio M; S405, for rising segment signals Find the second derivative to obtain the rising segment signal. The coordinates at which the maximum value is obtained are denoted as the starting point of change, index. 0； S406. Calculate the coordinate difference ΔT between the initial change point index0 and the initial fitting point index1; S407. Based on the coordinate difference ΔT, the absolute value of the ratio M, and the initial fitting point index... 1， Calculate the fitting endpoint The coordinates are obtained; the slope fitting interval is denoted as [ , ].

[0014] More preferably, in S407, the fitting endpoint The coordinates are calculated using the following formula: ; in, The coordinates of the fitting endpoint are given, where ΔT is the coordinate difference and M is the absolute value of the ratio.

[0015] More preferably, in S5, from the preprocessed test component detection signal, a signal segment within the slope fitting interval is extracted, and linear fitting is performed to obtain the subsequent slope feature value, including: Preprocessed test piece detection signal In the middle, the slope fitting interval is truncated. , The data within [ ] were linearly fitted using the least squares method, and the slope of the fitted line was used as the slope feature value for the later stage. ; ; Where slope is the slope function. This is the fitted curve of the detection signal of the test piece after preprocessing.

[0016] More preferably, in S6, calculating the corrosion thinning amount of the tested part based on the later slope characteristic value and determining the defect location includes: A calibration position is set at an uncorroded area of ​​the test piece, and the slope characteristic value of the calibration position is obtained later. ; Obtain the measured slope feature value at the detection location in the later stage. ; Based on the known thickness at the calibration location Calculate the thickness at the measured location using the following formula. : Calculate the thickness at the measured location. .

[0017] The present invention also provides an adaptive pulsed eddy current corrosion defect detection device for performing the above-described adaptive pulsed eddy current corrosion defect detection method, comprising: The signal excitation and acquisition module is used to generate a pulsed excitation magnetic field and acquire the induced voltage signal. The signal preprocessing and feature extraction module is used to process the acquired signal and extract the slope feature value in the later stage; The thickness calculation and output module is used to calculate and output the wall thickness or corrosion assessment results based on the characteristic values.

[0018] More preferably, the signal excitation and acquisition module includes: A pulse signal generator is used to generate pulse currents with a set frequency and duty cycle. An excitation coil is connected to the pulse signal generator to generate an electromagnetic field, which excites the device under test to generate a pulsed eddy current field. The receiving coil is used to sense the change in the composite magnetic field formed by the electromagnetic field generated by the excitation coil and the secondary magnetic field generated by the test device in the pulsed eddy current field, and outputs the corresponding electrical signal. The data acquisition unit is used to synchronously acquire the voltage signal output by the receiving coil to obtain the air domain reference signal. V air With the detection signal of the test piece V test .

[0019] The adaptive pulsed eddy current corrosion defect detection method and device disclosed in this application have the following advantages compared with the prior art: 1. It can adaptively determine the slope fitting interval in the later stage, avoiding the uncertainty caused by fixed time thresholds or preset time window rules.

[0020] This invention automatically determines the start and end points of the slope fitting interval by analyzing the variation characteristics of the pulsed eddy current detection signal and its differential signal. It does not rely on the feature decay time or a fixed number of time windows, thus overcoming the problems of fuzzy selection of fitting interval and poor versatility in the prior art.

[0021] 2. Effectively improves the stability and repeatability of slope feature extraction in the later stage.

[0022] By determining continuous fitting intervals based on signal change characteristics, linear fitting always operates on the effective attenuation phase of the signal, avoiding the introduction of noise-enhanced sections or signal-flattening sections into the fitting process, thereby reducing the impact of noise, system drift, and changes in detection conditions on the slope feature value.

[0023] 3. Reduce human experience intervention and improve the objectivity and consistency of the testing process.

[0024] This invention eliminates the need for manual setting of fitting start and end points or adjustment of fitting interval parameters. The extraction process of slope feature values ​​can be completed automatically by the program, which is beneficial for obtaining consistent detection results under different operators, different detection objects, and different detection environments.

[0025] 4. Applicable to various testing conditions, with strong engineering applicability.

[0026] Since the fitting interval is determined based on the signal's own variation characteristics, rather than depending on the specific wall thickness range of the test piece, the material's electromagnetic parameters, or the sensor's lift-off height, this invention has good adaptability under different thicknesses and different coating layers.

[0027] 5. The method has a clear structure, low computational cost, and is easy to implement in engineering and integrate into systems.

[0028] The signal processing steps used in this invention are all based on conventional digital signal processing methods. The algorithm flow is clear, the computational complexity is low, and it is suitable for implementation in embedded devices or online detection systems. This is beneficial for the automation and intelligent application of pulse eddy current detection technology. Attached Figure Description

[0029] Figure 1 This is a schematic flowchart of the adaptive pulsed eddy current corrosion defect detection method provided by the present invention.

[0030] Figure 2 This is a schematic diagram of the system apparatus for the pulsed eddy current detection method of the present invention.

[0031] Figure 3 This is a schematic diagram of the time-domain curves of the air domain reference signal and the test device detection signal after preprocessing.

[0032] Figure 4 This is a schematic diagram of the differential signal curve constructed from the air domain reference signal and the test signal of the device under test.

[0033] Figure 5 This is a schematic diagram of key feature points for determining the slope fitting interval in the later stage based on the differential signal.

[0034] Figure 6A comparative schematic diagram showing the adaptive fitting results of the later slope obtained by the method of this invention for test pieces of different thicknesses, 9mm and 12mm. Detailed Implementation

[0035] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] The following detailed descriptions are illustrative and intended to further illustrate this application. Unless otherwise specified, the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. Exemplary embodiments according to this application will now be described in more detail with reference to the accompanying drawings. However, these exemplary embodiments may be implemented in many different forms and should not be construed as being limited to the embodiments set forth herein.

[0037] like Figure 1 As shown, the adaptive pulsed eddy current corrosion defect detection method provided by one embodiment of the present invention includes the following steps: S1. Acquire data; including acquiring the detection signal of the pulse eddy current sensor when it is placed in air without load, as the reference signal in the air domain; and acquiring the detection signal of the test device when the pulse eddy current sensor is placed above the test device. During the testing process, a pulsed excitation current acts on the excitation coil, generating a pulsed eddy current field in the test piece. The receiving coil senses the combined magnetic field changes formed by the electromagnetic field generated by the excitation coil and the secondary magnetic field generated by the test piece in the pulsed eddy current field, and outputs a corresponding electrical signal to form a pulsed eddy current detection signal. The test piece is set as a Q235 steel plate with a certain thickness of cladding layer on top. The excitation signal is a square wave signal with a frequency of 2Hz and a duty cycle of 50%. Figure 2 As shown.

[0038] S2. After preprocessing the air domain reference signal and the measured object detection signal, differential calculation is performed. The preprocessing includes filtering and logarithmic transformation. The sensor is placed in the air and above the measured object, respectively, to acquire the air domain signal and the measured object signal. The acquired signals are processed by a filter, and a logarithmic transformation is performed on the filtered signals to convert the original exponentially decaying time-domain signal into an approximately linear form for subsequent feature extraction and linear fitting. The transformation formula is: in, The original signal, This is the transformed signal.

[0039] After the above preprocessing and analysis of the original signals, the detection signals of the tested component are obtained respectively. and air domain detection signal ,like Figure 3As shown, the logarithmically transformed detection signal and the air domain reference signal are differentially processed to obtain the differential signal. ,like Figure 4 As shown.

[0040] S3. Perform curve fitting on the differential signal, and take the coordinates of the maximum peak value in the fitted curve as the starting fitting point index1; S4. Starting from the initial fitting point index1, adaptively determine the fitting endpoint index2 based on the variation characteristics of the difference signal, and use the initial fitting point index1 and the fitting endpoint index2 as the boundaries of the subsequent slope fitting interval; for example... Figure 5 The specific steps shown are as follows: S401. Using the initial fitting point index1 as the boundary, divide the differential signal into rising segments. and falling segment signal ; S402. Using the sliding slope method to analyze the rising segment of the differential signal. and falling segment signal Perform local least squares linear fitting separately; S403, Calculate the rising segment signal maximum slope With falling segment signal minimum slope ;in: Maximum slope of the ascending segment:

[0041] Minimum slope of the descending segment: Where t is the sampling time, used to characterize the moment when the induced voltage signal changes with time during pulse eddy current detection, specifically the time corresponding to each sampling point after pulse excitation. This indicates the solution for the rising segment signal. The first derivative with respect to time t; This indicates the solution for the falling segment signal. The first derivative with respect to time t; S404, Calculate the maximum slope With minimum slope The absolute value of the ratio M; ; S405, for rising segment signals Find the second derivative to obtain the rising segment signal. The coordinates at which the maximum value is obtained are denoted as the starting point of change, index0. ; S406. Calculate the coordinate difference ΔT between the initial change point index0 and the initial fitting point index1; ;

[0042] S407. Based on the coordinate difference ΔT, the absolute value of the ratio M, and the initial fitting point index... 1， Calculate the coordinates of the fitting endpoint The slope fitting interval is denoted as [ , ].

[0043] The above slope fitting interval is as follows: Figure 5 As shown.

[0044] S5. In the preprocessed test component detection signal, extract the signal segment within the slope fitting interval and perform linear fitting to obtain the subsequent slope feature value. Preprocessed test signal of the test piece In the middle, the slope fitting interval is truncated. , The data within [ ] were linearly fitted using the least squares method, and the slope of the fitted line was used as the slope feature value for the later stage. ; ; Where slope is the slope function. This is the fitted curve of the detection signal of the test piece after preprocessing.

[0045] S6. Calculate the corrosion thinning amount of the tested part based on the later slope characteristic value, and determine the defect location.

[0046] A calibration position is set at an uncorroded area of ​​the test piece, and the slope characteristic value of the calibration position is obtained later. ; Obtain the measured slope feature value at the detection location in the later stage. ; Based on the known thickness at the calibration location Calculate the thickness at the measured location using the following formula. : .

[0047] To verify the effectiveness of the method of this invention, pulsed eddy current testing experiments were conducted on steel plate test pieces of different thicknesses. For example, the cladding layer height was set to 5 mm for a 9 mm thick test piece and 30 mm for a 12 mm thick test piece. Detection signals were acquired under different cladding layer thickness conditions, and the slope fitting interval for the later stage was automatically determined using the method of this invention. The characteristic value slope was calculated based on the fitted straight line, such as... Figure 6 As shown. Under the above conditions, the slope of the 9mm test piece was measured to be -20.12, and the slope of the 12mm test piece was measured to be -11.28.

[0048] Using a 12mm thick test piece as the calibration test piece and a 9mm thick test piece as the test piece, the characteristic value slopes of -11.28 and -20.12 were obtained from the above measurements. Based on the published formula... Calculate the thickness at the measured location. The calculated thickness of the tested component was 8.98 mm, with a relative error of 0.22%. This allows for the assessment of corrosion thinning of the component.

[0049] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method of adaptive pulsed eddy current corrosion defect detection, characterized by, The method comprises the following steps: S1, acquiring acquisition data; The method comprises the following steps: S2, performing differential calculation on the air domain reference signal and the measured part detection signal after pre-processing, wherein the pre-processing comprises filtering and single logarithmic transformation; S3, performing curve tracing on the differential signal, and marking the coordinates when the maximum peak value of the tracing curve as the starting fitting point index1; S4, taking the starting fitting point index1 as the starting point, adaptively determining the fitting end point index2 according to the change characteristics of the differential signal, and taking the starting fitting point index1 and the fitting end point index2 as the boundaries of the later slope fitting interval; S5, in the pre-processed measured part detection signal, intercepting the signal segment in the slope fitting interval, and performing linear fitting to obtain the later slope characteristic value; S6, calculating the corrosion thickness reduction of the measured part based on the later slope characteristic value, and determining the defect position.

2. The adaptive pulsed eddy current corrosion defect detection method of claim 1, wherein, In S2, the single logarithmic transformation adopts the following formula: wherein is the original signal, is the transformed signal.

3. The method of adaptive pulsed eddy current corrosion defect detection of claim 1, wherein, In S2, the data of the air domain reference signal after preprocessing is represented as The data of the measured piece detection signal after preprocessing is represented as The difference calculation is represented by the following formula: ； wherein represents a difference signal between the air domain reference signal and the measured piece detection signal.

4. The method of adaptive pulsed eddy current corrosion defect detection of claim 3, wherein, In S4, taking the starting fitting point index1 and the fitting end point index2 as the boundaries of the later slope fitting interval, comprising the following steps: S401、with the starting fitting point index1 as the demarcation, the differential signal is divided into the rising section signal and the falling section signal ​ S402、adopt sliding slope method to the rising section signal of difference signal and falling section signal respectively carry out local least square linear fitting; S403、calculating the maximum slope of the rising section signal ;​​​ S404、calculating the maximum slope with the minimum slope absolute value M; S405、obtain the rising section signal Obtain the second derivative to obtain the rising section signal Obtain the coordinate when the maximum value is obtained, and mark it as the starting change point index 0； S406, calculating the coordinate difference ΔT of the starting change point index0 and the starting fitting point index1; S407、According to the coordinate difference value ΔT and the ratio absolute value M and the starting fitting point index 1， Calculate the fitting end point The coordinates of the fitting end point are obtained; the slope fitting interval is recorded as [ ,  ].

5. The method of adaptive pulsed eddy current corrosion defect detection of claim 4, wherein, In S407, the fitting end point The coordinates of the fitting end point are calculated using the following formula: ； wherein is the coordinate of the end of the fit, AT is the coordinate difference; M is the absolute value of the ratio.

6. The method of adaptive pulsed eddy current corrosion defect detection of claim 4, wherein, In S5, in the pre-processed measured part detection signal, intercepting the signal segment in the slope fitting interval, and performing linear fitting to obtain the later slope characteristic value, comprising: Preprocessed test piece detection signal In the middle, the slope fitting interval is truncated. , The data within [ ] were linearly fitted using the least squares method, and the slope of the fitted line was used as the slope feature value for the later stage. ; ； wherein, slope is a slope function, is a fitting curve of the detected signal of the measured piece after pre-processing.

7. The method of adaptive pulsed eddy current corrosion defect detection of claim 4, wherein, In S6, calculating the corrosion thickness reduction of the measured part based on the later slope characteristic value, and determining the defect position, comprising: Setting a calibration position at a non-corroded position of the measured piece, and obtaining a late-stage slope characteristic value at the calibration position ; obtaining a measured late-stage slope characteristic value at the to-be-detected position ; According to the known thickness of the calibration position ; the thickness of the measured position is calculated according to the formula : calculating a thickness of the measured position .

8. An adaptive pulsed eddy current corrosion defect detection apparatus, characterized by, An adaptive pulsed eddy current corrosion defect detection device for performing the adaptive pulsed eddy current corrosion defect detection method according to any one of claims 1 to 7, comprising: a signal excitation and acquisition module for generating a pulsed excitation magnetic field and acquiring an induced voltage signal; a signal preprocessing and feature extraction module for processing the acquired signal and extracting a later slope characteristic value; a thickness calculation and output module for calculating and outputting the wall thickness or corrosion evaluation result according to the characteristic value.

9. The adaptive pulsed eddy current corrosion defect detection device according to claim 8, wherein the signal excitation and acquisition module comprises: a pulse signal generator for generating a pulsed current with a set frequency and duty cycle; an excitation coil connected to the pulse signal generator for generating an electromagnetic field to excite the measured part to generate a pulsed eddy current field; a receiving coil for sensing the composite magnetic field change formed by the electromagnetic field generated by the excitation coil and the secondary magnetic field generated by the measured part in the pulsed eddy current field, and outputting the corresponding electric signal; A data acquisition unit is configured to synchronously acquire a voltage signal output by the receiving coil to obtain an air domain reference signal V air a detection signal of the measured piece V test .

Images