Hierarchical transmission method of ultrasound data

Abstract

The present disclosure belongs to the technical field of industrial Internet of Things, and provides a hierarchical transmission method of ultrasonic data, which acquires full-time domain waveform data and a noise window without effective echo in the full-time domain waveform data; based on the amplitude statistical characteristics of the noise window, an effective echo interval is screened from the full-time domain waveform data, and local diagnostic waveform data containing the effective echo interval and original sampling values of the adjacent areas before and after the effective echo interval are extracted, the effective echo interval is used to calculate a thickness value of a measured component, and the thickness value is used to determine a thickness change characteristic of the measured component; in response to a target level determined based on the thickness change characteristic, the local diagnostic waveform data and a first communication link, a first communication link or a second communication link is called to perform data transmission, so that dynamic balance between data processing complexity and communication energy consumption can be achieved under the premise of ensuring the effectiveness of monitoring data.

Description

Technical Field

[0001] This disclosure belongs to the field of industrial Internet of Things (IoT) technology, and more specifically, relates to a hierarchical transmission method for ultrasonic data. Background Technology

[0002] Currently, online monitoring of corrosion / erosion in pressure-bearing equipment such as industrial storage tanks and pipelines largely relies on ultrasonic technology.

[0003] In related technologies, among the data transmission schemes for fixed-installation sensor nodes, one type of transmission scheme only uploads the thickness value. This scheme cannot support remote diagnosis of anomalies such as coupling failure and waveform distortion. Another type of transmission scheme uploads the entire time-domain waveform. This scheme will significantly occupy the bandwidth of the low-power wide-area communication link, significantly increase the node's energy consumption, and shorten the battery life.

[0004] Therefore, the data transmission schemes for fixed-installation sensor nodes in related technologies cannot balance data processing complexity and communication energy consumption. Summary of the Invention

[0005] The purpose of this disclosure is to provide a hierarchical transmission method for ultrasound data, which aims to solve the technical problem that data transmission schemes for fixed-installation sensor nodes in related technologies cannot simultaneously address the complexity of data processing and the energy consumption of communication.

[0006] To achieve the above objectives, in a first aspect, embodiments of this disclosure provide a method for hierarchical transmission of ultrasound data, the method comprising: Acquire the full time-domain waveform data, and the noise window in the full time-domain waveform data that does not contain valid echoes; Based on the amplitude statistical characteristics of the noise window, effective echo intervals are screened from the full time-domain waveform data, and local diagnostic waveform data containing the effective echo intervals and their preceding and following neighborhood original sample values ​​are extracted. The effective echo intervals are used to calculate the thickness value of the tested component, and the thickness value is used to determine the thickness variation characteristics of the tested component. In response to the target level determined based on the thickness variation characteristics, the local diagnostic waveform data, and the first communication link, the first communication link or the second communication link is invoked to perform data transmission; wherein, the first communication link is a low-power wide-area communication link, and the second communication link is a short-range high-speed communication link, and the content of the data transmission includes thickness values, local diagnostic waveform data, or full time-domain waveform data.

[0007] Secondly, embodiments of this disclosure provide an ultrasonic sensor node, comprising: The analog front end is used to transmit ultrasonic pulses and receive the corresponding full-time-domain waveform data. A processor, coupled to the analog front end, is configured to execute a computer program to implement the method as described in any one of the first aspects.

[0008] Thirdly, embodiments of this disclosure provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the electronic device performs the method as described in any of the first aspects.

[0009] Fourthly, embodiments of this disclosure provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method as described in any of the first aspects.

[0010] Fifthly, embodiments of this disclosure provide a computer program product that, when run on an electronic device, causes the electronic device to perform the method described in any one of the first aspects.

[0011] The beneficial effects of the embodiments disclosed herein compared to the prior art are as follows: This embodiment proposes a hierarchical transmission method for ultrasonic data. First, full-scale time-domain waveform data is acquired, and noise windows without valid echoes are identified within the full-scale time-domain waveform data. Second, based on the amplitude statistical characteristics of the noise windows, valid echo intervals are selected from the full-scale time-domain waveform data, and local diagnostic waveform data containing the original sampled values ​​of the valid echo interval and its preceding and following neighborhoods is extracted. Then, the thickness value and thickness variation characteristics of the measured component are calculated based on the valid echo intervals. Finally, given that the first communication link is a low-power wide-area network link, such as LoRa, and the second communication link is a short-range high-speed communication link, such as BLE, the first or second communication link is adaptively invoked to perform data transmission based on the thickness variation characteristics, local diagnostic waveform data, and the target level determined by the first communication link. Specifically, at least one of the following can be transmitted: thickness value, local diagnostic waveform data, or full-scale time-domain waveform data. This method effectively solves the technical problem that traditional fixed-installation sensor node data transmission schemes struggle to balance data processing complexity and communication energy consumption. While ensuring the validity of monitoring data, it achieves a dynamic balance between data processing complexity and communication energy consumption, thereby meeting the needs of unattended operation in industrial IoT scenarios. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic flowchart of a hierarchical transmission method for ultrasound data provided in an embodiment of this disclosure; Figure 2 This is a schematic flowchart of a hierarchical transmission method for ultrasound data provided in an embodiment of this disclosure; Figure 3 This is a schematic flowchart of a hierarchical transmission method for ultrasound data provided in an embodiment of this disclosure; Figure 4 This is a schematic flowchart of a hierarchical transmission method for ultrasound data provided in an embodiment of this disclosure; Figure 5 This is a schematic flowchart of a hierarchical transmission method for ultrasound data provided in an embodiment of this disclosure; Figure 6 This is a schematic diagram of the structure of a graded transmission device for ultrasonic data provided in an embodiment of this disclosure; Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0014] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, so as to provide a thorough understanding of the embodiments of this disclosure. However, those skilled in the art will understand that this disclosure may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this disclosure with unnecessary detail.

[0015] It should be understood that, when used in this disclosure and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0016] It should also be understood that, in the description of this disclosure, unless otherwise stated, the " / " used in this specification and the appended claims indicates that the related objects are in an "or" relationship. For example, A / B can mean A or B. The "and / or" in this disclosure is merely a description of the relationship between the related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. Furthermore, in the description of this disclosure, unless otherwise stated, "multiple" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0017] Furthermore, to facilitate a clear description of the technical solutions of the embodiments of this disclosure, the terms "first" and "second" are used in the embodiments of this disclosure to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, but are only used for distinguishing descriptions, and the terms "first" and "second" are not necessarily different, nor should they be construed as indicating or implying relative importance.

[0018] As used in this disclosure and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if [the described condition or event] is detected" may be interpreted, depending on the context, as "once determined," "in response to determination," "once [the described condition or event] is detected," or "in response to detection of [the described condition or event]."

[0019] References to "one embodiment" or "some embodiments" as described in this disclosure mean that one or more embodiments of this disclosure include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including, but not limited to," unless otherwise specifically emphasized.

[0020] First, some terms used in the embodiments of this disclosure will be explained to facilitate understanding by those skilled in the art.

[0021] Full-volume time-domain waveform data refers to the complete digital sampling sequence obtained by the analog front end performing continuous analog-to-digital conversion on the echo signal along the time axis after the ultrasonic transducer emits an excitation pulse.

[0022] Local diagnostic waveform data refers to a continuous segment of raw sampled values ​​extracted from the full time-domain waveform data, with the effective echo interval as the core. Unlike the traditional approach of extracting only peak values ​​or time of flight, this data retains the rising and falling edges, sidelobe structure, and nearby background noise characteristics of the effective echo, and is associated with location metadata such as the starting index and sampling rate.

[0023] A noise window is a pre-defined sampling section in the full-time waveform data that explicitly contains no valid ultrasonic echoes. It can be the silent recovery period after the transmitted pulse or the baseline region between adjacent valid echoes.

[0024] The first communication link refers to the low-power wide-area communication link (such as LoRa), which has the characteristics of long coverage distance, extremely low power consumption but limited bandwidth. It is mainly used for the periodic reporting of daily monitoring data.

[0025] The second communication link refers to a short-range high-speed communication link (such as BLE), which has the characteristics of high transmission rate and large data throughput but high power consumption. It is mainly activated in the field maintenance mode and is used to transmit full time-domain waveform data or perform firmware upgrades.

[0026] The above is a brief introduction to the terms used in the embodiments of this disclosure, and will not be repeated below.

[0027] This disclosure provides an example of a hierarchical transmission method for ultrasound data; please refer to... Figure 1 As shown, Figure 1 A schematic flowchart of a hierarchical transmission method for ultrasonic data provided in this disclosure is shown. This is by way of example and not limitation; the method can be applied to or implemented in electronic devices, such as sensor nodes in an Industrial Internet of Things (IIoT) scenario. The method includes: S101. Obtain the full time-domain waveform data, as well as the noise window in the full time-domain waveform data that does not contain valid echoes.

[0028] S102. Based on the amplitude statistics of the noise window, select the effective echo interval from the full time-domain waveform data, and extract the local diagnostic waveform data containing the effective echo interval and the original sampled values ​​of its immediate and next neighboring regions.

[0029] The effective echo range is used to calculate the thickness value of the measured component, and the thickness value is used to determine the thickness variation characteristics of the measured component.

[0030] S103. In response to the target level determined based on thickness variation characteristics, local diagnostic waveform data and the first communication link, the first communication link or the second communication link is invoked to perform data transmission.

[0031] The first communication link is a low-power wide-area communication link, and the second communication link is a short-distance high-speed communication link. The data transmitted includes thickness values, local diagnostic waveform data, or full time-domain waveform data.

[0032] The hierarchical transmission method for ultrasonic data provided in this disclosure can be specifically adapted to the long-term monitoring needs of fixed-installation ultrasonic sensor nodes (such as ultrasonic thickness sensor nodes). For example, it is designed for long-term online monitoring scenarios of infrastructure such as petrochemical storage tanks, subsea pipelines, and power towers. It aims to solve the technical problem that fixed-installation sensor nodes cannot balance data processing complexity and communication energy consumption under limited power supply conditions. It can balance data processing complexity and communication energy consumption while ensuring the validity of monitoring data, thus adapting to the unattended operation requirements in industrial IoT scenarios.

[0033] To facilitate understanding of the embodiments of this disclosure, steps S101 to S103 will be explained in detail below.

[0034] In this embodiment, the full-volume time-domain waveform data is the original sampled value sequence obtained by the analog front end of the ultrasonic sensor node after completing ultrasonic transmission by performing high-speed analog-to-digital conversion on the continuous echo signal received by the ultrasonic transducer. This sequence completely records the entire process from the end of the transmission dead zone to the complete attenuation of the echo signal. It can include not only the effective echo used to calculate the thickness of the tested component (such as industrial storage tanks, submarine pipelines, and power towers), but also environmental noise, clutter interference, and waveform distortion information caused by changes in probe coupling state.

[0035] In this embodiment, the noise window is a reference segment extracted from the full time-domain waveform data that clearly does not contain effective ultrasonic echoes. For example, it can be selected from the initial silence period after the excitation pulse is emitted, or it can be selected from the signal baseline period between two consecutive effective echoes. The data in this noise window characterizes the environmental noise floor characteristics under the current operating conditions.

[0036] In this embodiment, the amplitude statistical characteristics are a set of feature values ​​obtained by mathematical analysis of the amplitude of the sampling points within the noise window. These features may include parameters such as the average amplitude and the degree of amplitude dispersion. These features can reflect the noise fluctuation of the on-site environment in real time and serve as a benchmark for adaptively adjusting the detection sensitivity.

[0037] In this embodiment, the effective echo interval is a specific continuous sampling segment in the full time-domain waveform data whose amplitude is significantly higher than the background noise and conforms to the ultrasonic propagation law. The start and end positions of this effective echo interval directly determine the accuracy of the flight time calculation, and thus the accuracy of the thickness value. The thickness value is a quantitative value representing the current wall thickness of the measured component, calculated based on the parameters of the effective echo interval. It reflects the current state of the measured component and can be specifically derived from the product of the propagation speed of sound in the medium and the time delay. The thickness variation characteristics can be generated based on the thickness values ​​obtained from multiple consecutive measurements. It is used to characterize the trend of the wall thickness of the measured component changing over time, and can include the rate of change and amplitude of the thickness value. It is used to characterize whether the measured component is undergoing an evolution process of accelerated corrosion, erosion thinning, or other structural deterioration.

[0038] In this embodiment, the local diagnostic waveform is a continuous data segment containing the original sampled values ​​of the effective echo interval and its preceding and following neighborhoods. The number of sampling points in the preceding and following neighborhoods can be determined according to parameters such as the pulse width and sampling rate of the effective echo interval. The rising edge, peak shape, and trailing edge attenuation characteristics of the effective echo interval are preserved to support the server in remotely diagnosing echo quality, saturation distortion, or probe loosening. Furthermore, the data volume of the local diagnostic waveform data is much smaller than that of the full time-domain waveform data.

[0039] It should be noted that the first communication link is a low-power wide-area communication link, configured to communicate based on LoRa or NB-IoT protocols. It features wide coverage, low power consumption, and a relatively low transmission rate, making it suitable for small data transmission needs in daily monitoring scenarios. It can handle data reporting tasks under normal operating conditions and ensure the long-term battery life of sensor nodes. The second communication link is a short-range high-speed communication link, configured to communicate based on Bluetooth or Wi-Fi protocols. It features high transmission rate, large data transmission volume per transmission, and a relatively small coverage area. It is suitable for the full-time waveform data transmission needs of maintenance personnel in close-range maintenance scenarios and can meet the full-time waveform data backtracking needs under abnormal operating conditions.

[0040] In this embodiment, the target level is a transmission decision level that is comprehensively divided based on thickness variation characteristics, local diagnostic waveforms, and the operating status of the first communication link. Different levels correspond to different transmission data granularity and link selection strategies. In other words, the target level is a transmission decision label that is comprehensively evaluated based on the operating status of the tested component, data quality, and communication link resource status. It is used to dynamically match the optimal communication resources, which can realize the dynamic matching of transmission requirements and link capabilities, thereby optimizing energy consumption and bandwidth usage while ensuring the effectiveness of monitoring.

[0041] In practical implementation, sensor nodes can be deployed on the bottom outer wall of outdoor oil storage tanks. Due to variations in the internal medium and ambient temperature, the noise background varies with each measurement. For example, upon each wake-up, the sensor node first controls the analog front-end to acquire full-scale time-domain waveform data, then automatically locates the silent section after the transmitted pulse as the noise window. Based on the amplitude statistical characteristics of the noise window, effective echo intervals are adaptively selected, thickness values ​​are calculated, and local diagnostic waveforms are extracted. Then, combining the thickness change trend, waveform quality, and link resource status, the target level is determined. Appropriate communication links are selected as needed to transmit data at the corresponding granularity, achieving a balance between monitoring effectiveness and energy consumption.

[0042] For example, when acquiring full-scale time-domain waveform data, the analog front-end can perform sampling and quantization of the echo signal after each measurement trigger, storing the continuous sampled values ​​in chronological order as full-scale time-domain waveform data. Simultaneously, a silent segment following the transmitted excitation pulse can be extracted from the full-scale time-domain waveform data as a noise window, or a baseline segment between two adjacent valid echoes can be extracted as a noise window. When filtering valid echo intervals based on the amplitude statistical characteristics of the noise window, statistical analysis can be performed on the amplitude of the noise window to obtain statistical parameters such as the average amplitude and amplitude dispersion. A filtering threshold is generated based on the combination of statistical parameters. The full-scale time-domain waveform data is then traversed along the time axis, and continuous sampling sequences whose amplitudes meet the filtering rules are determined as valid echo intervals.

[0043] As an example, and not a limitation, when extracting local diagnostic waveforms, the start and end boundaries of the effective echo interval can be used as a reference. A preset number of sampling points can be extended forward and backward along the time axis, and the original sampled values ​​of the corresponding segments can be extracted as local diagnostic waveforms. The number of extended sampling points can be dynamically adjusted according to parameters such as the pulse width of the effective echo and the system sampling rate. When determining the target level, the thickness change characteristics can be combined to determine whether the state of the tested component is in a steady state, the quality of the local diagnostic waveform can be combined to assess the reliability of the effective echo interval, and the available air interface resources and transmission power consumption limit of the first communication link can be combined to classify different target levels. Alternatively, the target level can be classified based on a combination of two of these parameters.

[0044] When invoking a communication link to perform data transmission, if the target level corresponds to the first level with low data volume transmission requirements, the first communication link can be used to transmit thickness values. If the target level corresponds to the second level with medium data volume transmission requirements, the first communication link can be used to transmit local diagnostic waveforms. If the target level corresponds to the third level with high data volume transmission requirements or the resources of the first communication link are insufficient to support the corresponding data transmission, the second communication link can be activated directly to transmit full time-domain waveform data. Alternatively, before activating the second communication link, a maintenance request and device authentication information can be reported to the server through the first communication link. After receiving the activation command from the server, the transmission of full time-domain waveform data can be performed.

[0045] In this embodiment, by acquiring full time-domain waveform data and a noise window within the full time-domain waveform data that does not contain valid echoes, and by filtering valid echo intervals from the full time-domain waveform data based on the amplitude statistical characteristics of the noise window, the accuracy of echo identification can be improved under complex operating conditions, and the interference of environmental noise on measurement results can be reduced. By extracting local diagnostic waveforms that retain morphological features, sufficient diagnostic information can be preserved while reducing the amount of data transmitted, balancing communication bandwidth usage and diagnostic requirements. By combining thickness variation characteristics, waveform quality, and link resource status to select suitable communication links, communication transmission energy consumption can be significantly reduced while ensuring monitoring effectiveness, adapting to the operational needs of long-term unattended industrial monitoring scenarios.

[0046] In some embodiments, please refer to Figure 2 As shown, Figure 2 This illustration shows a schematic flowchart of a graded transmission method for ultrasound data provided in this disclosure. Based on the amplitude statistical characteristics of a noise window, it filters effective echo intervals from the full time-domain waveform data, including: S201. Based on the amplitude distribution characteristics of the noise window, determine the preset trigger threshold and the preset sustain threshold, wherein the preset trigger threshold is greater than the preset sustain threshold. S202. Traverse the entire time-domain waveform data along the time axis and determine the sequence of sampling points whose amplitude first crosses the preset trigger threshold and remains continuously above the preset maintenance threshold as candidate echoes. S203. Determine the effective echo interval based on the duration of at least one candidate echo obtained through traversal.

[0047] In step S201, the preset trigger threshold is an amplitude judgment benchmark set based on the amplitude distribution characteristics of the noise window. It serves as the initial threshold for triggering valid echo recognition. The preset trigger threshold is higher than the typical amplitude level of the noise window, which can initially filter out low-amplitude background noise. The preset maintenance threshold is another amplitude judgment benchmark set based on the amplitude distribution characteristics of the noise window and lower than the preset trigger threshold. Together with the preset trigger threshold, it forms a hysteresis comparison interval, preventing frequent switching of the recognition state when the echo amplitude fluctuates near the threshold, thus improving the stability of the echo recognition process.

[0048] In step S202, candidate echoes are sequences of continuous sampling points that meet the condition of amplitude crossing a high threshold and remaining consistently above a low threshold when traversing the full time-domain waveform data. These are candidate objects for the valid echo interval and need to be verified later to be confirmed as true valid echoes. The duration is the time length of the continuous sampling point sequence corresponding to a single candidate echo, and its function is to distinguish between true valid echoes that conform to the physical characteristics of the ultrasonic transducer and transient interference noise with extremely short duration.

[0049] For example, when determining the preset trigger threshold and the preset sustain threshold, the preset trigger threshold can be obtained based on the product of the average amplitude of the noise window and the first coefficient, and the preset sustain threshold can be obtained based on the product of the average amplitude of the noise window and the second coefficient. The first coefficient is greater than the second coefficient, so that a hysteresis interval adapted to the current noise level is formed between the two thresholds. Alternatively, the threshold can be dynamically adjusted by combining the standard deviation of the noise window amplitude, so that the threshold can be adaptively increased or decreased with the noise fluctuation amplitude, adapting to the working conditions where the noise level changes significantly.

[0050] When traversing the full time-domain waveform data along the time axis, the amplitude and the relationship between the two thresholds can be compared point by point starting from the first sampling point. When the amplitude of the current sampling point is detected to be greater than the preset trigger threshold for the first time, the sampling point is marked as a potential starting point. Then, it is checked whether the amplitude of multiple consecutive sampling points is greater than the preset maintenance threshold. If the amplitude drops briefly during the verification process but rises again within a preset number of sampling points, the drop can be ignored and the current candidate echo can be locked until the amplitude of multiple consecutive sampling points is lower than the preset maintenance threshold, and a valid termination point is confirmed.

[0051] In step S203, when determining the effective echo interval based on the duration of the candidate echo, a minimum pulse width reference that matches the center frequency of the ultrasonic transducer and the system sampling rate can be preset. When the duration of the candidate echo is greater than the minimum pulse width reference, the interval corresponding to the candidate echo is determined as the effective echo interval. If there are multiple candidate echoes that meet the duration requirement, the interval corresponding to the candidate echo with the largest amplitude can be selected as the effective echo interval. Alternatively, a secondary screening can be performed by combining whether the position of the echo matches the distribution pattern of historical high-confidence echoes.

[0052] By setting high and low thresholds and forming a hysteresis comparison interval based on the amplitude distribution characteristics of the noise window, the system can effectively suppress transient noise interference and false identification caused by amplitude fluctuations, improve the accuracy of effective echo interval positioning, and further eliminate pseudo echoes that do not conform to the physical characteristics of ultrasonic echoes by introducing a duration verification mechanism, thus adapting to the long-term stable operation requirements under complex industrial conditions.

[0053] In some embodiments, please refer to Figure 3 As shown, Figure 3 A schematic flowchart of a hierarchical transmission method for ultrasound data provided in this disclosure is shown, which determines the effective echo interval based on the duration of at least one candidate echo obtained through traversal, including: S301. In response to a candidate echo having a duration greater than a first preset pulse width, the interval in which the candidate echo is located is determined as a valid echo interval.

[0054] The first preset pulse width is obtained by predefining at least one of the center frequency of the ultrasonic transducer, the current sampling rate, and the on-site noise level.

[0055] S302. If the duration of two or more candidate echoes is greater than the first preset pulse width, then the interval where the candidate echo with the largest amplitude is located is selected as the effective echo interval.

[0056] In this embodiment, the first preset pulse width is a preset time length reference. This reference can be set based on the center frequency of the ultrasonic transducer to match the inherent vibration period of the transducer. It can also be calculated in combination with the current system sampling rate to ensure that the time length corresponds accurately to the number of sampling points. It can also be dynamically adjusted according to the noise level at the scene.

[0057] The duration of a candidate echo refers to the time span from the first sampling point above the preset sustaining threshold to the last sampling point above the preset sustaining threshold, which can distinguish between real ultrasound echoes and transient pulse interference. The effective echo interval is the continuous sampling segment that is ultimately confirmed as a real effective echo.

[0058] For example, when the duration of a candidate echo is determined to be greater than the first preset pulse width, the interval between the start and end points of the candidate echo can be directly mapped as the effective echo interval, and this signal is assumed to be the only effective reflection. When there are two or more candidate echoes whose durations are both greater than the first preset pulse width, the amplitude characteristics of each candidate echo can be compared one by one, and the interval where the candidate echo with the largest amplitude is located can be selected as the effective echo interval. This is because, in the absence of multiple reflections or complex structures inside the component under test, the first bottom reflection formed by the outer and inner walls of the component under test usually has the strongest energy.

[0059] In addition, if there are multiple candidate echoes that meet the criteria, the timing and location of the echoes can be used to assist in the judgment. For example, echoes that appear within the expected propagation time window can be prioritized, or the typical location distribution of effective echoes in historical measurement data can be matched and filtered to further improve the recognition accuracy in multi-echo interference environments.

[0060] By introducing a duration verification mechanism based on the physical characteristics of ultrasonic transducers, transient interference signals that do not conform to physical laws can be effectively filtered out, significantly improving the anti-interference capability of echo recognition. By selecting the signal with the largest amplitude as the effective echo in multi-echo scenarios, it is possible to ensure that the most representative reflection interface is locked in complex acoustic paths.

[0061] In some embodiments, please refer to Figure 4 As shown, Figure 4 A schematic flowchart of a hierarchical transmission method for ultrasound data provided in this disclosure is shown, which extracts local diagnostic waveform data containing the original sampled values ​​of the effective echo interval and its preceding and following neighborhoods, including: S401. Extend the first preset sampling point forward based on the starting boundary of the effective echo interval, and extend the second preset sampling point backward based on the ending boundary of the effective echo interval, and extract the continuous original sampled values ​​of the corresponding segment in the full time domain waveform data as the target waveform segment. S402. Generate positioning metadata. Positioning metadata represents the absolute starting index and total number of sampling points of the target waveform segment in the full time-domain waveform data, and is used to uniquely identify the position of the target waveform segment. S403. Encapsulate the target waveform segment and positioning metadata into local diagnostic waveform data.

[0062] In some embodiments, the first preset sampling point and the second preset sampling point are determined based on at least one of the peak width of the effective echo interval, the current sampling rate, and the amplitude standard deviation of the noise window, and the target waveform segment includes the rising edge, falling edge, and sidelobe characteristics of the effective echo interval.

[0063] In this embodiment, the first preset sampling point can be understood as a sampling point extending in the negative direction of the time axis based on the starting boundary of the effective echo interval. Its function is to completely preserve the rising edge characteristics of the effective echo. This rising edge reflects the incident and reflection characteristics of the ultrasonic wave at the medium interface and is an important basis for judging the probe coupling state and signal saturation. The second preset sampling point is a sampling point extending in the positive direction of the time axis based on the ending boundary of the effective echo interval. Its function is to completely preserve the falling edge and trailing edge attenuation characteristics of the effective echo. These characteristics help to analyze the attenuation characteristics of the internal structure of the material under test and whether there is multiple reflection interference.

[0064] It should be noted that the target waveform segment is a continuous sequence of raw sampled values ​​extracted from the full time-domain waveform data, containing valid echoes and information about their preceding and following neighborhoods. This sequence, while removing a large amount of invalid silent baseline from the full data, possesses all the key acoustic features required for diagnosis. The location metadata is index information associated with the target waveform segment, including its absolute starting index number in the full time-domain waveform data and the total number of sample points contained in the segment. Its purpose is to establish a mapping relationship between local data and the global raw record, enabling the server to uniquely determine the specific location of that data segment within the original full waveform after receiving the local diagnostic waveform data.

[0065] For example, when expanding along the boundary of the effective echo interval, the number of the first and second preset sampling points can be set based on the measured peak width of the effective echo interval. For instance, the peak width can be expanded forward by 0.5 times to fully capture the rising edge, and the peak width can be expanded backward by 1 time to cover the falling edge and part of the trailing edge. Alternatively, the number of sampling points expanded backward can be dynamically adjusted according to the noise level at the scene. In environments with strong noise, the number of sampling points expanded backward can be appropriately increased to retain more attenuation characteristics for signal-to-noise ratio analysis. In scenarios with a high system sampling rate, the number of expanded sampling points can also be increased accordingly to ensure that the capture time is sufficient to reflect the complete shape of the effective echo.

[0066] When generating location metadata, the absolute starting index can be represented by the sampling point sequence number in the full time-domain waveform data, while the total number of sampling points directly reflects the data scale of the target waveform segment. Together, they constitute a unique location fingerprint. When encapsulating local diagnostic waveform data, the location metadata can be placed at the beginning of the target waveform segment to form a data frame, or the location metadata can be encapsulated as an independent field along with the target waveform segment into a key-value pair structure for easy server-side parsing. Furthermore, during the expansion of sampling points, if the forward expansion touches the beginning of the full time-domain waveform data or the backward expansion touches the end, the data is truncated to the actual boundary, and the actual expansion range is recorded in the location metadata.

[0067] In this embodiment, a conversion logic from full-time waveform data to local diagnostic waveform data is provided. By adaptively expanding to the preceding and following neighborhoods of the effective echo interval and extracting the target waveform segment, the data volume can be significantly compressed while retaining key morphological features such as the rising edge, peak value, and falling edge of the effective echo interval, thus meeting the waveform quality requirements of remote diagnosis. By generating positioning metadata containing the absolute starting index and the total number of sampling points, accurate spatiotemporal correlation between local waveform data and the original full-time waveform data can be achieved.

[0068] In some embodiments, the target level includes a first level, a second level, and a third level; please refer to... Figure 5 As shown, Figure 5A schematic flowchart of a graded transmission method for ultrasound data provided in this disclosure is shown. In response to a target grade determined based on thickness variation characteristics, local diagnostic waveform data, and a first communication link, data transmission is performed by invoking either the first or second communication link, including: S501. In response to the thickness change characteristic being less than the preset steady-state threshold and the reliability characteristic of the local diagnostic waveform data being higher than the first threshold, the target level is determined to be the first level, and the thickness value is transmitted by calling the first communication link.

[0069] S502. In response to the thickness change characteristic being not less than the preset steady-state threshold, the confidence characteristic being higher than the second threshold, and the resource budget characteristic of the first communication link satisfying the preset packet sending constraint, the target level is determined to be the second level, and the first communication link is invoked to transmit local diagnostic waveform data.

[0070] S503. In response to the credibility feature being lower than the second threshold, or the resource budget feature not meeting the preset packet sending constraints, the target level is determined to be the third level, and the second communication link is activated to execute full time-domain waveform data transmission.

[0071] Among them, the first threshold is greater than the second threshold; the credibility feature is used to characterize at least one of the morphological validity of the effective echo in the local diagnostic waveform data, the reliability of the thickness calculation result, and the identification of the anomaly source; the resource budget feature is used to characterize at least one of the available air interface duration of the first communication link, the upper limit of radio frequency transmission energy consumption, and the concurrent access carrying capacity of the gateway to which it belongs.

[0072] In this embodiment, the target level is a transmission decision label that is comprehensively evaluated based on the operating status of the tested component, data quality, and communication link resource status. It is used to guide the system to select an appropriate transmission strategy under different operating conditions, thereby optimizing energy consumption and bandwidth usage while ensuring the effectiveness of monitoring.

[0073] Specifically, the preset steady-state threshold is a benchmark value for measuring whether the thickness change characteristics are within a smooth fluctuation range. Its function is to distinguish between the normal service state and the potential abnormal evolution state of the tested component. When the thickness change characteristics are lower than the preset steady-state threshold, it indicates that the wall thickness of the tested component has not changed significantly and is in a low-concern operation state.

[0074] The reliability feature is a comprehensive indicator for quantitatively evaluating the quality of local diagnostic waveform data. Its function is to reflect the reliability of the thickness value calculated based on the current waveform. This feature can characterize the morphological effectiveness of the effective echo, such as whether the echo leading edge is steep and the peak shape is symmetrical. It can also characterize the reliability of the thickness calculation result, such as whether the signal-to-noise ratio meets the minimum requirement. Furthermore, it can characterize the identification of the source of anomalies, such as whether there is waveform distortion caused by probe loosening or coupling agent drying. The higher the reliability feature, the more accurately the current measurement result reflects the physical state of the measured component.

[0075] It should be noted that the first threshold and the second threshold are two critical values ​​for dividing the credibility level. The first threshold is higher than the second threshold. For example, the two together divide the credibility feature into three intervals: high, medium and low, which correspond to three data quality states: no human intervention required, recommended attention, and mandatory verification.

[0076] In this embodiment, the measurement identifier is a unique identifier generated for each measurement process. Its function is to establish the correlation between thickness values, local diagnostic waveform data, and even full time-domain waveform data. The corresponding measurement identifier is sent when data is transmitted at any level on any communication link, which enables the server to accurately bind data of different granularities with specific measurement times and physical locations.

[0077] In this embodiment, the resource budget feature is a comprehensive set of parameters for evaluating the feasibility of the first communication link in transmitting data at the current moment. Its function is to predict whether the current transmission will exceed the energy consumption limit of the node or the carrying capacity of the network. This feature can characterize the available air interface duration, that is, the length of the time window that can be used for data transmission under the current channel occupancy. It can also characterize the upper limit of radio frequency transmission energy consumption, that is, the maximum electrical energy that a single transmission task is allowed to consume. Furthermore, it can characterize the concurrent access carrying capacity of the gateway, that is, the number of newly established transmission links that the gateway can accept in the current time period.

[0078] It should be noted that the preset packet sending constraint is an admission rule set for resource budget characteristics. Only when all indicators of resource budget characteristics are within the allowable range can it be determined that the preset packet sending constraint is met, that is, transmission is allowed through the first communication link.

[0079] For example, when the target level is determined to be the first level, if the detected thickness change feature is less than the preset steady-state threshold, it indicates that the thickness of the measured component has not changed significantly in the near future. At the same time, the reliability feature of the local diagnostic waveform data is higher than the first threshold, indicating that the current echo pattern is complete and the calculation accuracy is high. At this time, it is determined that the current working condition does not require high-frequency monitoring. It is only necessary to call the first communication link to transmit the simplified thickness value and the corresponding measurement identifier, thereby minimizing the communication power consumption.

[0080] When the target level is determined to be the second level, the detected thickness change feature is not less than the preset steady-state threshold, indicating that there may be abnormal trends such as corrosion thinning or erosion. At this time, if the confidence feature is higher than the second threshold but not the first threshold, and the resource budget feature of the first communication link meets the preset packet sending constraint, it is determined that it is necessary to upload more details for analysis. Therefore, the first communication link is called to transmit local diagnostic waveform data and measurement identifiers, so as to provide the server with echo patterns to help determine the nature of the anomaly under the premise of increasing limited transmission overhead.

[0081] When the target level is determined to be the third level, if the credibility feature is lower than the second threshold, it indicates that the current waveform has serious distortion or interference, the calculation result is unreliable, or the resource budget feature does not meet the preset packet sending constraints. For example, the estimated transmission energy consumption will exceed the limit of the power management module or the gateway load is too high. In this case, it is determined that the conventional link cannot meet the transmission requirements, and the second communication link, which is in a dormant or power-off state, is activated to use its high bandwidth feature to transmit the full time domain waveform data, ensuring that the original information is not lost.

[0082] It should also be noted that, in addition to the threshold-based judgment logic mentioned above, time dimension considerations can also be introduced when determining the target level. For example, if multiple consecutive measurement cycles are judged to be at the second level, even if the resource budget feature of the current cycle has a slight surplus, it can be directly upgraded to the third level to actively obtain full data for trend confirmation. Or, if the battery voltage is lower than the preset safety value, even if it is at the first level, it can be forcibly downgraded and only highly compressed feature values ​​can be transmitted.

[0083] Furthermore, this embodiment adaptively adjusts the transmission strategy based on the health status of the tested component and the actual network operating conditions. While ensuring that abnormal data is not missed, it minimizes invalid data transmission under normal operating conditions. By selectively choosing the first or second communication link, it can achieve the complementary advantages of low-power wide-area coverage and high-speed near-field diagnosis, significantly extending the battery life of sensor nodes and improving robustness.

[0084] In some embodiments, the resource budget characteristics satisfying preset packaging constraints include: Based on the modulation parameters of the current first communication link, the number of available payload bytes per packet, and the total amount of local diagnostic waveform data, predict the total number of data packets, total transmission time, and total transmission energy consumption required to transmit the local diagnostic waveform data. If the total number of data packets is less than the preset upper limit, the total transmission time is less than the preset upper limit, and the total transmission energy consumption is less than the preset upper limit, the resource budget characteristics are determined to meet the preset packet sending constraints.

[0085] In some embodiments, the resource budget characteristics satisfying the preset packet sending constraints are a prerequisite for determining whether the first communication link can carry local diagnostic waveform data transmission. Specifically, this can be based on a joint evaluation of the current link physical layer parameters and the size of the data to be transmitted, in order to avoid link blockage, data packet loss or node energy depletion caused by excessively long data packets, excessive transmission time or excessive energy consumption.

[0086] Specifically, modulation parameters are core variables configured in the physical layer of the first communication link. They can include one or a combination of spreading factor, coding rate, and bandwidth. These parameters directly determine the symbol rate and the energy consumption per bit of data. For example, a higher spreading factor can improve receiver sensitivity but will significantly increase the transmission time of a single packet. The number of usable payload bytes per packet refers to the byte capacity of each data packet that can carry effective application data after deducting the frame header, check bits, and network protocol mandatory overhead under the current modulation parameters. This value limits the upper limit of data transmission in a single air interface.

[0087] The total data volume of local diagnostic waveform data refers to the total length in bytes occupied by the target waveform segment to be transmitted and its location metadata. The total number of data packets is the number of physical frames required to transmit complete local diagnostic waveform data; its value is obtained by dividing the total data volume by the number of available payload bytes per packet and rounding up. Too many data packets increase the probability of channel contention and protocol handshake overhead. The total transmission time refers to the cumulative time required to complete the air interface transmission of all data packets, including the actual transmission time of each data packet and the necessary inter-frame interval. An excessively long transmission time increases the time window for node wake-up, increases overall power consumption, and increases the probability of collisions with other devices.

[0088] Total transmit energy consumption refers to the electrical energy expected to be consumed to complete this transmission task. It can be calculated based on transmit power, total transmit duration, and RF front-end efficiency, and is a key indicator for evaluating the impact of transmission behavior on node endurance. The preset packet count limit, preset duration limit, and preset energy consumption limit are system-preset safety thresholds. They are used to limit the maximum number of packets allowed to be split in a single transmission task, the longest air interface time allowed, and the maximum electrical energy allowed to be consumed, respectively. These thresholds can be set based on the gateway's concurrent processing capabilities, the total battery capacity, and the allowed depth of discharge.

[0089] By transforming abstract link resource states into specific packet counts, durations, and energy consumption metrics, and comparing them with preset security thresholds, accurate predictions of transmission feasibility are achieved. This prediction mechanism enables the autonomous assessment of the cost and risk of transmission behavior before invoking the first communication link to transmit local diagnostic waveform data, ensuring that transmission is only performed when resources are sufficient, thereby guaranteeing network stability and the sustainability of node lifecycles.

[0090] For example, when predicting the parameters required for transmission, the processor can read the modulation parameter configuration of the current first communication link and calculate it in combination with the total amount of data in the local diagnostic waveform data. If the current modulation parameter uses a higher spreading factor and the number of available payload bytes per packet is relatively small, the total number of data packets predicted will increase accordingly, and the total transmission time and total transmission energy consumption will also increase.

[0091] When determining whether preset packet sending constraints are met, a triple logical AND operation can be performed: the total number of data packets must be less than the preset packet count limit to prevent the gateway from overloading due to processing too many fragmented packets; the total transmission duration must be less than the preset duration limit to prevent nodes from being kicked out by the network due to prolonged channel occupation; and the total transmission energy consumption must be less than the preset energy consumption limit to prevent a single transmission from depleting the battery. In addition to the above determination based on the current instantaneous state, historical statistical data can be introduced to assist in decision-making. For example, if the average packet loss rate of the first communication link is high recently, even if the current predicted parameters meet the constraints, the preset packet count limit or preset energy consumption limit can be appropriately tightened, proactively downgrading to transmitting only thickness values ​​to reduce the probability of transmission failure. Conversely, if the current battery voltage is sufficient and the network load is light, the preset duration limit can be temporarily relaxed to allow the transmission of larger-scale local diagnostic waveform data.

[0092] In addition, when it is determined that the preset packet sending constraints are not met, a degraded transmission mechanism can be triggered. For example, lossy compression can be performed on the local diagnostic waveform data to reduce the total amount of data, or the resource budget characteristics can be re-evaluated during a network idle period. If the preset packet sending constraints still cannot be met, it is finally determined that the second communication link needs to be activated for transmission.

[0093] By predicting transmission costs based on modulation parameters, single packet payload, and total data volume, and by strictly constraining and judging according to preset multidimensional thresholds, it can ensure that the transmission behavior of local diagnostic waveform data is always within the safe range that the network and nodes can bear. This effectively avoids the risk of network congestion and node downtime caused by blind transmission. By dynamically adjusting the transmission strategy to adapt to changes in link resources, the robustness and survivability of ultrasonic sensing nodes in unstable network environments are significantly improved.

[0094] In some embodiments, triggering the activation of the second communication link to perform full time-domain waveform data transmission includes: The maintenance request and device authentication identifier are reported to the server via the first communication link. The server uses the device authentication identifier to verify the legitimate identity of the sensor node and generates an activation command based on the verification result. The activation command includes a limited duration. In response to the activation command being detected within the pre-configured downlink receiving window of the first communication link, the second communication link, which is in a dormant or power-off state, is activated within a limited time period, and the full time-domain waveform data is transmitted through the activated second communication link. In response to the completion of data transmission or the expiration of the time limit, the second communication link is closed, and the sensor node is controlled to switch to sleep mode.

[0095] Specifically, a maintenance request is a control signal sent by a sensor node to the server via the first communication link to request the activation of the second communication link for data interaction. The request may carry the current working status of the node, fault diagnosis requirements, or data backlog, and is used to prompt the server that deep data interaction is currently required.

[0096] For example, the device authentication identifier is a unique identity credential of the sensor node. Its function is to allow the server to verify whether the node initiating the request is a legitimate access device, preventing unauthorized nodes from maliciously occupying network resources or triggering unnecessary maintenance processes, and ensuring the security of the communication system.

[0097] Specifically, the activation command is a control command issued by the server to the sensor node via the first communication link after successfully verifying the device authentication identifier. This command not only grants the ultrasonic sensor node the permission to enable the second communication link, but also includes a key parameter, the time limit, which is used to constrain the maximum operating time of the second communication link. The time limit is the maximum time window during which the second communication link is allowed to be active, as specified in the activation command. Its purpose is to prevent the second communication link from remaining in a high-power state for an extended period due to software malfunctions, transmission failures, or human negligence.

[0098] It should be noted that the pre-configured downlink receiving window is a specific time segment pre-set in the periodic wake-up mechanism of the first communication link for listening to the downlink data of the server. Since the first communication link usually adopts a non-continuous receiving mode to save energy, the node must listen for the start command within the predetermined window period and cannot respond to the server's call in real time.

[0099] Sleep mode refers to a low-power mode in which the radio frequency front-end and related processing circuits of the second communication link maintain only a minimum power supply or partially shut down the clock. Power-off mode refers to a zero-power mode in which the power supply circuit of the second communication link is physically cut off. Both of these states are the normal state of the second communication link during non-task periods, in order to minimize the overall energy consumption of the node.

[0100] For example, when reporting a maintenance request and device authentication identifier via the first communication link, the node can immediately assemble a request message after determining that full time-domain waveform data needs to be transmitted, and send it to the server using the next available uplink window of the first communication link. The device authentication identifier can be a signature generated by a preset encryption key or a unique serial number. After the server verifies and generates an enable command, the node needs to keep the receiver of the first communication link on within its pre-configured downlink receiving window. If it detects an enable command containing a matching device authentication identifier and a limited duration, the node immediately activates its internal power management module to supply operating voltage to the second communication link or wakes up from sleep mode.

[0101] After activating the second communication link, the sensor node can establish a physical connection with the corresponding access point or server. It then transmits the cached full-time waveform data and corresponding measurement identifiers at high speed through this link. During transmission, the sensor node can monitor the data transmission progress in real time. Once it detects that all the full-time waveform data has been successfully transmitted, or the system runtime has reached the specified duration in the activation command, the sensor node will immediately cut off the power supply to the second communication link or put it back into sleep mode. It will then control the entire sensor node to enter deep sleep mode, awaiting wake-up for the next measurement cycle.

[0102] In addition to the time-threshold-based control described above, after data transmission is complete, the sensor node can also send a transmission completion confirmation signal to the server via the second communication link and receive an immediate sleep command from the server. If the sensor node fails to complete data transmission within the specified time, it can also automatically stop transmission and close the second communication link after the timeout, while reporting the unsuccessful transmission status information to the server via the first communication link, so as to arrange subsequent retransmission or maintenance plans.

[0103] Identity authentication and command issuance are performed through the first communication link, enabling out-of-band control of the high-power channel by the low-power channel. This avoids the significant energy consumption caused by the second communication link being in a standby listening state for extended periods. The second communication link focuses on a collaborative model for large-volume data transmission, fully leveraging its high bandwidth advantage while ensuring data transmission security. Furthermore, even in the event of transmission anomalies or server unresponsiveness, the second communication link will not consume energy indefinitely, effectively mitigating the risk of uncontrolled energy consumption due to anomalies in the high-speed communication module. This allows sensor nodes to obtain full waveform data when in-depth diagnostics are needed, while maintaining extremely low power consumption under normal conditions.

[0104] In some embodiments, the method further includes: Extract the signal-to-noise ratio, peak width consistency, and peak spacing stability within the effective echo range; Detect whether there is amplitude saturation / or clipping anomaly within the effective echo range to obtain waveform anomaly detection results; The credibility feature is obtained by weighted averaging based on parameters such as signal-to-noise ratio, peak width consistency, peak spacing stability, and waveform anomaly detection results.

[0105] In this embodiment, the signal-to-noise ratio (SNR) is the ratio of the effective echo signal power to the noise power within the noise window. Its function is to characterize the prominence of the effective echo in the background noise. A higher SNR indicates a purer echo signal, and the more reliable the thickness value calculated based on it. Peak width consistency refers to the degree of agreement between the pulse width of the current effective echo and the nominal pulse width of the ultrasonic transducer under standard operating conditions or the peak width of historical high-confidence echoes. Its function is to reflect whether the vibration characteristics of the ultrasonic transducer are normal. Significant peak width broadening or narrowing may indicate poor probe coupling, wafer aging, or changes in the condition of the measured surface.

[0106] In this embodiment, peak spacing stability is applicable to scenarios with multiple reflected echoes. It refers to the consistency of the time interval between two adjacent echoes, and its function is to verify the stability of sound wave propagation in the medium. A stable peak spacing indicates that the measured material is uniform and the measurement process is less affected by interference. Waveform anomaly detection is a process of checking whether the signal amplitude within the effective echo interval exceeds the acquisition range of the analog front end. Amplitude saturation or clipping anomalies refer to the phenomenon that the peak part of the echo signal is forcibly truncated because it exceeds the range of the analog-to-digital converter. This anomaly will cause the loss of the echo rising edge or peak characteristics, which will seriously affect the accuracy of thickness calculation. The results of waveform anomaly detection are usually presented in the form of Boolean values ​​or severity levels. Weighted average calculation refers to the mathematical operation of quantifying parameters such as signal-to-noise ratio, peak width consistency, peak spacing stability, and waveform anomaly detection results into a unified scoring scale, and then summing them according to preset weight coefficients. Through this operation, multi-dimensional waveform quality features can be fused into a single confidence feature value.

[0107] For example, when extracting the signal-to-noise ratio (SNR), the ratio of the squared average amplitude within the effective echo interval to the squared average amplitude within the noise window can be calculated and converted to a decibel value on a logarithmic scale. When evaluating peak width consistency, the full width of the effective echo at half the peak value can be measured and compared with the reference pulse width recorded during system calibration to calculate the relative deviation. When detecting amplitude saturation or clipping anomalies, sampling points within the effective echo interval can be scanned to check for multiple consecutive sampling points reaching the maximum positive or minimum negative quantization value of the analog-to-digital converter; if this occurs, clipping anomalies can be identified. When performing weighted average calculations, weights can be assigned based on the degree of influence of each parameter on the thickness calculation accuracy. For example, SNR and clipping anomaly detection results can be given higher weights because they have the most direct impact on measurement error, while peak width consistency and peak spacing stability can be given relatively moderate weights as auxiliary judgment criteria.

[0108] In addition to the simple weighting of the parameters mentioned above, a nonlinear mapping function can be introduced. For example, when a clipping anomaly is detected, the confidence feature can be set to a very low base score, or its contribution can be saturated when the signal-to-noise ratio is higher than a certain quality threshold. Furthermore, the weighting coefficients of the weighted average can be dynamically adjusted. For instance, when it is known that the tested component is in a high-temperature environment leading to increased sound wave attenuation, the expected value of the signal-to-noise ratio can be appropriately reduced, while the focus on peak width consistency can be increased to adapt to special operating conditions.

[0109] By integrating multiple dimensions such as signal-to-noise ratio, peak width consistency, peak spacing stability, and waveform anomaly detection results for weighted average calculation, a comprehensive, objective, and robust credibility feature is generated. This feature can not only accurately reflect the data quality of a single measurement, but also significantly improve the intelligence level and diagnostic accuracy of ultrasonic sensor nodes in complex industrial environments.

[0110] In some embodiments, a sequence of sampling points whose amplitude first crosses a preset trigger threshold and remains continuously above a preset sustain threshold is identified as a candidate echo, including: If the amplitude of the current sampling point is greater than a preset trigger threshold, it is marked as a potential starting point. Starting from the potential starting point, check whether the amplitude of the first preset number of subsequent consecutive sampling points is greater than the preset maintenance threshold. In response to the successful verification, the potential starting point is confirmed as a valid starting point, and the first sampling point with an amplitude lower than the preset maintenance threshold is confirmed as a valid ending point; The sequence of sampling points between the valid start point and the valid end point is determined as candidate echoes.

[0111] In this embodiment, candidate echoes are suspected valid echo sequences identified from the full time-domain waveform data based on a dual-threshold hysteresis comparison mechanism. These are the preliminary discrimination objects for determining the valid echo interval. The potential starting point is the first sampling point whose amplitude exceeds a preset trigger threshold when traversing the full time-domain waveform data; its function is to mark the possible starting position of the echo signal. The preset trigger threshold is a relatively high amplitude threshold set based on the amplitude statistical characteristics of the noise window, used to filter out low-amplitude background noise and trigger the echo identification process.

[0112] The preset sustain threshold is an amplitude threshold below the preset trigger threshold. Together with the preset trigger threshold, it forms a hysteresis interval used to verify the sustained stability of the echo signal and avoid misjudgments when the amplitude fluctuates near the threshold. The first preset quantity is the preset number of consecutive sampling points used to define the minimum duration of the echo sustaining state, ensuring that the identified signal conforms to the physical continuity characteristics of ultrasonic echoes. The valid start point is the echo start sampling point finally confirmed after continuity verification, marking the beginning of the valid echo interval. The valid end point is the first sampling point whose amplitude falls below the preset sustain threshold after the echo signal ends, marking the end of the valid echo interval.

[0113] For example, when traversing the entire time-domain waveform data along the time axis, the amplitude of each sample point is compared with a preset trigger threshold. When the amplitude of the current sample point is detected to be greater than the preset trigger threshold for the first time, this point is marked as a potential starting point, and the maintenance state verification process is immediately initiated. Starting from this potential starting point, subsequent sample points are continuously read, and it is verified whether the amplitudes of a first preset number of consecutive sample points are all greater than the preset maintenance threshold. If the amplitude of any sample point within this continuous verification window falls below the preset maintenance threshold, the trigger is deemed invalid, the potential starting point mark is cleared, and the traversal continues to find a new trigger point.

[0114] If the verification passes, meaning that the first preset number of consecutive sampling points are all above the preset maintenance threshold, then the potential starting point is confirmed as a valid starting point, and subsequent sampling points continue to be monitored. Thereafter, amplitude changes will be continuously tracked until the amplitude of a sampling point is detected to be below the preset maintenance threshold, and this state continues to meet the termination condition. At this point, the first sampling point below the preset maintenance threshold is confirmed as a valid termination point. In addition to the hard verification based on the number of consecutive sampling points, a dynamic window adjustment mechanism can be introduced during the verification process. For example, when the sampling rate changes, the size of the first preset number can be adjusted accordingly to maintain constant sensitivity to the echo pulse width. Furthermore, after confirming a valid starting point, if the amplitude briefly drops but recovers within a very short time, the locked state can be maintained according to preset anti-interference logic, without immediately determining termination, thereby further improving the robustness of recognition in complex noisy environments.

[0115] By employing a dual-threshold hysteresis identification method that combines initial trigger threshold crossing with subsequent continuous maintenance verification, transient noise pulses and valid ultrasonic echoes with sustained characteristics are effectively distinguished. First, a higher preset trigger threshold is used to capture the initial characteristics of the signal. Then, by checking whether subsequent sampling points are consistently higher than a lower preset maintenance threshold, interference signals that only briefly exceed the trigger threshold but cannot be maintained are filtered out. This accurately locks onto echo sequences that conform to physical laws, effectively suppressing momentary false triggers caused by circuit noise and electromagnetic interference, significantly improving the accuracy and reliability of echo identification by precisely defining the valid start and end points.

[0116] In some embodiments, the method further includes: after confirming a valid starting point, if the amplitude is detected to fall below a preset maintenance threshold within a first preset number of consecutive sampling points, but then rises above the preset maintenance threshold again within a time period less than the first preset number of sampling points, then the number of consecutive sampling points below the preset maintenance threshold reaches a first preset number, and the first sampling point below the preset maintenance threshold is confirmed as a valid termination point.

[0117] After confirming a valid starting point, the system enters echo sustaining monitoring mode, verifying the relationship between the amplitude of subsequent sampling points and the preset sustaining threshold point by point. If the amplitude is detected to fall below the preset sustaining threshold at a certain point, no immediate response is initiated; instead, a temporary observation counter is started. If this fall lasts for less than a first preset number of sampling points, and the amplitude subsequently rises back above the preset sustaining threshold, the fall is ignored, the observation counter is reset, and the lock on the current candidate echo continues. During this process, monitoring continues until, after a certain amplitude fall, the number of consecutive sampling points below the preset sustaining threshold truly reaches the first preset number. Only then is the echo signal deemed to have substantially ended, and the first sampling point in the sequence below the preset sustaining threshold is confirmed as a valid termination point.

[0118] In addition to determining the amplitude based on the number of consecutive sampling points, the depth of amplitude drop can also be used as an auxiliary factor. For example, when the amplitude drops briefly but does not fall below 50% of the preset maintenance threshold, even if the drop duration is close to the first preset number, the observation window can be extended as appropriate. Alternatively, when the drop depth is extremely large, the required number of consecutive sampling points below the threshold can be appropriately shortened to more flexibly adapt to different echo signal patterns. Furthermore, during the locked state, if the amplitude is detected to cross the high trigger threshold again, the echo identification process can be reset to address potential multiple echo overlaps.

[0119] In some embodiments, the amplitude statistics include the average amplitude; determining the preset trigger threshold and the preset sustain threshold based on the amplitude statistics of the noise window includes: The preset trigger threshold is determined based on the product of the average amplitude of the noise window and the first coefficient. The preset maintenance threshold is determined based on the product of the average amplitude of the noise window and the second coefficient, where the first coefficient is greater than the second coefficient.

[0120] In this embodiment, the first coefficient is a positive real-valued scaling factor greater than 1, which amplifies the average amplitude to a level sufficient to distinguish valid echoes from background noise, thereby setting a reasonable trigger sensitivity. The second coefficient is a positive real-valued scaling factor less than the first coefficient, which sets a sustaining baseline below the trigger threshold, forming a hysteresis interval with the preset trigger threshold to prevent misjudgment when the echo amplitude fluctuates near the threshold. The preset trigger threshold is a higher amplitude threshold determined based on the average amplitude and the first coefficient, used to capture the starting point of the echo signal. The preset sustaining threshold is a lower amplitude threshold determined based on the average amplitude and the second coefficient, used to verify the continuous stability of the echo signal.

[0121] For example, after acquiring the noise window, the processor can calculate the average amplitude of all sampling points within the window. This average is then multiplied by a first coefficient to obtain a preset trigger threshold, and simultaneously multiplied by a second coefficient to obtain a preset sustain threshold. Since the first coefficient is greater than the second coefficient, the preset trigger threshold is always higher than the preset sustain threshold, and the difference between the two constitutes the hysteresis interval for echo recognition. In practical applications, the first and second coefficients can be preset based on the acoustic characteristics of the material being tested, the sensitivity of the probe, and historical noise data of the environment. For example, in a relatively stable noise environment, a smaller coefficient difference can be selected to maintain higher recognition sensitivity; in an environment with more interference, the coefficient difference can be appropriately increased to improve anti-interference capability. Furthermore, this scheme can also work in conjunction with an automatic gain control circuit. When the overall echo signal is detected to be weak, the processor can dynamically adjust the values ​​of the first and second coefficients, appropriately lowering the threshold level to ensure that effective echoes can be triggered normally.

[0122] By determining high and low thresholds based on the product relationship between the average amplitude and different coefficients, the threshold can be adaptively adjusted with low computational cost, ensuring that the echo identification threshold always maintains a reasonable proportional relationship with the current noise baseline. This method has low computational complexity and is suitable for conventional industrial monitoring scenarios with relatively stable noise levels and small fluctuations. It can ensure identification accuracy and reduce computational load.

[0123] In some embodiments, the amplitude statistical characteristics include the amplitude standard deviation and the amplitude average; determining the preset trigger threshold and the preset sustain threshold based on the amplitude statistical characteristics of the noise window further includes: Calculate the standard deviation of the noise window amplitude; The preset trigger threshold and preset sustain threshold are determined based on a linear combination of the average amplitude and the standard deviation of the amplitude.

[0124] Both the preset trigger threshold and the preset sustain threshold increase with the increase of the amplitude standard deviation, and the increase rate of the preset trigger threshold is greater than the increase rate of the preset sustain threshold.

[0125] In this embodiment, the amplitude standard deviation is a measure of the dispersion of the amplitude of sampling points within the noise window from the average amplitude, used to quantify the severity of noise fluctuations. Linear combination refers to a mathematical operation that superimposes the average amplitude and the amplitude standard deviation according to a preset ratio. Its function is to incorporate both the baseline level and fluctuation characteristics of the noise into the threshold calculation model, enabling the generated threshold to simultaneously respond to both static and dynamic noise features. The increase rate refers to how quickly the threshold increases with the amplitude standard deviation. The increase rate of the preset trigger threshold is greater than the increase rate of the preset maintenance threshold. When noise fluctuations intensify, not only is the overall recognition threshold raised, but the hysteresis interval is also widened accordingly, thereby enhancing the anti-interference capability of the ultrasonic sensor node in harsh environments.

[0126] For example, the processor first calculates the average amplitude and standard deviation of the noise window. Then, it calculates a preset trigger threshold and a preset sustain threshold using a linear combination formula. For instance, the preset trigger threshold can be expressed as the sum of the products of the average amplitude, the standard deviation of the amplitude, and a first weighting coefficient; the preset sustain threshold can be expressed as the sum of the products of the average amplitude, the standard deviation of the amplitude, and a second weighting coefficient. The first weighting coefficient is greater than the second weighting coefficient, ensuring that the rate of increase of the preset trigger threshold is greater than the rate of increase of the preset sustain threshold.

[0127] When environmental noise is relatively stable and the amplitude standard deviation is small, the preset trigger threshold and preset sustain threshold are mainly determined by the average amplitude, maintaining high sensitivity. When environmental noise fluctuates drastically and the amplitude standard deviation increases, the preset trigger threshold and preset sustain threshold will increase significantly with the increase in standard deviation, and the increase in the preset trigger threshold is more pronounced, thus effectively avoiding the peak noise region. Furthermore, the weighting coefficients in this scheme can be pre-configured according to different application scenarios. For example, in scenarios with high real-time measurement requirements, the weighting coefficients can be appropriately reduced to speed up the response; in scenarios with extremely high measurement accuracy requirements, the weighting coefficients can be appropriately increased to ensure the stability of the recognition results.

[0128] By introducing the amplitude standard deviation and determining the threshold based on a linear combination of the mean and standard deviation, the preset trigger threshold and preset sustain threshold can adaptively increase with the increase of noise fluctuation amplitude, and the difference between the two can be widened synchronously. This effectively suppresses false triggering in environments with drastic noise changes, and significantly enhances the adaptability and measurement reliability of ultrasonic sensor nodes in complex and harsh industrial environments.

[0129] In some embodiments, during the traversal of the full time-domain waveform data along the time axis, the following search window strategy is used to search for candidate echoes: Obtain the reference echo position and construct a priority search window based on the reference echo position, where the reference echo position is the position of the valid echo identified during the historical confidence measurement process; First, candidate echoes are searched within the priority search window. If no candidate echo that meets the preset confidence condition is identified within the priority search window, candidate echoes are searched within the full-range time window. The search interval corresponding to the priority search window is smaller than the search interval corresponding to the full-range time window.

[0130] In this embodiment, differentiated time windows are introduced during the traversal of the full time-domain waveform data to adapt to the long-term monitoring needs of fixed-installation ultrasonic sensor nodes. The reference echo position is the coordinate of the valid echo identified during historical high-confidence measurements on the time axis. This coordinate can be fixed based on the high-precision echo position acquired during the initial installation and commissioning phase, or it can be dynamically updated statistically based on the echo positions of high-confidence measurements within the most recent preset number of measurements. For example, the arithmetic mean or median can be used as a representative value to provide empirical prior information on the location of valid echoes of the tested component under current operating conditions.

[0131] The priority search window is a finite time interval centered on the reference echo position. Its width is much smaller than the full-range time window. Its function is to focus the echo search range within a reasonable fluctuation area near the historical valid echo position, significantly reducing the amount of computation and the probability of false identification. The full-range time window is a search interval covering the entire time span from the ultrasonic pulse emission until the echo completely attenuates and disappears. Its function is to ensure that even when the thickness of the measured component changes significantly, causing the echo position to shift out of the priority search window, a valid echo can still be captured through a large-scale search.

[0132] It should be noted that the preset confidence condition is a comprehensive index system used to judge whether the quality of candidate echoes meets the standards. It may include signal-to-noise ratio threshold, peak width consistency threshold, or waveform integrity index. Only candidate echoes that meet this condition can be recognized as valid search results.

[0133] For example, when acquiring the reference echo position, a high-precision valid echo position can be identified and stored during the initial installation and calibration phase through manual intervention or an automatic algorithm. In subsequent routine monitoring, this reference echo position is first retrieved, and using this position as the midpoint, a first adaptation length is extended forward along the time axis, and a second adaptation length is extended backward, thereby constructing a priority search window. The first and second adaptation lengths can be set based on the maximum pulse width of historical echoes, the system sampling rate, and the maximum allowable wall thickness change rate of the measured component.

[0134] When traversing the full time-domain waveform data, the processor first concentrates its computing resources on the sampling points within a priority search window. A dual-threshold hysteresis algorithm is then used within this priority search window to search for candidate echoes that meet preset confidence criteria. If a candidate echo that meets the criteria is successfully identified within the priority search window, the search is considered successful, and further large-scale traversal is not performed.

[0135] If no candidate echoes meeting the confidence criteria are found within the priority search window—for example, in cases where severe corrosion leads to a sudden decrease in wall thickness—the window restriction is automatically lifted, and the search begins again from the beginning within the full-range time window to ensure that no valid echoes that have experienced positional shifts are missed. In addition to the confidence-based criteria, a consecutive failure count mechanism can be introduced. That is, a full-range time window search is triggered only if no valid echo is identified within the priority search window within a preset number of consecutive measurement cycles, thereby avoiding frequent full-window scans caused by occasional noise interference.

[0136] By adopting a hierarchical search strategy that combines a priority search window based on the reference echo position with a full-range time window, the computational efficiency of echo identification can be greatly improved by making full use of the physical characteristics of the measured component, and the average power consumption of the system can be reduced. At the same time, when facing extreme working conditions with significant changes in wall thickness, the continuity and reliability of monitoring can still be guaranteed through full-range search, significantly enhancing the adaptive capability and lifespan of ultrasonic sensor nodes in long-term unattended scenarios.

[0137] In some embodiments, a priority search window is constructed based on the reference echo position, including: Using the reference echo position as a reference, the first adaptation length is extended forward along the time axis, and the second adaptation length is extended backward to form a priority search window; wherein, the first adaptation length and the second adaptation length are determined based on at least one of the following: the maximum pulse width of historical effective echoes, the sampling rate, and the maximum allowable wall thickness change rate of the measured component.

[0138] For example, when constructing a priority search window based on a reference echo position, the processor can read the maximum pulse width of pre-stored historical valid echoes and use this pulse width as a basic extension unit to extend forward by a first adaptation length and backward by a second adaptation length along the time axis, thereby forming a symmetrical search interval centered on the reference echo position. The first and second adaptation lengths can be set directly based on a multiple of the maximum pulse width. For example, extending forward by a preset multiple of the maximum pulse width can fully capture the rising edge, and extending backward by a preset multiple of the maximum pulse width can cover the falling edge and part of the trailing edge. In addition to setting based on historical pulse widths, the extension length in the time dimension can also be converted into the corresponding number of sampling points by combining the sampling rate to adapt to the accuracy requirements of different hardware platforms.

[0139] Furthermore, the first and second adapter lengths can be dynamically adjusted according to the maximum allowable wall thickness change rate of the measured component. For example, when the measured object is in normal operation, the wall thickness change rate is low, and a smaller extension length can be used to improve search efficiency. When it is detected that the measured object may be in an abnormal corrosion or erosion state, the extension length can be appropriately increased to ensure that no effective echoes are missed. In some embodiments, the first and second adapter lengths can be asymmetrically configured, for example, the backward extension length is greater than the forward extension length, to adapt to the characteristic that the echo attenuation at the trailing edge is relatively long when ultrasound propagates in the measured medium.

[0140] By dynamically constructing a priority search window based on multi-dimensional parameters such as the maximum pulse width of historical effective echoes, sampling rate, and the maximum allowable wall thickness change rate of the tested component, the range of the search window can always match the echo distribution characteristics under actual working conditions, maximizing search efficiency while ensuring that no effective echoes are missed.

[0141] In some embodiments, the method further includes: when a valid echo that meets a preset confidence condition is identified within the full-range time window, updating the position of the valid echo to a new reference echo position, and adjusting the range of the priority search window based on the new reference echo position.

[0142] For example, when a valid echo meeting the preset confidence condition is successfully identified within the full-range time window, the processor can extract the measured location of the valid echo and replace the original reference echo location for storage. Subsequently, the priority search window can be reconstructed based on the new reference echo location and according to the aforementioned adaptation length determination rules, so that the initial search range of the next measurement cycle can focus on the area where the updated valid echo may be distributed.

[0143] In practical applications, to avoid contaminating the reference position with occasional abnormal data, a smooth update strategy can be adopted. For example, a weighted average can be calculated between the new valid echo position and the original reference echo position, or a sliding window can be used to maintain a set of echo positions from the most recent high-confidence measurements and calculate its statistical representative value. When multiple consecutive measurements fail to identify valid echoes within the full-range time window, the initial range of the priority search window can be gradually expanded, or the trigger sensitivity for the full-range search can be increased to adapt to conditions where the measured object is in a rapidly changing phase. Conversely, when the recognition success rate within the priority search window recovers to a high level, the range of the priority search window can be gradually narrowed, returning to the highly efficient small-window search mode.

[0144] By introducing a dynamic update mechanism for the reference echo position, the priority search window can adaptively follow the evolution of the physical state of the measured component. Under conditions such as slow changes in wall thickness or sudden abnormal thinning, it can obtain real and effective echoes through full-range search and calibrate the search reference in a timely manner, thereby ensuring the accuracy of echo identification and efficient utilization of computing resources during long-term monitoring.

[0145] The following specific embodiment will be used to uniformly describe and explain in detail the relevant content involved in the embodiments of this disclosure: For example, when a sensor node performs a routine pulse-echo measurement, the obtained full-time waveform data is denoted as X={x(0),x(1),…,x(N)}. 1)}, where N is the total number of complete sampling points. This full-volume time-domain waveform data is a time-domain echo sampling sequence acquired along the time axis after the ultrasonic transducer emits a short pulse. The local diagnostic waveform data contains continuous raw sampled values ​​copied from the full-volume time-domain waveform data, rather than consisting only of characteristic parameters such as peak value and arrival time. Each local segment is associated with at least the measurement identifier, timestamp, segment start index, segment length, total number of points in the full-volume time-domain waveform data, sampling interval, quantization method, effective peak index, temperature and wall thickness results, and can carry waveform features such as peak amplitude, peak width, inter-peak time, signal-to-noise ratio, threshold, saturation / clipping flag, and diagnostic confidence. Based on the above metadata, the server can map the local diagnostic waveform data to the correct position in the full time coordinate, thereby preserving the echo profile, side lobes, and noise floor morphology.

[0146] In the effective echo identification and adaptive extraction of local original waveforms, firstly, full-time waveform data, temperature, battery status, and timestamps are acquired, and a unique measurement identifier is generated for this measurement. The full-time waveform data is preferably written to non-volatile memory first. Then, one or more operations are performed on the full-time waveform data, including DC removal, rectification and absolute value extraction, envelope extraction, smoothing, or bandpass processing, to obtain the amplitude sequence y(i) used for judgment, while retaining the original sampled values ​​to form local diagnostic waveform data. Next, the excitation dead zone and candidate echo time windows are determined. The excitation dead zone is used to exclude the transmission pulse, transducer ringing, and circuit recovery stages. The candidate echo time windows can be determined based on preset minimum / maximum wall thickness, material sound velocity, the previous effective thickness, or historical peak positions.

[0147] During the baseline establishment phase, after personnel complete sensor fixation, coupling, and initial measurement confirmation, the node or debugging terminal analyzes a set of full-time-domain waveform data, recording the reference sampling point indices of the first and second effective backwall echoes, denoted as i1,ref and i2,ref, respectively. N_pre,ref sampling points are reserved before the first echo, and N_post,ref sampling points are reserved after the second echo to cover the rising edge, falling edge, sidelobes, and necessary background noise of the echoes. This yields the reference start point, reference end point, and maximum local return length for subsequent priority searches and local return.

[0148] i_start,ref=max(0,i1,ref N_pre,ref); i_end,ref=min(N 1, i2, ref+N_post, ref); L_max=i_end,ref i_start,ref+1; Where i_start,ref is the starting sampling point index of the reference local waveform; i_end,ref is the ending sampling point index of the reference local waveform; N is the total number of sampling points in one complete A-scan signal; N_pre,ref and N_post,ref are the number of sampling points reserved before the first echo and after the second echo when the mounting reference is established, respectively; L_max is the maximum number of sampling points allowed to be included in the reference local waveform. max(0,·) and min(N 1,·) is used to limit the index from exceeding the valid range of the complete A-scan data.

[0149] In the noise statistics and adaptive threshold determination steps, a sampling region known to lack valid ultrasonic echoes is selected from the full time-domain waveform data as a noise window. This noise window can be located outside the candidate echo time window, or it can be a region without valid echoes predetermined during equipment installation or initialization. The set of sampling point indices contained within the noise window is denoted as: Let W_n be the set of sampling point indices contained in the noise window, and let N_n be the total number of sampling points in the noise window, where N_n>1; let y(i) be the i-th sampling point after absolute value taking, envelope extraction, smoothing or other amplitude processing.

[0150] μ_n=(1 / N_n)·Σy(i),i∈W_n; where μ_n is the average noise amplitude, used to represent the average level of the signal noise floor when there is no effective echo.

[0151] σ_n=√{[Σ(y(i) μ_n)²] / (N_n 1)}, i∈W_n; Where σ_n is the standard deviation of the noise amplitude, which is used to represent the degree to which the noise sampling points fluctuate around the average noise value; the more unstable the noise, the larger σ_n is, and the more stable the noise, the smaller σ_n is.

[0152] T_H=μ_n+k_H·σ_n T_L=μ_n+k_L·σ_n k_H>k_L>0; Wherein, T_H is a preset trigger threshold used to confirm the start of a candidate echo; T_L is a preset sustain threshold used to confirm when a candidate echo that has started ends; k_H is a high threshold coefficient, and k_L is a low threshold coefficient. k_H and k_L can be determined through pre-testing, calibration, or statistical analysis of historical data, based on the transducer type, analog gain, sampling rate, ambient noise level, and required detection sensitivity. For example, under the condition that k_H > k_L, k_H can be 4–6, and k_L can be 2–4, but this invention is not limited to these ranges. A larger k_H results in a lower probability of false noise triggering, but may decrease the detection sensitivity for weak echoes; a smaller k_H makes it easier to detect weak echoes, but may increase the probability of false noise triggering.

[0153] Since k_H > k_L, therefore T_H > T_L. Within the candidate echo time window, when the signal amplitude first exceeds the preset trigger threshold T_H, and subsequently remains above the preset sustain threshold T_L for at least a preset number of consecutive sampling points, a candidate echo is determined to have started. After a candidate echo starts, even if the signal amplitude drops below T_H, as long as it remains above T_L, it is still considered part of the same candidate echo; only when the signal amplitude falls below T_L for a preset number of consecutive sampling points is the candidate echo determined to have ended. This high and low dual-threshold hysteresis determination can reduce false triggering by single-point noise and repeated segmentation of the same echo when the signal fluctuates around a single threshold. In other embodiments, the high and preset sustain thresholds can also be determined using the root mean square noise multiple, noise amplitude percentile, historical noise statistics, or a combination thereof.

[0154] Within the candidate echo time window, a high threshold crossing point is searched. A candidate echo is formed only when the amplitude exceeds the preset trigger threshold T_H, and subsequently, at least L_min consecutive sampling points remain above the preset sustaining threshold T_L. Here, L_min is the minimum number of continuous sampling points for a candidate echo, used to exclude transient spikes where only a single or a few sampling points cross the threshold; its value can be preset or calibrated based on the sampling rate, transducer bandwidth, typical echo duration, and ambient noise conditions. Furthermore, transient spikes and erroneous echoes are excluded based on peak width, peak amplitude, saturation state, interval with adjacent echoes, and deviation from historical valid peak positions.

[0155] For each confirmed valid echo, echo boundaries are searched forward and backward along the time axis. When the amplitude is below a preset maintenance threshold T_L for Q_end consecutive sampling points, the corresponding position is determined as the basic boundary in that direction. Here, Q_end is the number of consecutive low-threshold sampling points required to confirm that the echo has left the threshold region, used to prevent premature truncation caused by brief drops or single-point fluctuations within the echo. Subsequently, N_pre sampling points are extended before the basic starting point and N_post sampling points are extended after the basic ending point to preserve the rising edge, falling edge, noise floor, and adjacent sidelobes of the echo. N_pre and N_post can be fixed values ​​or adaptively determined based on peak width, sampling rate, transducer bandwidth, or the current LoRa resource budget. Here, N_pre and N_post are used to represent the number of extended points before and after, to distinguish them from N, which represents the total number of sampling points for a complete A-scan.

[0156] When the interval between two valid echo segments is less than the merging threshold, they are merged into a single continuous local segment; otherwise, multiple segments can be formed. For implementations that calculate thickness using the interval between adjacent backwall echoes, it is preferable to retain at least two corresponding echoes used for thickness calculation and their surrounding original sampled values.

[0157] The original sampled values ​​within the above-mentioned segment are combined with the location metadata to form a local diagnostic waveform. The 12-bit ADC samples can be packed into bits in a manner where every two 12-bit samples occupy 3 bytes, or they can be stored as 16-bit unsigned numbers; this encoding method does not change the physical meaning of the local original waveform.

[0158] In some embodiments, diagnostic confidence, or confidence features, are used to evaluate whether the current local waveform is sufficient to support remote judgment. Confidence features can be obtained by rule-based decision-making, weighted scoring, or machine learning classification, preferably considering at least the following metrics: the signal-to-noise ratio of the effective echo window; the duration of the candidate peak; whether the spacing between the two echoes used for thickness measurement is consistent; the deviation of the peak spacing from historical trends; whether the peak is saturated or clipped; whether the number of effective echoes is unique; whether the local segment boundary is intact; and whether the temperature, gain, and battery status are within acceptable ranges.

[0159] Signal quality can be evaluated using the average power of the effective echo window and the noise window. Let W_n be the set of sampling point indices for the noise window, and N_n be the number of sampling points; let W_s be the set of sampling point indices for the effective echo window, and N_s be the number of sampling points. The average power P_n of the noise window and the total average power P_e of the effective echo window can be calculated using the following formulas: P_n=(1 / N_n)·Σy(i)²,i∈W_n P_e=(1 / N_s)·Σy(i)²,i∈W_s P_sig=max(P_e P_n,ε) SNR=10·log 10 (P_sig / P_n); Where P_n is the average power of the noise window; P_e is the total average power formed by the signal and noise within the effective echo window; P_sig is the effective signal power estimated after deducting the noise power; ε is the power to prevent P_e from being emitted. P_n is a very small positive number set to be equal to zero or a non-positive value; SNR is an estimated signal-to-noise ratio expressed in decibels. If P_e / P_n is used directly without deducting noise power, then this ratio should be expressed as the power ratio of the echo window to the noise window, rather than the pure signal-to-noise ratio in a strict sense.

[0160] An exemplary comprehensive diagnostic reliability can be represented using a weighted scoring method as follows: C = Σw_j·Q_j, j = 1, 2, ..., m Σw_j=1, j=1,2,…,m; w_j≥0; 0≤Q_j≤1; Where C represents the overall diagnostic reliability; m represents the total number of quality indicators involved in the evaluation; Q_j represents the j-th normalized quality indicator; and w_j represents the weight corresponding to the j-th quality indicator. Each quality indicator is preferably homogenized so that a larger Q_j indicates higher signal quality or diagnostic sufficiency. For example, signal-to-noise ratio, peak spacing consistency, and boundary integrity can directly constitute positive quality indicators; saturation degree, clipping degree, and candidate peak ambiguity degree can be converted into non-saturated quality indicators, non-clipping quality indicators, and peak uniqueness indicators before being weighted. Weights can be determined through experimental calibration, expert experience, or training with historical data. This formula is only one embodiment and does not limit specific indicators, weights, or scoring methods. A first reliability threshold C1 and a second reliability threshold C2 can be set, where C1 > C2.

[0161] The sensor node or server determines the maximum allowed number of application layer payload bytes, B_max(DR,Region), based on the region frequency plan used by the sensor node and the current LoRaWAN data rate class (DR). After deducting the application layer frame header, measurement identifier, location metadata, and application layer check fields, the number of bytes per packet that can be used to carry local raw sampled values, B_data, is obtained. B_data = B_max(DR,Region) B_header B_meta B_check; Here, B_header represents the number of bytes occupied by the application layer frame header; B_meta represents the number of bytes occupied by metadata such as measurement identifier, segment index, complete point count, sampling interval, and quantization mode; B_check represents the number of bytes occupied by the application layer check field; and B_data represents the number of bytes in a single LoRa application layer data packet that can actually be used for waveform sampling values. B_data should be greater than zero; otherwise, the DR and message format cannot carry waveform sampling values.

[0162] When each sample value is tightly packed into a continuous bit stream of q bits, the maximum number of sample points N_pkt that a data packet can hold and the total number of data packets K required for local diagnostic waveforms can be estimated by the following formula: N_pkt = floor(8·B_data / q) K = ceil(N_local / N_pkt); Where q is the number of bits actually occupied by each sample value after encoding; N_local is the total number of local original sample values ​​to be sent this time; N_pkt is the maximum number of sample points that a LoRa application layer data packet can hold; K is the total number of data packets required to send all local sample values; floor(·) means round down, and ceil(·) means round up. If the sample values ​​are not packed with consecutive bits, but each sample value is stored as b_s complete bytes, then N_pkt=floor(B_data / b_s) can be used to calculate the number of sample points per packet.

[0163] The communication scheduling unit can also calculate the air interface transmission time of each data packet based on the spreading factor, bandwidth, coding rate, and payload length, and further calculate the cumulative air interface time T_air and the expected RF transmission energy E_tx for transmitting all local waveform data. T_air=ΣT_packet,r,r=1,2,…,K E_tx = V_tx·I_tx·T_air; Where K is the total number of data packets in this partial waveform transmission; r is the data packet sequence number; T_packet,r is the air interface transmission time of the r-th data packet; T_air is the cumulative air interface transmission time of all data packets; V_tx is the supply voltage during LoRa transmission; I_tx is the average transmission current under the corresponding transmission power and operating mode; and E_tx is the estimated LoRa RF transmission energy. This energy is an estimated value of communication transmission energy consumption and by default does not include additional energy consumption generated by MCU operations, memory read / write, receive window, and other peripheral circuits. If any of K, T_air, or E_tx exceeds the preset resource budget, the partial waveform will no longer be forced to be transmitted via LoRa.

[0164] LoRa local waveform messages preferably include: frame type, measurement identifier, current segment or data packet sequence number, total number of segments or packets, starting index of the segment in the complete waveform, number of complete points, sampling interval, quantization method, local raw sample value, and check field. After each data packet is successfully sent or acknowledged, a task progress index is written to non-volatile memory so that transmission can continue from the next packet after hibernation, power failure, or link interruption.

[0165] In some embodiments, based on the diagnostic confidence and resource budget assessment results described above, the following three-level data transmission strategy can be implemented: Level 1: When the wall thickness result is within the normal range, the thickness change rate does not exceed the trend threshold, and the diagnostic confidence is not lower than the first confidence threshold C1, only the wall thickness, temperature, battery level, status code, and confidence summary are sent through the first communication link. Full time-domain waveform data is retained locally or overwritten cyclically according to the storage strategy.

[0166] Level 2: When the wall thickness result is abnormal, the trend changes, or the server actively requests auxiliary diagnosis, but the diagnostic confidence is still not lower than the second confidence threshold C2, and the number of packets, air interface time, and expected power consumption of the local diagnostic waveform data do not exceed the LoRa budget, the local diagnostic waveform data and location metadata are sent through the first communication link. The remote end can see the true shape around the effective echo without transmitting all background sampling points.

[0167] Level 3: When no valid echo is detected, multiple indistinguishable candidate peaks appear, the signal-to-noise ratio is too low, waveform saturation / cropping occurs, local segment boundaries are incomplete, confidence level is lower than C2, local data exceeds the LoRa budget, or the server explicitly requests complete data, the ultrasonic sensor node (also referred to as the sensor node) reports an alarm, measurement identifier, and availability status of the second communication link via the first communication link. The server issues or pre-configures authorization to enable the second communication link within the LoRa receiving window, and the node broadcasts the activation of the second communication link within a limited time. After completing identity verification, the on-site mobile terminal downloads the full time-domain waveform data, historical data, or performs a firmware upgrade. Furthermore, the selection of the communication link is no longer solely determined by file size, but rather by a combination of factors including whether there is an anomaly, whether the local waveform is sufficient for diagnosis, and whether the low-bandwidth link can handle the load.

[0168] During the BLE authorization and high-speed on-site maintenance phases, the ultrasonic sensor node remains in a deep sleep state with the RTC as the wake-up source during normal monitoring, and the second communication link remains powered off or in deep sleep. Upon triggering Level 3, the ultrasonic sensor node can first report a maintenance request via the first communication link; remote users can also actively select a specific ultrasonic sensor node on the application server, which will then send a "force BLE enable" command to that ultrasonic sensor node via the LoRaWAN network server. For LoRaWAN Class A nodes, this command is preferably sent within the RX1 or RX2 receive window after the node's next uplink. The enable command may include a session identifier, validity period, allowed operation types, and authentication information.

[0169] The second communication link broadcasts only within its valid time period and automatically closes upon timeout or completion of the operation, after which the ultrasonic sensor node re-enters deep sleep. During full time-domain waveform data download, the mobile terminal reads the specified complete data based on the measurement identifier and performs segmentation, verification, and retransmission. During firmware upgrades, the ultrasonic sensor node receives the firmware image, verifies the version, length, hash value, or digital signature, writes it to the upgrade area, and reports the version and status via the first communication link after the upgrade is complete.

[0170] In some embodiments, the reference thickness d_ref can be determined using a calibrated ultrasonic instrument or other reliable method at a location where the inner wall of the material under test is flat and free from significant corrosion. At multiple stable calibration temperatures T_i, the time interval Δt_i between two adjacent corresponding backwall echoes is acquired, and the material sound velocity v_i corresponding to each calibration temperature is calculated. v_i = 2·d_ref / Δt_i; Where d_ref is the known thickness at the reference location; T_i is the i-th calibration temperature; Δt_i is the time interval between two adjacent backwall echoes at temperature T_i; and v_i is the calibration velocity of sound at temperature T_i.

[0171] A temperature-velocity of sound mapping table is established using multiple temperature points. When the current measured temperature T lies between the adjacent calibration temperatures T_i and T_(i+1), i.e., T_i≤T≤T_(i+1), linear interpolation can be used to obtain the corrected velocity of sound v(T) corresponding to the current temperature: v(T) = v_i + [(v_(i+1)] v_i) / (T_(i+1) T_i)]·(T T_i); Where v_i and v_(i+1) are the speeds of sound corresponding to adjacent calibration temperatures T_i and T_(i+1), respectively; T is the current measured temperature; and v(T) is the current temperature-corrected speed of sound obtained from the mapping relationship between temperature and speed of sound.

[0172] During measurement, the time interval between two adjacent corresponding backwall echoes in the current A-scan signal is denoted as Δt_e, and the thickness d of the measured component is calculated using the following formula: d = v(T)·Δt_e / 2; Where Δt_e is the propagation time interval between two adjacent corresponding backwall echoes in the current measurement, and d is the calculated thickness of the measured component. It should be noted that Δt_e represents the time difference between echoes, which is different from the aforementioned time interval Δt_s between adjacent ADC sampling points.

[0173] The beneficial effects of this embodiment include, but are not limited to: preserving the original sampled values ​​around the valid echo, rather than only retaining abstract features such as peak values, thereby significantly reducing the amount of data while still preserving echo contours, peak widths, sidelobes, noise, and clipping information, improving remote anomaly diagnosis capabilities; using blind zones, candidate time windows, adaptive dual thresholds, continuous point counts, and inter-peak relationships to jointly identify valid echoes, reducing the probability of mistakenly truncating excitation ringing or transient noise as valid waveforms; local waveforms carrying start indexes, complete point counts, and sampling intervals, allowing the server to accurately display local data in complete coordinates, avoiding the loss of time or distance positional relationships after truncation; and improving diagnostic reliability and... LoRa resource budget jointly determines the transmission level, avoiding meaningless transmission of complete data and preventing the continued transmission of incomplete information when the local waveform is insufficient for diagnosis; the full time-domain waveform data and firmware are carried by the high-speed link of the second communication link in the field, shortening the RF working time, reducing the shared air interface occupation of the first communication link, and reducing the risk of congestion in large-scale node deployments; the second communication link is normally dormant and is only triggered by authorization or specified conditions of the first communication link, providing high-speed maintenance capabilities without disassembly while maintaining daily low-power wide-area monitoring; dynamic packet subdivision and breakpoint resumption reduce repeated transmissions caused by data rate changes, dormancy, power failure, and link interruption.

[0174] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this disclosure.

[0175] According to an embodiment of this disclosure, an ultrasonic sensor node is provided, comprising: an analog front end for transmitting ultrasonic pulses and receiving corresponding full-time-domain waveform data; and a processor coupled to the analog front end, the processor being configured to execute a computer program to implement a hierarchical transmission method for ultrasonic data as described in any of the first aspects.

[0176] Ultrasonic sensor nodes can be deployed in industrial sites for long-term periodic monitoring of the wall thickness of the tested components. They are suitable for online sensing of corrosion and erosion conditions in facilities such as petrochemical storage tanks, long-distance pipelines, and power towers. They can independently complete data acquisition, processing, and transmission without human intervention.

[0177] The analog front end can be coupled between the processor and the ultrasonic transducer. For example, it can be a signal processing circuit to realize the ultrasonic transmission excitation and echo signal reception conditioning. Its function is to convert the digital control signal of the processor into a high-voltage pulse to drive the transducer to vibrate, amplify, filter and perform analog-to-digital conversion on the weak echo signal received by the transducer, and output full time-domain waveform data that can be directly processed by the processor. It can also cooperate with the processor to collect noise window data in the silent section after the transmission pulse or the baseline section between adjacent echoes.

[0178] As the computing unit of the ultrasonic sensor node, the processor can connect to the analog front end, communication module, and power management module. It is configured to run a pre-set computer program to realize the hierarchical transmission method logic of ultrasonic data. Its functions include parsing the full-volume time-domain waveform data and noise window data acquired by the analog front end, executing the echo filtering algorithm, calculating the thickness value and thickness variation characteristics of the measured component, evaluating the reliability of the waveform data, deciding the transmission strategy based on the resource status of the communication link, and controlling the power supply and working status switching of each functional module.

[0179] For example, the processor can wake up the analog front end according to a preset cycle, send a transmission trigger signal to the analog front end, and the analog front end responds to the trigger signal to generate a high-voltage pulse to drive the ultrasonic transducer to transmit ultrasound to the component under test. Then, the receiving link is turned on to condition the echo signal returned by the transducer and the converted full time-domain waveform data is stored in the processor's memory.

[0180] For example, the processor can extract a noise window located in the silent section after the transmitted pulse from the full time-domain waveform data, determine the dual thresholds for echo identification based on the amplitude statistical characteristics of the noise window, traverse the full time-domain waveform data along the time axis to filter the effective echo intervals, calculate the thickness value of the measured component based on the effective echo intervals, and extract local diagnostic waveform data containing the effective echo intervals and their surrounding sampled values.

[0181] For example, the processor can also generate thickness change features based on thickness values ​​obtained from multiple consecutive measurements, combine the reliability features of local diagnostic waveform data and the resource budget features of the first communication link, determine the corresponding target level, and call the appropriate communication link to transmit data of the corresponding granularity according to the target level. Under normal steady-state conditions, the processor can call the first communication link to transmit simplified thickness values, under abnormal fluctuation conditions, the processor can call the first communication link to transmit local diagnostic waveform data, and when the data reliability is insufficient or the resources of the first communication link are insufficient, the processor can trigger the second communication link to transmit full time-domain waveform data.

[0182] The analog front end can also be in a low-power state during non-task periods, only starting to work when the processor triggers a data acquisition task, in order to reduce the overall energy consumption of the sensor node.

[0183] By integrating an analog front-end with a processor configured to execute a hierarchical transmission method, ultrasonic sensor nodes can independently complete the entire process from data acquisition to transmission decision-making in unattended industrial scenarios. While ensuring measurement accuracy and diagnostic effectiveness, they minimize communication and computing power consumption, adapt to the application requirements of long-term continuous monitoring, and improve the reliability and economy of industrial equipment condition monitoring.

[0184] In some embodiments, the ultrasonic sensor node can be deployed in an industrial field to achieve long-term online monitoring of the wall thickness of the component under test. In addition to the analog front end, namely the ultrasonic sensor and analog front end and the processor, the ultrasonic sensor node may also include a temperature sensing unit, a non-volatile storage unit, a communication scheduling unit, a first wireless communication unit, a second wireless communication unit, and a power management and RTC wake-up unit.

[0185] The analog front end is used to transmit ultrasonic pulses to the component under test and receive the echo signal reflected from the interface. The analog front end includes a high-voltage pulse drive circuit, a low-noise amplifier, a variable gain amplifier and a bandpass filter. It can amplify and filter the weak echo signal as necessary to obtain a high-quality analog echo signal. Its thickness measurement excitation method can adopt the pulse echo method or be adapted to other thickness measurement mechanisms based on the ultrasonic propagation principle.

[0186] Temperature sensing units are used to sense the thermodynamic state of the environment. They can be placed on the surface of the component being measured to measure the body temperature, or placed inside the ultrasonic sensor node to measure the ambient temperature. The acquired temperature sensing data can be used to perform temperature compensation correction on the propagation speed of ultrasound in the medium, or it can be used as an auxiliary parameter to participate in the judgment of abnormal states, or it can be recorded as metadata along with the measurement results.

[0187] Non-volatile storage units are storage media with the ability to retain data even when power is lost. They are used to store and manage various types of data during node operation. They can save complete time-domain waveform data, local diagnostic waveform data, calculated thickness values, measurement identifiers that uniquely identify measurement events, system firmware images, and progress indexes of transmission tasks that were not completed due to network anomalies, ensuring that data is not lost after system reset or power failure.

[0188] The communication scheduling unit is specifically used for network access decisions and data stream forwarding. It can dynamically determine the level of the current transmission task based on the current measurement status, the diagnostic reliability of the data, and the estimated wireless resource budget, and control the wake-up and sleep timing and data routing path of different communication modules accordingly. The communication scheduling unit can be integrated with the processor in the same microcontroller chip to achieve a high degree of integration design, or multiple microcontrollers with mutual communication capabilities can respectively undertake the processing and scheduling functions.

[0189] The first wireless communication unit can be understood as a wide-area communication module that undertakes routine low-power data transmission tasks. For example, it can employ low-power wide-area communication technology based on LoRa or LoRaWAN protocols, suitable for performing periodic thickness result reporting, abnormal event alarms, packet transmission of local diagnostic waveform data, and signaling transmission for authorization of the second communication unit within a short window period after waking from deep sleep. The second wireless communication unit can be understood as a near-field communication module that undertakes high-speed, large-data-volume transmission tasks. For example, it can specifically employ Bluetooth Low Energy communication technology, mainly used in field maintenance scenarios to allow maintenance personnel to download full-volume time-domain waveform data, historical stored data, or perform wireless firmware upgrade operations via handheld terminals.

[0190] The power management and RTC wake-up unit can realize energy efficiency management of the ultrasonic sensor node. It can independently control the power supply circuits of the ultrasonic front end, memory, first wireless communication unit and second wireless communication unit. For example, during normal monitoring intervals, this unit can cut off the power to most circuits except for the real-time clock or low-power wake-up circuit, so that the ultrasonic sensor node enters a deep sleep state to minimize energy consumption. When the preset measurement trigger time is reached, it can accurately wake up the relevant circuits to complete the data acquisition, processing and transmission tasks, and control the ultrasonic sensor node to enter a deep sleep state again after the task is completed.

[0191] Specifically, the power supply of the second wireless communication unit can be strictly limited within the maintenance time window authorized by the server, avoiding unnecessary power consumption during non-task periods. Since the LoRa gateway is a field-deployed wireless network infrastructure, it receives radio frequency frames emitted by ultrasonic sensor nodes and converts them into Internet Protocol (IP) data packets for forwarding to the backend network server. The LoRaWAN network server is responsible for authenticating the identity of access terminals, performing packet deduplication, running adaptive data rate algorithms to optimize air interface efficiency, and scheduling uplink and downlink communication resources.

[0192] The application server, running in the cloud or locally, is responsible for receiving thickness measurement results and local diagnostic waveform data from the web server. It then reconstructs the waveform's horizontal axis using algorithms and performs in-depth remote fault diagnosis. The mobile terminal used by maintenance personnel can be any type of portable human-machine interface tool. After obtaining authorization from the server, the mobile terminal can establish a connection with the sensor node via Bluetooth Low Energy communication link to perform operations such as full data download, parameter configuration, or firmware updates.

[0193] By integrating the above functional units, the ultrasonic sensor node can operate for a long time in ultra-low power mode. It can use the first wireless communication unit to complete the routine monitoring data reporting, and provide the ability to download the full amount of raw data through the controlled activated second wireless communication unit when necessary. Together with the back-end server and the mobile terminal on site, it forms a complete monitoring and diagnostic ecosystem, effectively solving the contradiction between power supply limitation, bandwidth limitation and diagnostic depth in long-term monitoring in industrial sites.

[0194] Corresponding to the graded transmission method of ultrasound data in the above embodiment, Figure 6 This is a schematic diagram of the structure of a graded transmission device for ultrasound data provided in an embodiment of this disclosure. This device can be implemented as part or all of a computer device by software, hardware, or a combination of both. The computer device can be... Figure 7 The electronic device shown.

[0195] Reference Figure 6 The graded transmission device for ultrasound data includes: Acquisition unit 601 is used to acquire full time-domain waveform data and a noise window in the full time-domain waveform data that does not contain valid echoes; Extraction unit 602 is used to filter effective echo intervals from full time-domain waveform data based on the amplitude statistical characteristics of noise windows, and extract local diagnostic waveform data containing the effective echo intervals and their previous and next neighboring original sampled values. The effective echo intervals are used to calculate the thickness value of the measured component, and the thickness value is used to determine the thickness variation characteristics of the measured component. The calling unit 603 is used to call the first communication link or the second communication link to perform data transmission in response to the target level determined based on the thickness change characteristics, local diagnostic waveform data and the first communication link; wherein the first communication link is a low-power wide-area communication link and the second communication link is a short-distance high-speed communication link, and the content of the data transmission includes thickness values, local diagnostic waveform data or full time-domain waveform data.

[0196] It is understood that the embodiments of the graded transmission device for ultrasound data and any implementation thereof correspond to the embodiments of the graded transmission method for ultrasound data and any implementation thereof. The technical effects corresponding to the embodiments of the graded transmission device for ultrasound data and any implementation thereof can be found in the technical effects corresponding to the aforementioned embodiments of the graded transmission method for ultrasound data and any implementation thereof, and will not be repeated here.

[0197] It should be noted that the ultrasonic data hierarchical transmission device provided in the above embodiments is only an example of the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0198] The functional units and modules in the above embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of the embodiments of this disclosure.

[0199] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this disclosure. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0200] This disclosure also provides an electronic device, which includes one or more processors and a memory; The memory is coupled to one or more processors. The memory is used to store computer program code, which includes computer instructions. The one or more processors invoke the computer instructions to cause the electronic device to perform the aforementioned method for hierarchical transmission of ultrasound data.

[0201] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. The electronic device 700 can be a mobile phone, smart screen, tablet computer, wearable electronic device, in-vehicle electronic device, augmented reality (AR) device, virtual reality (VR) device, laptop computer, ultra-mobile personal computer (UMPC), netbook, personal digital assistant (PDA), projector, or a communication device such as a server, storage device, or base station, or a smart car, etc. This disclosure does not limit the specific type of electronic device.

[0202] The memory 701 can be used to store computer programs 702 and modules. The processor 703 executes various functional applications and data processing of the electronic device by running the software programs and modules stored in the memory 701. The memory 701 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device (such as audio data, telephone directory, etc.). In addition, the memory 701 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0203] The processor 703 may include one or more processors such as a central processing unit (CPU), an application processor (AP), and a baseband processor. The processor can serve as the nerve center and command center of the wireless router. The processor 703 can generate operation control signals based on instruction opcodes and timing signals to control instruction fetching and execution. The memory 701 can be used to store executable program code, including instructions. The processor 703 executes various functional applications and data processing of the network device by running the instructions stored in the memory. The memory 701 may include a program storage area and a data storage area, such as storing data for audio signals to be played. For example, the memory may be Double Data Rate Synchronous Dynamic Random Access Memory (DDR) or Flash memory.

[0204] This disclosure also provides a computer-readable storage medium storing computer instructions; when the computer-readable storage medium is used on an electronic device, it causes the electronic device to perform the aforementioned graded transmission method for ultrasonic data.

[0205] The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access, or it can include one or more data storage devices such as servers or data centers that can be integrated with media. The available medium can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media, or semiconductor media (e.g., solid-state disks (SSDs)).

[0206] This disclosure also provides a computer program product containing computer instructions that, when run on an electronic device, enables the electronic device to execute the aforementioned method for graded transmission of ultrasonic data.

[0207] The computer storage medium and computer program product provided in the above-described embodiments are used to execute the methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects corresponding to the methods provided above, and will not be repeated here.

[0208] In the above embodiments, implementation can also be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this disclosure are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer, or a data storage device such as a server or data center that integrates one or more available media. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc., and the storage medium can also include combinations of the above types of memory.

[0209] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0210] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments claimed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0211] In the embodiments provided in this disclosure, it should be understood that the disclosed apparatus / network devices and methods can be implemented in other ways. For example, the apparatus / network device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0212] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0213] The technical features of the various embodiments described above in this disclosure can be combined arbitrarily without conflict. For the sake of brevity, this specification does not describe all possible combinations, but as long as these combinations do not violate the technical spirit of this disclosure, they should all be considered within the scope of this disclosure. Based on the content disclosed in this disclosure, those skilled in the art can reasonably combine, delete, or replace the technical features of the above embodiments according to actual needs, and such modifications and variations all fall within the protection scope of this disclosure.

[0214] The above-described embodiments are only used to illustrate the technical solutions of this disclosure, and are not intended to limit it. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure, and should all be included within the protection scope of this disclosure.

Claims

1. A method for hierarchical transmission of ultrasonic data, characterized in that, include: Acquire the full time-domain waveform data, and the noise window in the full time-domain waveform data that does not contain valid echoes; Based on the amplitude statistical characteristics of the noise window, effective echo intervals are screened from the full time-domain waveform data, and local diagnostic waveform data containing the effective echo intervals and their preceding and following neighborhood original sample values ​​are extracted. The effective echo intervals are used to calculate the thickness value of the tested component, and the thickness value is used to determine the thickness variation characteristics of the tested component. In response to the target level determined based on the thickness variation characteristics, the local diagnostic waveform data, and the first communication link, the first communication link or the second communication link is invoked to perform data transmission; wherein, the first communication link is a low-power wide-area communication link, and the second communication link is a short-range high-speed communication link, and the content of the data transmission includes thickness values, local diagnostic waveform data, or full time-domain waveform data.

2. The method according to claim 1, characterized in that, The step of filtering valid echo intervals from the full time-domain waveform data based on the amplitude statistical characteristics of the noise window includes: Based on the amplitude distribution characteristics of the noise window, a preset trigger threshold and a preset sustain threshold are determined, wherein the preset trigger threshold is greater than the preset sustain threshold; Traverse the full time-domain waveform data along the time axis, and determine the sequence of sampling points whose amplitude first crosses the preset trigger threshold and remains continuously above the preset sustain threshold as candidate echoes; The effective echo interval is determined based on the duration of at least one of the candidate echoes obtained through the traversal.

3. The method according to claim 2, characterized in that, Determining the effective echo interval based on the duration of at least one of the candidate echoes obtained through traversal includes: In response to a candidate echo having a duration greater than a first preset pulse width, the interval in which the candidate echo is located is determined as the effective echo interval, wherein the first preset pulse width is predefined based on at least one of the center frequency of the ultrasonic transducer, the current sampling rate, and the ambient noise level. If the duration of two or more candidate echoes is greater than the first preset pulse width, the interval containing the candidate echo with the largest amplitude is selected as the effective echo interval.

4. The method according to claim 2, characterized in that, The extraction of local diagnostic waveform data containing the effective echo interval and its preceding and following neighborhood original sample values ​​includes: Based on the starting boundary of the effective echo interval, extend the first preset sampling point forward, and based on the ending boundary of the effective echo interval, extend the second preset sampling point backward, and extract the continuous original sampled values ​​of the corresponding segment in the full time domain waveform data as the target waveform segment; Generate positioning metadata, which represents the absolute starting index and total number of sampling points of the target waveform segment in the full time-domain waveform data, and is used to uniquely identify the position of the target waveform segment. The target waveform segment and the location metadata are encapsulated into the local diagnostic waveform data.

5. The method according to any one of claims 1 to 4, characterized in that, The target level includes a first level, a second level, and a third level; the step of responding to the target level determined based on the thickness variation characteristics, the local diagnostic waveform data, and the first communication link, and invoking the first communication link or the second communication link to perform data transmission includes: In response to the thickness change feature being less than a preset steady-state threshold and the confidence feature of the local diagnostic waveform data being higher than a first threshold, the target level is determined to be the first level, and the first communication link is invoked to transmit the thickness value; In response to the thickness change feature being no less than the preset steady-state threshold, the confidence feature being higher than the second threshold, and the resource budget feature of the first communication link satisfying the preset packet sending constraint, the target level is determined to be the second level, and the first communication link is invoked to transmit the local diagnostic waveform data; In response to the credibility feature being lower than the second threshold, or the resource budget feature not meeting the preset packet sending constraint, the target level is determined to be the third level, triggering the activation of the second communication link to perform full time-domain waveform data transmission; Wherein, the first threshold is greater than the second threshold; the credibility feature is used to characterize at least one of the morphological validity of the effective echo in the local diagnostic waveform data, the reliability of the thickness calculation result, and the identifiability of the anomaly source; the resource budget feature is used to characterize at least one of the available air interface duration of the first communication link, the upper limit of radio frequency transmission energy consumption, and the concurrent access carrying capacity of the gateway to which it belongs.

6. The method according to claim 5, characterized in that, The resource budget characteristics that satisfy the preset package issuance constraints include: Based on the modulation parameters of the current first communication link, the number of available payload bytes per packet, and the total amount of data in the local diagnostic waveform data, predict the total number of data packets, the total transmission time, and the total transmission energy consumption required to transmit the local diagnostic waveform data. In response to the total number of data packets being less than a preset packet count limit, the total transmission duration being less than a preset duration limit, and the total transmission energy consumption being less than a preset energy consumption limit, it is determined that the resource budget characteristics satisfy the preset packet sending constraints.

7. The method according to claim 5, characterized in that, The triggering activation of the second communication link to perform full time-domain waveform data transmission includes: The maintenance request and device authentication identifier are reported to the server via the first communication link. The server is used to verify the legitimate identity of the sensor node based on the device authentication identifier and generate an activation command based on the verification result. The activation command includes a limited duration. In response to the activation command being detected within the pre-configured downlink receiving window of the first communication link, the second communication link, which is in a dormant or power-off state, is activated within the specified time period, and the full time-domain waveform data is transmitted via the activated second communication link. In response to the completion of data transmission or the expiration of the specified time, the second communication link is closed, and the sensor node is controlled to switch to sleep mode.

8. The method according to claim 5, characterized in that, The method further includes: Extract the signal-to-noise ratio, peak width consistency, and peak spacing stability within the effective echo interval; Detect whether there is amplitude saturation / or clipping anomaly within the effective echo interval to obtain waveform anomaly detection results; The credibility feature is obtained by weighted averaging the signal-to-noise ratio, peak width consistency, peak spacing stability parameter, and waveform anomaly detection result.

9. The method according to claim 2, characterized in that, The step of determining the sequence of sampling points whose amplitude first crosses the preset trigger threshold and remains continuously above the preset sustain threshold as candidate echoes includes: If the amplitude of the current sampling point is greater than the preset trigger threshold, it is marked as a potential starting point; Starting from the potential starting point, check whether the amplitude of the subsequent first preset number of sampling points is greater than the preset maintenance threshold. In response to the successful verification, the potential starting point is confirmed as a valid starting point, and the first sampling point with an amplitude lower than the preset maintenance threshold is confirmed as a valid ending point. The sequence of sampling points between the valid start point and the valid end point is determined as the candidate echo.

10. The method according to claim 2, characterized in that, During the process of traversing the full time-domain waveform data along the time axis, the following search window strategy is used to search for candidate echoes: Obtain the reference echo position and construct a priority search window based on the reference echo position, wherein the reference echo position is the position of the historical valid echo identified during the historical confidence measurement process; First, the candidate echo is searched within the priority search window. If no candidate echo that meets the preset confidence condition is identified within the priority search window, the candidate echo is searched within the full-range time window. The search interval corresponding to the priority search window is smaller than the search interval corresponding to the full-range time window.

Images