Corrosion detection method and system for inner wall of water cooling wall of boiler enclosure wall

Abstract

The embodiment of the invention provides a corrosion detection method and system for the inner wall of a water cooling wall of a boiler enclosure wall, and the method comprises the steps: configuring detection parameters, controlling execution equipment to position sensing equipment to a calibration region, and receiving and processing sensing data to establish a detection reference; a scanning instruction is issued to the execution device based on the detection reference, the sensing device is controlled to move in the axial direction of the pipeline, and the collected ultrasonic signals are received in real time; processing and analyzing the received ultrasonic signal in real time, and generating an analysis result; and based on the analysis result, generating a detection report, controlling execution equipment to return to an initial position, and performing data storage and archiving. According to the embodiment of the invention, the problems of poor reliability, low efficiency, dependence on manpower, difficulty in tracing and the like in the traditional detection are systematically solved, and the transformation of the boiler water wall corrosion detection from discretization and qualification to systematization, intelligence and quantifiability is realized.

Description

Technical Field

[0001] This application relates to the technical field of boiler corrosion detection, specifically to a method and system for detecting corrosion on the inner wall of the water-cooled wall of a boiler. Background Technology

[0002] Boilers are the core equipment of thermal power generating units, and their water-cooled walls are the most important heat-receiving surfaces. These walls are composed of multiple parallel, tightly arranged tubes arranged around the furnace and on the inner wall of the tail flue. High-temperature, high-pressure water or steam flows inside them, while they are directly subjected to flame radiation and the scouring of high-temperature flue gas. Their main functions are to absorb heat, generate steam, and protect the furnace walls. Due to long-term exposure to harsh environments of high temperature, high pressure, corrosion, and stress loads, their inner walls are prone to defects such as flow-accelerated corrosion, hydrogen damage, and microcracks. These defects are the main causes of boiler leaks, unplanned shutdowns, and even tube rupture accidents.

[0003] Generally, traditional boiler wall water-cooled wall inspection mainly includes visually inspecting the tube wall after shutdown by entering the furnace through the manhole to check for macroscopic defects such as bulging, discoloration, and cracks. Alternatively, a conventional single-crystal ultrasonic thickness gauge can be used to sample and measure the thickness of the tube wall to assess corrosion thinning. Magnetic particle testing or penetrant testing can be performed on suspicious areas to detect surface cracks, or conventional ultrasonic testing can be used to spot-check internal defects.

[0004] However, inspections can only be conducted after the unit has been shut down and cooled down. This results in a short inspection window, low efficiency, and the fact that the furnace is so large that only sampling inspections are possible, making it easy to miss tiny or undiscovered defects. Visual inspections and routine UTs heavily rely on the experience of the inspectors, making it difficult to accurately quantify and record parameters such as the depth of corrosion pits and the length of microcracks. Furthermore, the lack of precise digital coordinate positioning and historical data comparison makes it difficult to scientifically assess and predict the rate of defect development and the remaining lifespan of components. Finally, there are safety risks associated with inspectors entering the high-temperature, high-dust furnace.

[0005] Therefore, there is an urgent need for a detection scheme for corrosion of the inner wall of the boiler water-cooled wall to solve the aforementioned technical problems. Summary of the Invention

[0006] This application proposes a method and system for detecting corrosion on the inner wall of the water-cooled wall of a boiler casing, in order to overcome the deficiencies of the prior art.

[0007] According to a first aspect of the embodiments of this application, a method for detecting corrosion on the inner wall of a boiler cladding water-cooled wall is provided, comprising: Configure detection parameters and control the execution device to position the sensing device to the calibration area, receive and process sensing data to establish a detection benchmark; Based on the detection benchmark, a scanning command is sent to the execution device to control the sensing device to move along the pipeline axis and receive the collected ultrasonic signals in real time. The received ultrasound signals are processed and analyzed in real time, and analysis results are generated. Based on the analysis results, a detection report is generated, and the execution device is controlled to return to its initial position, and data is stored and archived.

[0008] In some embodiments, the real-time processing and analysis of the received ultrasound signal includes: When the ultrasound signal is identified as an abnormal signal, a re-inspection command is generated to control the execution device to perform a fine re-scan and obtain re-collected data; The re-collected data is analyzed and processed, and an intelligent recognition algorithm is used to classify and quantify the defects in the re-collected data and generate evaluation parameters. When the evaluation parameters meet the defect threshold, historical records are invoked for trend analysis, and life prediction is performed in conjunction with material parameters. When the ultrasound signal is identified as non-abnormal or the evaluation parameters do not meet the defect threshold, the current operating status is maintained.

[0009] In some embodiments, configuring the detection parameters and controlling the execution device to position the sensing device to the calibration area includes: Configure the detection parameters by receiving user input through the human-computer interaction interface; A first control instruction is generated according to the configuration instruction. The first control instruction is used to control the execution device to drive the sensing device to move to a calibration area that has been manually verified to be defect-free.

[0010] In some implementations, receiving and processing the sensor data to establish a detection benchmark includes: The sensor receives background noise and ultrasonic signals collected in the calibration area by the sensor device. The ultrasonic signal is processed and the average sound velocity and attenuation coefficient of the calibration area are calculated; The average sound velocity and the attenuation coefficient are stored as reference values, and a special test template for water-cooled wall tubes is generated.

[0011] In some embodiments, the real-time processing and analysis of the received ultrasound signal includes: The real-time uploaded ultrasound signal is processed using a full-focusing algorithm to generate a real-time ultrasound image; The signal-to-noise ratio (SNR) of the ultrasound image is analyzed. When the SNR drops below a set threshold relative to the baseline value, it is determined to be an abnormal signal.

[0012] In some implementations, the generation of re-inspection instructions controls the execution device to perform refined re-scanning and obtain re-collected data, including: A second control command containing precise coordinates is generated. The second control command is used to control the execution device to reduce the scanning speed to the re-inspection speed, retreat to the upstream of the suspicious point, and rescan. The control sensor device starts a high-frequency sampling mode at the suspicious point location to collect signals multiple times, and the suspicious point location is the location where the abnormal signal is detected.

[0013] In some implementations, the step of using an intelligent recognition algorithm to classify and quantify defects in the re-collected data and generate evaluation parameters includes: A trained deep learning model is used to identify and classify defects in the re-collected data after analysis and processing. If it is identified as a corrosion defect, its depth is calculated using time-difference diffraction. If the defect is identified as a crack, its length and height are calculated using the end diffraction wave method. Otherwise, the re-collected data is marked as defect-free or unidentifiable, and a judgment result is generated.

[0014] In some implementations, the step of calling historical records for trend analysis based on the evaluation parameters meeting the defect threshold includes: Retrieve historical inspection records for the current pipe section from the historical database; Calculate the trend of change in the number or density of defects per unit area; When the trend of change exceeds a set threshold, a high-risk warning message is generated; When the change trend does not exceed the set threshold, a marker information indicating that the pipe section is in a stable state is generated.

[0015] In some implementations, the lifetime prediction based on material parameters includes: Based on the current size and historical development trend of the defect, the corrosion or crack propagation rate is calculated using the Paris formula. The remaining life prediction result of the current pipe section is generated by querying the material performance parameter database and combining it with the expansion rate.

[0016] According to a second aspect of this application, a corrosion detection system for the inner wall of a boiler cladding water-cooled wall is provided, comprising: The detection benchmark establishment module is used to configure detection parameters and control the execution device to position the sensing device to the calibration area, and to receive and process sensing data to establish a detection benchmark. An ultrasonic signal acquisition module is used to send scanning commands to the execution device based on the detection benchmark, control the sensing device to move along the pipeline axis, and receive the acquired ultrasonic signals in real time. The analysis result generation module is used to process and analyze the received ultrasound signals in real time and generate analysis results. The test report generation module is used to generate a test report based on the analysis results, control the execution device to return to the initial position, and perform data storage and archiving.

[0017] The beneficial effects of the corrosion detection method and system for the inner wall of the boiler casing water-cooled wall according to the embodiments of this application include at least the following: This application embodiment effectively eliminates the interference of different material properties and environmental background noise on the detection signal by establishing an acoustic benchmark in a defect-free calibration area in advance, providing a unified and reliable comparison standard for subsequent full-area scanning, thereby ensuring the accuracy and consistency of the detection results from the source. By driving the execution device to perform automated axial scanning based on benchmark parameters, full-coverage data acquisition of water-cooled wall pipes is achieved, avoiding the blind spots and missed detection risks inherent in traditional manual sampling methods. At the same time, the real-time upload mechanism ensures strict synchronization and integrity of massive detection data and spatial coordinates. Through the embedded real-time processing unit to perform full-focus imaging and intelligent analysis of ultrasonic signals, abnormal signals can be quickly identified online and decision-making processes can be triggered, significantly improving defect identification efficiency and the degree of detection automation, and reducing reliance on the subjective experience of operators. Finally, through the automatic generation of structured reports and data archiving functions, a standardized and traceable detection closed loop is formed, which not only greatly improves the efficiency of report generation, but also provides a solid data foundation for long-term status monitoring, trend analysis, and predictive maintenance of equipment. In summary, the embodiments of this application systematically solve the problems of poor reliability, low efficiency, reliance on manual labor, and difficulty in traceability in traditional detection by constructing a complete closed loop of benchmark establishment, automatic scanning, real-time analysis, and intelligent archiving. It realizes the transformation of boiler water-cooled wall corrosion detection from discrete and qualitative to systematic, intelligent, and quantifiable, and ultimately provides core technical support for the safe operation and predictive maintenance of equipment. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of a corrosion detection method for the inner wall of a boiler casing water-cooled wall according to an embodiment of this application. Figure 2 This is a flowchart illustrating a specific implementation process of the corrosion detection method for the inner wall of the boiler casing water-cooled wall according to an embodiment of this application. Figure 3 A schematic diagram of the corrosion detection system for the inner wall of the boiler casing water-cooled wall according to an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] In the description of this specification, the references to terms such as "an embodiment," "example," "specific example," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the embodiments of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0021] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed embodiments of the present application, but merely to illustrate selected embodiments of the present application. Other embodiments obtained by those skilled in the art based on the embodiments of the present application without inventive effort are all within the scope of protection of the embodiments of the present application.

[0022] It can be noted that similar reference numerals and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it will not be further defined and explained in subsequent figures according to the embodiments of this application.

[0023] This application discloses a method for detecting corrosion on the inner wall of a boiler cladding water-cooled wall. The purpose of this method is to provide a way to detect corrosion on the inner wall of a boiler cladding water-cooled wall, thereby solving the problems mentioned in the background art. (See attached document.) Figure 1 As shown, the method includes steps 110-140.

[0024] Step 110: Configure detection parameters and control the execution device to position the sensing device in the calibration area, receive and process sensing data to establish a detection benchmark.

[0025] Before commencing the formal scanning, this embodiment first establishes a detection benchmark. Step 110 configures initial detection parameters through the human-computer interaction interface of the analysis layer and controls the magnetic adsorption crawling robot of the execution layer to precisely position the integrated high-frequency ultrasonic phased array sensor (i.e., the sensing device) to a calibration area that has been manually verified to be defect-free. Subsequently, the system drives the sensing device to move within this area and collects the raw ultrasonic signals and background noise data of all array elements.

[0026] Furthermore, by processing and analyzing these raw signals, the system calculates key benchmark parameters such as the average sound velocity and signal attenuation coefficient of the material in that specific region. These parameters are stored in a database, and a dedicated inspection template for that water-cooled wall tube is generated. This aims to eliminate the influence of differences in the acoustic properties of the material itself and environmental background noise on subsequent inspection results, providing an accurate and reliable quantitative comparison benchmark for subsequent fully automated scanning and defect identification, thereby fundamentally ensuring the accuracy and consistency of the inspection results.

[0027] In some implementations, configuring the detection parameters and controlling the execution device to position the sensing device to the calibration area includes: receiving a user-inputted detection parameter configuration instruction through a human-machine interface and configuring it; generating a first control instruction based on the configuration instruction, wherein the first control instruction is used to control the execution device to drive the sensing device to move to the calibration area that has been manually verified to be defect-free.

[0028] In some embodiments, receiving and processing sensor data to establish a detection benchmark includes: receiving background noise and ultrasonic signals collected by the sensing device in a calibration area; processing the ultrasonic signals and calculating the average sound velocity and attenuation coefficient of the calibration area; storing the average sound velocity and attenuation coefficient as benchmark values ​​and generating a dedicated detection template for water-cooled wall tubes.

[0029] In some embodiments, receiving and processing sensor data to establish a detection benchmark includes: receiving background noise and ultrasonic signals collected by the sensing device in a calibration area; processing the ultrasonic signals and calculating the average sound velocity and attenuation coefficient of the calibration area; storing the average sound velocity and attenuation coefficient as benchmark values ​​and generating a dedicated detection template for water-cooled wall tubes.

[0030] Step 120: Based on the detection benchmark, a scanning command is sent to the execution device to control the sensing device to move along the pipeline axis and receive the collected ultrasonic signals in real time.

[0031] After establishing the detection benchmark, this embodiment of the application enters the fully automated scanning stage. Based on the stored dedicated detection template parameters, the analysis layer sends scanning commands to the servo drive control system of the execution layer. These commands control a magnetically adsorbed crawling robot carrying the sensing device to move smoothly and continuously along the axial direction of the boiler water-cooled wall tubes at a preset constant speed.

[0032] During this process, the high-frequency ultrasonic phased array probe integrated into the sensing device is continuously excited and acquires raw ultrasonic A-scan signals at predetermined intervals. This massive amount of raw data, synchronized with the precise position information fed back in real time by the encoder, is uploaded in real time to the signal processing unit of the analysis layer for buffering and storage via a high-bandwidth communication link. This provides a complete, continuous, and location-tagged data source for subsequent real-time imaging and intelligent analysis. This step achieves 100% coverage detection of the water-cooled wall tube bank, laying the foundation for high-precision, quantifiable data assessment.

[0033] Step 130: The received ultrasound signal is processed and analyzed in real time, and analysis results are generated.

[0034] In this embodiment, the received raw ultrasound signal is processed and analyzed in real time online. This process first utilizes a full-focusing algorithm to perform high-speed computation and beamforming on massive amounts of raw data, dynamically generating a high-resolution image to visualize the detection area. Simultaneously, the algorithm's core calculates the signal-to-noise ratio in real time and compares it with a pre-established detection benchmark.

[0035] Furthermore, when system analysis reveals that the signal-to-noise ratio (SNR) at a certain location is significantly lower than the baseline threshold, and this anomaly is correlated between adjacent array element channels, it is marked as a suspicious point. The analysis layer simultaneously retrieves the acoustic coupling state data at that location for verification. If the coupling is good, it is confirmed as a potential defect signal. Finally, the system generates an analysis result instruction containing precise coordinates and re-inspection parameters, providing a decision-making basis for triggering subsequent refined re-inspection processes. This real-time processing and analysis capability is the core of achieving efficient, accurate, and automated detection.

[0036] In some embodiments, the real-time processing and analysis of the received ultrasonic signal includes: when the ultrasonic signal is identified as an abnormal signal, generating a re-inspection command to control the execution device to perform a refined re-scan and obtain re-collected data; analyzing and processing the re-collected data, using an intelligent recognition algorithm to classify and quantify defects in the re-collected data and generate evaluation parameters; when the evaluation parameters meet the defect threshold, calling historical records for trend analysis and combining material parameters for life prediction; when the ultrasonic signal is identified as a non-abnormal signal or the evaluation parameters do not meet the defect threshold, maintaining the current operating state.

[0037] In some embodiments, the real-time processing and analysis of the received ultrasound signal includes: processing the real-time uploaded ultrasound signal using a full-focusing algorithm to generate a real-time ultrasound image; analyzing the signal-to-noise ratio (SNR) change of the ultrasound image, and determining an abnormal signal when the SNR drops below a set threshold relative to the reference value.

[0038] In some implementations, the step of generating a re-inspection command to control the execution device to perform a refined re-scan and obtain re-collected data includes: generating a second control command containing precise coordinate positions, the second control command being used to control the execution device to reduce the scanning speed to the re-inspection speed, retreat to the upstream of the suspicious point, and re-scan; and controlling the sensing device to start a high-frequency sampling mode at the suspicious point position to collect signals multiple times, the suspicious point position being the location where the abnormal signal was detected.

[0039] In some implementations, the step of using an intelligent identification algorithm to classify and quantify defects in the re-collected data and generate evaluation parameters includes: using a trained deep learning model to identify and classify defects in the analyzed and processed re-collected data; if the defect is identified as corrosion, the depth is calculated using time-difference diffraction; if the defect is identified as crack, the length and height are calculated using end diffraction wave method; otherwise, the re-collected data is marked as defect-free or unidentifiable, and the ultrasonic signal, the judgment result, and the corresponding coordinate position are saved to a database for subsequent historical data comparison and model optimization.

[0040] In some implementations, the step of calling historical records for trend analysis based on the evaluation parameters meeting the defect threshold includes: calling historical inspection records of the current pipe segment from the historical database; calculating the trend of change in the number or density of defects per unit area; generating high-risk warning information when the trend exceeds a set threshold; and archiving and storing the re-collected data and the corresponding historical data in the historical inspection records when the trend does not exceed the set threshold, and generating a marker information indicating that the pipe segment is in a stable state.

[0041] In some implementations, the life prediction based on material parameters includes: calculating the corrosion or crack propagation rate using the Paris formula based on the current size and historical development trend of the defect; querying a material performance parameter library and generating a prediction result of the remaining life of the current pipe segment based on the propagation rate.

[0042] Step 140: Based on the analysis results, generate a detection report and control the execution device to return to the initial position, and perform data storage and archiving.

[0043] In this embodiment of the application, after the testing process is completed, the system automatically generates a structured comprehensive evaluation report based on the final defect analysis, trend comparison, and lifespan prediction results. This report fully includes corrosion patterns, defect quantification parameters, historical change curves, predicted remaining lifespan, and maintenance recommendations, and is converted into a standard format file.

[0044] Simultaneously, the analysis layer sends a return command to the execution device, controlling the magnetically adsorbed crawling robot to return to the detection starting point carrying the sensing device and automatically release the magnetic adsorption state. All raw data, processing procedures, and final reports are encrypted and compressed, assigned a unique identifier, and automatically uploaded to the power plant's data center server for long-term archiving and storage, forming a complete, traceable, and reproducible testing archive, providing data support for the full lifecycle management of the equipment.

[0045] In some implementations, generating the inspection report further includes completing data storage and archiving, specifically including: generating a complete structured assessment report containing defect parameters, images, trend analysis, and life prediction; associating the report with the original data and packaging it into an encrypted file with a unique identifier; and controlling the uploading of the encrypted file to a data server for long-term archiving and storage.

[0046] In some implementations, the method further includes communication and interaction, such as communicating with sensing layer devices and execution layer devices through a standard data interface to receive acquired data and issue control commands; and displaying processing results, ultrasound images and system status in real time through a human-machine interface.

[0047] This application embodiment effectively eliminates the interference of different material properties and environmental background noise on the detection signal by establishing an acoustic benchmark in a defect-free calibration area in advance, providing a unified and reliable comparison standard for subsequent full-area scanning, thereby ensuring the accuracy and consistency of the detection results from the source. By driving the execution device to perform automated axial scanning based on benchmark parameters, full-coverage data acquisition of water-cooled wall pipes is achieved, avoiding the blind spots and missed detection risks inherent in traditional manual sampling methods. At the same time, the real-time upload mechanism ensures strict synchronization and integrity of massive detection data and spatial coordinates. Through the embedded real-time processing unit to perform full-focus imaging and intelligent analysis of ultrasonic signals, abnormal signals can be quickly identified online and decision-making processes can be triggered, significantly improving defect identification efficiency and the degree of detection automation, and reducing reliance on the subjective experience of operators. Finally, through the automatic generation of structured reports and data archiving functions, a standardized and traceable detection closed loop is formed, which not only greatly improves the efficiency of report generation, but also provides a solid data foundation for long-term status monitoring, trend analysis, and predictive maintenance of equipment. In summary, the embodiments of this application systematically solve the problems of poor reliability, low efficiency, reliance on manual labor, and difficulty in traceability in traditional detection by constructing a complete closed loop of benchmark establishment, automatic scanning, real-time analysis, and intelligent archiving. It realizes the transformation of boiler water-cooled wall corrosion detection from discrete and qualitative to systematic, intelligent, and quantifiable, and ultimately provides core technical support for the safe operation and predictive maintenance of equipment.

[0048] See attached document Figure 2As shown in the embodiment, the corrosion detection method for the inner wall of the boiler water-cooled wall disclosed in this application operates based on a sensing layer, an execution layer, and an analysis layer architecture. It should be noted that data from the sensing layer is uploaded to the analysis layer, processed by the analysis layer, and then sends control commands to the execution layer to form a closed-loop detection. The sensing layer collects pipe wall status data through an ultrasonic sensor; the execution layer receives commands and executes scanning actions through a motion mechanism; the analysis layer performs real-time signal processing through an imager, identifies and assesses defects through industrial machine-integrated intelligence, and interacts with the power plant management system through a data interface. In this embodiment, the specific steps include S1-S8: S1. By setting detection parameters through the analysis layer, the control execution layer positions the sensing device in a calibration area that has been manually verified to be defect-free, and collects background noise data as well as the average sound velocity and attenuation coefficient of that area.

[0049] In some implementations, the execution layer also includes a servo motor drive controller, a constant force pressure regulator, and an emergency brake; the servo motor drive controller is installed inside the magnetic adsorption crawling drive device to receive speed commands and control the movement of the track; the constant force pressure regulator is installed between the sensing device and the drive device to maintain a constant pressure of 400N±5N between the probe and the pipe wall; the emergency brake is integrated at both ends of the drive device to immediately lock the mechanism in case of demagnetization or overspeed.

[0050] In some implementations, the analysis layer also includes an explosion-proof industrial tablet PC, an ultrasonic imaging processing module, and a data analysis server; the explosion-proof industrial tablet PC is used for human-computer interaction and real-time image display; the ultrasonic imaging processing module is used for phased array acoustic beam synthesis and image generation; and the data analysis server is deployed for defect identification and classification.

[0051] In some implementations, step S1 includes the following sub-steps S11-S14: S11. Set the detection parameters using the explosion-proof industrial tablet PC: excitation voltage is 80V, pulse repetition frequency is set to 2kHz, and acquisition depth is 15mm.

[0052] S12. The analysis layer controls the execution layer to position the sensor in the calibration area and control the servo motor to move 100mm at a speed of 5mm / s.

[0053] S13. During this process, the sensing layer acquires the raw ultrasonic signals of all array elements, and the imaging processing module generates an average sound velocity of 5900 m / s and an attenuation coefficient of 0.8 dB / mm in this region.

[0054] S14. Store these parameters as baseline values ​​in the database of the data analysis server, and establish a dedicated testing template for the water-cooled wall tube.

[0055] S2. After the baseline is established, the analysis layer sends a full scan command to the execution layer. The execution layer drives the sensing device to move along the pipeline axis, continuously collecting raw ultrasonic signals and uploading them in real time.

[0056] In some implementations, step S2 includes the following sub-steps S21-S24: S21. After the baseline is established, the full scan program is started through the explosion-proof industrial panel PC. The imaging processing module sends a uniform motion command of 15mm / s to the servo motor drive controller.

[0057] S22. The encoder positioning module starts recording the movement distance, triggering data acquisition every 0.5mm, and the sensing layer acquires ultrasonic signals from all 64 channels.

[0058] S23. The acoustic coupling monitoring module monitors the first wave amplitude of each channel in real time. When the amplitude fluctuation exceeds the set threshold, an abnormal coupling status flag is generated.

[0059] S24. All acquired data is transmitted in real time to the ultrasound imaging processing module for buffering and storage via armored network cable.

[0060] S3. The analysis layer performs real-time full-focus processing on the uploaded signal.

[0061] In some implementations, step S3 includes the following sub-steps S31-S34: S31. The imaging processing module performs real-time full-focus processing on the transmitted data to generate real-time B-scan and C-scan images.

[0062] S32. When image processing detects an abnormal signal at a certain location, and the signal-to-noise ratio comparison benchmark value of the signal drops by 6dB, it is marked as a suspicious point.

[0063] S33. The acoustic coupling monitoring module synchronously verifies the coupling status at this point. If the coupling status is good, it is confirmed as a real defect signal.

[0064] S34. The analysis layer generates a deceleration re-inspection command containing precise coordinates, which interrupts the current scan via the servo motor drive controller.

[0065] S4. After receiving the instruction, the execution layer immediately performs a reciprocating scan at the suspicious point location, and the sensing layer starts the high-frequency sampling mode.

[0066] In some embodiments, the sensing layer further includes a 64-element high-frequency linear ultrasonic phased array probe, an acoustic coupling monitoring module, and an encoder positioning module. The 64-element high-frequency linear ultrasonic phased array probe is installed inside the high-temperature resistant ultrasonic phased array sensing device, with a center frequency of 10MHz and an adjustable pulse repetition frequency range of 100Hz-5kHz. The acoustic coupling monitoring module is integrated inside the probe and determines the coupling state in real time by monitoring the amplitude of the echo signal. The encoder positioning module is installed on the drive wheel shaft of the magnetic adsorption crawling drive device and is used to accurately record the scanning position with a positioning accuracy better than 0.1mm.

[0067] In some implementations, step S4 includes the following sub-steps S41-S44: S41. After receiving the instruction, the servo motor drive controller immediately reduces the scanning speed to 5mm / s.

[0068] S42. Control the sensor to retract to a position 50mm upstream of the suspicious point and rescan the suspicious point area at a speed of 5mm / s.

[0069] S43. The sensing layer starts the high-frequency acquisition mode and performs 128 signal acquisitions at this location.

[0070] S44. The imaging processing module averages the 128 acquired signals, significantly improving the signal-to-noise ratio.

[0071] S5. The analysis layer processes the re-collected data and uses a deep learning model for defect classification.

[0072] In some implementations, step S5 includes the following sub-steps S51-S54: S51. Defect identification is performed on high-quality data that has undergone averaging.

[0073] S52. If the defect is identified as a corrosion defect, its depth shall be accurately measured by time-difference diffraction, with a measurement accuracy of 0.1 mm.

[0074] S53. If the defect is identified as a crack, its length and height shall be measured using the end diffraction wave method.

[0075] S54. All defect parameters and images are saved to the database and correlated with historical inspection records.

[0076] S6. After the defect is confirmed, the analysis layer calls the historical database to perform a year-on-year analysis.

[0077] In some implementations, step S6 includes the following sub-steps S61-S64: S61. The data analysis server retrieves the three most recent detection records for this pipe segment.

[0078] S62. Calculate the trend of the number of defects per unit area. If the defect density growth rate exceeds 200%.

[0079] S63. Generate a high-risk early warning report and send it to the power plant management system via the data interface.

[0080] S64. Mark the pipe section with a flashing red icon on the explosion-proof industrial panel PC to indicate the need for enhanced monitoring.

[0081] S7. The analysis layer uses the Paris formula for corrosion propagation rate based on defect data and development trends, and calculates the remaining life in combination with material performance parameters.

[0082] In some implementations, step S7 includes the following sub-steps S71-S74: S71. Based on defect size data and development trends, the corrosion propagation rate is obtained using the Paris formula.

[0083] S72. Based on the material performance parameters, generate the remaining life of the pipe section under different operating conditions.

[0084] S73. Generate a complete assessment report including corrosion patterns, expansion curves, and lifetime predictions.

[0085] S74. The report is uploaded to the power plant management system in a standard data format, and the recommended maintenance cycle is provided.

[0086] S8. After the detection is completed, the analysis layer controls the execution layer to return to the initial position and release the magnetic adsorption. All detection data is packaged, stored, and a uniquely identified report file is generated.

[0087] In some implementations, step S7 includes the following sub-steps S81-S84: S81. After the test is completed, the explosion-proof industrial panel PC issues a return command.

[0088] S82. The servo motor drive controller controls the drive device to return to the starting point at a speed of 20mm / s.

[0089] S83. After reaching the starting point, the electromagnetic adsorption system is powered off to release the magnetic force.

[0090] S84. All test data are packaged into an encrypted report file and uploaded to the server for long-term storage.

[0091] This application embodiment sets detection parameters through an analysis layer and controls the execution layer to position the sensing device in a manually verified defect-free calibration area, collecting background noise data and the average sound velocity and attenuation coefficient of that area. This application embodiment uses a magnetic adsorption crawling robot for automated scanning, achieving comprehensive coverage detection of water-cooled wall tube banks, greatly improving detection efficiency. By fusing high-frequency ultrasonic phased array technology with a full-focusing algorithm, it achieves accurate identification and quantitative assessment of internal wall corrosion and microcracks. Through a deep learning model for intelligent defect classification and combined with historical data trend analysis, it enables risk warning and lifespan prediction, adapting to complex boiler environments and ensuring operational safety. Ultimately, it solves the problems of traditional detection relying on shutdown, sampling inspection, subjective judgment, and lack of quantification.

[0092] This application also discloses a specific embodiment of a corrosion detection method for the inner wall of a boiler water-cooled wall, including the following steps: First, the analysis layer sets the detection parameters, and the execution layer controls the sensor to be positioned in a manually verified, defect-free calibration area. It moves 100mm at a speed of 5mm / s, collects background noise data, and calculates the average sound velocity and attenuation coefficient of the area to establish a reference value. The calculation uses the time-of-flight method formula: ; In the formula, The nominal thickness of the pipe wall. This represents the average round-trip time of the ultrasonic wave. Attenuation coefficient The following is obtained by calculating the logarithmic attenuation rate of the signal amplitude: ; In the formula, The initial amplitude, This is the amplitude of a bottom wave; This step aims to eliminate the influence of material background noise and sound velocity differences on the detection.

[0093] Automatic scanning and data acquisition: After the baseline is established, the analysis layer sends a fully automatic scanning command to the execution layer. The execution layer drives the sensing device to move along the pipeline axis at a scanning speed of 15 mm / s. The sensing layer continuously collects the original ultrasonic signals and uploads them in real time.

[0094] The encoder provides precise position information, enabling a strict correspondence between the ultrasonic scanning signal and spatial coordinates, thus laying the foundation for imaging.

[0095] Real-time signal processing and suspicious point identification: The analysis layer processes the uploaded signal using a real-time total focusing algorithm (TFM). When a single array channel signal-to-noise ratio (SNR) is detected... When the signal drops by more than 6dB and the signals of adjacent channels are normal, mark the point as a point of coupling abnormality or suspected defect and generate a speed-down re-inspection command. Total Focusing Method (TFM): The core formula for sound beam synthesis at each imaging point is: ; In the formula, It is the imaging point The strength, It is a launch array element Receiver array element The signal Sound waves from Time Back to The duration of transmission; Signal-to-noise ratio calculate: ; In the formula, The amplitude of the defect signal. This represents the background noise amplitude.

[0096] A 6dB drop is equivalent to the signal amplitude decreasing by half.

[0097] Refined re-inspection and data optimization: After receiving the command, the execution layer immediately reduces the scanning speed to 5mm / s and performs reciprocating scanning at the suspicious point location. The sensing layer starts a high-frequency sampling mode, collecting 128 signals and averaging them to improve the signal-to-noise ratio. The signal averaging method is achieved by superimposing multiple acquisitions, and its signal-to-noise ratio improvement factor is N, where N is the number of averaging (in this example, N=128). (An increase of approximately 11 times).

[0098] This step effectively suppresses random noise and highlights the true defect signals.

[0099] Defect Confirmation and Quantification: The analysis layer processes the re-collected data using advanced algorithms and employs a deep learning model for defect classification. If corrosion pits deeper than 0.5 mm or microcracks longer than 2 mm are confirmed on the inner wall, their coordinates, dimensions, and burial depth are recorded. Defect depth measurement primarily uses Time-of-Flight Diffraction (TOFD), with the following formula: ; In the formula, For the speed of sound, The time difference of the diffracted wave at the defect end. The length is half the probe spacing. The length is measured using the 6dB drop method. The deep learning model is a classifier based on a convolutional neural network (CNN), trained using a large number of labeled corrosion, crack, and noise samples to achieve automatic defect type identification.

[0100] After defect confirmation, the analysis layer calls the historical database for year-on-year analysis. If the defect density in the area increases by more than 200% within three months, a high-risk warning is generated and the pipe section is marked as a key monitoring target. The defect density growth rate is then used to determine the risk level. The calculation formula is: ; In the formula, This represents the defect density (number of defects / area) for the current inspection. This represents the defect density from the previous inspection. Lifetime Assessment and Report Generation: Based on defect data and development trends, the analysis layer calculates the corrosion propagation rate using the Paris formula and combines it with material performance parameters to calculate the remaining life, generating a complete assessment report. For crack-type defects, the Paris formula is: ; In the formula, This represents the crack propagation rate. denoted as the stress intensity factor amplitude, and C and m as material constants.

[0101] Predict the remaining cycle number or running time through integral calculations.

[0102] System Exit and Data Archiving: After the test is completed, the analysis layer controls the execution layer to return to the initial position and automatically releases the magnetic adsorption. All test data is packaged, stored, and a uniquely identified report file is generated.

[0103] The report contains raw data, processing results, images, and evaluation conclusions. It is stored in an encrypted format and uploaded to the power plant's data center for archiving.

[0104] Through the above steps, this embodiment of the application uses a magnetic adsorption crawling robot for automated scanning, achieving comprehensive coverage inspection of water-cooled wall tube banks, greatly improving inspection efficiency. By integrating high-frequency ultrasonic phased array technology with a full-focusing algorithm, it achieves accurate identification and quantitative assessment of internal wall corrosion and microcracks. Through deep learning models for intelligent defect classification and combined with historical data trend analysis, it achieves risk warning and life prediction, adapts to complex boiler environments and ensures operational safety. Ultimately, it solves the problems of traditional inspection relying on shutdown, sampling inspection, subjective judgment and lack of quantification, realizing the transformation from passive maintenance to proactive prediction.

[0105] This application also discloses a corrosion detection system for the inner wall of a boiler cladding water-cooled wall. (See attached document.) Figure 3 As shown, the system includes: a detection benchmark establishment module 310, an ultrasonic signal acquisition module 320, an analysis result generation module 330, and a detection report generation module 340.

[0106] For example, the detection benchmark establishment module 310 is used to configure detection parameters and control the execution device to position the sensing device to the calibration area, and to receive and process sensing data to establish a detection benchmark.

[0107] For example, the ultrasonic signal acquisition module 320 is used to send a scanning command to the execution device based on the detection benchmark, control the sensing device to move along the pipeline axis, and receive the acquired ultrasonic signals in real time.

[0108] For example, the analysis result generation module 330 is used to process and analyze the received ultrasound signal in real time and generate analysis results.

[0109] For example, the detection report generation module 340 is used to generate a detection report based on the analysis results, control the execution device to return to the initial position, and perform data storage and archiving.

[0110] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of this application, and this application is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this application, and these modifications and improvements are also considered to be within the scope of protection of this application.

Claims

1. A method for detecting corrosion on the inner wall of a boiler's water-cooled wall, characterized in that, include: Configure detection parameters and control the execution device to position the sensing device to the calibration area, receive and process sensing data to establish a detection benchmark; Based on the detection benchmark, a scanning command is sent to the execution device to control the sensing device to move along the pipeline axis and receive the collected ultrasonic signals in real time. The received ultrasound signals are processed and analyzed in real time, and analysis results are generated. Based on the analysis results, a detection report is generated, and the execution device is controlled to return to its initial position, and data is stored and archived.

2. The method according to claim 1, characterized in that, The real-time processing and analysis of the received ultrasound signals includes: When the ultrasound signal is identified as an abnormal signal, a re-inspection command is generated to control the execution device to perform a fine re-scan and obtain re-collected data; The re-collected data is analyzed and processed, and an intelligent recognition algorithm is used to classify and quantify the defects in the re-collected data and generate evaluation parameters. When the evaluation parameters meet the defect threshold, historical records are invoked for trend analysis, and life prediction is performed in conjunction with material parameters. When the ultrasound signal is identified as non-abnormal or the evaluation parameters do not meet the defect threshold, the current operating status is maintained.

3. The method according to claim 2, characterized in that, The configuration of detection parameters and control of the execution device to position the sensing device to the calibration area includes: Configure the detection parameters by receiving user input through the human-computer interaction interface; A first control instruction is generated according to the configuration instruction. The first control instruction is used to control the execution device to drive the sensing device to move to a calibration area that has been manually verified to be defect-free.

4. The method according to claim 3, characterized in that, The process of receiving and processing sensor data to establish a detection benchmark includes: The sensor receives background noise and ultrasonic signals collected in the calibration area by the sensor device. The ultrasonic signal is processed and the average sound velocity and attenuation coefficient of the calibration area are calculated; The average sound velocity and the attenuation coefficient are stored as reference values, and a special test template for water-cooled wall tubes is generated.

5. The method according to claim 2, characterized in that, The real-time processing and analysis of the received ultrasound signals includes: The real-time uploaded ultrasound signal is processed using a full-focusing algorithm to generate a real-time ultrasound image; The signal-to-noise ratio (SNR) of the ultrasound image is analyzed. When the SNR drops below a set threshold relative to the baseline value, it is determined to be an abnormal signal.

6. The method according to claim 5, characterized in that, The generation of re-inspection instructions controls the execution device to perform a refined re-scan and obtain re-collected data, including: A second control command containing precise coordinates is generated. The second control command is used to control the execution device to reduce the scanning speed to the re-inspection speed, retreat to the upstream of the suspicious point, and rescan. The control sensor device starts a high-frequency sampling mode at the suspicious point location to collect signals multiple times, and the suspicious point location is the location where the abnormal signal is detected.

7. The method according to claim 2, characterized in that, The process of using an intelligent recognition algorithm to classify and quantify defects in the re-collected data and generate evaluation parameters includes: A trained deep learning model is used to identify and classify defects in the re-collected data after analysis and processing. If it is identified as a corrosion defect, its depth is calculated using time-difference diffraction. If the defect is identified as a crack, its length and height are calculated using the end diffraction wave method. Otherwise, the re-collected data is marked as defect-free or unidentifiable, and a judgment result is generated.

8. The method according to claim 2, characterized in that, The step of calling historical records for trend analysis when the evaluation parameters meet the defect threshold includes: Retrieve historical inspection records for the current pipe section from the historical database; Calculate the trend of change in the number or density of defects per unit area; When the trend of change exceeds a set threshold, a high-risk warning message is generated; When the change trend does not exceed the set threshold, a marker information indicating that the pipe section is in a stable state is generated.

9. The method according to claim 2, characterized in that, The life prediction based on material parameters includes: Based on the current size and historical development trend of the defect, the corrosion or crack propagation rate is calculated using the Paris formula. The remaining life prediction result of the current pipe section is generated by querying the material performance parameter database and combining it with the expansion rate.

10. A corrosion detection system for the inner wall of a boiler's water-cooled wall, characterized in that, include: The detection benchmark establishment module is used to configure detection parameters and control the execution device to position the sensing device to the calibration area, and to receive and process sensing data to establish a detection benchmark. An ultrasonic signal acquisition module is used to send scanning commands to the execution device based on the detection benchmark, control the sensing device to move along the pipeline axis, and receive the acquired ultrasonic signals in real time. The analysis result generation module is used to process and analyze the received ultrasound signals in real time and generate analysis results. The test report generation module is used to generate a test report based on the analysis results, control the execution device to return to the initial position, and perform data storage and archiving.

Images