A stainless steel pipe quality evaluation method and system based on detection response

Abstract

The present application relates to the technical field of intelligent sensor, specifically to a stainless steel pipe quality evaluation method and system based on detection response, comprising: constructing and emitting a specific phase encoding excitation flow, and collecting the response after the propagation in the stainless steel pipe; generating the response spectrum available for analysis by processing the response signal; determining two key physical performance parameters representing the material by analyzing the response spectrum: the bulk nonlinear acoustic characteristic value reflecting the microscopic physical properties, and the interface modulation acoustic characteristic value reflecting the structural integrity; finally, based on the comprehensive evaluation of the two determined physical performance parameters, the precise differentiation of the material physical property degradation and the internal structure discontinuity is realized. The present application can realize the synchronous analysis and evaluation of multiple physical performances of the material through one nondestructive testing.

Description

Technical Field

[0001] This invention relates to the field of intelligent sensor technology, specifically to a method and system for evaluating the quality of stainless steel pipes based on detection response. Background Technology

[0002] Multiphase alloy media, due to their heterogeneous grain structure with significant differences in acoustic impedance, exhibit typical strong scattering and attenuation characteristics acoustically. The core challenge in quality assessment of such media lies in accurately resolving the physical features characterizing the intrinsic states of the medium from complex wave-field interactions.

[0003] Current linear ultrasonic testing primarily relies on the principle of sound wave reflection, but it faces signal resolution challenges when evaluating strongly scattering media: random scattering from coarse grain boundaries within the medium generates high-amplitude coherent noise, severely masking effective information; simultaneously, closed structural damage interfaces exhibit extremely high acoustic transmittance, resulting in weak linear response characteristics and making them prone to missed detections. Furthermore, traditional ultrasonic testing front-ends typically function only as passive data acquisition probes, lacking intelligent sensing and adaptive adjustment capabilities, and cannot directly modulate complex excitation flows or perform preliminary background noise filtering at the sensor end. This necessitates uploading large amounts of noisy raw data to a host computer for processing, introducing transmission delays and making it difficult to cope with the annihilation of weak nonlinear signals by transient electromagnetic interference in industrial environments.

[0004] While nonlinear ultrasound technology, introduced in recent years, offers high sensitivity, the responses received by the detection system are often multi-source mixed signals. Existing analytical methods cannot decouple the different physical components, meaning they cannot distinguish whether the nonlinear modulation energy originates from volume nonlinear effects due to changes in the microscopic properties of the medium, or from contact acoustic nonlinear effects at locally closed interfaces. This ambiguity of the signal source prevents the evaluation system from outputting accurate qualitative and quantitative quality assessment conclusions.

[0005] Therefore, there is a need for an evaluation method based on smart sensor technology that can suppress background noise and physically decouple hybrid nonlinear responses.

[0006] Therefore, this invention proposes a method and system for evaluating the quality of stainless steel pipes based on detection response. Summary of the Invention

[0007] The purpose of this invention is to provide a method and system for evaluating the quality of stainless steel pipes based on detection response. By using phase-encoded excitation and differential acoustic response index analysis, multi-source nonlinear responses are decoupled, enabling accurate differentiation and evaluation of material degradation and structural damage in stainless steel pipes.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for quality assessment of stainless steel pipes based on detection response, comprising:

[0010] Obtain the basic ultrasonic testing parameters of the stainless steel pipe to be evaluated, and construct a phase-coded excitation flow containing axial polarization shear wave sequences and circumferential polarization shear wave sequences based on the basic ultrasonic testing parameters. The axial polarization shear wave sequences and circumferential polarization shear wave sequences contain positive-phase bursts and negative-phase bursts.

[0011] The original detection response of the stainless steel tube to be evaluated under the action of phase-encoded excitation current is collected. The positive-phase burst response and the anti-phase burst response in the original detection response are linearly superimposed to generate the structural noise suppression response spectrum.

[0012] Analyze the structural noise suppression response spectrum and extract the bulk nonlinear acoustic eigenvalues ​​of the corresponding axially polarized shear wave sequence and the interface modulation acoustic eigenvalues ​​of the circumferentially polarized shear wave sequence.

[0013] The ratio of the interface modulation acoustic eigenvalue to the bulk nonlinear acoustic eigenvalue is calculated to generate the differential acoustic response index.

[0014] The differential acoustic response index and the volume nonlinear acoustic eigenvalue are input into the comprehensive quality evaluation model. When the differential acoustic response index is within the benchmark range and the volume nonlinear acoustic eigenvalue increases abnormally, the model outputs the quality assessment result of the constitutive anomaly of the medium. When the differential acoustic response index is higher than the benchmark range, the model outputs the quality assessment result of the structural continuity anomaly.

[0015] Preferably, the process of constructing a phase-encoded excitation flow containing axially polarized shear wave sequences and circumferentially polarized shear wave sequences based on ultrasonic testing fundamental parameters includes: obtaining the wall thickness parameters and sound velocity anisotropy coefficient of the stainless steel pipe to be evaluated as ultrasonic testing fundamental parameters; loading the ultrasonic testing fundamental parameters through a smart sensor to drive an electromagnetic ultrasonic transducer array integrated within the smart sensor; using the smart sensor to control the axial coil group in the electromagnetic ultrasonic transducer array to generate shear waves with polarization directions parallel to the pipe axis as the axially polarized shear wave sequence; using the smart sensor to control the circumferential coil group in the electromagnetic ultrasonic transducer array to generate shear waves with polarization directions perpendicular to the pipe axis as the circumferentially polarized shear wave sequence; applying zero-degree phase modulation to the axially polarized shear wave sequence and the circumferentially polarized shear wave sequence respectively to generate the positive-phase burst sound, and applying 180-degree phase modulation to generate the anti-phase burst sound, and combining them to construct the phase-encoded excitation flow.

[0016] Preferably, before constructing the phase-encoded excitation flow containing axial polarization shear wave sequences and circumferential polarization shear wave sequences based on the basic parameters of ultrasonic testing, the method further includes: selecting a standard sample tube with the same material and in good condition as the stainless steel tube to be evaluated; testing the standard sample tube using the phase-encoded excitation flow to obtain the standard body nonlinear acoustic characteristic value and the standard interface modulation acoustic characteristic value; and setting a numerical threshold for determining whether the differential acoustic response index is within a reference range based on the statistical distribution data of the standard body nonlinear acoustic characteristic value and the standard interface modulation acoustic characteristic value.

[0017] Preferably, the process of generating a structural noise suppression response spectrum by linearly superimposing the positive-phase burst response and the negative-phase burst response in the original detection response includes: reading the time-domain waveform data of the positive-phase burst response and the time-domain waveform data of the negative-phase burst response; performing multi-cycle time-domain synchronous averaging processing on the positive-phase burst response and the negative-phase burst response respectively to suppress random background noise; performing point-to-point amplitude addition on the processed positive-phase burst response and the negative-phase burst response, removing the fundamental component using the phase cancellation principle, and generating a nonlinear time-domain residual signal from the merged data; performing a fast Fourier transform on the nonlinear time-domain residual signal to map the time-domain information to the frequency domain, generating spectral data containing the fundamental component and higher harmonic components; and extracting the energy distribution data from the spectral data to define the structural noise suppression response spectrum.

[0018] Preferably, the process of analyzing the structural noise suppression response spectrum and extracting the bulk nonlinear acoustic eigenvalues ​​of the corresponding axially polarized shear wave sequence and the interface modulation acoustic eigenvalues ​​of the circumferentially polarized shear wave sequence includes: locking the second harmonic frequency band range corresponding to the center frequency of the phase-encoded excitation current in the structural noise suppression response spectrum; performing cumulative calculation on the spectral amplitude within the second harmonic frequency band range for the structural noise suppression response spectrum generated by the axially polarized shear wave sequence to obtain the total energy of higher harmonics, and performing relative intensity conversion in combination with the fundamental energy, defining the obtained dimensionless value as the bulk nonlinear acoustic eigenvalue, which is used to characterize the energy distortion caused by the accumulation of lattice anharmonicity when the sound wave propagates in the medium; calculating the amplitude integral within the second harmonic frequency band range for the structural noise suppression response spectrum generated by the circumferentially polarized shear wave sequence, and defining the normalized integral result as the interface modulation acoustic eigenvalue, which is used to characterize the local spectral modulation of the sound wave at the contact interface due to the breathing effect.

[0019] Preferably, the step of normalizing the integration result and defining it as the volume nonlinear acoustic characteristic value, and the step of normalizing the integration result and defining it as the interface modulation acoustic characteristic value, specifically includes: locking the fundamental frequency band range corresponding to the center frequency of the phase-coded excitation flow in the structural noise suppression response spectrum; calculating the fundamental amplitude integral within the fundamental frequency band range, and defining the square of the fundamental amplitude integral as the energy normalization factor; performing a division operation with the amplitude integral within the second harmonic frequency band range as the numerator and the energy normalization factor as the denominator, and defining the operation results as the volume nonlinear acoustic characteristic value and the interface modulation acoustic characteristic value, respectively, and using the energy normalization factor to eliminate the characteristic value deviation caused by the initial energy fluctuation of the phase-coded excitation flow.

[0020] Preferably, the process of calculating the ratio of the interface modulation acoustic feature value to the bulk nonlinear acoustic feature value to generate the differential acoustic response index includes: using the interface modulation acoustic feature value as the numerator and the bulk nonlinear acoustic feature value as the denominator, performing a division operation; defining the quotient obtained by the operation as the differential acoustic response index; the differential acoustic response index is used to characterize the anisotropy of the nonlinear source in the original detection response; when the differential acoustic response index approaches a reference value, it indicates that the nonlinear source is an isotropic source; when the differential acoustic response index is higher than the reference value, it indicates that the nonlinear source is an interface contact source with anisotropic characteristics.

[0021] Preferably, the quality comprehensive evaluation model is a two-dimensional state-space logical classification algorithm. The process of inputting the differential acoustic response index and the volume nonlinear acoustic feature value into the quality comprehensive evaluation model includes: constructing a two-dimensional acoustic state space containing two orthogonal dimensions, wherein the first dimension represents the degree of microscopic distortion of the medium, and the second dimension represents the degree of anisotropy of the nonlinear source; dividing the two-dimensional acoustic state space into a preset safety reference region, a volume distribution nonlinearity-dominant region, and an interface contact nonlinearity-dominant region; using the volume nonlinear acoustic feature value as the first feature component and the differential acoustic response index as the second feature component, combining them to form a feature vector; projecting the feature vector onto the two-dimensional acoustic state space, determining the coordinate position, and identifying the region in which the coordinate position falls within the two-dimensional acoustic state space.

[0022] Preferably, the process of outputting the structural continuity anomaly quality assessment result includes: when the state point falls into the volume distribution nonlinearity-dominant region, determining that the nonlinear response of the stainless steel tube to be evaluated originates from the degradation of the microscopic physical properties of the medium, and outputting the volume distribution nonlinearity-dominant state identification result; when the state point falls into the interface contact nonlinearity-dominant region, determining that the nonlinear response of the stainless steel tube to be evaluated originates from closed interface contact friction, and outputting the interface contact nonlinearity-dominant state identification result.

[0023] A stainless steel pipe quality assessment system based on detection response, comprising:

[0024] An excitation construction module is used to obtain the basic ultrasonic testing parameters of the stainless steel pipe to be evaluated, and to construct a phase-encoded excitation flow containing an axial polarization shear wave sequence and a circumferential polarization shear wave sequence based on the basic ultrasonic testing parameters. The axial polarization shear wave sequence and the circumferential polarization shear wave sequence contain positive-phase bursts and negative-phase bursts.

[0025] The signal processing module is used to acquire the original detection response of the stainless steel tube to be evaluated under the action of the phase-encoded excitation current, and to perform linear superposition of the positive-phase burst response and the negative-phase burst response in the original detection response to generate the structural noise suppression response spectrum.

[0026] The feature analysis module is used to analyze the structural noise suppression response spectrum and extract the bulk nonlinear acoustic feature values ​​of the corresponding axial polarization shear wave sequence and the interface modulation acoustic feature values ​​of the circumferential polarization shear wave sequence.

[0027] The index calculation module is used to calculate the ratio of the interface modulation acoustic characteristic value to the bulk nonlinear acoustic characteristic value, and generate the differential acoustic response index.

[0028] The state recognition module is used to input the differential acoustic response index and the volume nonlinear acoustic characteristic value into the comprehensive quality evaluation model. When the differential acoustic response index is within the benchmark range and the volume nonlinear acoustic characteristic value is abnormally increased, it outputs the medium constitutive anomaly quality assessment result; when the differential acoustic response index is higher than the benchmark range, it outputs the structural continuity anomaly quality assessment result.

[0029] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0030] 1. This invention utilizes the phase cancellation principle to cancel out the linear reflected echoes and random scattered coherent noise (fundamental component) generated by coarse grain boundaries inside the medium. Thus, without the need for complex filtering algorithms, it can accurately extract weak nonlinear high-order harmonic signals from the strongly attenuated stainless steel tube medium, effectively solving the problem of effective information being masked by high-amplitude noise.

[0031] 2. This invention utilizes the directional specificity of the interaction between shear waves and the medium to compare the bulk nonlinear acoustic characteristic values ​​that characterize the intrinsic state of the medium with the interface modulation acoustic characteristic values ​​that characterize the state of the contact interface. This successfully distinguishes between the degradation of microscopic properties of isotropic media (such as lattice distortion) and the damage of closed structures with anisotropic characteristics (such as microcracks), thus solving the problem of qualitative evaluation difficulties caused by the ambiguity of signal sources in traditional methods.

[0032] 3. This invention abandons the limitations of a single threshold judgment and expands the evaluation dimensions to two orthogonal dimensions: "degree of microscopic distortion of the medium" and "degree of anisotropy of nonlinear source". It can simultaneously output two completely different evaluation results: constitutive anomaly of the medium (material aging) and structural continuity anomaly (physical damage), which significantly improves the accuracy and robustness of the quality evaluation of stainless steel pipes in complex service environments. Attached Figure Description

[0033] Figure 1 This is a flowchart of a stainless steel pipe quality assessment method based on detection response according to the present invention;

[0034] Figure 2 This is a flowchart of the quality assessment decision-making process based on two-dimensional acoustic state space according to an embodiment of the present invention;

[0035] Figure 3 This is a block diagram of a stainless steel pipe quality assessment system based on detection response according to the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Other embodiments obtained by those skilled in the art based on the ideas in this specification without creative effort all fall within the protection scope of this invention.

[0037] Reference Figures 1 to 3 This invention provides a method and system for evaluating the quality of stainless steel pipes based on detection response, the specific technical solution of which is as follows:

[0038] Example 1:

[0039] A method for quality assessment of stainless steel pipes based on detection response, referring to Figure 1 ,include:

[0040] Obtain the basic ultrasonic testing parameters of the stainless steel pipe to be evaluated, and construct a phase-coded excitation flow containing axial polarization shear wave sequences and circumferential polarization shear wave sequences based on the basic ultrasonic testing parameters. The axial polarization shear wave sequences and circumferential polarization shear wave sequences contain positive-phase bursts and negative-phase bursts.

[0041] The original detection response of the stainless steel tube to be evaluated under the action of phase-encoded excitation current is collected. The positive-phase burst response and the anti-phase burst response in the original detection response are linearly superimposed to generate the structural noise suppression response spectrum.

[0042] Analyze the structural noise suppression response spectrum and extract the bulk nonlinear acoustic eigenvalues ​​of the corresponding axially polarized shear wave sequence and the interface modulation acoustic eigenvalues ​​of the circumferentially polarized shear wave sequence.

[0043] The ratio of the interface modulation acoustic eigenvalue to the bulk nonlinear acoustic eigenvalue is calculated to generate the differential acoustic response index.

[0044] The differential acoustic response index and the volume nonlinear acoustic eigenvalue are input into the comprehensive quality evaluation model. When the differential acoustic response index is within the benchmark range and the volume nonlinear acoustic eigenvalue increases abnormally, the model outputs the quality assessment result of the constitutive anomaly of the medium. When the differential acoustic response index is higher than the benchmark range, the model outputs the quality assessment result of the structural continuity anomaly.

[0045] Furthermore, the process of constructing a phase-encoded excitation flow containing axially polarized shear wave sequences and circumferentially polarized shear wave sequences based on ultrasonic testing fundamental parameters includes: obtaining the wall thickness parameters and sound velocity anisotropy coefficient of the stainless steel pipe to be evaluated as ultrasonic testing fundamental parameters; loading the ultrasonic testing fundamental parameters through a smart sensor to drive an electromagnetic ultrasonic transducer array integrated within the smart sensor; using the smart sensor to control the axial coil group in the electromagnetic ultrasonic transducer array to generate shear waves with polarization directions parallel to the pipe axis as the axially polarized shear wave sequence; using the smart sensor to control the circumferential coil group in the electromagnetic ultrasonic transducer array to generate shear waves with polarization directions perpendicular to the pipe axis as the circumferentially polarized shear wave sequence; applying zero-degree phase modulation to the axially polarized shear wave sequence and the circumferentially polarized shear wave sequence respectively to generate the positive-phase burst sound, and applying 180-degree phase modulation to generate the anti-phase burst sound, combining them to construct the phase-encoded excitation flow.

[0046] Specifically, the wall thickness and sound velocity anisotropy coefficient of the stainless steel pipe to be evaluated are obtained as the basic parameters for ultrasonic testing. Before conducting a quality assessment on a section of in-service high-pressure fluid transport stainless steel pipe, its basic physical parameters are obtained by consulting its design data or by conducting on-site measurements using a standard thickness gauge. In this embodiment, the obtained wall thickness is 5.0 mm and the sound velocity anisotropy coefficient is 1.05.

[0047] The microprocessor control unit of the smart sensor reads and loads the basic ultrasonic detection parameters, calculates the timing logic of the excitation signal based on these parameters, and controls the built-in signal generation circuit and power amplification circuit to drive the electromagnetic ultrasonic transducer array integrated inside the smart sensor. The previously obtained values ​​of 5.0 mm pipe wall thickness and 1.05 sound velocity anisotropy coefficient are input to the control unit of the smart sensor through a human-machine interface. After receiving these parameters, the internal microprocessor of the smart sensor automatically calculates and sets the optimal excitation signal parameters based on a preset acoustic model, such as optimizing the center frequency to 5 MHz. This setting process is called "loading". Subsequently, the control unit of the smart sensor, according to the optimized parameters, begins to provide the necessary power and control commands to its physical probe (i.e., the electromagnetic ultrasonic transducer array integrated inside the sensor).

[0048] The intelligent sensor drives a double-layer flexible circuit board coil array integrated inside the probe, controlling the axial coil group to generate shear waves with polarization direction parallel to the tube axis as the axial polarized shear wave sequence. The internal signal generation unit of the intelligent sensor, according to pre-loaded timing control instructions, generates a series of high-frequency alternating current pulses with specific phase encoding characteristics under the precise control of its microprocessor, and outputs them directionally to the axial coil group, which employs a racetrack-shaped folding structure. The long-side conductors of this coil group are arranged perpendicular to the axis of the stainless steel tube. When high-frequency alternating current passes through, closed circumferential eddy currents are induced on the surface of the stainless steel tube through electromagnetic induction coupling. These circumferential eddy currents interact with the radially biased static magnetic field provided by the transducer permanent magnet, generating a Lorentz force along the tube axis based on the left-hand rule, thereby exciting shear waves with polarization direction strictly parallel to the tube axis.

[0049] The intelligent sensor controls a circumferential coil group orthogonally stacked with the axial coil group to generate a shear wave with a polarization direction perpendicular to the tube axis, which serves as the circumferentially polarized shear wave sequence. To ensure the correlation of multidimensional acoustic features in the time domain, within a preset microsecond-level protection interval after axial excitation, the internal control logic of the intelligent sensor automatically switches the output channel of the signal generation unit to the circumferential coil group with a butterfly structure. The effective conductor segment of this coil group is arranged parallel to the axis of the stainless steel tube. At this time, the signal generation unit injects high-frequency alternating current into the circumferential coil group, inducing eddy currents flowing along the tube axis on the surface of the stainless steel tube. These axial eddy currents interact with the radially biased static magnetic field provided by the permanent magnet of the transducer, generating a Lorentz force along the tangential direction of the tube circumference on the tube wall surface, thereby exciting a shear wave with a polarization direction strictly perpendicular to the tube axis. The entire switching and excitation process is locked within a single detection command cycle and is automatically and continuously completed by the intelligent sensor to ensure that the acquired axial and circumferential responses reflect the physical state of the measured object at the same moment.

[0050] The intelligent sensor applies zero-degree phase modulation to both the axially polarized shear wave sequence and the circumferentially polarized shear wave sequence to generate the positive-phase burst, and applies 180-degree phase modulation to generate the anti-phase burst, combining them to construct the phase-coded excitation stream. During the generation of the electrical excitation signals for the two sets of shear waves, the digital signal processing unit of the intelligent sensor, under the instruction of the microprocessor, performs digital phase modulation on the sine wave signal packet containing eight cycles to be output. Zero-degree phase modulation is applied to generate the excitation signal for the positive-phase burst; 180-degree phase modulation is applied to generate the excitation signal for the anti-phase burst. Finally, the intelligent sensor precisely controls the sequential transmission of these four modulated excitation signals according to its internally fixed timing logic of [axial positive-phase - axial anti-phase - circumferential positive-phase - circumferential anti-phase], which together form a complete phase-coded excitation stream.

[0051] By having intelligent sensors uniformly execute the entire process of parameter loading, waveform generation, coil control, and phase modulation, the constructed phase-encoded excitation stream is ensured to have extremely high accuracy and consistency in timing, phase, and polarization state, laying a solid foundation for subsequent signal processing and accurate evaluation.

[0052] As a preferred embodiment, before constructing the phase-encoded excitation flow containing axially polarized shear wave sequences and circumferentially polarized shear wave sequences based on the ultrasonic testing fundamental parameters, the method further includes: performing narrowband frequency sweep pre-excitation around the initial center frequency determined by the ultrasonic testing fundamental parameters within a preset frequency range; acquiring the fundamental response energy of the stainless steel tube to be evaluated under the narrowband frequency sweep pre-excitation in real time; determining the frequency point corresponding to the maximum value of the fundamental response energy as the dynamically calibrated working center frequency; and constructing the phase-encoded excitation flow based on the dynamically calibrated working center frequency.

[0053] Specifically, firstly, based on the obtained basic ultrasonic testing parameters such as the pipe wall thickness (5.0 mm) and the sound velocity anisotropy coefficient (1.05), a theoretical initial center frequency is calculated, which is 5.0 MHz in this embodiment. This frequency serves as the reference for the self-calibration process.

[0054] Subsequently, before formally transmitting the phase-coded excitation stream used for damage assessment, the smart sensor automatically executes a dynamic self-calibration process. This process controls the signal generation unit to sequentially transmit a series of low-energy, short-duration single-frequency ultrasonic pulses in 0.01 MHz steps within a preset narrowband frequency range (4.8 MHz to 5.2 MHz in this embodiment). This process is known as narrowband frequency sweep pre-excitation.

[0055] Simultaneously with the transmission at each pre-excitation frequency point, the receiving and processing unit of the intelligent sensor synchronously acquires the echo signal after penetrating the pipe wall, and performs a Fast Fourier Transform only for the fundamental frequency band to calculate the energy amplitude of the fundamental response in real time. All frequency points and their corresponding energy amplitudes are recorded to form an energy-frequency response curve.

[0056] Finally, the response curve is analyzed to automatically find the frequency point where the energy amplitude reaches its peak. In this embodiment, the received fundamental response energy is strongest when the excitation frequency is 5.07 MHz. This frequency point is determined as the optimal acoustic transmission frequency under the actual physical conditions (including temperature, stress, and coupling conditions) of the current measurement point. This method then locks 5.07 MHz as the working center frequency after dynamic calibration and constructs the phase-encoded excitation stream for subsequent formal testing based on this frequency.

[0057] This scheme, through a rapid energy feedback closed loop, enables the excitation frequency to match the optimal acoustic window of the measurement point in real time, significantly improving the effective acoustic energy injected into the tube wall and the signal-to-noise ratio of the final response signal, thus providing the best raw data foundation for subsequent high-precision feature extraction.

[0058] Furthermore, before constructing the phase-encoded excitation flow containing axial polarization shear wave sequences and circumferential polarization shear wave sequences based on the basic parameters of ultrasonic testing, the method further includes: selecting a standard sample tube with the same material and in good condition as the stainless steel tube to be evaluated; testing the standard sample tube using the phase-encoded excitation flow to obtain the standard body nonlinear acoustic characteristic value and the standard interface modulation acoustic characteristic value; and setting a numerical threshold for determining whether the differential acoustic response index is within the reference range based on the statistical distribution data of the standard body nonlinear acoustic characteristic value and the standard interface modulation acoustic characteristic value.

[0059] Specifically, a standard sample tube of the same material and in good condition as the stainless steel tube to be evaluated is selected. Before conducting a formal evaluation of a batch of in-service stainless steel tubes, a benchmark calibration process is performed. This process uses a brand-new standard sample tube from the same manufacturing batch as the tube to be evaluated, confirmed to be free of any macroscopic or microscopic defects. This standard sample tube represents the ideal acoustic response benchmark for this type of stainless steel tube in a healthy and intact state.

[0060] The standard sample tube was tested using the phase-encoded excitation stream to obtain standard volume nonlinear acoustic characteristic values ​​and standard interface modulation acoustic characteristic values. The same entire method as the subsequent formal testing was applied to the standard sample tube. That is, the same phase-encoded excitation stream was emitted, its response signal was acquired and processed, and its volume nonlinear acoustic characteristic values ​​and interface modulation acoustic characteristic values ​​were calculated according to the methods defined in the aforementioned embodiments. To obtain statistically significant benchmark data, this testing process was repeated 20 times at different locations on the standard sample tube, resulting in 20 independent pairs of characteristic value data. These values ​​measured from the intact sample tube were defined as standard volume nonlinear acoustic characteristic values ​​and standard interface modulation acoustic characteristic values, respectively.

[0061] Based on the statistical distribution data of the standard body nonlinear acoustic characteristic values ​​and the standard interface modulation acoustic characteristic values, a numerical threshold is set to determine whether the differential acoustic response index is within the benchmark range. Using the 20 sets of standard characteristic value data obtained in the previous step, 20 corresponding standard differential acoustic response indices are calculated. Subsequently, statistical analysis is performed on the sample dataset composed of these 20 standard differential acoustic response indices to calculate their mean and standard deviation. In this embodiment, the calculated mean is 1.05 and the standard deviation is 0.25. To set a robust boundary that can effectively distinguish between normal fluctuations and true anomalies, the numerical threshold is set to the mean plus three times the standard deviation, i.e., 1.05 + (3 * 0.25) = 1.8. This calculated value of 1.8 is used as the final decision threshold for determining whether the differential acoustic response index is higher than the benchmark range in all subsequent evaluations. Being within the benchmark range means that the index value is less than or equal to 1.8.

[0062] By introducing this pre-calibration process based on standard sample tubes, an objective and reproducible basis for setting key judgment thresholds is provided, eliminating the subjectivity and uncertainty brought about by setting parameters based on experience, thereby significantly improving the accuracy and reliability of the entire quality assessment method.

[0063] Furthermore, the process of generating a structural noise suppression response spectrum by linearly superimposing the positive-phase burst response and the negative-phase burst response in the original detection response includes: reading the time-domain waveform data of the positive-phase burst response and the time-domain waveform data of the negative-phase burst response; performing multi-cycle time-domain synchronous averaging processing on the positive-phase burst response and the negative-phase burst response respectively to suppress random background noise; performing point-to-point amplitude addition on the processed positive-phase burst response and the negative-phase burst response, removing the fundamental component using the phase cancellation principle, and generating a nonlinear time-domain residual signal from the merged data; performing a fast Fourier transform on the nonlinear time-domain residual signal to map the time-domain information to the frequency domain, generating spectral data containing the fundamental component and higher harmonic components; and extracting the energy distribution data from the spectral data to define the structural noise suppression response spectrum.

[0064] Specifically, the time-domain waveform data of the positive-phase burst response and the negative-phase burst response are read, and multi-cycle time-domain synchronous averaging is performed on both responses to suppress random background noise. After the aforementioned phase-coded excitation current is applied to the stainless steel tube, the receiving coil of the electromagnetic ultrasonic transducer array captures a series of echo signals, i.e., the original detection response. The data acquisition card samples and performs analog-to-digital conversion on these analog signals at a preset sampling rate (100 MHz in this embodiment) to generate discrete time-domain waveform data. To improve the signal-to-noise ratio, the aforementioned four-segment phase-coded excitation current is repeatedly transmitted 256 times. Subsequently, the waveform data of the 256 positive-phase burst responses are aligned and averaged point by point. Similarly, the waveform data of the 256 negative-phase burst responses undergo the same time-domain synchronous averaging process. This processing utilizes the statistical inconsistency of random background noise, effectively reducing its amplitude through multiple averaging, thereby enhancing the weak acoustic response signal submerged in noise. This step is performed independently for the axially polarized and circumferentially polarized responses, respectively.

[0065] Point-to-point amplitude addition is performed on the processed positive-phase burst response and negative-phase burst response. The fundamental component is removed using the phase cancellation principle, and the merged data generates a nonlinear time-domain residual signal. The positive-phase burst response waveform data, after time-domain synchronous averaging, is linearly superimposed with the negative-phase burst response waveform data. This operation is a point-to-point addition, directly adding the amplitude values ​​of the two time series at every identical sampling point. According to the physical principles of sound wave propagation, the phase of the linear response component (fundamental wave) will synchronously reverse with the phase of the excitation source. Therefore, the fundamental component in the positive-phase response and the fundamental component in the negative-phase response have opposite phases and approximately equal amplitudes, canceling each other out after addition. The generation of nonlinear response components (such as second harmonics) is proportional to the square of the excitation source amplitude, and their phase does not reverse with the phase reversal of the excitation source. Therefore, they are enhanced by in-phase superposition after addition. The merged data obtained after the operation is the nonlinear time-domain residual signal, whose main component is the enhanced higher harmonic signal, while the fundamental component is significantly suppressed.

[0066] A Fast Fourier Transform (FFT) is performed on the nonlinear time-domain residual signal to map the time-domain information to the frequency domain, generating spectral data containing the fundamental component and higher harmonic components. A standard FFT algorithm is then used to process the nonlinear time-domain residual signal generated in the preceding steps. This transformation is a conventional technique for converting a signal from a time-domain representation to a frequency-domain representation. After processing, a spectral data structure is generated, which describes the distribution of signal energy at different frequencies. In this spectrum, although the fundamental component (center frequency 5 MHz) has been significantly suppressed by time-domain processing, residual energy may still exist; while the energy of higher harmonic components (such as the second harmonic, center frequency 10 MHz) becomes clearly discernible due to in-phase superposition.

[0067] The energy distribution data extracted from the spectral data is defined as the structural noise suppression response spectrum. The complete spectral data generated in the previous step is truncated, retaining only the frequency range of interest, for example, from 1 MHz to 20 MHz. The frequency distribution data and corresponding amplitude (or energy) within this range are collectively defined as the structural noise suppression response spectrum. Here, structural noise specifically refers to the strong fundamental wave signal generated by the linear response of the material that interferes with nonlinear signals. After the above series of processing steps, the signal-to-noise ratio of the nonlinear harmonic components of the resulting response spectrum is greatly improved, providing a high-quality data foundation for the subsequent accurate extraction of acoustic feature values. This process ultimately generates two independent structural noise suppression response spectra: one derived from axially polarized shear waves and the other from circumferentially polarized shear waves.

[0068] By executing the above signal processing procedure, the extremely weak, damage-related nonlinear information in the original detection response is effectively separated and enhanced from the strong background noise and linear structural response, forming a frequency domain representation with a high signal-to-noise ratio that can be used for quantitative analysis.

[0069] Further, the process of analyzing the structural noise suppression response spectrum and extracting the bulk nonlinear acoustic eigenvalues ​​of the corresponding axially polarized shear wave sequence and the interface modulation acoustic eigenvalues ​​of the circumferentially polarized shear wave sequence includes: locking the second harmonic frequency band range corresponding to the center frequency of the phase-encoded excitation flow in the structural noise suppression response spectrum; performing cumulative calculation on the spectral amplitude within the second harmonic frequency band range for the structural noise suppression response spectrum generated by the axially polarized shear wave sequence to obtain the total energy of higher harmonics, and performing relative intensity conversion in combination with the fundamental energy, defining the obtained dimensionless value as the bulk nonlinear acoustic eigenvalue, which is used to characterize the energy distortion caused by the accumulation of lattice anharmonicity when the sound wave propagates in the medium; calculating the amplitude integral within the second harmonic frequency band range for the structural noise suppression response spectrum generated by the circumferentially polarized shear wave sequence, and defining the normalized integral result as the interface modulation acoustic eigenvalue, which is used to characterize the local spectral modulation of the sound wave at the contact interface due to the breathing effect.

[0070] Specifically, the second harmonic frequency band corresponding to the center frequency of the phase-encoded excitation flow is locked in the structural noise suppression response spectrum. This is performed on the two structural noise suppression response spectra generated in the preceding steps, one originating from axial polarization and the other from circumferential polarization. Given that the center frequency of the excitation signal is 5 MHz, its second harmonic center frequency is theoretically located at 10 MHz. To include all relevant nonlinear energy and tolerate slight frequency shifts, a specific frequency band is set for analysis; in this embodiment, the second harmonic frequency band is locked to 9.5 MHz to 10.5 MHz. All subsequent feature extraction calculations are performed within this preset frequency band.

[0071] For the structural noise suppression response spectrum generated by the axially polarized shear wave sequence, a cumulative calculation is performed on the spectral amplitude within the second harmonic frequency band to obtain the total energy of higher harmonics. This energy is then combined with the fundamental energy for relative intensity conversion, and the resulting dimensionless value is defined as the bulk nonlinear acoustic characteristic value. In this axial spectrum, the spectral amplitude values ​​at all discrete frequency points within the 9.5 MHz to 10.5 MHz band are summed to obtain a cumulative value representing the total energy of the second harmonics. Subsequently, the total fundamental energy within the fundamental frequency band (4.5 MHz to 5.5 MHz in this embodiment) is calculated in the same manner. By dividing the total second harmonic energy by a normalization factor derived from the total fundamental energy, a relative intensity conversion is performed to obtain a dimensionless value representing the eliminated excitation energy fluctuation effect. This value is defined as the bulk nonlinear acoustic characteristic value, and its physical meaning lies in quantifying the intensity of the nonlinear effect caused by the lattice anharmonicity resulting from the accumulated micro-damage (such as dislocations and precipitates) inside the material when the sound wave propagates through the thickness of the stainless steel pipe wall.

[0072] For the structural noise suppression response spectrum generated by the circumferentially polarized shear wave sequence, the amplitude integral within the second harmonic frequency band is calculated. The normalized integral result is defined as the interface modulation acoustic characteristic value. In this circumferential spectrum, an integration operation is performed on the spectral curve within the 9.5 MHz to 10.5 MHz frequency band. This operation, on the discrete spectrum, is represented by summing the amplitude values ​​at all frequency points within the band. The integral result is also normalized based on the fundamental wave energy to ensure the stability of the characteristic value. The normalized result is defined as the interface modulation acoustic characteristic value, whose physical meaning lies in quantifying the local nonlinear modulation intensity generated by the "breathing effect" (alternating opening and closing under acoustic pressure) of the interface when the sound wave interacts with potential interface-characteristic defects (such as closed microcracks) in the pipe wall.

[0073] This process reduces the two complex structural noise suppression response spectra to two single values ​​with clear physical orientations, providing quantitative and distinguishable feature inputs for subsequent differentiation of the two damage modes: material constitutive anomaly and structural continuity anomaly.

[0074] Further, the step of normalizing the integration result and defining it as the volume nonlinear acoustic eigenvalue, and the step of normalizing the integration result and defining it as the interface modulation acoustic eigenvalue, specifically includes: locking the fundamental frequency band range corresponding to the center frequency of the phase-coded excitation flow in the structural noise suppression response spectrum; calculating the fundamental amplitude integral within the fundamental frequency band range, and defining the square of the fundamental amplitude integral as the energy normalization factor; performing a division operation with the amplitude integral within the second harmonic frequency band range as the numerator and the energy normalization factor as the denominator, and defining the operation results as the volume nonlinear acoustic eigenvalue and the interface modulation acoustic eigenvalue, respectively, and using the energy normalization factor to eliminate the eigenvalue deviation caused by the initial energy fluctuation of the phase-coded excitation flow.

[0075] Specifically, the fundamental frequency band range corresponding to the center frequency of the phase-encoded excitation current is locked within the structural noise suppression response spectrum. This operation is performed independently for the two structural noise suppression response spectra originating from axially polarized waves and circumferentially polarized waves. Given that the center frequency of the excitation signal is 5 MHz, and considering the potential for spectral broadening or slight frequency shifts in actual detection, to ensure complete capture of the fundamental energy, the fundamental frequency band range is set to fluctuate by 10% above and below the center frequency, i.e., 4.5 MHz to 5.5 MHz. Locking this frequency band is a prerequisite for subsequent accurate energy calculations.

[0076] The fundamental amplitude integral within the fundamental frequency band is calculated, and the square of the fundamental amplitude integral is defined as the energy normalization factor. Within the locked 4.5 MHz to 5.5 MHz frequency band, the amplitude values ​​corresponding to all frequency points of the discrete spectrum data are summed to obtain the fundamental amplitude integral within this band, which is internally labeled as parameter A1. Given that in acoustic physics, the energy intensity of a sound wave is proportional to the square of the sound pressure amplitude, to ensure that subsequent normalization processing accurately reflects the relative changes in energy, parameter A1 is multiplied (i.e., the square of A1 is calculated). This multiplied value is defined as the energy normalization factor, which, as a scalar, accurately characterizes the total fundamental energy that penetrates the stainless steel pipe wall and is effectively captured by the receiver, providing a unified reference standard for eliminating energy fluctuations during the detection process.

[0077] The amplitude integral within the second harmonic frequency band is used as the dividend, and the energy normalization factor is used as the divisor to perform a numerical division operation. This step is the core calculation step for achieving dimensionless nonlinear characteristics. First, the amplitude integral calculated in the previous step for the 9.5 MHz to 10.5 MHz second harmonic frequency band is read, marked as parameter A2, and set as the numerator of the operation; simultaneously, the energy normalization factor calculated in the previous step (i.e., the square value of parameter A1) is called and set as the denominator of the operation. Then, the numerical division operation is performed to obtain the quotient of parameter A2 and the square value of parameter A1. This operation is physically equivalent to calculating the ratio of the second harmonic response intensity to the fundamental frequency energy intensity, thereby mathematically eliminating the influence of the input excitation energy difference and generating a standard quantitative parameter that can independently characterize the nonlinearity of the medium.

[0078] The calculation results are defined as the bulk nonlinear acoustic eigenvalue and the interface modulation acoustic eigenvalue, respectively. The energy normalization factor is used to eliminate the eigenvalue deviation caused by the initial energy fluctuation of the phase-encoded excitation flow. The quotient obtained from the above division operation is the final eigenvalue. When A1 and A2 used in the calculation are both derived from the response spectrum of axially polarized waves, the quotient is defined as the bulk nonlinear acoustic eigenvalue. When A1 and A2 are derived from the response spectrum of circumferentially polarized waves, the quotient is defined as the interface modulation acoustic eigenvalue. The key function of this normalization method is that during field testing, any change in initial acoustic wave energy caused by changes in transducer contact pressure or fluctuations in instrument transmission power will proportionally affect A2 and A1. 2 .

[0079] By performing division, this common, damage-independent fluctuation is effectively canceled out, thereby ensuring that the final output eigenvalues ​​faithfully reflect only the intrinsic nonlinear properties of the material or interface, greatly improving the stability and repeatability of the evaluation results.

[0080] As a preferred embodiment, the process of analyzing the structural noise suppression response spectrum further includes: calculating at least one spectral morphology parameter of the spectrum for the second harmonic frequency band, wherein the spectral morphology parameter is spectral skewness or spectral kurtosis; and using the spectral morphology parameter together with the bulk nonlinear acoustic feature value and the interface modulation acoustic feature value as a combined feature for distinguishing damage modes.

[0081] Specifically, after calculating the bulk nonlinear acoustic eigenvalues ​​and interface modulation acoustic eigenvalues, the calculation of the differential exponent is not immediately initiated. Instead, an additional spectral morphology analysis subprocess is launched. This subprocess analyzes the structural noise suppression response spectrum originating from circumferentially polarized shear waves, focusing on its second harmonic frequency band (9.5 MHz to 10.5 MHz).

[0082] In this analysis sub-process, this method calls a standard statistical calculation function to process discrete spectral data within the second harmonic frequency band and calculate its spectral skewness. Spectral skewness is a dimensionless statistic used to describe the asymmetry of spectral energy distribution.

[0083] The physical significance of this spectral parameter lies in the fact that the nonlinear response generated by the homogeneous degradation of the material's lattice anharmonicity should theoretically exhibit a symmetrical distribution of second harmonic energy around the center frequency (10 MHz), with a spectral skewness close to 0. However, the contact acoustic nonlinearity generated by the opening and closing of microcrack interfaces under the action of sound waves (breathing effect) has a more complex physical process, leading to asymmetrical energy leakage to the high-frequency end, thus causing the second harmonic spectrum to exhibit a positive skewness, with a significantly positive skewness value.

[0084] Therefore, this spectral skewness value will be introduced as an independent feature dimension during the final evaluation. For example, if the interface modulation acoustic feature value of a measurement point is abnormally high and its second harmonic spectral skewness is calculated to be 0.8 (significantly greater than the preset threshold of 0.5), this method will determine it as a structural continuity anomaly dominated by interface contact with a higher confidence level.

[0085] This scheme, by introducing quantitative analysis of the "shape" of the harmonic spectrum, provides a second dimension of information independent of energy magnitude for identifying the microscopic physical mechanisms of nonlinear sources, greatly enhancing the accuracy of identifying damage modes in fuzzy and critical states.

[0086] Further, the process of calculating the ratio of the interface modulation acoustic feature value to the bulk nonlinear acoustic feature value to generate the differential acoustic response index includes: taking the interface modulation acoustic feature value as the numerator and the bulk nonlinear acoustic feature value as the denominator, and performing a division operation; defining the quotient obtained by the operation as the differential acoustic response index; the differential acoustic response index is used to characterize the anisotropy of the nonlinear source in the original detection response. When the differential acoustic response index approaches a reference value, it indicates that the nonlinear source is an isotropic source; when the differential acoustic response index is higher than the reference value, it indicates that the nonlinear source is an interface contact source with anisotropic characteristics.

[0087] Specifically, the interface-modulated acoustic characteristic value is used as the numerator, and the bulk nonlinear acoustic characteristic value is used as the denominator, and a division operation is performed. The two independent, dimensionless scalar values ​​obtained in the preceding steps are used as inputs for this step. Specifically, the interface-modulated acoustic characteristic value, characterizing the intensity of the interaction between the sound wave and the potential interface, is placed in the numerator of the division operation; the bulk nonlinear acoustic characteristic value, characterizing the degree of distortion caused by the constitutive nonlinearity of the sound wave propagating in the material matrix, is placed in the denominator, and a numerical division operation is performed.

[0088] The quotient obtained from the calculation is defined as the differential acoustic response index. The quotient obtained from the aforementioned division operation is given a specific technical name. The differential acoustic response index is a relative ratio obtained by comparing the nonlinear response intensity excited by circumferentially polarized shear waves (sensitive to interfaces) and axially polarized shear waves (sensitive to bulk phases). The purpose of this index is to amplify the difference in sensitivity between the two orthogonally polarized acoustic waves when detecting different types of defects, thereby providing a quantitative basis for qualitatively determining the source of the nonlinear response.

[0089] The differential acoustic response index is used to characterize the anisotropy of the nonlinear source in the original detection response. The physical meaning of this index lies in revealing the geometric characteristics of the nonlinear source. When the damage mode of the stainless steel pipe is uniform material fatigue or micro-phase transformation, its physical property degradation is approximately uniform in all directions, constituting an isotropic nonlinear source. In this case, the nonlinear response intensity excited by axial and circumferential shear waves is of the same order of magnitude, resulting in the bulk nonlinear acoustic characteristic value being similar in magnitude to the interface modulation acoustic characteristic value. The calculated differential acoustic response index will approach a reference value (for example, theoretically close to 1.0 for healthy materials). Conversely, when the damage mode is an axially propagating microcrack, the crack interface constitutes an anisotropic nonlinear source with a clear direction. The circumferentially polarized shear wave generated by controlling the circumferential coil group with intelligent sensors can effectively drive the crack interface to open and close because the direction of particle vibration is perpendicular to the opening direction of the axial crack, thus generating a strong interface nonlinear effect and causing a significant increase in the interface modulation acoustic characteristic value. On the other hand, the axially polarized shear wave, because the direction of particle vibration is parallel to the crack direction, mainly undergoes interlaminar slip rather than opening and closing, and is therefore not sensitive to this type of crack, and its bulk nonlinear acoustic characteristic value does not change significantly.

[0090] By calculating the ratio of the interface to the volumetric nonlinear acoustic eigenvalues, this step generates a single index with high discriminative power for the geometry of nonlinear sources. This simplifies the complex dual-feature analysis into a threshold judgment of a single index, providing a direct and physically meaningful decision-making basis for accurately distinguishing between constitutive anomalies of the medium and structural continuity anomalies.

[0091] Furthermore, referring to Figure 2The quality comprehensive evaluation model is a two-dimensional state-space logical classification algorithm. The process of inputting the differential acoustic response index and the volume nonlinear acoustic feature value into the quality comprehensive evaluation model includes: constructing a two-dimensional acoustic state space containing two orthogonal dimensions, wherein the first dimension represents the degree of microscopic distortion of the medium, and the second dimension represents the degree of anisotropy of the nonlinear source; dividing the two-dimensional acoustic state space into a preset safety reference region, a volume distribution nonlinearity-dominant region, and an interface contact nonlinearity-dominant region; using the volume nonlinear acoustic feature value as the first feature component and the differential acoustic response index as the second feature component, combining them to form a feature vector; projecting the feature vector onto the two-dimensional acoustic state space, determining the coordinate position, and identifying the region in which the coordinate position falls within the two-dimensional acoustic state space.

[0092] Specifically, a two-dimensional acoustic state space with two orthogonal dimensions is constructed. The first dimension characterizes the degree of microscopic distortion of the medium, and the second dimension characterizes the degree of anisotropy of the nonlinear source. This step defines a mathematical space for logical classification. This space is a two-dimensional Cartesian coordinate system, where the horizontal axis (first dimension) is designated as the volume nonlinear acoustic eigenvalue, the magnitude of which directly reflects the degree of microstructural distortion caused by factors such as fatigue accumulation within the material. The vertical axis (second dimension) is designated as the differential acoustic response index, the magnitude of which quantifies the anisotropy intensity of the detected nonlinear signal source in the geometric direction.

[0093] In a two-dimensional acoustic state space, a preset safety reference region, a volume distribution nonlinearity-dominant region, and an interface contact nonlinearity-dominant region are defined. This step involves dividing the aforementioned two-dimensional space into three non-overlapping regions based on preset numerical boundaries. Each region corresponds to a specific stainless steel pipe quality state. These numerical boundaries are set based on statistical data obtained from testing a large number of intact sample pipes and sample pipes with known defects. In this embodiment, a numerical threshold of 5.0 is set to determine whether the volume nonlinear acoustic characteristic value is abnormally elevated, and a numerical threshold of 1.8 is set to determine whether the differential acoustic response index is higher than the reference range. Based on these two thresholds, the region division rules are as follows:

[0094] Safety reference area: This area is defined by a two-dimensional spatial range in which the volumetric nonlinear acoustic eigenvalue does not exceed the numerical threshold of 5.0 and the differential acoustic response index does not exceed the numerical threshold of 1.8.

[0095] Volumetric nonlinearity-dominant region: This region is defined by the spatial range in which the volumetric nonlinear acoustic eigenvalues ​​have exceeded the numerical threshold of 5.0, but the differential acoustic response index has not yet exceeded the numerical threshold of 1.8.

[0096] Nonlinearity-dominant region of interface contact: This region is defined by all spatial ranges where the differential acoustic response index exceeds the numerical threshold of 1.8.

[0097] To construct an evaluation input with strict physical correspondence and consistent data dimensions, the bulk nonlinear acoustic feature values ​​calculated for a specific measuring point of the stainless steel pipe to be evaluated within a single detection command cycle, and the differential acoustic response index calculated based on the same set of original detection response data, are first extracted. To ensure that these two physically different parameters are mathematically equivalent when constructing the feature vector, they need to undergo strict dimensionless standardization.

[0098] The specific processing steps are as follows: First, read the currently detected volumetric nonlinear acoustic feature value, subtract the average value of the feature value recorded in the historical baseline statistical model, and then divide the difference by the standard deviation of the feature value in the historical baseline statistical model to obtain the dimensionless first feature component; simultaneously, read the currently detected differential acoustic response index, subtract the average value of the index recorded in the historical baseline statistical model, and then divide the difference by the standard deviation of the index in the historical baseline statistical model to obtain the dimensionless second feature component.

[0099] After ensuring that both the first and second feature components have been converted into relative deviation values ​​on the same statistical scale, a dimension concatenation operation is performed. The first feature component is mapped to the first dimension (usually the horizontal axis), and the second feature component is mapped to the second dimension (usually the vertical axis), combining them to form a dimensionless two-dimensional feature vector that only represents the degree of state deviation. This serves as the digital representation of the acoustic state of the measurement point.

[0100] The feature vector is projected into a two-dimensional acoustic state space to determine its coordinate position and identify the region within which the coordinate position falls. The coordinate point (7.2, 1.3) generated in the previous step is mapped onto the pre-divided two-dimensional acoustic state space. Through logical judgment, since the first dimension component 7.2 of this point exceeds the threshold of 5.0 set for the volume nonlinear acoustic feature value, while its second dimension component 1.3 does not exceed the threshold of 1.8 set for the differential acoustic response index, this point is determined to fall into the region dominated by volume nonlinearity. This identification result will serve as the direct input for the final quality assessment conclusion. The two-dimensional state space logical classification algorithm refers to the method of classifying and determining the quality status of stainless steel pipes through the complete process of constructing space, dividing regions, projecting feature vectors, and determining the region.

[0101] By mapping two independent acoustic feature values ​​to a predefined two-dimensional state space, this method transforms a complex numerical analysis problem into an intuitive geometric region judgment problem, making the classification decision logically clear and robust.

[0102] Furthermore, referring to Figure 2 The process of outputting the structural continuity anomaly quality assessment result includes: when the state point falls into the volume distribution nonlinearity-dominant region, determining that the nonlinear response of the stainless steel tube to be evaluated originates from the degradation of the microscopic physical properties of the medium, and outputting the volume distribution nonlinearity-dominant state identification result; when the state point falls into the interface contact nonlinearity-dominant region, determining that the nonlinear response of the stainless steel tube to be evaluated originates from the closed interface contact friction, and outputting the interface contact nonlinearity-dominant state identification result.

[0103] Specifically, when the state point falls within the volume distribution nonlinearity-dominant region, the nonlinear response of the stainless steel tube to be evaluated is determined to originate from the degradation of the microscopic physical properties of the medium, and the volume distribution nonlinearity-dominant state identification result is output. This is the final physical interpretation and conclusion output of the aforementioned classification results. Continuing the example in the previous embodiment, the feature vector coordinate point (7.2, 1.3) is determined to fall within the volume distribution nonlinearity-dominant region. The logical basis for this determination is that its differential acoustic response index (1.3) is within the preset benchmark range (not exceeding the threshold of 1.8), indicating that the source of nonlinearity does not have significant anisotropy in the direction; at the same time, its volume nonlinear acoustic characteristic value (7.2) has abnormally increased (exceeding the threshold of 5.0), indicating that the overall nonlinear response intensity of the material has significantly increased. The physical root of this combination of features is determined to be the degradation of the uniformly distributed microscopic physical properties of the medium caused by long-term service, such as the general increase in lattice anharmonicity. Therefore, the method finally outputs the medium constitutive anomaly as the quality assessment result of the stainless steel tube.

[0104] When the state point falls within the interface contact nonlinearity-dominant region, the nonlinear response of the stainless steel pipe to be evaluated is determined to originate from closed interface contact friction, and the interface contact nonlinearity-dominant state identification result is output. As an example of another case, suppose the feature vector coordinate point obtained from another detection is (4.5, 2.5). This point will be determined to fall within the interface contact nonlinearity-dominant region because its differential acoustic response index (2.5) is higher than the preset benchmark range (exceeding the threshold 1.8). This result indicates that the source of the nonlinear response has a strong direction dependence, i.e., anisotropy. The physical mechanism of this phenomenon is determined to be that there is a defect with interface characteristics (such as a closed microcrack) in the pipe wall, which generates strong contact friction and "breathing effect" under the action of circumferential polarized shear waves, thereby generating an interface modulation nonlinear signal that is much stronger than the bulk nonlinear effect. Therefore, the method finally outputs the structural continuity anomaly as the quality assessment result of the stainless steel pipe.

[0105] This step accurately transforms the location of abstract regions in two-dimensional space into a final quality assessment conclusion with clear physical meaning and engineering guidance value, realizing the automated and quantitative differentiation of the two core early damage modes.

[0106] In a preferred embodiment, the comprehensive quality evaluation model is also used to: store and update historical feature vectors for specific measurement points of the stainless steel pipe to be evaluated, so as to construct a dynamic baseline statistical model characterizing its health status; calculate the Mahalanobis distance between the currently detected feature vector and the dynamic baseline statistical model; and output an early performance degradation warning when the Mahalanobis distance exceeds a preset statistical confidence threshold, even if it has not yet fallen into any preset abnormal area.

[0107] Specifically, firstly, a model building warm-up period is set for a specific measuring point (e.g., measuring point A-01 on the field pipeline). During the warm-up period (e.g., the first thirty measurements), only the feature vectors obtained from each measurement are stored to build a historical sample dataset, without performing bias calculations. Once the number of samples reaches a preset statistical significance threshold, the initial mean vector and covariance matrix are calculated based on this dataset, thereby constructing an initial health baseline statistical model specific to that measuring point.

[0108] Secondly, in each routine inspection after the warm-up period (e.g., six months later), this method invokes the existing health baseline statistical model and uses its inverse covariance matrix to calculate the Mahalanobis distance between the newly measured eigenvector (e.g., [3.5, 1.2]) and the model's mean center. After distance calculation and evaluation, a recursive update strategy is used to integrate the new eigenvector into the historical database, dynamically correcting the model's mean vector and covariance matrix. In this way, the model can adaptively follow the natural aging drift of stainless steel pipes in a healthy state while maintaining sensitivity to sudden anomalies.

[0109] In each detection step, this method compares the calculated Mahalanobis distance with a pre-set threshold based on statistical confidence (e.g., 95%). Mahalanobis distance, as an effective statistical distance that considers inter-feature correlations and is scale-insensitive, can capture subtle multidimensional changes. If the Mahalanobis distance of the current point exceeds this threshold, even if its coordinates remain within the safe baseline region in the two-dimensional state space, its state is considered to have undergone a statistically significant drift. In this case, the method will output an additional early performance degradation warning message next to the main evaluation conclusion.

[0110] This solution upgrades the evaluation model from a static fault diagnosis classifier to a dynamic health trend monitor, enabling it to capture slowly accumulating performance degradation far below a fixed fault threshold, thus achieving a leap from passive diagnosis to proactive early warning.

[0111] The method disclosed in this invention significantly improves the detection sensitivity for early, weak, nonlinear damage signals by constructing a phase-encoded excitation flow containing orthogonally polarized shear waves and utilizing the superposition technique of positive and negative phase responses to efficiently suppress strong structural noise. Furthermore, by extracting acoustic features specifically sensitive to bulk material and interface defects and calculating their differential response exponents, this method successfully achieves accurate qualitative differentiation between two key damage modes: constitutive anomalies and structural continuity anomalies, which exhibit drastically different properties. This solves the technical challenge of traditional detection methods in identifying the source of damage. Simultaneously, the method employs a normalization process based on fundamental wave energy and a pre-calibration procedure using standard sample tubes, effectively eliminating uncertainties caused by fluctuations in measurement conditions and ensuring that the final evaluation conclusion possesses high sensitivity, high reliability, and clear physical directionality.

[0112] Example 2:

[0113] This invention provides a stainless steel pipe quality assessment system based on detection response. This system can be materialized as an integrated non-destructive testing device. Its core typically includes an electromagnetic ultrasonic transducer probe coupled to the surface of the stainless steel pipe to be evaluated, and a portable industrial computer with built-in data processing and analysis software as the system host. The various functional modules of the system and their collaborative workflow are shown below, with reference to... Figure 3 .

[0114] The excitation construction module is functionally implemented by the system host's control software and a high-precision arbitrary waveform generator. The operator first inputs the basic parameters of the stainless steel pipe to be evaluated through the host's human-machine interface, such as design document data showing a pipe wall thickness of 5.0 mm and a sound velocity anisotropy coefficient of 1.05. After receiving these parameters, the module's internal logic calculates the optimal excitation signal center frequency (5 MHz in this example) and pulse width. Subsequently, the control software instructs the waveform generator to generate a precise, four-segment phase-coded electrical signal sequence. This sequence is sent to the electromagnetic ultrasonic transducer probe, sequentially driving its axial and circumferential coil groups to generate axial and circumferential polarized shear waves containing a positive-phase burst at zero degrees and an anti-phase burst at 180 degrees, respectively. This physical ultrasonic sequence constitutes the phase-coded excitation current injected into the stainless steel pipe to be evaluated.

[0115] The signal processing module is functionally comprised of the receiving coil of the electromagnetic ultrasonic transducer probe, a high-speed data acquisition card, and signal processing software within the system host. When the phase-encoded excitation current propagates within the tube wall and interacts with the internal structure, the resulting echo is induced as a weak voltage signal by the receiving coil; this is the original detection response. The data acquisition card digitizes this analog signal at a sampling rate of 100 MHz and transmits it to the host. The signal processing software then starts. First, it performs time-domain synchronous averaging on the 256 repeatedly acquired response data sets to suppress random electromagnetic noise. Next, the software performs point-to-point linear superposition of the averaged positive-phase and negative-phase response waveforms, using the phase cancellation principle to remove the fundamental component and generate a nonlinear time-domain residual signal. Finally, the module performs a fast Fourier transform on the residual signal, converting it to the frequency domain to generate two independent structural noise suppression response spectra, corresponding to the axial and circumferential responses respectively, which are then passed as data structures to the next module.

[0116] The feature analysis module is a pure software module running on the system host. This module receives two structure noise suppression response spectra generated by the signal processing module as input. Its internal algorithm first locks the fundamental frequency band (4.5-5.5 MHz) with a center frequency of 5 MHz and the second harmonic frequency band (9.5-10.5 MHz) with a center frequency of 10 MHz in the two spectra according to preset rules. Then, it calculates the fundamental amplitude integral A1 and the second harmonic amplitude integral A2 in each spectrum. Based on A2 / (A1... 2 The module uses a normalization formula to define the calculation results derived from the axial response spectrum as volume nonlinear acoustic eigenvalues ​​and the calculation results derived from the circumferential response spectrum as interface modulation acoustic eigenvalues. These two final calculated dimensionless scalar values ​​are then packaged and output.

[0117] The index calculation module is also a pure software module running on the system host. It receives the volume nonlinear acoustic eigenvalues ​​and interface modulation acoustic eigenvalues ​​output by the feature analysis module. The sole task of this module is to perform a simple division operation: using the interface modulation acoustic eigenvalue as the numerator and the volume nonlinear acoustic eigenvalue as the denominator. The resulting quotient is defined by this module as the differential acoustic response index and passed to the state recognition module.

[0118] The state recognition module is the core software module for system decision-making. This module receives the differential acoustic response index generated by the index calculation module and the volume nonlinear acoustic eigenvalues ​​generated by the feature analysis module. Internally, it incorporates a two-dimensional state-space logical classification algorithm, which includes pre-calibrated decision thresholds using standard sample tubes (e.g., a threshold of 1.8 for the differential acoustic response index and 5.0 for the volume nonlinear acoustic eigenvalues). When the input differential acoustic response index is within the baseline range (≤1.8) and the volume nonlinear acoustic eigenvalue is abnormally high (>5.0), the module determines that the state point falls within the region dominated by volume distribution nonlinearity and ultimately outputs a quality assessment result of medium constitutive anomaly on the system host's display screen. When the input differential acoustic response index is higher than the baseline range (>1.8), the module determines that the state point falls within the region dominated by interface contact nonlinearity and outputs a quality assessment result of structural continuity anomaly.

[0119] In summary, the system in this embodiment automates the complex physical detection, signal processing, and feature analysis process through the orderly collaboration of its modules, ultimately providing operators with an intuitive, clear, and reliable assessment of the internal damage patterns of stainless steel pipes.

[0120] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the scope of protection defined in the claims.

Claims

1. A method for quality assessment of stainless steel pipes based on detection response, characterized in that, include: Obtain the basic ultrasonic testing parameters of the stainless steel pipe to be evaluated, and construct a phase-coded excitation flow containing axial polarization shear wave sequences and circumferential polarization shear wave sequences based on the basic ultrasonic testing parameters. The axial polarization shear wave sequences and circumferential polarization shear wave sequences contain positive-phase bursts and negative-phase bursts. The original detection response of the stainless steel tube to be evaluated under the action of phase-encoded excitation current is collected. The positive-phase burst response and the anti-phase burst response in the original detection response are linearly superimposed to generate the structural noise suppression response spectrum. Analyze the structural noise suppression response spectrum and extract the bulk nonlinear acoustic eigenvalues ​​of the corresponding axially polarized shear wave sequence and the interface modulation acoustic eigenvalues ​​of the circumferentially polarized shear wave sequence. The ratio of the interface modulation acoustic eigenvalue to the bulk nonlinear acoustic eigenvalue is calculated to generate the differential acoustic response index. The differential acoustic response index and the volume nonlinear acoustic eigenvalue are input into the comprehensive quality evaluation model. When the differential acoustic response index is within the benchmark range and the volume nonlinear acoustic eigenvalue increases abnormally, the model outputs the quality assessment result of the constitutive anomaly of the medium. When the differential acoustic response index is higher than the benchmark range, the model outputs the quality assessment result of the structural continuity anomaly.

2. The method for quality assessment of stainless steel pipes based on detection response according to claim 1, characterized in that, The process of constructing a phase-encoded excitation flow containing axially polarized shear wave sequences and circumferentially polarized shear wave sequences based on ultrasonic testing fundamental parameters includes: obtaining the wall thickness parameters and sound velocity anisotropy coefficient of the stainless steel pipe to be evaluated as ultrasonic testing fundamental parameters; loading the ultrasonic testing fundamental parameters through a smart sensor to drive an electromagnetic ultrasonic transducer array integrated within the smart sensor; using the smart sensor to control the axial coil group in the electromagnetic ultrasonic transducer array to generate shear waves with polarization directions parallel to the pipe axis as the axially polarized shear wave sequence; using the smart sensor to control the circumferential coil group in the electromagnetic ultrasonic transducer array to generate shear waves with polarization directions perpendicular to the pipe axis as the circumferentially polarized shear wave sequence; applying zero-degree phase modulation to the axially polarized shear wave sequence and the circumferentially polarized shear wave sequence respectively to generate the positive-phase burst sound, and applying 180-degree phase modulation to generate the anti-phase burst sound, combining them to construct the phase-encoded excitation flow.

3. The method for quality assessment of stainless steel pipes based on detection response according to claim 1, characterized in that, Before constructing a phase-encoded excitation flow containing axially polarized shear wave sequences and circumferentially polarized shear wave sequences based on the basic parameters of ultrasonic testing, the process further includes: selecting a standard sample tube with the same material and in good condition as the stainless steel tube to be evaluated; testing the standard sample tube using the phase-encoded excitation flow to obtain standard body nonlinear acoustic characteristic values ​​and standard interface modulation acoustic characteristic values; and setting a numerical threshold for determining whether the differential acoustic response index is within a reference range based on the statistical distribution data of the standard body nonlinear acoustic characteristic values ​​and the standard interface modulation acoustic characteristic values.

4. The method for quality assessment of stainless steel pipes based on detection response according to claim 1, characterized in that, The process of generating a structural noise suppression response spectrum by linearly superimposing the positive-phase burst response and the negative-phase burst response in the original detection response includes: reading the time-domain waveform data of the positive-phase burst response and the time-domain waveform data of the negative-phase burst response; performing multi-cycle time-domain synchronous averaging processing on the positive-phase burst response and the negative-phase burst response respectively to suppress random background noise; performing point-to-point amplitude addition on the processed positive-phase burst response and the negative-phase burst response, removing the fundamental component using the phase cancellation principle, and generating a nonlinear time-domain residual signal from the merged data; performing a fast Fourier transform on the nonlinear time-domain residual signal to map the time-domain information to the frequency domain, generating spectral data containing the fundamental component and higher harmonic components; and extracting the energy distribution data from the spectral data to define the structural noise suppression response spectrum.

5. The method for quality assessment of stainless steel pipes based on detection response according to claim 1, characterized in that, The process of analyzing the structural noise suppression response spectrum and extracting the bulk nonlinear acoustic eigenvalues ​​of the corresponding axially polarized shear wave sequence and the interface modulation acoustic eigenvalues ​​of the circumferentially polarized shear wave sequence includes: locking the second harmonic frequency band range corresponding to the center frequency of the phase-encoded excitation current in the structural noise suppression response spectrum; performing cumulative calculation on the spectral amplitude within the second harmonic frequency band range for the structural noise suppression response spectrum generated by the axially polarized shear wave sequence to obtain the total energy of higher harmonics, and performing relative intensity conversion in combination with the fundamental energy, defining the obtained dimensionless value as the bulk nonlinear acoustic eigenvalue, which is used to characterize the energy distortion caused by the accumulation of lattice anharmonicity when the sound wave propagates in the medium; calculating the amplitude integral within the second harmonic frequency band range for the structural noise suppression response spectrum generated by the circumferentially polarized shear wave sequence, and defining the normalized integral result as the interface modulation acoustic eigenvalue, which is used to characterize the local spectral modulation of the sound wave at the contact interface due to the breathing effect.

6. The method for quality assessment of stainless steel pipes based on detection response according to claim 5, characterized in that, The process of normalizing the integration result and defining it as the volume nonlinear acoustic eigenvalue and the process of normalizing the integration result and defining it as the interface modulation acoustic eigenvalue specifically includes: locking the fundamental frequency band range corresponding to the center frequency of the phase-coded excitation flow in the structural noise suppression response spectrum; calculating the fundamental amplitude integral within the fundamental frequency band range and defining the square of the fundamental amplitude integral as the energy normalization factor; performing a division operation with the amplitude integral within the second harmonic frequency band range as the numerator and the energy normalization factor as the denominator, and defining the operation results as the volume nonlinear acoustic eigenvalue and the interface modulation acoustic eigenvalue, respectively, and using the energy normalization factor to eliminate the eigenvalue deviation caused by the initial energy fluctuation of the phase-coded excitation flow.

7. The method for quality assessment of stainless steel pipes based on detection response according to claim 1, characterized in that, The process of generating a differential acoustic response index by calculating the ratio of the interface modulation acoustic feature value to the bulk nonlinear acoustic feature value includes: performing a division operation with the interface modulation acoustic feature value as the numerator and the bulk nonlinear acoustic feature value as the denominator; defining the quotient as the differential acoustic response index; the differential acoustic response index is used to characterize the anisotropy of the nonlinear source in the original detection response; when the differential acoustic response index approaches a reference value, it indicates that the nonlinear source is an isotropic source; when the differential acoustic response index is higher than the reference value, it indicates that the nonlinear source is an interface contact source with anisotropic characteristics.

8. The method for quality assessment of stainless steel pipes based on detection response according to claim 1, characterized in that, The comprehensive quality evaluation model is a two-dimensional state-space logical classification algorithm. The process of inputting the differential acoustic response index and the volume nonlinear acoustic feature value into the comprehensive quality evaluation model includes: constructing a two-dimensional acoustic state space containing two orthogonal dimensions, wherein the first dimension represents the degree of microscopic distortion of the medium, and the second dimension represents the degree of anisotropy of the nonlinear source; dividing the two-dimensional acoustic state space into a preset safety reference region, a volume distribution nonlinearity-dominant region, and an interface contact nonlinearity-dominant region; using the volume nonlinear acoustic feature value as the first feature component and the differential acoustic response index as the second feature component, combining them to form a feature vector; projecting the feature vector onto the two-dimensional acoustic state space, determining the coordinate position, and identifying the region in which the coordinate position falls within the two-dimensional acoustic state space.

9. The method for quality assessment of stainless steel pipes based on detection response according to claim 8, characterized in that, The process of outputting the structural continuity anomaly quality assessment result includes: when the state point falls into the volume distribution nonlinearity-dominant region, determining that the nonlinear response of the stainless steel tube to be evaluated originates from the degradation of the microscopic physical properties of the medium, and outputting the volume distribution nonlinearity-dominant state identification result; when the state point falls into the interface contact nonlinearity-dominant region, determining that the nonlinear response of the stainless steel tube to be evaluated originates from closed interface contact friction, and outputting the interface contact nonlinearity-dominant state identification result.

10. A stainless steel pipe quality assessment system based on detection response, characterized in that, include: An excitation construction module is used to obtain the basic ultrasonic testing parameters of the stainless steel pipe to be evaluated, and to construct a phase-encoded excitation flow containing an axial polarization shear wave sequence and a circumferential polarization shear wave sequence based on the basic ultrasonic testing parameters. The axial polarization shear wave sequence and the circumferential polarization shear wave sequence contain positive-phase bursts and negative-phase bursts. The signal processing module is used to acquire the original detection response of the stainless steel tube to be evaluated under the action of the phase-encoded excitation current, and to perform linear superposition of the positive-phase burst response and the negative-phase burst response in the original detection response to generate the structural noise suppression response spectrum. The feature analysis module is used to analyze the structural noise suppression response spectrum and extract the bulk nonlinear acoustic feature values ​​of the corresponding axial polarization shear wave sequence and the interface modulation acoustic feature values ​​of the circumferential polarization shear wave sequence. The index calculation module is used to calculate the ratio of the interface modulation acoustic characteristic value to the bulk nonlinear acoustic characteristic value, and generate the differential acoustic response index. The state recognition module is used to input the differential acoustic response index and the volume nonlinear acoustic characteristic value into the comprehensive quality evaluation model. When the differential acoustic response index is within the benchmark range and the volume nonlinear acoustic characteristic value is abnormally increased, it outputs the medium constitutive anomaly quality assessment result; when the differential acoustic response index is higher than the benchmark range, it outputs the structural continuity anomaly quality assessment result.

Images