A pipe thickness detection method and system based on ultrasonic waves

Abstract

The application discloses a pipe thickness detection method and system based on ultrasonic waves, stores original waveform data to a database, carries out normalization conversion on the original waveform data, maps the original waveform data to a fixed amplitude reference and a standard time coordinate system, obtains multiple sets of normalized waveforms, extracts waveform time domain features and amplitude features from the multiple sets of normalized waveforms, identifies abnormal waveforms through waveform time domain similarity analysis, eliminates the abnormal waveforms, screens effective waveform sample sets, calls a pre-constructed pipe echo feature matching model, compares features of each waveform in the effective waveform sample sets, accurately locates feature time points of pipe outer surface echoes and pipe bottom inner wall echoes, calculates pipe thickness according to a time difference between the outer surface echoes and the bottom inner wall echoes and in combination with a pipe material sound velocity, and statistically corrects effective sample calculation results to output final pipe wall thickness measurement results. The application has the characteristics of high accuracy and strong applicability.

Description

Technical Field

[0001] This invention relates to the field of pipe thickness detection technology, specifically to a pipe thickness detection method and system based on ultrasound. Background Technology

[0002] Pipe fittings are used to transport fluids and have a wide range of applications. When high-temperature, high-pressure, or corrosive fluids are transported inside pipe fittings, their durability is tested. Therefore, to ensure that manufactured pipe fittings meet requirements, it is usually necessary to use measuring tools to measure aspects such as the pipe wall thickness.

[0003] Existing technologies for pipe thickness detection based on ultrasonic feedback waveforms are prone to inconsistencies in the data waveforms during the detection process. This means the data waveforms fluctuate, and existing technologies cannot accurately analyze these waveform changes, leading to inconsistent measurement results for the same pipe thickness. This results in low accuracy in pipe thickness measurement. Furthermore, existing technologies often use complementary equipment for pipe thickness measurement. Since the data output from different ultrasonic devices is inconsistent, existing technologies cannot accurately analyze the data waveforms, leading to measurement errors. Therefore, it is essential to design a more accurate and versatile ultrasonic-based pipe thickness detection method and system. Summary of the Invention

[0004] The purpose of this invention is to provide an ultrasonic-based method and system for detecting pipe thickness, in order to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a pipe thickness detection method based on ultrasound, comprising: An ultrasonic pulse is emitted to the pipe under test, and multiple sets of raw waveform data of ultrasonic feedback echoes reflected from the pipe wall are received and stored in the database. The original waveform data is normalized and mapped to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms. Waveform time-domain features and amplitude features are extracted from multiple sets of normalized waveforms. Abnormal waveforms are identified through waveform time-domain similarity analysis, and abnormal waveforms are removed to obtain a valid waveform sample set. The pre-built pipe echo feature matching model is invoked to compare the features of each waveform in the effective waveform sample set, accurately locate the characteristic time points of the echo on the outer surface of the pipe and the echo on the inner wall of the bottom surface of the pipe. Based on the time difference between the echo on the outer surface and the echo on the inner wall of the bottom surface, combined with the sound velocity of the pipe material, the pipe thickness is calculated. The calculation results of the effective samples are statistically corrected, and the final pipe wall thickness measurement result is output.

[0006] According to the above technical solution, the normalization transformation of the original waveform data, mapping the original waveform data to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms, includes the following steps: Extracting the amplitude sequence of a single raw echo waveform , The total number of sampling points for a single waveform is used to scan the original waveform and identify the maximum amplitude. and minimum amplitude Calculate the amplitude range ; Identify waveform change characteristics, segment the waveform, mark the no-signal sampling interval after ultrasonic wave emission and before the arrival of effective echo as the baseline noise region, and mark the signal interval of pulse echo as the effective echo region; Calculate the mean amplitude of the baseline noise region Calculate the waveform asymmetry coefficient The signal power in the baseline noise region and the effective echo region is identified respectively, and the waveform signal-to-noise ratio (SNR) is calculated. When the waveform signal-to-noise ratio is higher than the preset threshold and the waveform asymmetry coefficient is greater than the preset threshold, the waveform is determined to be peak signal dominant and the maximum value normalization working mode is enabled; when the waveform signal-to-noise ratio is lower than the preset threshold and the waveform asymmetry coefficient is less than the preset threshold, the waveform is determined to have baseline offset interference and the minimum value normalization working mode is enabled. If the waveform amplitude range is extremely large If the waveform is less than the preset minimum threshold, it is determined that the current waveform is an invalid DC waveform with no effective echo. Normalization is not performed, and the waveform is marked and removed.

[0007] According to the above technical solution, the normalization transformation of the original waveform data, mapping the original waveform data to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms, further includes the following steps: Maximum value normalization working mode: adopts the normalization calculation formula. , in, Indicates the first element in the original waveform. The original amplitude of each sampling point , This represents the total number of sampling points for this echo waveform, calculated using the global maximum value for a single waveform. Using this as the sole scaling reference, the amplitudes of all sampling points of the waveform are scaled proportionally, while the peak amplitude of the waveform is fixed and normalized to 1. The corresponding scaling factor is then used to map the remaining sampling points to their respective amplitudes. interval; Minimum value normalization working mode: uses the normalization calculation formula The waveform amplitude range is retrieved from the database to obtain the corresponding stretching ratio. The waveform is then stretched and normalized according to the stretching ratio. Finally, all sampling points of Jiang Baixing are mapped to their respective amplitudes according to the stretching ratio. interval; When abnormal waveform extreme values ​​or amplitude ranges are detected, the normalization operation is terminated, the waveform is marked as invalid, and deleted. Based on the system's preset standard sampling time interval The actual single sampling time interval of the identification device Based on linear interpolation, the time axis of the original waveform is resampled and interpolated to map the time coordinates of the original waveforms from different devices to a standard equally spaced time axis, resulting in a waveform aligned based on the time dimension. The aligned waveform is then mapped to a two-dimensional coordinate system to obtain a normalized waveform.

[0008] According to the above technical solution, the step of extracting waveform time-domain features and amplitude features from multiple sets of normalized waveforms, identifying abnormal waveforms through waveform time-domain similarity analysis, removing abnormal waveforms, and filtering to obtain a valid waveform sample set includes the following steps: All normalized waveforms collected from the same measuring point are combined into a waveform set. , Each waveform consists of a standard time axis and a normalized amplitude sequence; The arithmetic mean of the amplitudes of all waveforms within the waveform set is calculated at each time point to generate a ensemble-averaged reference waveform that characterizes the echo pattern at that measurement point. ; The normalized waveform sequence to be evaluated is labeled as... , After normalization of a single original echo waveform, it is mapped to... The number of points in the interval, marking the sequence of the ensemble average reference waveform as... The construction size is Euclidean distance matrix Calculate the waveform to be evaluated. The sampling point and the reference waveform The amplitude difference of each sampling point ; The amplitude difference is mapped to a two-dimensional coordinate system to construct an amplitude difference curve. The variation characteristics of the amplitude difference curve are identified, and the minimum cumulative DTW distance is calculated iteratively point by point based on these variation characteristics. The iterative formula is as follows: , thus obtaining the minimum cumulative distance value; Using the exponential decay mapping formula Convert DTW distance to Waveform similarity of intervals, where Preset smoothing constraint coefficients; The system presets a fixed similarity threshold. If the calculated waveform similarity Sim < The waveforms were marked as abnormal fluctuations and removed. like The waveform is marked as a valid waveform and stored in the valid waveform sample set.

[0009] According to the above technical solution, the step of calling the pre-constructed pipe echo feature matching model to perform feature comparison on each waveform in the effective waveform sample set, accurately locating the characteristic time points of the pipe outer surface echo and the pipe bottom inner wall echo, calculating the pipe thickness based on the time difference between the outer surface echo and the bottom inner wall echo, combined with the sound velocity of the pipe material, performing statistical correction on the calculation results of the effective samples, and outputting the final pipe wall thickness measurement result includes the following steps: A standard echo feature matching model adapted for pipe testing is constructed. The model incorporates the standard time-domain features, amplitude features, rising edge gradient features, and peak span features of the incident echo from the outer surface of the pipe and the reflected echo from the bottom surface of the inner wall. Retrieve the effective normalized waveform, identify its features, and compare it segment by segment with the model's standard features to pinpoint the characteristic time point corresponding to the first abrupt peak in the echo from the outer surface of the anchor pipe. Characteristic time points corresponding to the stable peak values ​​of the echoes from the inner wall of the bottom surface of the pipe. ; Based on the round-trip propagation law of ultrasonic waves in pipe media, the time difference between the inner and outer echoes is calculated. Retrieve the inherent standard ultrasonic velocity of the pipe material to be tested. The wall thickness of the pipe is calculated frame by frame using a formula. The independent wall thickness detection value corresponding to a single valid waveform is obtained and stored in the initial wall thickness dataset. Obtain the initial wall thickness dataset and calculate the mean of the initial wall thickness data. with unbiased standard deviation Construct an effective data range ,in The parameters are adjusted to the set effective data range to map the initial wall thickness dataset to the range. If the initial wall thickness data is not mapped to the range, the initial wall thickness data is marked as abnormal data; otherwise, the system continues to detect. Calculate pipe wall thickness using a weighted formula ,in For the final detection of wall thickness, To determine the number of effective wall thickness data, To determine the similarity of individual waveforms, To correspond to the wall thickness detection value of a single frame, output the final pipe wall thickness measurement result.

[0010] According to the above technical solution, the ultrasonic-based pipe thickness detection system includes a data acquisition module, a data processing module, an anomaly analysis module, and a pipe wall thickness calculation module. The data acquisition module is used to send ultrasonic pulses to the pipe under test, receive multiple sets of raw waveform data of ultrasonic feedback echoes reflected from the pipe wall, and store the raw waveform data in the database. The data processing module is used to normalize the original waveform data and map the original waveform data to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms. The anomaly analysis module is used to extract waveform time-domain features and amplitude features from multiple sets of normalized waveforms, identify abnormal waveforms through waveform time-domain similarity analysis, remove abnormal waveforms, and filter to obtain a valid waveform sample set. The pipe wall thickness calculation module is used to call the pre-built pipe echo feature matching model, compare the features of each waveform in the effective waveform sample set, accurately locate the characteristic time points of the pipe outer surface echo and the pipe bottom inner wall echo, calculate the pipe thickness based on the time difference between the outer surface echo and the bottom inner wall echo, and combine the sound velocity of the pipe material. Statistical correction is performed on the calculation results of the effective samples, and the final pipe wall thickness measurement result is output.

[0011] According to the above technical solution, the data processing module is further used for: Extracting the amplitude sequence of a single raw echo waveform , The total number of sampling points for a single waveform is used to scan the original waveform and identify the maximum amplitude. and minimum amplitude Calculate the amplitude range ; Identify waveform change characteristics, segment the waveform, mark the no-signal sampling interval after ultrasonic wave emission and before the arrival of effective echo as the baseline noise region, and mark the signal interval of pulse echo as the effective echo region; Calculate the mean amplitude of the baseline noise region Calculate the waveform asymmetry coefficient The signal power in the baseline noise region and the effective echo region is identified respectively, and the waveform signal-to-noise ratio (SNR) is calculated. When the waveform signal-to-noise ratio is higher than the preset threshold and the waveform asymmetry coefficient is greater than the preset threshold, the waveform is determined to be peak signal dominant and the maximum value normalization working mode is enabled; when the waveform signal-to-noise ratio is lower than the preset threshold and the waveform asymmetry coefficient is less than the preset threshold, the waveform is determined to have baseline offset interference and the minimum value normalization working mode is enabled. If the waveform amplitude range is extremely large If the waveform is less than the preset minimum threshold, it is determined that the current waveform is an invalid DC waveform with no effective echo. Normalization is not performed, and the waveform is marked and removed.

[0012] According to the above technical solution, the data processing module is further used for: Maximum value normalization working mode: adopts the normalization calculation formula. ,in, Indicates the first element in the original waveform. The original amplitude of each sampling point , This represents the total number of sampling points for this echo waveform, calculated using the global maximum value for a single waveform. Using this as the sole scaling reference, the amplitudes of all sampling points of the waveform are scaled proportionally, while the peak amplitude of the waveform is fixed and normalized to 1. The corresponding scaling factor is then used to map the remaining sampling points to their respective amplitudes. interval; Minimum value normalization working mode: uses the normalization calculation formula The waveform amplitude range is retrieved from the database to obtain the corresponding stretching ratio. The waveform is then stretched and normalized according to the stretching ratio. Finally, all sampling points of Jiang Baixing are mapped to their respective amplitudes according to the stretching ratio. interval; When abnormal waveform extreme values ​​or amplitude ranges are detected, the normalization operation is terminated, the waveform is marked as invalid, and deleted. Based on the system's preset standard sampling time interval The actual single sampling time interval of the identification device Based on linear interpolation, the time axis of the original waveform is resampled and interpolated to map the time coordinates of the original waveforms from different devices to a standard equally spaced time axis, resulting in a waveform aligned based on the time dimension. The aligned waveform is then mapped to a two-dimensional coordinate system to obtain a normalized waveform.

[0013] According to the above technical solution, the anomaly analysis module is used for: All normalized waveforms collected from the same measuring point are combined into a waveform set. , Each waveform consists of a standard time axis and a normalized amplitude sequence; The arithmetic mean of the amplitudes of all waveforms within the waveform set is calculated at each time point to generate a ensemble-averaged reference waveform that characterizes the echo pattern at that measurement point. ; The normalized waveform sequence to be evaluated is labeled as... , After normalization of a single original echo waveform, it is mapped to... The number of points in the interval, marking the sequence of the ensemble average reference waveform as... The construction size is Euclidean distance matrix Calculate the waveform to be evaluated. The sampling point and the reference waveform The amplitude difference of each sampling point ; The amplitude difference is mapped to a two-dimensional coordinate system to construct an amplitude difference curve. The variation characteristics of the amplitude difference curve are identified, and the minimum cumulative DTW distance is calculated iteratively point by point based on these variation characteristics. The iterative formula is as follows: , thus obtaining the minimum cumulative distance value; Using the exponential decay mapping formula Convert DTW distance to Waveform similarity of intervals, where Preset smoothing constraint coefficients; The system presets a fixed similarity threshold. If the calculated waveform similarity Sim < The waveforms were marked as abnormal fluctuations and removed. like The waveform is marked as a valid waveform and stored in the valid waveform sample set.

[0014] According to the above technical solution, the pipe wall thickness calculation module is used for: A standard echo feature matching model adapted for pipe testing is constructed. The model incorporates the standard time-domain features, amplitude features, rising edge gradient features, and peak span features of the incident echo from the outer surface of the pipe and the reflected echo from the bottom surface of the inner wall. Retrieve the effective normalized waveform, identify its features, and compare it segment by segment with the model's standard features to pinpoint the characteristic time point corresponding to the first abrupt peak in the echo from the outer surface of the anchor pipe. Characteristic time points corresponding to the stable peak values ​​of the echoes from the inner wall of the bottom surface of the pipe. ; Based on the round-trip propagation law of ultrasonic waves in pipe media, the time difference between the inner and outer echoes is calculated. Retrieve the inherent standard ultrasonic velocity of the pipe material to be tested. The wall thickness of the pipe is calculated frame by frame using a formula. The independent wall thickness detection value corresponding to a single valid waveform is obtained and stored in the initial wall thickness dataset. Obtain the initial wall thickness dataset and calculate the mean of the initial wall thickness data. with unbiased standard deviation Construct an effective data range ,in The parameters are adjusted to the set effective data range to map the initial wall thickness dataset to the range. If the initial wall thickness data is not mapped to the range, the initial wall thickness data is marked as abnormal data; otherwise, the system continues to detect. Calculate pipe wall thickness using a weighted formula ,in For the final detection of wall thickness, To determine the number of effective wall thickness data, To determine the similarity of individual waveforms, To correspond to the wall thickness detection value of a single frame, output the final pipe wall thickness measurement result.

[0015] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention analyzes waveform data characteristics and its effective data range, retains the relative amplitude difference of waveform peaks, and the morphological characteristics of rising and falling edges before and after peaks, avoiding the interference of minimum value jitter caused by sound on the normalization results. It can adapt to different detection devices, achieving higher applicability. Through calibration analysis based on data sampling time intervals, it eliminates baseline systematic errors of different devices and detection periods, strengthens the amplitude increment characteristics of the effective echo signal relative to the noise floor, and highlights the difference between stable flat waveform ranges and abrupt echo ranges. By locally stretching, compressing, and elastically aligning the waveform time axis, this system adaptively compensates for non-pathological time shifts caused by probe jitter, uneven coupling, equipment micro-vibration, and propagation delay fluctuations. Through a flexible matching mechanism, it corrects only minor time shifts under normal operating conditions, fully preserving the amplitude characteristics and echo morphology of the effective waveform. This significantly reduces the probability of false or missed waveform rejection, improving detection stability and repeatability. By analyzing the similarity between the collected waveform data and standard reference data, the system filters the data and corrects and compensates for pipe wall thickness based on the similarity, resulting in more accurate pipe wall thickness measurement results. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the method steps of the present invention.

[0017] Figure 2 This is a schematic diagram of the system module composition of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 The present invention provides a technical solution: a method for detecting pipe thickness based on ultrasound, comprising: An ultrasonic pulse is emitted to the pipe under test, and multiple sets of raw waveform data of ultrasonic feedback echoes reflected from the pipe wall are received and stored in the database. The original waveform data is normalized and mapped to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms. Waveform time-domain features and amplitude features are extracted from multiple sets of normalized waveforms. Abnormal waveforms are identified through waveform time-domain similarity analysis, and abnormal waveforms are removed to obtain a valid waveform sample set. The pre-built pipe echo feature matching model is invoked to compare the features of each waveform in the effective waveform sample set, accurately locate the characteristic time points of the echo on the outer surface of the pipe and the echo on the inner wall of the bottom surface of the pipe. Based on the time difference between the echo on the outer surface and the echo on the inner wall of the bottom surface, combined with the sound velocity of the pipe material, the pipe thickness is calculated. The calculation results of the effective samples are statistically corrected, and the final pipe wall thickness measurement result is output.

[0020] In some preferred embodiments, the original waveform data is normalized and mapped to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms. This further includes the following steps: Extracting the amplitude sequence of a single raw echo waveform , The total number of sampling points for a single waveform is used to scan the original waveform and identify the maximum amplitude. and minimum amplitude Calculate the amplitude range ; Identify waveform change characteristics, segment the waveform, mark the no-signal sampling interval after ultrasonic wave emission and before the arrival of effective echo as the baseline noise region, and mark the signal interval of pulse echo as the effective echo region; Calculate the mean amplitude of the baseline noise region Calculate the waveform asymmetry coefficient The signal power in the baseline noise region and the effective echo region is identified respectively, and the waveform signal-to-noise ratio (SNR) is calculated. When the waveform signal-to-noise ratio is higher than the preset threshold and the waveform asymmetry coefficient is greater than the preset threshold, the waveform is determined to be peak signal dominant and the maximum value normalization working mode is enabled; when the waveform signal-to-noise ratio is lower than the preset threshold and the waveform asymmetry coefficient is less than the preset threshold, the waveform is determined to have baseline offset interference and the minimum value normalization working mode is enabled. If the waveform amplitude range is extremely large If the waveform is less than the preset minimum threshold, it is determined that the current waveform is an invalid DC waveform with no effective echo. Normalization is not performed, and the waveform is marked and removed.

[0021] In some preferred embodiments, the original waveform data is normalized and mapped to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms. The method further includes the following steps: Maximum value normalization working mode: adopts the normalization calculation formula. ,in, Indicates the first element in the original waveform. The original amplitude of each sampling point , This represents the total number of sampling points for this echo waveform, calculated using the global maximum value for a single waveform. Using this as the sole scaling reference, the amplitudes of all sampling points of the waveform are scaled proportionally, while the peak amplitude of the waveform is fixed and normalized to 1. The corresponding scaling factor is then used to map the remaining sampling points to their respective amplitudes. In this range, the mode fully preserves the relative amplitude difference of the waveform peaks, the rising and falling edge morphology characteristics before and after the peaks, avoids the interference of minimum value jitter caused by noise on the normalization results, and can adapt to different detection devices to achieve higher applicability. Minimum value normalization working mode: uses the normalization calculation formula The waveform amplitude range is retrieved from the database to obtain the corresponding stretching ratio. The waveform is then stretched and normalized according to the stretching ratio. Finally, all sampling points of Jiang Baixing are mapped to their respective amplitudes according to the stretching ratio. interval; When abnormal waveform extreme values ​​or amplitude ranges are detected, the normalization operation is terminated, the waveform is marked as invalid, and deleted. Based on the system's preset standard sampling time interval The actual single sampling time interval of the identification device Based on linear interpolation, the original waveform time axis is resampled and interpolated to uniformly map the original waveform time coordinates of different devices to a standard equally spaced time axis, resulting in a waveform aligned based on the time dimension. This aligned waveform is then mapped to a two-dimensional coordinate system to obtain a normalized waveform. This eliminates baseline system errors from different devices and detection periods, enhances the amplitude increment characteristics of the effective echo signal relative to the noise floor, and highlights the difference between the stable flat region and the abrupt echo region of the waveform. The normalized amplitude output by both modes strictly converges to... The standard amplitude range ensures the uniformity of waveform data dimensions and seamless compatibility of the entire algorithm process across multiple devices and operating conditions.

[0022] In some preferred embodiments, waveform time-domain features and amplitude features are extracted from multiple sets of normalized waveforms, abnormal waveforms are identified through waveform time-domain similarity analysis, abnormal waveforms are removed, and a valid waveform sample set is obtained by screening. The process further includes the following steps: All normalized waveforms collected from the same measuring point are combined into a waveform set. , Each waveform consists of a standard time axis and a normalized amplitude sequence; The arithmetic mean of the amplitudes of all waveforms within the waveform set is calculated at each time point to generate a ensemble-averaged reference waveform that characterizes the echo pattern at that measurement point. ; The normalized waveform sequence to be evaluated is labeled as... , After normalization of a single original echo waveform, it is mapped to... The number of points in the interval, marking the sequence of the ensemble average reference waveform as... The construction size is Euclidean distance matrix Calculate the waveform to be evaluated. The sampling point and the reference waveform The amplitude difference of each sampling point ; The amplitude difference is mapped to a two-dimensional coordinate system to construct an amplitude difference curve. The variation characteristics of the amplitude difference curve are identified, and the minimum cumulative DTW distance is calculated iteratively point by point based on these variation characteristics. The iterative formula is as follows: This allows for obtaining the minimum cumulative distance value, enabling local stretching, compression, and elastic alignment of the waveform time axis, and adaptive compensation for non-pathological time offsets caused by probe jitter, uneven coupling, equipment micro-vibration, and propagation delay fluctuations. Using the exponential decay mapping formula Convert DTW distance to Waveform similarity of intervals, where Preset smoothing constraint coefficients; The system presets a fixed similarity threshold. If the calculated waveform similarity Sim < The waveforms were marked as abnormal fluctuations and removed. like The algorithm marks the waveform as a valid waveform and stores it in the valid waveform sample set. Through a flexible matching mechanism, it only corrects the small time offset of the waveform under normal working conditions, and fully preserves the amplitude characteristics and echo shape of the valid waveform. This greatly reduces the probability of false rejection and false rejection of waveforms, and improves the detection stability and repeatability.

[0023] In some preferred embodiments, a pre-built pipe echo feature matching model is invoked to perform feature comparison on each waveform in the effective waveform sample set, accurately locating the characteristic time points of the echo from the outer surface of the pipe and the echo from the inner wall of the bottom surface. Based on the time difference between the echo from the outer surface and the echo from the inner wall of the bottom surface, combined with the sound velocity of the pipe material, the pipe thickness is calculated. Statistical correction is performed on the calculation results of the effective samples, and the final pipe wall thickness measurement result is output. This further includes the following steps: A standard echo feature matching model adapted for pipe testing is constructed. The model incorporates the standard time-domain features, amplitude features, rising edge gradient features, and peak span features of the incident echo from the outer surface of the pipe and the reflected echo from the bottom surface of the inner wall. Retrieve the effective normalized waveform, identify its features, and compare it segment by segment with the model's standard features to pinpoint the characteristic time point corresponding to the first abrupt peak in the echo from the outer surface of the anchor pipe. Characteristic time points corresponding to the stable peak values ​​of the echoes from the inner wall of the bottom surface of the pipe. To avoid interference from clutter, secondary reflections, and noise spurious peaks, ensuring the uniqueness and accuracy of echo time point extraction; Based on the round-trip propagation law of ultrasonic waves in pipe media, the time difference between the inner and outer echoes is calculated. Retrieve the inherent standard ultrasonic velocity of the pipe material to be tested. The wall thickness of the pipe is calculated frame by frame using a formula. The independent wall thickness detection value corresponding to a single valid waveform is obtained and stored in the initial wall thickness dataset. Obtain the initial wall thickness dataset and calculate the mean of the initial wall thickness data. with unbiased standard deviation Construct an effective data range ,in The parameters are adjusted to the set effective data range, and the initial wall thickness dataset is mapped to the range. If the initial wall thickness data is not mapped to the range, the initial wall thickness data is marked as abnormal data. Otherwise, the system continues to detect and judges the discrete out-of-range wall thickness data outside the range as gross error anomalies and removes them directly. This eliminates the accidental extreme errors caused by single waveform positioning deviation and instantaneous noise disturbance, and purifies the wall thickness data sample set. Calculate pipe wall thickness using a weighted formula ,in For the final detection of wall thickness, To determine the number of effective wall thickness data, To determine the similarity of individual waveforms, To correspond to the single-frame wall thickness detection value, the final pipe wall thickness measurement result is output. The higher the similarity and the closer the waveform shape is to the standard waveform, the greater the weight of the detection data. The weight of the critical effective waveform with low similarity is adaptively reduced, so as to achieve the effect of giving priority to high-quality data and weakening the interference of poor data.

[0024] Using the same inventive concept as the above embodiments, this application also provides an ultrasonic-based pipe thickness detection system, including a data acquisition module, a data processing module, an anomaly analysis module, and a pipe wall thickness calculation module; The data acquisition module is used to send ultrasonic pulses to the pipe under test, receive multiple sets of raw waveform data of ultrasonic feedback echoes reflected from the pipe wall, and store the raw waveform data in the database. The data processing module is used to normalize the raw waveform data and map the raw waveform data to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms. The anomaly analysis module is used to extract waveform time-domain features and amplitude features from multiple sets of normalized waveforms, identify abnormal waveforms through waveform time-domain similarity analysis, remove abnormal waveforms, and filter to obtain a valid waveform sample set; The pipe wall thickness calculation module is used to call the pre-built pipe echo feature matching model, compare the features of each waveform in the effective waveform sample set, accurately locate the characteristic time points of the echo on the outer surface of the pipe and the echo on the inner wall of the bottom surface of the pipe, calculate the pipe thickness based on the time difference between the echo on the outer surface and the echo on the inner wall of the bottom surface, and combine the sound velocity of the pipe material. Statistical correction is performed on the calculation results of the effective samples, and the final pipe wall thickness measurement result is output.

[0025] The data processing module is also used for: Extracting the amplitude sequence of a single raw echo waveform , The total number of sampling points for a single waveform is used to scan the original waveform and identify the maximum amplitude. and minimum amplitude Calculate the amplitude range ; Identify waveform change characteristics, segment the waveform, mark the no-signal sampling interval after ultrasonic wave emission and before the arrival of effective echo as the baseline noise region, and mark the signal interval of pulse echo as the effective echo region; Calculate the mean amplitude of the baseline noise region Calculate the waveform asymmetry coefficient The signal power in the baseline noise region and the effective echo region is identified respectively, and the waveform signal-to-noise ratio (SNR) is calculated. When the waveform signal-to-noise ratio is higher than the preset threshold and the waveform asymmetry coefficient is greater than the preset threshold, the waveform is determined to be peak signal dominant and the maximum value normalization working mode is enabled; when the waveform signal-to-noise ratio is lower than the preset threshold and the waveform asymmetry coefficient is less than the preset threshold, the waveform is determined to have baseline offset interference and the minimum value normalization working mode is enabled. If the waveform amplitude range is extremely large If the waveform is less than the preset minimum threshold, it is determined that the current waveform is an invalid DC waveform with no effective echo. Normalization is not performed, and the waveform is marked and removed.

[0026] The data processing module is also used for: Maximum value normalization working mode: adopts the normalization calculation formula. ,in, Indicates the first element in the original waveform. The original amplitude of each sampling point , This represents the total number of sampling points for this echo waveform, calculated using the global maximum value for a single waveform. Using this as the sole scaling reference, the amplitudes of all sampling points of the waveform are scaled proportionally, while the peak amplitude of the waveform is fixed and normalized to 1. The corresponding scaling factor is then used to map the remaining sampling points to their respective amplitudes. interval; Minimum value normalization working mode: uses the normalization calculation formula The waveform amplitude range is retrieved from the database to obtain the corresponding stretching ratio. The waveform is then stretched and normalized according to the stretching ratio. Finally, all sampling points of Jiang Baixing are mapped to their respective amplitudes according to the stretching ratio. interval; When abnormal waveform extreme values ​​or amplitude ranges are detected, the normalization operation is terminated, the waveform is marked as invalid, and deleted. Based on the system's preset standard sampling time interval The actual single sampling time interval of the identification device Based on linear interpolation, the time axis of the original waveform is resampled and interpolated to map the time coordinates of the original waveforms from different devices to a standard equally spaced time axis, resulting in a waveform aligned based on the time dimension. The aligned waveform is then mapped to a two-dimensional coordinate system to obtain a normalized waveform.

[0027] The anomaly analysis module is used for: All normalized waveforms collected from the same measuring point are combined into a waveform set. , Each waveform consists of a standard time axis and a normalized amplitude sequence; The arithmetic mean of the amplitudes of all waveforms within the waveform set is calculated at each time point to generate a ensemble-averaged reference waveform that characterizes the echo pattern at that measurement point. ; The normalized waveform sequence to be evaluated is labeled as... , After normalization of a single original echo waveform, it is mapped to... The number of points in the interval, marking the sequence of the ensemble average reference waveform as... The construction size is Euclidean distance matrix Calculate the waveform to be evaluated. The sampling point and the reference waveform The amplitude difference of each sampling point ; The amplitude difference is mapped to a two-dimensional coordinate system to construct an amplitude difference curve. The variation characteristics of the amplitude difference curve are identified, and the minimum cumulative DTW distance is calculated iteratively point by point based on these variation characteristics. The iterative formula is as follows: , thus obtaining the minimum cumulative distance value; Using the exponential decay mapping formula Convert DTW distance to Waveform similarity of intervals, where Preset smoothing constraint coefficients; The system presets a fixed similarity threshold. If the calculated waveform similarity Sim < The waveforms were marked as abnormal fluctuations and removed. like The waveform is marked as a valid waveform and stored in the valid waveform sample set.

[0028] The pipe wall thickness calculation module is used for: A standard echo feature matching model adapted for pipe testing is constructed. The model incorporates the standard time-domain features, amplitude features, rising edge gradient features, and peak span features of the incident echo from the outer surface of the pipe and the reflected echo from the bottom surface of the inner wall. Retrieve the effective normalized waveform, identify its features, and compare it segment by segment with the model's standard features to pinpoint the characteristic time point corresponding to the first abrupt peak in the echo from the outer surface of the anchor pipe. Characteristic time points corresponding to the stable peak values ​​of the echoes from the inner wall of the bottom surface of the pipe. ; Based on the round-trip propagation law of ultrasonic waves in pipe media, the time difference between the inner and outer echoes is calculated. Retrieve the inherent standard ultrasonic velocity of the pipe material to be tested. The wall thickness of the pipe is calculated frame by frame using a formula. The independent wall thickness detection value corresponding to a single valid waveform is obtained and stored in the initial wall thickness dataset. Obtain the initial wall thickness dataset and calculate the mean of the initial wall thickness data. with unbiased standard deviation Construct an effective data range ,in The parameters are adjusted to the set effective data range to map the initial wall thickness dataset to the range. If the initial wall thickness data is not mapped to the range, the initial wall thickness data is marked as abnormal data; otherwise, the system continues to detect. Calculate pipe wall thickness using a weighted formula ,in For the final detection of wall thickness, To determine the number of effective wall thickness data, To determine the similarity of individual waveforms, To correspond to the wall thickness detection value of a single frame, output the final pipe wall thickness measurement result.

[0029] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0030] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting pipe thickness based on ultrasound, characterized in that: The method includes: An ultrasonic pulse is emitted to the pipe under test, and multiple sets of raw waveform data of ultrasonic feedback echoes reflected from the pipe wall are received and stored in the database. The original waveform data is normalized and mapped to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms. Waveform time-domain features and amplitude features are extracted from multiple sets of normalized waveforms. Abnormal waveforms are identified through waveform time-domain similarity analysis, and abnormal waveforms are removed to obtain a valid waveform sample set. The pre-built pipe echo feature matching model is invoked to compare the features of each waveform in the effective waveform sample set, accurately locate the characteristic time points of the echo on the outer surface of the pipe and the echo on the inner wall of the bottom surface of the pipe. Based on the time difference between the echo on the outer surface and the echo on the inner wall of the bottom surface, combined with the sound velocity of the pipe material, the pipe thickness is calculated. The calculation results of the effective samples are statistically corrected, and the final pipe wall thickness measurement result is output.

2. The method for detecting pipe thickness based on ultrasound according to claim 1, characterized in that: The normalization transformation of the original waveform data, mapping the original waveform data to a fixed amplitude reference and standard time coordinate system, to obtain multiple sets of normalized waveforms, includes the following steps: Extracting the amplitude sequence of a single raw echo waveform , The total number of sampling points for a single waveform is used to scan the original waveform and identify the maximum amplitude. and minimum amplitude Calculate the amplitude range ; Identify waveform change characteristics, segment the waveform, mark the no-signal sampling interval after ultrasonic wave emission and before the arrival of effective echo as the baseline noise region, and mark the signal interval of pulse echo as the effective echo region; Calculate the mean amplitude of the baseline noise region Calculate the waveform asymmetry coefficient The signal power in the baseline noise region and the effective echo region is identified respectively, and the waveform signal-to-noise ratio (SNR) is calculated. When the waveform signal-to-noise ratio is higher than the preset threshold and the waveform asymmetry coefficient is greater than the preset threshold, the waveform is determined to be peak signal dominant and the maximum value normalization working mode is enabled; when the waveform signal-to-noise ratio is lower than the preset threshold and the waveform asymmetry coefficient is less than the preset threshold, the waveform is determined to have baseline offset interference and the minimum value normalization working mode is enabled. If the waveform amplitude range is extremely large If the waveform is less than the preset minimum threshold, it is determined that the current waveform is an invalid DC waveform with no effective echo. Normalization is not performed, and the waveform is marked and removed.

3. The method for detecting pipe thickness based on ultrasound according to claim 2, characterized in that: The normalization transformation of the original waveform data, mapping the original waveform data to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms, also includes the following steps: Maximum value normalization working mode: adopts the normalization calculation formula. ,in, Indicates the first element in the original waveform. The original amplitude of each sampling point , This represents the total number of sampling points for this echo waveform, calculated using the global maximum value for a single waveform. Using this as the sole scaling reference, the amplitudes of all sampling points of the waveform are scaled proportionally, while the peak amplitude of the waveform is fixed and normalized to 1. The corresponding scaling factor is then used to map the remaining sampling points to their respective amplitudes. interval; Minimum value normalization working mode: uses the normalization calculation formula The waveform amplitude range is retrieved from the database to obtain the corresponding stretching ratio. The waveform is then stretched and normalized according to the stretching ratio. Finally, all sampling points of Jiang Baixing are mapped to their respective amplitudes according to the stretching ratio. interval; When abnormal waveform extreme values ​​or amplitude ranges are detected, the normalization operation is terminated, the waveform is marked as invalid, and deleted. Based on the system's preset standard sampling time interval The actual single sampling time interval of the identification device Based on linear interpolation, the time axis of the original waveform is resampled and interpolated to map the time coordinates of the original waveforms from different devices to a standard equally spaced time axis, resulting in a waveform aligned based on the time dimension. The aligned waveform is then mapped to a two-dimensional coordinate system to obtain a normalized waveform.

4. The method for detecting pipe thickness based on ultrasound according to claim 3, characterized in that: The process of extracting waveform time-domain features and amplitude features from multiple sets of normalized waveforms, identifying abnormal waveforms through waveform time-domain similarity analysis, eliminating abnormal waveforms, and filtering to obtain a valid waveform sample set includes the following steps: All normalized waveforms collected from the same measuring point are combined into a waveform set. , Each waveform consists of a standard time axis and a normalized amplitude sequence; The arithmetic mean of the amplitudes of all waveforms within the waveform set is calculated at each time point to generate a ensemble-averaged reference waveform that characterizes the echo pattern at that measurement point. ; The normalized waveform sequence to be evaluated is labeled as... , After normalization of a single original echo waveform, it is mapped to... The number of points in the interval, marking the sequence of the ensemble average reference waveform as... The construction size is Euclidean distance matrix Calculate the waveform to be evaluated. The sampling point and the reference waveform The amplitude difference of each sampling point ; The amplitude difference is mapped to a two-dimensional coordinate system to construct an amplitude difference curve. The variation characteristics of the amplitude difference curve are identified, and the minimum cumulative DTW distance is calculated iteratively point by point based on these variation characteristics. The iterative formula is as follows: , thus obtaining the minimum cumulative distance value; Using the exponential decay mapping formula Convert DTW distance to Waveform similarity of intervals, where Preset smoothing constraint coefficients; The system presets a fixed similarity threshold. If the calculated waveform similarity Sim < The waveforms were marked as abnormal fluctuations and removed. like The waveform is marked as a valid waveform and stored in the valid waveform sample set.

5. The method for detecting pipe thickness based on ultrasound according to claim 4, characterized in that: The process involves calling a pre-built pipe echo feature matching model to perform feature comparison on each waveform in the effective waveform sample set. This accurately locates the characteristic time points of the echo from the outer surface of the pipe and the echo from the inner wall of the bottom surface. Based on the time difference between the echo from the outer surface and the echo from the inner wall of the bottom surface, and combined with the sound velocity of the pipe material, the pipe thickness is calculated. Statistical corrections are then performed on the calculation results of the effective samples, and the final pipe wall thickness measurement result is output. This includes the following steps: A standard echo feature matching model adapted for pipe testing is constructed. The model incorporates the standard time-domain features, amplitude features, rising edge gradient features, and peak span features of the incident echo from the outer surface of the pipe and the reflected echo from the bottom surface of the inner wall. Retrieve the effective normalized waveform, identify its features, and compare it segment by segment with the model's standard features to pinpoint the characteristic time point corresponding to the first abrupt peak in the echo from the outer surface of the anchor pipe. Characteristic time points corresponding to the stable peak values ​​of the echoes from the inner wall of the bottom surface of the pipe. ; Based on the round-trip propagation law of ultrasonic waves in pipe media, the time difference between the inner and outer echoes is calculated. Retrieve the inherent standard ultrasonic velocity of the pipe material to be tested. The wall thickness of the pipe is calculated frame by frame using a formula. The independent wall thickness detection value corresponding to a single valid waveform is obtained and stored in the initial wall thickness dataset. Obtain the initial wall thickness dataset and calculate the mean of the initial wall thickness data. With unbiased standard deviation Construct an effective data range ,in The parameters are adjusted to the set effective data range to map the initial wall thickness dataset to the range. If the initial wall thickness data is not mapped to the range, the initial wall thickness data is marked as abnormal data; otherwise, the system continues to detect. Calculate pipe wall thickness using a weighted formula ,in For the final detection of wall thickness, To determine the number of effective wall thickness data, To determine the similarity of individual waveforms, To correspond to the wall thickness detection value of a single frame, output the final pipe wall thickness measurement result.

6. A pipe thickness detection system based on ultrasound, characterized in that: The system includes a data acquisition module, a data processing module, an anomaly analysis module, and a pipe wall thickness calculation module. The data acquisition module is used to send ultrasonic pulses to the pipe under test, receive multiple sets of raw waveform data of ultrasonic feedback echoes reflected from the pipe wall, and store the raw waveform data in the database. The data processing module is used to normalize the original waveform data and map the original waveform data to a fixed amplitude reference and standard time coordinate system to obtain multiple sets of normalized waveforms. The anomaly analysis module is used to extract waveform time-domain features and amplitude features from multiple sets of normalized waveforms, identify abnormal waveforms through waveform time-domain similarity analysis, remove abnormal waveforms, and filter to obtain a valid waveform sample set. The pipe wall thickness calculation module is used to call the pre-built pipe echo feature matching model, compare the features of each waveform in the effective waveform sample set, accurately locate the characteristic time points of the pipe outer surface echo and the pipe bottom inner wall echo, calculate the pipe thickness based on the time difference between the outer surface echo and the bottom inner wall echo, and combine the sound velocity of the pipe material. Statistical correction is performed on the calculation results of the effective samples, and the final pipe wall thickness measurement result is output.

7. The ultrasonic-based pipe thickness detection system according to claim 6, characterized in that: The data processing module is also used for: Extracting the amplitude sequence of a single raw echo waveform , The total number of sampling points for a single waveform is used to scan the original waveform and identify the maximum amplitude. and minimum amplitude Calculate the amplitude range ; Identify waveform change characteristics, segment the waveform, mark the no-signal sampling interval after ultrasonic wave emission and before the arrival of effective echo as the baseline noise region, and mark the signal interval of pulse echo as the effective echo region; Calculate the mean amplitude of the baseline noise region Calculate the waveform asymmetry coefficient The signal power in the baseline noise region and the effective echo region is identified respectively, and the waveform signal-to-noise ratio (SNR) is calculated. When the waveform signal-to-noise ratio is higher than the preset threshold and the waveform asymmetry coefficient is greater than the preset threshold, the waveform is determined to be peak signal dominant and the maximum value normalization working mode is enabled; when the waveform signal-to-noise ratio is lower than the preset threshold and the waveform asymmetry coefficient is less than the preset threshold, the waveform is determined to have baseline offset interference and the minimum value normalization working mode is enabled. If the waveform amplitude range is extremely large If the waveform is less than the preset minimum threshold, it is determined that the current waveform is an invalid DC waveform with no effective echo. Normalization is not performed, and the waveform is marked and removed.

8. The ultrasonic-based pipe thickness detection system according to claim 7, characterized in that: The data processing module is also used for: Maximum value normalization working mode: adopts the normalization calculation formula. ,in, Indicates the first element in the original waveform. The original amplitude of each sampling point , This represents the total number of sampling points for this echo waveform, calculated using the global maximum value for a single waveform. Using this as the sole scaling reference, the amplitudes of all sampling points of the waveform are scaled proportionally, while the peak amplitude of the waveform is fixed and normalized to 1. The corresponding scaling factor is then used to map the remaining sampling points to their respective amplitudes. interval; Minimum value normalization working mode: uses the normalization calculation formula The waveform amplitude range is retrieved from the database to obtain the corresponding stretching ratio. The waveform is then stretched and normalized according to the stretching ratio. Finally, all sampling points of Jiang Baixing are mapped to their respective amplitudes according to the stretching ratio. interval; When abnormal waveform extreme values ​​or amplitude ranges are detected, the normalization operation is terminated, the waveform is marked as invalid, and deleted. Based on the system's preset standard sampling time interval The actual single sampling time interval of the identification device Based on linear interpolation, the time axis of the original waveform is resampled and interpolated to map the time coordinates of the original waveforms from different devices to a standard equally spaced time axis, resulting in a waveform aligned based on the time dimension. The aligned waveform is then mapped to a two-dimensional coordinate system to obtain a normalized waveform.

9. The ultrasonic-based pipe thickness detection system according to claim 8, characterized in that: The anomaly analysis module is used for: All normalized waveforms collected from the same measuring point are combined into a waveform set. , Each waveform consists of a standard time axis and a normalized amplitude sequence; The arithmetic mean of the amplitudes of all waveforms within the waveform set is calculated at each time point to generate a ensemble-averaged reference waveform that characterizes the echo pattern at that measurement point. ; The normalized waveform sequence to be evaluated is labeled as... , After normalization of a single original echo waveform, it is mapped to... The number of points in the interval, marking the sequence of the ensemble average reference waveform as... The construction size is Euclidean distance matrix Calculate the waveform to be evaluated. The sampling point and the reference waveform The amplitude difference of each sampling point ; The amplitude difference is mapped to a two-dimensional coordinate system to construct an amplitude difference curve. The variation characteristics of the amplitude difference curve are identified, and the minimum cumulative DTW distance is calculated iteratively point by point based on these variation characteristics. The iterative formula is as follows: , thus obtaining the minimum cumulative distance value; Using the exponential decay mapping formula Convert DTW distance to Waveform similarity of intervals, where Preset smoothing constraint coefficients; The system presets a fixed similarity threshold. If the calculated waveform similarity Sim < The waveforms were marked as abnormal fluctuations and removed. like The waveform is marked as a valid waveform and stored in the valid waveform sample set.

10. The ultrasonic-based pipe thickness detection system according to claim 9, characterized in that: The pipe wall thickness calculation module is used for: A standard echo feature matching model adapted for pipe testing is constructed. The model incorporates the standard time-domain features, amplitude features, rising edge gradient features, and peak span features of the incident echo from the outer surface of the pipe and the reflected echo from the bottom surface of the inner wall. Retrieve the effective normalized waveform, identify its features, and compare it segment by segment with the model's standard features to pinpoint the characteristic time point corresponding to the first abrupt peak in the echo from the outer surface of the anchor pipe. Characteristic time points corresponding to the stable peak values ​​of the echoes from the inner wall of the bottom surface of the pipe. ; Based on the round-trip propagation law of ultrasonic waves in pipe media, the time difference between the inner and outer echoes is calculated. Retrieve the inherent standard ultrasonic velocity of the pipe material to be tested. The wall thickness of the pipe is calculated frame by frame using a formula. The independent wall thickness detection value corresponding to a single valid waveform is obtained and stored in the initial wall thickness dataset. Obtain the initial wall thickness dataset and calculate the mean of the initial wall thickness data. with unbiased standard deviation Construct an effective data range ,in The parameters are adjusted to the set effective data range to map the initial wall thickness dataset to the range. If the initial wall thickness data is not mapped to the range, the initial wall thickness data is marked as abnormal data; otherwise, the system continues to detect. Calculate pipe wall thickness using a weighted formula ,in For the final detection of wall thickness, To determine the number of effective wall thickness data, To determine the similarity of individual waveforms, To correspond to the wall thickness detection value of a single frame, output the final pipe wall thickness measurement result.

Images