Method for detecting oxide skin in stainless steel tube by electromagnetic induction technology

Abstract

The invention relates to the technical field of metal pipe nondestructive testing, and discloses a method for detecting an oxide skin in a stainless steel pipe by an electromagnetic induction technology, which comprises the following steps: S1, generating multi-frequency excitation signals which comprise a high-frequency signal and a low-frequency signal which are used for respectively detecting the surface characteristics and deep distribution of the oxide skin in the stainless steel pipe; s2, applying the excitation signal to the stainless steel tube through an excitation coil, wherein the excitation coil adopts a multi-layer structure to ensure uniform distribution of a magnetic field; and S3, stainless steel pipe induction signals under the action of the excitation signals are collected, wherein the induction signals comprise response signals of the surface oxide skin and interference signals of the base body. According to the invention, the multi-frequency excitation technology is combined with signal decomposition and fusion processing, so that comprehensive detection of the thickness and distribution of the oxide skin is realized. Based on high sensitivity of high-frequency signals to surface characteristics and deep penetrability of low-frequency signals, detection results can accurately reflect characteristics and thickness distribution of the oxide skin at different depths.

Description

Technical Field

[0001] The invention relates to the technical field of nondestructive testing of metal pipes, in particular to a method for detecting oxide scale inside a stainless steel pipe using electromagnetic induction technology. Background Art

[0002] During the manufacturing and subsequent use of stainless steel pipes, an oxide layer is easily formed on the inner surface. The thickness and distribution characteristics of the oxide layer have a direct impact on the corrosion resistance, subsequent processing and final use performance of the product. In order to ensure product quality, it is necessary to quickly and accurately detect the oxide layer inside the stainless steel pipe on the industrial production line so as to adjust the manufacturing and processing technology in time, thereby improving product quality and production efficiency.

[0003] Existing stainless steel pipe internal oxide scale detection technologies are mainly based on single-frequency electromagnetic detection or optical detection. Under simpler production conditions, this type of technology can identify the oxide scale thickness within a certain range and achieve a basic assessment of the inner surface state of the pipe under specific frequency bands or optical conditions. This type of technology has practical value under low precision requirements and stable working conditions, which helps to optimize product quality and reduce unnecessary rework costs to a local extent.

[0004] However, when production requirements increase, the environment becomes more complex, or the characteristics of the oxide scale become more variable, single-frequency detection and limited signal analysis methods are difficult to meet the requirements of accuracy and comprehensiveness. Due to the lack of multi-frequency excitation, comprehensive data analysis, and real-time linkage with production parameters, the existing technology is limited in detection accuracy, range, and subsequent process dynamic optimization capabilities. It is difficult to provide all-round, high-precision oxide scale detection and dynamic process adjustment support for stainless steel pipe production under complex working conditions. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a method for detecting oxide scale inside stainless steel pipes using electromagnetic induction technology, which solves the problems in the prior art of low detection accuracy of oxide scale inside stainless steel pipes, incomplete detection results, inability to flexibly respond to complex working conditions and lack of real-time linkage optimization capabilities with production parameters.

[0006] To achieve the above purpose, the present invention is implemented by the following technical scheme: A method for detecting oxide scale in a stainless steel pipe by electromagnetic induction technology, comprising the following steps:

[0007] S1, generating a multi-frequency excitation signal, wherein the excitation signal includes a high-frequency signal and a low-frequency signal, which are used to respectively detect the surface characteristics and deep distribution of the oxide scale in the stainless steel pipe;

[0008] S2, applying the excitation signal to the stainless steel tube through an excitation coil, wherein the excitation coil adopts a multi-layer structure to ensure uniform distribution of the magnetic field;

[0009] S3, collecting the induction signal of the stainless steel pipe under the action of the excitation signal, wherein the induction signal includes the response signal of the surface oxide scale and the interference signal of the matrix;

[0010] S4, processing the sensing signal, including signal decomposition, noise suppression and signal fusion, to obtain a characteristic signal of the oxide scale;

[0011] S5. Based on the characteristic signal, predict the thickness and distribution of the oxide scale in the stainless steel pipe and generate an oxide scale distribution map;

[0012] S6. Output real-time alarm information and adjust production parameters according to the results of scale detection.

[0013] Preferably, the frequency range in S1 includes:

[0014] The high frequency band is 500kHz to 5MHz, which is used to capture the surface characteristics of the oxide scale;

[0015] The low frequency band is 50Hz~500kHz, which is used to penetrate the matrix to detect the distribution characteristics of deep oxide scale.

[0016] Preferably, the magnetic core material of the excitation coil in S2 is soft ferrite with a magnetic permeability higher than 800 to enhance the magnetic field strength and reduce magnetic flux leakage. The inner layer turns of the excitation coil are 100 to 500 turns, and the outer layer turns are 50 to 200 turns. The turns ratio is dynamically adjusted according to different signal frequency requirements.

[0017] Preferably, the signal acquisition in S3 comprises the following steps:

[0018] The sampling frequency is 2.5 times of the highest frequency signal to ensure the integrity and accuracy of the collected signal;

[0019] A broadband amplifier is used to pre-process weak signals, and the amplification factor is set at 20 to 100 times;

[0020] Synchronous sampling technology is used to avoid signal delay and distortion, and the dynamic range of the collected signal is -90dB to 10dB.

[0021] Preferably, the signal processing in S4 comprises the following steps:

[0022] The induced signals are decomposed into high-frequency surface signals and low-frequency deep-layer signals by short-time Fourier transform;

[0023] Adopting adaptive filter based on statistical characteristics to suppress noise;

[0024] The high-frequency and low-frequency signal features are integrated to generate a comprehensive characteristic signal related to the oxide scale thickness.

[0025] Preferably, in S5, the prediction of the oxide scale thickness is based on a mathematical model constructed based on a characteristic signal, wherein the input signal parameters include frequency response amplitude, phase change and harmonic characteristics, the model output is the thickness value of the oxide scale, and a two-dimensional thickness distribution map of the oxide scale is generated in combination with an image processing algorithm.

[0026] Preferably, in S5, the oxide scale thickness distribution map displays the thickness range of the oxide scale through a two-dimensional heat map, and at the same time, a three-dimensional modeling technology is combined to construct an oxide scale spatial distribution model to display the distribution characteristics of the oxide scale in the length direction and the circumferential direction.

[0027] Preferably, in S5, the electromagnetic properties of different tube wall thicknesses and oxide scale components are experimentally measured to establish a correction model for the matrix signal; during detection, the correction model is used to compensate the induced signal to eliminate interference from the matrix signal.

[0028] Preferably, in S6, the real-time alarm function includes:

[0029] When the detected oxide scale thickness exceeds the set threshold, visual and sound alarms are triggered;

[0030] Generate alarm report, record alarm location, thickness exceeding value and corresponding time;

[0031] Transmit alarm information to the production control center to achieve real-time adjustment of the production line.

[0032] Preferably, in S6, the oxide scale detection result is used to dynamically optimize the production process, including:

[0033] Adjust the heating temperature range to ±2°C of the target temperature;

[0034] Modify the cooling rate to control the rapid formation of oxide scale. The optimal range of cooling rate is between 10℃ / s and 50℃ / s.

[0035] Adjust the production line speed to the recommended range of the inspection data to ensure product consistency and surface quality.

[0036] The present invention provides a method for detecting oxide scale inside a stainless steel pipe using electromagnetic induction technology. It has the following beneficial effects:

[0037] 1. The present invention realizes comprehensive detection of oxide scale thickness and distribution through multi-frequency excitation technology combined with signal decomposition and fusion processing. Based on the high sensitivity of high-frequency signals to surface characteristics and the deep penetration of low-frequency signals, the detection results can accurately reflect the characteristics and thickness distribution of oxide scale at different depths.

[0038] 2. The present invention realizes intelligent regulation of heating temperature, cooling rate and production speed through a dynamic process optimization mechanism driven by real-time alarm and detection data, ensuring the uniformity of oxide thickness and consistency of product surface quality, while improving production efficiency and control accuracy.

[0039] 3. The present invention combines two-dimensional thermal maps with three-dimensional modeling technology to intuitively display the spatial distribution characteristics of oxide scale thickness, making it easy for operators to quickly locate high-risk areas and guide process improvements, greatly improving the efficiency of interpreting test results and quality control capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] Please refer to the attached Figure 1 The embodiment of the present invention provides a method for detecting oxide scale in a stainless steel pipe using electromagnetic induction technology, comprising the following steps:

[0043] S1. Generate a multi-frequency excitation signal, which includes a high-frequency signal and a low-frequency signal, to detect the surface characteristics and deep distribution of the oxide scale in the stainless steel pipe respectively. The multi-frequency excitation method ensures the effective extraction of the surface and deep signals, avoids the signal omission caused by single-frequency detection, and improves the comprehensiveness of the detection range;

[0044] S2. Apply an excitation signal to the stainless steel pipe through an excitation coil. The excitation coil adopts a multi-layer structure to ensure uniform distribution of the magnetic field. The excitation coil design improves the uniformity and intensity of the magnetic field and enhances the sensitivity of the detection signal, especially for the inner wall of the stainless steel pipe with curved surface and complex shape.

[0045] S3, collect the induction signal of the stainless steel pipe under the action of the excitation signal. The induction signal includes the response signal of the surface oxide scale and the interference signal of the matrix. The high sensitivity design of the acquisition module significantly improves the integrity and accuracy of the signal, providing a reliable data basis for subsequent signal processing;

[0046] S4, processing the sensing signal, including signal decomposition, noise suppression and signal fusion, to obtain the characteristic signal of the oxide scale. The signal processing step effectively enhances the reliability and resolution of the signal, ensuring the accuracy of the subsequent thickness prediction;

[0047] S5. Based on the characteristic signal, the thickness and distribution of the oxide scale in the stainless steel pipe are predicted, and the oxide scale distribution map is generated. The thickness prediction and distribution map generation provide comprehensive and visual results for oxide scale detection, which facilitates the rapid identification of high-risk areas and the adoption of corresponding measures.

[0048] S6. According to the results of oxide scale detection, real-time alarm information is output and production parameters are adjusted. The real-time alarm and parameter adjustment mechanism significantly reduces the risk of excessive oxide scale thickness, while improving the surface quality and consistency of the product.

[0049] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, the frequency range in S1 includes:

[0050] The high frequency band is 500kHz~5MHz, which is used to capture the surface characteristics of the oxide scale. Since the high frequency signal has a shorter wavelength, its electromagnetic penetration depth is small, usually below the micron level. This feature enables the high frequency signal to focus on the oxide scale surface area in the stainless steel pipe and obtain the electromagnetic response characteristics of the area. The use of high frequency signals improves the ability to detect the details of the oxide scale surface, especially when the oxide scale is thin, the detection sensitivity and accuracy are greatly improved, providing high-quality data for subsequent signal analysis;

[0051] The low frequency band is 50Hz~500kHz, which is used to penetrate the matrix to detect the distribution characteristics of deep oxide scale. The wavelength of the low frequency signal is longer and the electromagnetic penetration depth is greater. It can penetrate the stainless steel pipe wall and act on the deeper oxide scale area. Due to the interference of the matrix signal, the characteristics of the low frequency signal need to be combined with the high frequency signal to analyze the distribution of the deep oxide scale through relative changes. The introduction of the low frequency signal enables the distribution characteristics of the deep oxide scale to be accurately detected, especially in stainless steel pipes with thicker walls. The signal has strong penetration ability and can effectively make up for the shortcomings of the high frequency signal and achieve comprehensive detection.

[0052] Please refer to the attached Figure 1In a preferred embodiment of the present invention, the magnetic core material of the excitation coil in S2 is selected from soft ferrite with a magnetic permeability higher than 800 to enhance the magnetic field strength and reduce magnetic flux leakage. The number of turns of the inner layer of the excitation coil is 100 to 500 turns, and the number of turns of the outer layer is 50 to 200 turns. The turns ratio is dynamically adjusted according to the requirements of different signal frequencies. The use of soft ferrite materials with high magnetic permeability can significantly improve the magnetic field strength, while effectively reducing the leakage and loss of the magnetic field, thereby improving the effectiveness of the excitation signal, improving the detection sensitivity and signal stability; in addition, the high number of turns design of the inner coil increases the strength of the magnetic field, so that the high-frequency signal can act more effectively on the surface oxide. The lower number of turns design of the outer coil reduces the inductive reactance, improves the transmission efficiency of the low-frequency signal, and enhances the signal penetration ability. The double-layer structure coil design enhances the uniformity and strength of the magnetic field, meets the different requirements of high-frequency and low-frequency signal excitation, provides a more flexible signal transmission method, and improves the detection efficiency and accuracy.

[0053] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, the signal acquisition in S3 includes the following steps:

[0054] The sampling frequency is 2.5 times the highest frequency signal, ensuring the integrity and accuracy of the collected signal. Setting the sampling frequency to 2.5 times provides a margin for high-precision restoration of the signal. Through high-frequency sampling and high-bit depth quantization, it can avoid distortion while maintaining signal integrity, providing high-quality raw data for subsequent signal analysis, which is particularly suitable for capturing weak oxide skin signals.

[0055] A broadband amplifier is used to pre-process weak signals, and the amplification factor is set at 20 to 100 times. The pre-processing of the broadband amplifier significantly improves the effectiveness of the signal, especially in terms of weak signal capture capability and signal-to-noise ratio optimization, making oxide scale detection more sensitive and reliable;

[0056] Synchronous sampling technology is used to avoid signal delay and distortion, and the dynamic range of the collected signal is -90dB to 10dB. Synchronous sampling technology can avoid signal distortion or phase drift problems caused by timing deviation, thereby ensuring that high-frequency signals and low-frequency signals can be properly aligned during the processing process; in addition, the dynamic range design of -90dB to 10dB can capture very weak oxide scale signals while avoiding nonlinear distortion caused by strong signals, thereby improving the accuracy and adaptability of signal acquisition, and providing higher stability and consistency for oxide scale detection under complex conditions.

[0057] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, the signal processing in S4 includes the following steps:

[0058] The sensing signal is decomposed into high-frequency surface signal and low-frequency deep signal through short-time Fourier transform. STFT is a time-frequency analysis tool that can simultaneously analyze the time domain and frequency domain characteristics of the signal. Through the weighted processing of the window function, STFT can extract high-frequency and low-frequency signal features while maintaining the time resolution, thereby improving the layering and integrity of oxide skin detection;

[0059] Adopting an adaptive filter based on statistical characteristics to suppress noise, the adaptive filter can adjust the filter weights according to the dynamic changes of noise, without the need for a priori fixed filter parameters. Through the LMS algorithm, the adaptive filter can dynamically adjust the weights during signal processing to minimize the output error. The use of the adaptive filter significantly improves the signal-to-noise ratio of the signal, so that weak oxide skin signals can be effectively extracted, while avoiding the problem of signal loss caused by traditional fixed filtering methods.

[0060] The high-frequency and low-frequency signal features are fused to generate a comprehensive feature signal related to the thickness of the oxide scale. The fusion algorithm is based on the characteristic differences between high-frequency and low-frequency signals and uses a weighted strategy to combine the advantages of the two. The high-frequency signal has a higher resolution and is sensitive to surface characteristics; the low-frequency signal has a strong penetrating ability and is more sensitive to deep distribution characteristics. Through the fusion processing of features, the generated comprehensive signal can comprehensively reflect the thickness and distribution characteristics of the oxide scale.

[0061] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, in S5, the prediction of the oxide scale thickness is based on the construction of a mathematical model based on the characteristic signal, wherein the input signal parameters include the frequency response amplitude, phase change and harmonic characteristics, the model output is the oxide scale thickness value, and the two-dimensional thickness distribution map of the oxide scale is generated in combination with the image processing algorithm;

[0062] Specifically: a mathematical model for predicting oxide scale thickness is constructed, and a multivariate linear regression method is used to establish a functional relationship between input signal parameters and oxide scale thickness. The input signal parameters of the model include:

[0063] Frequency response amplitude: the amplitude of the signal at different frequencies, reflecting the response strength of the oxide thickness to the excitation signal; Phase change: the phase shift of the signal when passing through the oxide, reflecting the change of the signal propagation path; Harmonic characteristics: the high-order harmonics generated by the signal after the nonlinear effect, reflecting the non-uniformity and thickness distribution characteristics of the oxide;

[0064] In the model training phase, the signal response of oxide scales of different thicknesses is measured experimentally, and the regression relationship between the signal characteristic parameters and thickness is established to generate a mathematical model in the following form:

[0065] t 氧化皮 =β0+β1F 幅值 +β2F相位 +β3F 谐波

[0066] where t 氧化皮 To predict the thickness, F 幅值 、F 相位 and F 谐波 is the input characteristic parameter, β0~β3 are regression coefficients, and the oxide scale thickness directly affects the propagation characteristics of the excitation signal, including amplitude attenuation, phase shift and harmonic intensity. By regressing these signal characteristic parameters, the oxide scale thickness can be accurately predicted. The multivariate linear regression method is simple and easy to explain. At the same time, it has a high prediction accuracy when the parameters are properly selected, so it is suitable for stainless steel pipes of different specifications and oxidation states.

[0067] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, in S5, the oxide scale thickness distribution map displays the thickness range of the oxide scale through a two-dimensional heat map, and at the same time, a three-dimensional modeling technology is combined to construct an oxide scale spatial distribution model for displaying the distribution characteristics of the oxide scale in the length direction and the circumferential direction. The two-dimensional heat map intuitively displays the variation range of the oxide scale thickness through color. The gridding and interpolation processing of the thickness data ensure the continuity and visual effect of the heat map. The color segmentation makes it easy to distinguish different thickness areas in the heat map, thereby quickly locating abnormal areas. The three-dimensional modeling technology generates a three-dimensional distribution model by expanding the thickness data into the spatial coordinate system of the pipe. By combining the three-dimensional modeling technology, it can not only display the distribution of the oxide scale in the circumferential direction, but also display its thickness variation in the length direction, which is suitable for in-depth analysis of complex oxide scale distribution.

[0068] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, in S5, the electromagnetic properties of different pipe wall thicknesses and scale components are experimentally measured to establish a correction model for the matrix signal; in the detection, the correction model is used to compensate the induced signal to eliminate the interference of the matrix signal. The matrix thickness of the stainless steel pipe will have a significant attenuation effect on the induced signal, and the different scale components will affect the phase change and harmonic intensity of the signal. The interference of the matrix signal is mainly caused by the conductivity and magnetic permeability of the pipe wall. By decomposing the total induced signal and establishing a correction model for the matrix signal, the influence of the matrix signal can be effectively eliminated, so that the scale signal can be highlighted, thereby improving the detection accuracy. The interference component of the matrix signal is calculated in real time by the correction model and deducted from the total signal, which can improve the purity of the scale signal. This process does not require additional detection steps and is completely implemented by the algorithm, with both efficiency and accuracy.

[0069] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, in S6, the real-time alarm function includes:

[0070] When the detected oxide scale thickness exceeds the set threshold, visual and sound alarms are triggered. The detected thickness value is compared with the set threshold in real time through the threshold comparator. Once it exceeds the range, a trigger signal is sent to the alarm module. The combination of visual and sound alarms can quickly attract the attention of operators and help to respond to abnormal situations in a timely manner.

[0071] Generate an alarm report, record the alarm location, thickness exceeding the limit value and the corresponding time, and use the data capture module to record the information of abnormal points in real time and output it as a report in a structured manner to ensure that all abnormal information is completely saved. Automatic report generation avoids the omission of manual records and provides a detailed basis for subsequent quality analysis;

[0072] The alarm information is transmitted to the production control center to realize real-time adjustment of the production line. Through the real-time communication function of the industrial network, the alarm information can be quickly transmitted to the production control center. The automatic adjustment of production parameters is realized through a closed-loop control mechanism, avoiding delays or errors caused by human intervention.

[0073] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, in S6, the oxide scale detection result is used to dynamically optimize the production process, including:

[0074] Adjust the heating temperature range to ±2°C of the target temperature. The heating temperature is a key factor affecting the formation of oxide scale. Through the feedback adjustment of the PID controller, the temperature fluctuation can be controlled within the target range, effectively reducing the thickness of the oxide scale and improving its uniformity. At the same time, the problem of uneven oxide scale thickness caused by overheating or underheating is reduced, and the quality consistency of the product surface is improved;

[0075] Modify the cooling rate to control the rapid formation of oxide scale. The optimization range of cooling rate is between 10℃ / s and 50℃ / s. The cooling rate directly affects the structure and thickness of the oxide scale. A lower cooling rate helps to generate a uniform and thin oxide scale, while a higher cooling rate can avoid excessive oxidation reaction. Dynamic adjustment of the cooling rate can meet the needs of different test results, thereby improving the uniformity and stability of the oxide scale and avoiding quality problems caused by improper cooling process.

[0076] Adjust the production line speed to the recommended range of the test data to ensure product consistency and surface quality. The production line speed is directly related to the heating and cooling time. Reducing the speed can extend the heat treatment time, reduce the fluctuation of the oxide scale thickness, and improve the surface quality. When the oxide scale thickness is qualified, increasing the speed can improve production efficiency. Dynamically adjusting the production speed achieves the best balance between quality and efficiency, avoids surface quality problems caused by excessive speed, and improves the flexibility of production rhythm.

[0077] In order to better understand the present invention, the above contents are described in detail below in conjunction with specific embodiments.

[0078] Embodiment 1: No multi-frequency excitation signal is added, only single-frequency detection is used

[0079] In this embodiment, the detection system only uses a fixed single-frequency excitation signal with a frequency set to 500kHz, which is mainly used to detect the oxide thickness on the surface of the stainless steel pipe. The oxide thickness characteristics are extracted by traditional amplitude and phase analysis in the signal processing stage.

[0080] Effect: Single-frequency signal detection has a certain sensitivity to the surface details of the oxide scale, but it cannot penetrate the pipe wall for deep detection. At the same time, for areas where the oxide scale is unevenly distributed, the detection results are not accurate and comprehensive enough.

[0081] Example 2: Adding multi-frequency excitation signals and feature fusion analysis

[0082] This embodiment uses multi-frequency excitation technology to generate high-frequency (500kHz-5MHz) and low-frequency (50Hz-500kHz) signals to extract the surface characteristics and deep distribution characteristics of the oxide scale. Short-time Fourier transform and adaptive filtering technology are introduced in the signal processing stage to decompose, reduce noise and fuse features of the signal.

[0083] Effect: Multi-frequency signals cover the surface and deep distribution characteristics of the oxide scale, improving the comprehensiveness and accuracy of detection, especially in complex distribution scenarios.

[0084] Example 3: Adding dynamic production parameter optimization based on Example 2

[0085] This embodiment further adjusts production parameters in real time based on the results of the oxide scale detection. The heating temperature is accurately adjusted to ±2°C of the target temperature through a PID controller; the cooling rate is dynamically optimized between 10°C / second and 50°C / second; and the production line speed is adjusted in real time based on the detection results.

[0086] Effect: The linkage adjustment of test results and production parameters realizes dynamic control of oxide scale, ensures the consistency of product surface quality, and improves production efficiency and stability.

[0087] Comparative experiment 1: Comparative experiment on the detection accuracy of oxide scale thickness

[0088] Experimental purpose: To verify the high accuracy and comprehensiveness of the present invention in oxide scale thickness detection through multi-frequency excitation and feature fusion processing.

[0089] Experimental setup

[0090] Sample setup

[0091] Sample 1: Evenly distributed oxide scale with a thickness of 10μm.

[0092] Sample 2: Gradient distribution of oxide scale, thickness ranges from 10μm to 50μm.

[0093] Sample 3: Irregularly distributed oxide scale with a mixed distribution of thickness ranging from 10μm, 20μm and 50μm.

[0094] Experimental and control groups

[0095] Experimental group (the present invention)

[0096] The oxide scale thickness is detected by using multi-frequency excitation (high frequency 500kHz~5MHz, low frequency 50Hz~500kHz) combined with feature fusion analysis.

[0097] Control group 1 (traditional single frequency detection)

[0098] The oxide thickness is detected by amplitude analysis using a fixed frequency signal (500kHz).

[0099] Control group 2 (multi-frequency no fusion analysis)

[0100] A multi-frequency signal (high frequency 500kHz~5MHz, low frequency 50Hz~500kHz) is applied, but the high-frequency and low-frequency signals are analyzed separately without feature fusion.

[0101] Experimental procedures

[0102] Sample preparation: Prepare oxide scale samples with different thickness and distribution characteristics, numbered as sample 1, sample 2 and sample 3.

[0103] Signal application: Apply the excitation signals corresponding to the experimental group and the control group to each sample.

[0104] Data collection: record the oxide scale thickness value of each group of tests.

[0105] Error calculation: Compare the detected value with the actual thickness value and calculate the detection error rate.

[0106] Data analysis: Summarize the experimental data and analyze the detection performance of the experimental group and the control group under different sample conditions.

[0107] The comparative experimental data are shown in Table 1:

[0108] Table 1 Oxide scale thickness test results

[0109]

[0110] From the data in Table 1, we can get:

[0111] This experiment verified the comprehensiveness and high precision of the present invention in scale thickness detection through data comparison. Especially under complex distribution conditions, the experimental group of the present invention was significantly better than control group 1 and control group 2, reflecting the advantages of multi-frequency excitation and feature fusion processing.

[0112] Comparative experiment 2: Comparative experiment on deep scale distribution detection capability

[0113] Experimental purpose: To verify the ability of the present invention to detect the distribution of deep oxide scale when the pipe wall is thick.

[0114] Experimental setup

[0115] Sample setup

[0116] Sample 1: tube wall thickness 2mm, oxide scale thickness 50μm, evenly distributed.

[0117] Sample 2: tube wall thickness 5mm, oxide scale thickness 50μm, irregular distribution.

[0118] Sample 3: The tube wall thickness is 10 mm, the oxide scale thickness is 50 μm, and the oxide scale is distributed locally.

[0119] Experimental and control groups

[0120] Experimental group (the present invention)

[0121] Multi-frequency excitation (high frequency 500kHz~5MHz, low frequency 50Hz~500kHz) is used in combination with signal feature fusion analysis to detect the distribution of deep oxide scale.

[0122] Control group 1 (single frequency high frequency detection)

[0123] A single high-frequency signal (500kHz) is used for detection, mainly analyzing the surface and shallow layer signal characteristics.

[0124] Control group 2 (single frequency low frequency detection)

[0125] A single low-frequency signal (50 Hz) is used for detection, penetrating the pipe wall but without deep signal fusion analysis.

[0126] Experimental procedures

[0127] Sample preparation: pipe samples with wall thickness of 2 mm, 5 mm and 10 mm were prepared, and the inner wall of the pipe was coated with uniform, irregular and locally distributed oxide scale.

[0128] Signal application: Apply the excitation signals corresponding to the experimental group and the control group to the samples respectively, and record the deep-layer induction signal characteristics.

[0129] Data acquisition: record the scale distribution signal intensity and distribution characteristics obtained by the detection system.

[0130] Distribution consistency analysis: Compare the test results with the actual scale distribution to analyze the signal distribution consistency and effective signal strength.

[0131] The comparative experimental data are shown in Table 2:

[0132] Table 2 Signal intensity and distribution consistency of deep oxide scale detection

[0133]

[0134] From the data in Table 2, we can get:

[0135] The experimental results show that the present invention is significantly superior to single-frequency detection technology in terms of deep detection signal strength and distribution consistency through multi-frequency excitation and feature fusion analysis. Especially in samples with thicker walls, the present invention can accurately reflect the deep distribution characteristics of oxide scale.

[0136] Comparative experiment 3: Dynamic production process linkage optimization comparative experiment

[0137] Experimental purpose: To verify whether the real-time detection and process parameter linkage optimization of the present invention can effectively reduce the excess rate of oxide scale thickness.

[0138] Experimental setup

[0139] Production parameter settings

[0140] Heating temperature target value: 1050℃, allowable range is ±2℃.

[0141] Cooling rate range: 10℃ / sec~50℃ / sec.

[0142] Production line speed: dynamically adjusted according to test data.

[0143] Experimental group and control group

[0144] Experimental group (the present invention)

[0145] Using multi-frequency detection technology, the oxide thickness data is fed back in real time, and the heating temperature, cooling rate and production speed are dynamically adjusted.

[0146] Control group 1 (no test adjustment)

[0147] Fixed process parameters were used (heating temperature 1050°C, cooling rate 30°C / s, fixed production speed), without detection feedback and process adjustment.

[0148] Control group 2 (single frequency detection adjustment)

[0149] Based on the oxide thickness data fed back by single frequency detection (500kHz), limited process adjustments are made.

[0150] Experimental procedures

[0151] Sample preparation: Process stainless steel pipe products on the production line and observe the formation of oxide scale.

[0152] Parameter setting: Different detection and process optimization schemes were used in the experimental group and the control group.

[0153] Data collection: record the oxide scale thickness exceeding the standard rate, surface quality consistency and production efficiency (number of shutdowns and single batch production time).

[0154] Data analysis: Compare the performance of the experimental group and the control group on different indicators.

[0155] The comparative experimental data are shown in Table 3:

[0156] Table 3 Dynamic production process linkage optimization experimental results

[0157]

[0158]

[0159] From the data in Table 3, we can get:

[0160] The experimental group significantly improved the scale quality and production efficiency through real-time feedback and comprehensive process adjustment, which was far superior to the control group with no detection adjustment and single-frequency detection adjustment, verifying the advantages of the present invention in the dynamic production process linkage optimization.

[0161] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for detecting oxide scale inside a stainless steel pipe using electromagnetic induction technology, characterized in that: The following steps are involved: S1, generating a multi-frequency excitation signal, wherein the excitation signal includes a high-frequency signal and a low-frequency signal, which are used to respectively detect the surface characteristics and deep distribution of the oxide scale in the stainless steel pipe; S2, applying the excitation signal to the stainless steel tube through an excitation coil, wherein the excitation coil adopts a multi-layer structure to ensure uniform distribution of the magnetic field; S3, collecting the induction signal of the stainless steel pipe under the action of the excitation signal, wherein the induction signal includes the response signal of the surface oxide scale and the interference signal of the matrix; S4, processing the sensing signal, including signal decomposition, noise suppression and signal fusion, to obtain a characteristic signal of the oxide scale; S5. Based on the characteristic signal, predict the thickness and distribution of the oxide scale in the stainless steel pipe and generate an oxide scale distribution map; S6. Output real-time alarm information and adjust production parameters according to the results of scale detection.

2. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: The frequency range in S1 includes: The high frequency band is 500kHz to 5MHz, which is used to capture the surface characteristics of the oxide scale; The low frequency band is 50Hz~500kHz, which is used to penetrate the matrix to detect the distribution characteristics of deep oxide scale.

3. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: The magnetic core material of the excitation coil in S2 is soft ferrite with a magnetic permeability higher than 800 to enhance the magnetic field strength and reduce magnetic flux leakage. The inner layer of the excitation coil has 100 to 500 turns, and the outer layer has 50 to 200 turns. The turns ratio is dynamically adjusted according to different signal frequency requirements.

4. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: The signal acquisition in S3 includes the following steps: The sampling frequency is 2.5 times of the highest frequency signal to ensure the integrity and accuracy of the collected signal; A broadband amplifier is used to pre-process weak signals, and the amplification factor is set at 20 to 100 times; Synchronous sampling technology is used to avoid signal delay and distortion, and the dynamic range of the collected signal is -90dB to 10dB.

5. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: The signal processing in S4 comprises the following steps: The induced signals are decomposed into high-frequency surface signals and low-frequency deep-layer signals by short-time Fourier transform; Adopting adaptive filter based on statistical characteristics to suppress noise; The high-frequency and low-frequency signal features are integrated to generate a comprehensive characteristic signal related to the oxide scale thickness.

6. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: In S5, the prediction of oxide scale thickness is based on the construction of a mathematical model based on the characteristic signal, wherein the input signal parameters include the frequency response amplitude, phase change and harmonic characteristics, the model output is the thickness value of the oxide scale, and the two-dimensional thickness distribution map of the oxide scale is generated in combination with the image processing algorithm.

7. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: In S5, the oxide scale thickness distribution diagram displays the thickness range of the oxide scale through a two-dimensional heat map, and at the same time, a three-dimensional modeling technology is combined to construct an oxide scale spatial distribution model to display the distribution characteristics of the oxide scale in the length direction and the circumferential direction.

8. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: In S5, the electromagnetic characteristics of different tube wall thicknesses and oxide scale components are experimentally measured to establish a correction model for the matrix signal; during detection, the correction model is used to compensate the induced signal to eliminate interference from the matrix signal.

9. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: In S6, the real-time alarm function includes: When the detected oxide scale thickness exceeds the set threshold, visual and sound alarms are triggered; Generate alarm report, record alarm location, thickness exceeding value and corresponding time; Transmit alarm information to the production control center to achieve real-time adjustment of the production line.

10. The method for detecting oxide scale in stainless steel pipes by electromagnetic induction technology according to claim 1, characterized in that: In S6, the oxide scale detection results are used to dynamically optimize the production process, including: Adjust the heating temperature range to ±2°C of the target temperature; Modify the cooling rate to control the rapid formation of oxide scale. The optimal range of cooling rate is between 10℃ / s and 50℃ / s. Adjust the production line speed to the recommended range of the inspection data to ensure product consistency and surface quality.

Images