A method and apparatus for ultrasonic scanning imaging early warning of regional corrosion state
By combining ultrasonic scanning imaging with ARIMA/ARIMAX models, the accuracy issues of corrosion rate calculation and trend prediction in pipeline corrosion detection have been resolved, enabling dynamic monitoring and reliable early warning of pipeline corrosion, thus improving prediction accuracy and equipment safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHENYANG ZKWELL CORROSION CONTROL TECH
- Filing Date
- 2026-05-14
- Publication Date
- 2026-07-03
AI Technical Summary
Existing technologies for pipeline corrosion detection are sensitive to noise in corrosion rate calculation, have simple data processing, cannot identify unreliable data, and provide inaccurate corrosion trend predictions. They also ignore the influence of operating parameters, resulting in low corrosion prediction accuracy.
The ultrasonic scanning imaging method for regional corrosion is adopted. The wall thickness data is collected by a two-dimensional scanning mechanism, and sound velocity compensation and data correction are performed. The corrosion trend is predicted by combining the ARIMA or ARIMAX model. In addition, the operating parameters are introduced to construct a self-closed-loop data quality control mechanism for graded early warning.
It improves the accuracy and robustness of corrosion rate calculation, significantly enhances the accuracy of corrosion trend prediction, realizes dynamic monitoring and reliable early warning of corrosion process, and reduces false alarms.
Smart Images
Figure CN122329211A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the technical field of industrial non-destructive testing, specifically to an ultrasonic scanning imaging early warning method and device for regional corrosion conditions. Background Technology
[0002] Pipeline corrosion is one of the main factors leading to equipment failure and safety accidents in industries such as petroleum, chemical, and energy. Regularly inspecting pipeline wall thickness is an important means of assessing pipeline safety and preventing corrosion leaks.
[0003] Currently, simply obtaining the current wall thickness value is insufficient to meet the growing demand for intelligent operation and maintenance. While some pipeline inspection devices capable of two-dimensional scanning have emerged in the existing technology, their data processing methods remain relatively simple, mainly exhibiting the following technical shortcomings: Corrosion rates are often calculated by differentially comparing wall thickness maps of the same pipeline at different time points to determine the corrosion thinning rate in a specific area. This method is extremely sensitive to noise and errors in single measurements; a deviation in a measurement can severely distort the calculated corrosion rate. Furthermore, existing technologies cannot automatically identify and process unreliable data when measurement data fluctuates abnormally due to sensor malfunctions, environmental interference, or poor coupling, potentially leading to misleading corrosion rate calculations and maintenance decisions.
[0004] Secondly, corrosion trend prediction methods are relatively primitive. Most systems can only perform simple linear extrapolation or make rough estimates by treating the current corrosion rate as a constant value. In actual industrial environments, corrosion processes are often non-stationary, and the corrosion rate may accelerate or slow down with changes in time and operating conditions. Simple linear models are unable to accurately capture the dynamic characteristics of the corrosion process.
[0005] Finally, existing technologies rarely incorporate operating parameters (such as temperature, pressure, flow rate, and medium pH) into corrosion prediction models. In reality, the corrosion rate of pipelines is closely related to operating conditions. Ignoring the influence of operating parameters makes it difficult to distinguish whether changes in wall thickness are caused by fluctuations in normal operating conditions or by material degradation trends, thus limiting prediction accuracy. Summary of the Invention
[0006] To address the aforementioned problems, specifically those raised in the background section, this invention proposes a method and apparatus for ultrasonic scanning imaging early warning of regional corrosion conditions, comprising the following steps: S1. Control the two-dimensional scanning mechanism to scan along the preset path and collect the original wall thickness data, position coordinates and temperature parameters of each measuring point; S2. Compensate for the sound velocity according to the temperature parameter, and correct the original wall thickness data to the compensated wall thickness value. S3. Associate the compensated wall thickness value with the position coordinates to generate a wall thickness distribution heat map of the measured area; S4. Based on the compensated wall thickness values from multiple historical scans of each measuring point, calculate the corrosion rate of each measuring point and evaluate its data confidence level. S5. Extract the minimum wall thickness value of the entire monitoring area from each historical scan, arrange them in chronological order of scan time, and use ARIMA or ARIMAX models to predict the future trend of the minimum wall thickness of the area and the expected time to reach the safety threshold. S6. Based on the corrosion rate, data confidence level, and trend prediction results, a graded early warning is issued.
[0007] A further provision of the present invention is that step S4 specifically includes the following sub-steps: S41. For each measuring point, extract its historical wall thickness data from the most recent N valid scans, where N≥5, and construct a time and wall thickness data point set {(t1,d1),(t2,d2),…,(t...}}. n ,d n )}; S42. The least squares method is used to perform linear regression fitting on the data point set to obtain the trend line d(t)=a·t+b of the wall thickness change with time. The absolute value of the slope a of the fitted line is calculated as the corrosion rate CR=|a|. S43. Calculate the determination coefficient R of the linear regression fit. 2 , where R 2 =1-[Σ(di- i) 2 ] / [Σ(di- ) 2 ]; S44, R 2 Compared with the preset confidence threshold R 2 Compare R² with R_th. If R² ≥ R²_th, the corrosion rate is deemed valid and output; otherwise, R_th is not valid. 2 <R 2 If _th is encountered, the low confidence handling mechanism will be triggered.
[0008] A further provision of the present invention is that step S5, which uses the ARIMA model for prediction, specifically includes the following sub-steps: S5A1. Extract the minimum wall thickness values of each historical scan from the historical wall thickness database, arrange them in chronological order of scan time, and form time series data with a series length of at least 30 time points. S5A2. Perform an augmented Dickie-Fowler stationarity test on the time series data to determine the difference order d, and then successively differiate the non-stationary series until it becomes stationary. S5A3. Using the minimum information criterion that combines the Akaike information criterion and the Bayesian information criterion, the autoregression order p and the moving average order q are determined within the preset search range p∈[0,5] and q∈[0,5], and an ARIMA(p,d,q) model is constructed. S5A4. Perform residual white noise test on the constructed ARIMA model to verify the model's applicability; S5A5: Use the trained ARIMA model to predict the trend of regional minimum wall thickness change over a specified number of days in the future, and calculate the prediction confidence interval to evaluate the reliability of the prediction results. Based on the prediction results, calculate the time when the regional minimum wall thickness is expected to reach the safety threshold.
[0009] A further provision of the present invention is that the construction of the ARIMAX model in step S5 further includes the step of treating the operating condition parameters as exogenous variables. S5B1. Collect operating parameters aligned with the timestamps of each historical scan, wherein the operating parameters include at least one of temperature, pressure, flow rate, pH value, and chloride ion concentration; S5B2. Normalize each working condition parameter and perform correlation analysis with the minimum wall thickness sequence in the region to screen exogenous variables whose absolute value of the correlation coefficient is greater than the preset threshold. S5B3. Introduce the selected exogenous variables as regression terms into the ARIMA model to construct the ARIMAX(p,d,q) model. S5B4. Solve the model parameters using conditional maximum likelihood estimation or least squares estimation. The parameters include autoregressive coefficients, moving average coefficients, and regression coefficients of exogenous variables. S5B5. Use the trained ARIMAX model to predict the trend of regional minimum wall thickness change over a specified number of days in the future, and calculate the prediction confidence interval to evaluate the reliability of the prediction results. Based on the prediction results, calculate the time when the regional minimum wall thickness is expected to reach the safety threshold.
[0010] A further provision of the present invention is that the low confidence processing mechanism in step S44 includes at least one of the following operations: S441. The corrosion rate data of the measuring point is labeled with a low confidence level, and the area where the measuring point is located is marked with a preset exclusive color in the wall thickness distribution heat map. S442. Issue a retest control command to the two-dimensional scanning mechanism to perform a denser scan on the measurement point and its adjacent area in the next scanning cycle; the scanning step size of the denser scan is reduced by 50% compared with the conventional scanning step size, and the number of repeated acquisitions of a single point is set to 3-5 times; S443. Retrieve the device self-test log of the ultrasonic sensor and check at least one of the sensor excitation voltage, echo amplitude, and coupling status parameters; if any parameter exceeds the normal operating range, trigger and output sensor status warning information. S444. Retrieve the temperature and humidity environmental parameters for the same monitoring period as the measuring point, and analyze the impact of environmental parameter fluctuations on the wall thickness detection data; if the environmental parameter fluctuations exceed the preset normal fluctuation threshold range, then mark the measuring point as a suspected environmental interference point. S445. For a measurement point whose determination coefficient R2 is lower than the preset threshold for two consecutive scanning cycles, the system automatically increases the monitoring frequency of the measurement point from the preset default monitoring cycle to the preset encrypted monitoring cycle.
[0011] The invention is further configured to include an online model update step: after each new scan is completed and a new minimum wall thickness value for the region is obtained, the latest data is added to the time series, while the oldest data that exceeds the preset time window is removed, and the stationarity test, model order determination and parameter estimation are automatically re-executed to generate an updated prediction model; the prediction errors of the new and old models are compared, and if the root mean square error of the new model is reduced by more than a preset percentage compared with the old model, the new model is deployed.
[0012] A further setting of the present invention is that the graded early warning in step S6 includes a prediction early warning based on the prediction results of the ARIMA or ARIMAX model: when the prediction result shows that the minimum wall thickness of the region will approach the safety threshold in the first preset time period in the future, a first-level early warning is triggered; when the prediction result shows that the minimum wall thickness of the region will be lower than the safety threshold in the second preset time period in the future, a second-level early warning is triggered; and when the prediction result shows that the minimum wall thickness of the region will exceed the danger threshold in the third preset time period in the future, a third-level early warning is triggered, wherein the first preset time is greater than the second preset time, and the second preset time is greater than the third preset time.
[0013] An ultrasonic scanning monitoring device for performing the method of said area corrosion status includes a two-dimensional scanning mechanism installed on the outer wall of the pipe under test. The two-dimensional scanning mechanism includes an X-axis module arranged along the axial direction of the pipe under test and a Y-axis module arranged along the circumferential direction of the pipe under test. An ultrasonic sensor for collecting the original wall thickness data of each measuring point of the pipe under test is fixedly installed at the moving end of the two-dimensional scanning mechanism. An encoding positioning module for outputting position signals is also fixedly installed at the moving end of the two-dimensional scanning mechanism. A temperature sensor for collecting temperature parameters of the measured area is fixedly installed at the moving end of the two-dimensional scanning mechanism or on the outer wall of the pipe under test. The device also includes a controller. The ultrasonic sensor, the encoding positioning module and the temperature sensor are electrically connected to the controller. The controller is configured to acquire the position signal of the encoding positioning module and calculate it into position coordinates, acquire the original wall thickness data of the ultrasonic sensor, perform sound velocity compensation correction on the original wall thickness data according to the temperature parameters collected by the temperature sensor to generate a corrected wall thickness value, and associate and bind the corrected wall thickness value with the corresponding position coordinates.
[0014] A further configuration of the present invention is as follows: the Y-axis module includes an annular curved guide rail fixedly installed on the outer wall of the pipe being tested, and a Y-axis slider slidably fitted on the annular curved guide rail; the X-axis module includes a linear guide rail fixedly installed on the Y-axis slider, and an X-axis slider slidably fitted on the linear guide rail; the ultrasonic sensor, the coding positioning module, and the temperature sensor are fixedly installed on the X-axis slider.
[0015] A further feature of the present invention is that the encoding positioning module is an integrated position encoder that integrates a linear displacement sensing unit and an angular displacement sensing unit; the linear displacement sensing unit cooperates with a linear guide rail to output an X-axis axial position signal; the angular displacement sensing unit cooperates with an annular curved guide rail to output a Y-axis circumferential position signal; and a circumferential reference mark is fixed on the outer wall of the pipe under test to provide a circumferential zero-point positioning signal.
[0016] The beneficial technical effects of this invention are as follows: This invention uses the least squares method to perform linear regression fitting on multiple historical scan data of a single measuring point, and uses the absolute value of the slope of the fitted line as the corrosion rate. Compared with the two-point difference method commonly used in the prior art, this effectively suppresses the influence of single measurement noise on the calculation results, and improves the accuracy and robustness of corrosion rate calculation. Simultaneously, a coefficient of determination is introduced as an evaluation index of goodness of fit. When it falls below a preset threshold, a low-confidence linkage processing mechanism is automatically triggered, including data quality marking, automatic retesting, sensor self-checking, operating condition correlation analysis, and dynamic adjustment of monitoring frequency. This forms a self-closed-loop control of data quality, effectively avoiding false alarms caused by measurement noise, sensor failure, or environmental interference. This invention uses an ARIMA time series model to model and predict historical data of the minimum wall thickness in the region. Specifically, the parameter selection method of the model has been optimized to address the non-stationary and nonlinear variation characteristics of corrosion monitoring data. Furthermore, by incorporating operating parameters such as temperature, pressure, flow rate, and medium pH as exogenous variables into the ARIMA model, an ARIMAX model was constructed. This model can distinguish whether changes in corrosion rate are caused by fluctuations in operating conditions or by material degradation trends, significantly improving prediction accuracy. During the prediction process, prediction confidence intervals were calculated to assess the reliability of the prediction results, providing a quantitative basis for early warning decisions. Attached Figure Description
[0017] Figure 1 A schematic diagram of the structure of the present invention is shown.
[0018] Figure 2 The corrosion history curve and early warning analysis diagram of the present invention are shown.
[0019] Figure 3 A schematic diagram of the corrosion state thermogram of the present invention is shown.
[0020] Figure 4 A three-dimensional schematic diagram of the corrosion state of the present invention is shown.
[0021] Reference numerals: 1. Two-dimensional scanning mechanism; 2. Ultrasonic sensor; 3. Encoding and positioning module; 4. Temperature sensor; 5. Annular curved guide rail; 6. Y-axis slider; 61. Y-axis drive motor; 62. First gear; 63. Annular rack; 7. Linear guide rail; 8. X-axis slider; 81. X-axis drive motor; 82. Second gear; 83. Rack; 9. Fixture. Detailed Implementation
[0022] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.
[0023] like Figure 1 As shown, an ultrasonic scanning monitoring device for regional corrosion includes a two-dimensional scanning mechanism 1 fixedly installed on the outer wall of the pipe being tested. The two-dimensional scanning mechanism 1 includes an X-axis module arranged along the axial direction of the pipe and a Y-axis module arranged along the circumferential direction of the pipe. An ultrasonic sensor 2 and an encoding positioning module 3 are fixedly installed at the moving end of the two-dimensional scanning mechanism 1, and a temperature sensor 4 is fixedly installed at the moving end of the two-dimensional scanning mechanism 1 or on the outer wall of the pipe being tested. A controller is also included, and the ultrasonic sensor 2, the encoding positioning module 3, and the temperature sensor 4 are electrically connected to the controller.
[0024] The specific structure of the two-dimensional scanning mechanism 1 is as follows: The Y-axis module includes an annular curved guide rail 5 fixedly installed on the outer wall of the pipe being measured, and a Y-axis slider 6 slidably fitted on the annular curved guide rail 5. The X-axis module includes a linear guide rail 7 fixedly installed on the Y-axis slider 6, and an X-axis slider 8 slidably fitted on the linear guide rail 7. The ultrasonic sensor 2, the encoding positioning module 3, and the temperature sensor 4 are all fixedly installed on the X-axis slider 8 and move together with the X-axis slider 8.
[0025] The driving methods for the Y-axis and X-axis sliders are as follows: The Y-axis module also includes a Y-axis drive motor 61 that drives the Y-axis slider 6 to move circumferentially. The output shaft of the Y-axis drive motor 61 is connected to a first gear 62, which meshes with a ring rack 63 fixed on the annular curved guide rail 5. The X-axis module also includes an X-axis drive motor 81 that drives the X-axis slider 8 to move axially. The output shaft of the X-axis drive motor 81 is connected to a second gear 82, which meshes with a rack 83 fixed on the X-axis module. Precise position control can be achieved through this gear and rack meshing transmission.
[0026] The coding and positioning module 3 employs an integrated position encoder that combines a linear displacement sensing unit and an angular displacement sensing unit. This module is fixedly mounted on the X-axis slider 8. The linear displacement sensing unit, in conjunction with the linear guide rail 7, outputs the X-axis axial position signal; the angular displacement sensing unit, in conjunction with the annular curved guide rail 5, outputs the Y-axis circumferential position signal. To establish the absolute zero point of the circumferential position, a circumferential reference mark is also fixed to the outer wall of the measured pipe to provide the circumferential zero-point positioning signal.
[0027] In addition, the linear guide 7 is equipped with a travel limit component, which includes limit switches at both ends of the linear guide 7 and a trigger element linked with the X-axis slider 8, used to limit the axial movement range of the X-axis slider 8 and prevent overtravel movement from causing equipment damage.
[0028] To accommodate different pipe diameters, the annular curved guide rail 5 adopts a segmented opening and closing guide rail structure, specifically comprising at least two arc-shaped guide rail segments, each segment being spliced and fixed by detachable connectors. Depending on the outer diameter of the pipe being tested, operators can select an appropriate number of arc-shaped guide rail segments for splicing and assembly, thus adapting to the testing needs of pipes with different diameters. The annular curved guide rail 5 is coaxially fitted onto the outer wall of the pipe via a fixing frame 9. The fixing frame 9 is spaced circumferentially along the outer wall of the pipe and has a radially adjustable structure, capable of adapting to pipes with different outer diameter specifications.
[0029] The ultrasonic sensor 2 in this device is preferably an electromagnetic ultrasonic sensor. Electromagnetic ultrasonic sensors do not require coupling agents, provide non-contact detection, and are particularly suitable for corrosion detection in high-temperature pipelines or pipelines with rough surfaces.
[0030] The two-dimensional scanning mechanism 1 of this device can be fixed to the outer wall of the pipe being measured by either a mechanical locking structure or a magnetic adsorption structure. For ferromagnetic pipes, the magnetic adsorption structure is easy to install; for non-ferromagnetic pipes or in situations with special installation requirements, a mechanical locking structure can be used to achieve reliable fixation.
[0031] The X-axis and Y-axis modules of this device can flexibly select different types of guide rails according to the geometry of the pipe being tested. Specifically, when monitoring straight pipe sections, the X-axis module uses a linear guide rail to move along the pipe's axial direction, while the Y-axis module uses an arc-shaped guide rail to move along the pipe's circumference. The combination of these two components forms a cylindrical scanning area, achieving full coverage scanning of the outer wall of the straight pipe section. When monitoring elbow sections, since the elbow surface is a spatial curved surface, both the X-axis and Y-axis modules use spatial curved guide rails with a radius of curvature matching the surface of the elbow being tested. This allows the scanning mechanism to closely conform to the elbow surface, completing accurate detection of corrosion areas in the elbow. This modular design makes the same scanning device compatible with both straight and elbow pipe structures, significantly improving the equipment's versatility and applicability.
[0032] This device also includes a communication module, which is electrically connected to the controller. The wireless communication module is used to send the processed detection data to a cloud server, enabling remote viewing and data management of the detection results.
[0033] The following is a detailed description of the operation process of this device: Before conducting the scanning inspection, the equipment is first configured according to the type of pipeline being tested. If the object being tested is a straight pipe section, a linear guide rail is selected for the X-axis module and an arc-shaped guide rail is selected for the Y-axis module; if the object being tested is a bend, both the X-axis and Y-axis modules are selected with spatial curve guide rails that match the radius of curvature of the bend. Subsequently, the two-dimensional scanning mechanism 1 is fixedly installed on the outer wall of the pipeline being tested, ensuring that the scanning mechanism maintains an appropriate distance from the surface of the pipeline being tested.
[0034] After installation, the device is started. The controller first receives the position signal output by the encoding positioning module 3, calculates the axial position coordinates of the current X-axis slider 8 and the circumferential position coordinates of the current Y-axis slider 6. Simultaneously, the controller acquires the original wall thickness data at the measuring point collected by the ultrasonic sensor 2, and collects the temperature parameters of the current detection area in real time through the temperature sensor 4. Since the propagation speed of ultrasound in the measured medium is affected by temperature, the controller uses a preset sound velocity-temperature correction formula to compensate for the sound velocity based on the temperature parameters collected by the temperature sensor 4, thereby correcting the original wall thickness data to the compensated wall thickness value under actual temperature conditions. After correction, the controller associates and binds the compensated wall thickness value with the position coordinates of the current measuring point to generate a wall thickness distribution heat map of the measured area.
[0035] During the scanning process, the controller controls the X-axis drive motor 81 and the Y-axis drive motor 61 to move in tandem, driving the ultrasonic sensor 2 on the X-axis slider 8 to scan point by point along a preset path. When the X-axis slider 8 moves to the end of the linear guide rail 7, the travel limit component triggers the limit switch. After receiving the limit signal, the controller controls the X-axis drive motor 81 to stop or reverse its movement to prevent overtravel. The corrected wall thickness data and corresponding position coordinates of each measuring point are uploaded to the cloud server in real time via the wireless communication module.
[0036] The entire scanning process is completed automatically. Users can access real-time and historical data via corrosion monitoring software on their office computers or mobile devices, and view the wall thickness distribution map of each measuring point on the tested pipeline. Figure 3 This is a thermal map of pipeline corrosion status, which visually presents the wall thickness distribution of the measured area in a two-dimensional plane. The wall thickness values are distinguished by color gradients, and the location and severity of two typical corrosion defects of different sizes and depths are clearly identified. Figure 4 This is a three-dimensional oblique view of the pipeline defect. The three-dimensional shape of two corrosion defects is restored in the form of a three-dimensional curved surface, which intuitively shows the depth, outline and spatial distribution characteristics of the defect, and realizes the visualized and quantitative presentation of corrosion defects.
[0037] Based on the historical wall thickness data collected above, this method further performs corrosion rate calculation, trend prediction, and graded early warning. Specifically, it includes the following steps: S1. Control the two-dimensional scanning mechanism to scan along the preset path and collect the original wall thickness data, position coordinates and temperature parameters of each measuring point; S2. Compensate for the sound velocity according to the temperature parameter, and correct the original wall thickness data to the compensated wall thickness value. S3. Associate the compensated wall thickness value with the position coordinates to generate a wall thickness distribution heat map of the measured area; S4. Based on the compensated wall thickness values from multiple historical scans at each measuring point, calculate the corrosion rate at each measuring point and evaluate its data confidence level, specifically including the following sub-steps: S41. For each measuring point, extract its historical wall thickness data from the most recent N valid scans, where N≥5, and construct a time and wall thickness data point set {(t1,d1),(t2,d2),…,(t...}}. n ,d n )}; S42. The least squares method is used to perform linear regression fitting on the data point set to obtain the trend line d(t)=a·t+b of the wall thickness change with time. The absolute value of the slope a of the fitted line is calculated as the corrosion rate CR=|a|. S43. Calculate the determination coefficient R of the linear regression fit. 2 , where R 2 =1-[Σ(di- i) 2 ] / [Σ(di- ) 2 ]; S44, R 2 Compared with the preset confidence threshold R 2 The comparison is performed using _th (default is 0.85, user-adjustable). If R... 2 ≥R 2 If _th, then the corrosion rate is determined to be valid and output; if R 2 <R 2 If the system determines that the data fluctuation is too large or the erosion rate is unreliable, it will trigger a low-confidence processing mechanism, which includes at least one of the following operations: (1) Data quality marking The system attaches a low-confidence label to the corrosion rate data at this location and marks the area in the wall thickness distribution heatmap with a special color (such as gray) or a special style (such as diagonal fill) to remind users to refer to the corrosion rate data at this location with caution.
[0038] (2) Automatically trigger retest The system sends a retest command to the on-site monitoring host to conduct a more intensive scan of the point and its adjacent area during the next available scanning cycle. The specific method of the intensive scan is as follows: the scanning step is halved, that is, the original scanning step length is reduced by 50%; the number of repeated samplings is increased to 3 to 5 times to verify the consistency of the data and reduce random measurement errors.
[0039] (3) Sensor status self-test The system automatically retrieves the self-test log of the scanning and monitoring terminal to check the following parameters of the ultrasonic sensor: whether the excitation voltage is within the normal range, whether the echo amplitude is sufficient, and whether the coupling state is good. If any parameter is detected as abnormal, a sensor status warning is generated, prompting maintenance personnel to check whether the sensor is loose, whether there is dirt accumulation on the surface, or whether the electromagnetic interference is too high.
[0040] (4) Working condition correlation analysis The system retrieves environmental parameters such as temperature and humidity from locations with low confidence levels over the same time period to analyze the impact of environmental fluctuations on the data. (R) is then determined. 2 The reduction is determined to be caused by drastic environmental fluctuations. If the changes in environmental parameters exceed the preset normal fluctuation range, the location is marked as potentially affected by environmental disturbances, and the sound velocity after temperature compensation is used for refitting in subsequent calculations to eliminate interference from environmental factors.
[0041] (5) Dynamically adjust the monitoring frequency R calculated for two consecutive scanning cycles 2For locations where all readings are below the preset threshold, the system will automatically increase the monitoring frequency of that location from the default cycle (e.g., once a week) to an encrypted cycle (e.g., once a day), until the R value of that location reaches the preset threshold. 2 The risk level is restored to above the threshold to ensure continuous monitoring of the risk point.
[0042] (6) Generate diagnostic report The system automatically generates a low-confidence diagnostic report, which includes the following: low R... 2 Spatial distribution map of the monitoring points, scatter plot of historical wall thickness and fitting curve, sensor self-test data, environmental parameter curves, and suggested treatment measures are provided for technical personnel to refer to and analyze.
[0043] (7) Continuous labeling processing After completing the retest and sensor check, the system recalculates R based on the newly acquired data. 2 If data quality is restored (R... 2 If R ≥ the threshold, then the corrosion rate is recalculated using new data, and the low-confidence label is removed; if R 2 If the data is still below the threshold, the corrosion rate data at that location will remain marked with a low confidence level, and the situation at that location will be explained separately in the monitoring report to remind users to use the relevant data with caution.
[0044] S5. Extract the minimum wall thickness value of the entire monitoring area from each historical scan, arrange them in chronological order of scan time, and use ARIMA or ARIMAX models to predict the future trend of the minimum wall thickness of the area and the expected time to reach the safety threshold. Figure 2 The curves show the trend of pipe wall thickness over time and its life prediction. The left side shows historical wall thickness monitoring data and corrosion trend lines, while the right side shows the wall thickness decay prediction results based on a linear corrosion model. This visually displays the current wall thickness, corrosion rate, and the expected time point when the wall thickness drops to the warning / defense wall thickness.
[0045] This invention provides two parallel and independent prediction model implementation methods: ARIMA model and ARIMAX model. Users can choose one of them based on actual data conditions and the availability of operating parameters.
[0046] Method 1: Using the ARIMA model for prediction: First, perform the augmented Dickey-Fowler stationarity test (ADF test) on the time series data to determine the differencing order d. If the series is non-stationary, perform first-order differencing successively until it becomes stationary, and record the number of differencing iterations d. Since wall thickness series usually have a monotonically decreasing trend, d is usually taken as 1.
[0047] Secondly, the minimum information criterion, combining the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC), is used to determine the autoregressive order p and the moving average order q within a preset search range p∈[0,5] and q∈[0,5], thus constructing an ARIMA(p,d,q) model. The (p,q) combination that minimizes the weighted sum of AIC and BIC is selected as the optimal model parameters.
[0048] Then, the constructed ARIMA model is subjected to a residual white noise test (Ljung-Box Q test) to verify the model's applicability and ensure that the residual sequence of the selected model is free from autocorrelation.
[0049] Finally, the trained ARIMA model is used to predict the trend of regional minimum wall thickness change over a specified number of days in the future, and the prediction confidence interval is calculated to evaluate the reliability of the prediction results. Based on the prediction results, the expected time for the regional minimum wall thickness to reach the safety threshold is calculated.
[0050] Method 2: Use the ARIMAX model for prediction.
[0051] When the system is configured with an interface for acquiring operating parameters and historical operating data is available, the ARIMAX model can be used. ARIMAX is an extended model based on the ARIMA model that introduces exogenous variables, enabling the incorporation of operating parameters affecting corrosion rates into the prediction.
[0052] First, collect operating parameters aligned with the timestamps of each historical scan, including at least one of temperature, pressure, flow rate, pH value, and chloride ion concentration.
[0053] Secondly, each working condition parameter is normalized (Z-score standardization or Min-Max normalization) and correlated with the minimum wall thickness sequence in the region (Pearson correlation coefficient). Exogenous variables with absolute values of correlation coefficients greater than a preset threshold (such as 0.3) are selected to avoid introducing irrelevant variables that could lead to overfitting.
[0054] Then, the selected exogenous variables are introduced as regression terms into the ARIMA model to construct the ARIMAX(p,d,q) model.
[0055] Finally, conditional maximum likelihood estimation or least squares estimation is used to solve for the model parameters, including autoregressive coefficients, moving average coefficients, and exogenous variable regression coefficients. The trained ARIMAX model is used to predict the trend of regional minimum wall thickness change over a specified number of days in the future, and the prediction confidence interval is calculated to assess the reliability of the prediction results. Based on the prediction results, the expected time for the regional minimum wall thickness to reach the safe threshold is calculated.
[0056] It should be noted that the ARIMA and ARIMAX models mentioned above are two parallel and independent implementation methods. Users can choose one according to their actual application scenario, and the two do not need to be executed simultaneously. When historical operating data is sufficient and significantly correlated with wall thickness changes, the ARIMAX model usually has higher prediction accuracy; when operating data is unavailable or has weak correlation, the ARIMA model can meet basic prediction needs.
[0057] The corrosion process may change over time (e.g., the corrosion rate may accelerate or slow down), and prediction models with fixed parameters will gradually become ineffective. Therefore, this method introduces an online learning mechanism.
[0058] After each new scan is completed and a new minimum wall thickness value for the region is obtained, the system adds the latest data to the time series and removes the oldest data that is outside a preset time window (e.g., the last 365 days). It then automatically re-executes the stationarity test, model order determination, and parameter estimation to generate an updated prediction model. The system compares the prediction errors (e.g., root mean square error RMSE) of the new and old models. If the RMSE of the new model is reduced by more than a preset percentage (e.g., 5%) compared to the old model, the new model is deployed; otherwise, the original model is maintained to avoid overfitting.
[0059] S6. Based on the corrosion rate, data confidence level, and trend prediction results, a graded early warning is issued.
[0060] The tiered early warning system includes prediction-based early warning: when the prediction shows that the minimum wall thickness in the area will approach the safety threshold within a first preset time period, a Level 1 early warning is triggered and it is recommended to increase the monitoring frequency; when the prediction shows that the minimum wall thickness in the area will fall below the safety threshold within a second preset time period, a Level 2 early warning is triggered and a maintenance work order is automatically generated; when the prediction shows that the minimum wall thickness in the area will exceed the danger threshold within a third preset time period, a Level 3 early warning is triggered and an emergency alarm is sent, wherein the first preset time period is longer than the second preset time period, and the second preset time period is longer than the third preset time period. Preferably, the first preset time period is 90 days, the second preset time period is 30 days, and the third preset time period is 7 days.
[0061] Although the invention has been described with reference to preferred embodiments, various modifications can be made thereto and components can be replaced with equivalents without departing from the scope of the invention. In particular, the technical features mentioned in the various embodiments can be combined in any manner, provided there is no structural conflict. The invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.
[0062] In the description of this invention, terms such as "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," which indicate direction or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. These are used merely for ease of description and do not indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation; therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.
[0063] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.
[0064] The term "comprising" or any other similar term is intended to cover non-exclusive inclusion, such that a process, article, or apparatus / device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to those processes, articles, or apparatus / devices.
[0065] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.
Claims
1. A method for early warning of regional corrosion using ultrasonic scanning imaging, characterized in that: Includes the following steps: S1. Control the two-dimensional scanning mechanism to scan along the preset path and collect the original wall thickness data, position coordinates and temperature parameters of each measuring point; S2. Compensate for the sound velocity according to the temperature parameter, and correct the original wall thickness data to the compensated wall thickness value. S3. Associate the compensated wall thickness value with the position coordinates to generate a wall thickness distribution heat map of the measured area; S4. Based on the compensated wall thickness values from multiple historical scans of each measuring point, calculate the corrosion rate of each measuring point and evaluate its data confidence level. S5. Extract the minimum wall thickness value of the entire monitoring area from each historical scan, arrange them in chronological order of scan time, and use ARIMA or ARIMAX models to predict the future trend of the minimum wall thickness of the area and the expected time to reach the safety threshold. S6. Based on the corrosion rate, data confidence level, and trend prediction results, a graded early warning is issued.
2. The ultrasonic scanning imaging early warning method for regional corrosion status according to claim 1, characterized in that: Step S4 specifically includes the following sub-steps: S41, for each measuring point, extract its historical wall thickness data in the last N effective scans, where N≥5, and construct a time and wall thickness data point set {(t1, d1), (t2, d2), …, (tN, dN)}. n n} S42. The least squares method is used to perform linear regression fitting on the data point set to obtain the trend line d(t)=a·t+b of the wall thickness change with time. The absolute value of the slope a of the fitted line is calculated as the corrosion rate CR=|a|. S43. Calculate the determination coefficient R of the linear regression fit. 2 , where R 2 =1-[Σ(di- i) 2 ] / [Σ(di- ) 2 ]; S44, R 2 Compared with the preset confidence threshold R 2 Compare R² with R_th. If R² ≥ R²_th, the corrosion rate is deemed valid and output; otherwise, R_th is not valid. 2 <R 2 If _th is encountered, the low confidence handling mechanism will be triggered.
3. The ultrasonic scanning imaging early warning method for regional corrosion status according to claim 1, characterized in that: Step S5, which uses the ARIMA model for prediction, includes the following sub-steps: S5A1. Extract the minimum wall thickness values of each historical scan from the historical wall thickness database, arrange them in chronological order of scan time, and form time series data with a series length of at least 30 time points. S5A2. Perform an augmented Dickie-Fowler stationarity test on the time series data to determine the difference order d, and then successively differiate the non-stationary series until it becomes stationary. S5A3. Using the minimum information criterion that combines the Akaike information criterion and the Bayesian information criterion, the autoregression order p and the moving average order q are determined within the preset search range p∈[0,5] and q∈[0,5], and an ARIMA(p,d,q) model is constructed. S5A4. Perform residual white noise test on the constructed ARIMA model to verify the model's applicability; S5A5: Use the trained ARIMA model to predict the trend of regional minimum wall thickness change over a specified number of days in the future, and calculate the prediction confidence interval to evaluate the reliability of the prediction results. Based on the prediction results, calculate the time when the regional minimum wall thickness is expected to reach the safety threshold.
4. The ultrasonic scanning imaging early warning method for regional corrosion status according to claim 1, characterized in that: The construction of the ARIMAX model in step S5 also includes the step of treating the operating condition parameters as exogenous variables: S5B1. Collect operating parameters aligned with the timestamps of each historical scan, wherein the operating parameters include at least one of temperature, pressure, flow rate, pH value, and chloride ion concentration; S5B2. Normalize each working condition parameter and perform correlation analysis with the minimum wall thickness sequence in the region to screen exogenous variables whose absolute value of the correlation coefficient is greater than the preset threshold. S5B3. Introduce the selected exogenous variables as regression terms into the ARIMA model to construct the ARIMAX(p,d,q) model. S5B4. Solve the model parameters using conditional maximum likelihood estimation or least squares estimation. The parameters include autoregressive coefficients, moving average coefficients, and regression coefficients of exogenous variables. S5B5. Use the trained ARIMAX model to predict the trend of regional minimum wall thickness change over a specified number of days in the future, and calculate the prediction confidence interval to evaluate the reliability of the prediction results. Based on the prediction results, calculate the time when the regional minimum wall thickness is expected to reach the safety threshold.
5. The ultrasonic scanning imaging early warning method for regional corrosion status according to claim 2, characterized in that: The low-confidence processing mechanism in step S44 includes at least one of the following operations: S441. The corrosion rate data of the measuring point is labeled with a low confidence level, and the area where the measuring point is located is marked with a preset exclusive color in the wall thickness distribution heat map. S442. Issue a retest control command to the two-dimensional scanning mechanism to perform a denser scan on the measurement point and its adjacent area in the next scanning cycle; the scanning step size of the denser scan is reduced by 50% compared with the conventional scanning step size, and the number of repeated acquisitions of a single point is set to 3-5 times; S443. Retrieve the device self-test log of the ultrasonic sensor and check at least one of the sensor excitation voltage, echo amplitude, and coupling status parameters; if any parameter exceeds the normal operating range, trigger and output sensor status warning information. S444. Retrieve the temperature and humidity environmental parameters for the same monitoring period as the measuring point, and analyze the impact of environmental parameter fluctuations on the wall thickness detection data; if the environmental parameter fluctuations exceed the preset normal fluctuation threshold range, then mark the measuring point as a suspected environmental interference point. S445. The coefficient of determination R for two consecutive scanning cycles. 2 For all measurement points below the preset threshold, the system will automatically increase the monitoring frequency of the measurement point from the preset default monitoring cycle to the preset encrypted monitoring cycle.
6. The ultrasonic scanning imaging early warning method for regional corrosion status according to claim 1, characterized in that: It also includes an online model update step: after each new scan is completed and a new minimum wall thickness value for the region is obtained, the latest data is added to the time series, while the oldest data that exceeds the preset time window is removed. The stationarity test, model order determination and parameter estimation are automatically re-executed to generate an updated prediction model. The prediction errors of the new and old models are compared. If the root mean square error of the new model is reduced by more than a preset percentage compared with the old model, the new model is deployed.
7. The ultrasonic scanning imaging early warning method for regional corrosion status according to claim 1, characterized in that: The graded early warning in step S6 includes prediction early warning based on the prediction results of ARIMA or ARIMAX models: when the prediction result shows that the minimum wall thickness of the area will approach the safety threshold in the first preset time period in the future, a first-level early warning is triggered; when the prediction result shows that the minimum wall thickness of the area will be lower than the safety threshold in the second preset time period in the future, a second-level early warning is triggered; and when the prediction result shows that the minimum wall thickness of the area will exceed the danger threshold in the third preset time period in the future, a third-level early warning is triggered, wherein the first preset time is greater than the second preset time, and the second preset time is greater than the third preset time.
8. An ultrasonic scanning monitoring device for regional corrosion status, wherein the ultrasonic scanning monitoring device for regional corrosion status is used to execute the ultrasonic scanning imaging early warning method for regional corrosion status according to any one of claims 1-7, characterized in that: The system includes a two-dimensional scanning mechanism (1) installed on the outer wall of the pipe under test. The two-dimensional scanning mechanism (1) includes an X-axis module arranged along the axial direction of the pipe under test and a Y-axis module arranged along the circumferential direction of the pipe under test. An ultrasonic sensor (2) for collecting the original wall thickness data of each measuring point of the pipe under test is fixedly installed on the moving end of the two-dimensional scanning mechanism (1). An encoding positioning module (3) for outputting position signals is also fixedly installed on the moving end of the two-dimensional scanning mechanism (1). A temperature sensor (4) for collecting the temperature parameters of the measured area is fixedly installed on the moving end of the two-dimensional scanning mechanism (1) or on the outer wall of the pipe under test. The system also includes a controller. The ultrasonic sensor (2), the encoding positioning module (3) and the temperature sensor (4) are electrically connected to the controller. The controller is configured to acquire the position signal of the coding positioning module (3) and calculate it into position coordinates, acquire the original wall thickness data of the ultrasonic sensor (2), perform sound velocity compensation correction on the original wall thickness data according to the temperature parameters collected by the temperature sensor (4) to generate a corrected wall thickness value, and associate and bind the corrected wall thickness value with the corresponding position coordinates.
9. The ultrasonic scanning monitoring device for regional corrosion status according to claim 8, characterized in that: The Y-axis module includes an annular curved guide rail (5) fixedly installed on the outer wall of the pipe being tested, and a Y-axis slider (6) slidably fitted on the annular curved guide rail (5); the X-axis module includes a linear guide rail (7) fixedly installed on the Y-axis slider (6), and an X-axis slider (8) slidably fitted on the linear guide rail (7); the ultrasonic sensor (2), the coding positioning module (3), and the temperature sensor (4) are fixedly installed on the X-axis slider (8).
10. The ultrasonic scanning monitoring device for regional corrosion status according to claim 9, characterized in that: The encoding positioning module (3) is an integrated position encoder that integrates a linear displacement sensing unit and an angular displacement sensing unit; the linear displacement sensing unit cooperates with the linear guide rail (7) to output the X-axis axial position signal; the angular displacement sensing unit cooperates with the annular curved guide rail (5) to output the Y-axis circumferential position signal; the outer wall of the pipe under test is fixed with a circumferential reference mark to provide a circumferential zero-point positioning signal.