Leakage detection method, device and electronic equipment for tubular air preheater

Abstract

This specification discloses a leakage detection method, apparatus, and electronic equipment for a tubular air preheater. The method includes acquiring infrared thermal imaging images of each pipe of the tubular air preheater in response to pressure information indicating a negative pressure in the flue gas region of the tubular air preheater. The method further includes generating a first command in response to the presence of a temperature anomaly region in the infrared thermal imaging image corresponding to a target pipe, and receiving single-end sealing information indicating one end of the target pipe is sealed. Furthermore, the method includes generating a first detection result for the target pipe based on the outline shape of the flame in a flame image. In this embodiment, leaking pipes can be progressively screened without manual assistance or with only minimal manual assistance, efficiently and accurately identifying leaking pipes without consuming excessive manpower, and promptly addressing the leaks to ensure that boiler combustion efficiency meets expectations.

Description

Technical Field

[0001] The embodiments in this specification pertain to the field of tubular air preheater detection, and specifically relate to a method, apparatus, and electronic equipment for detecting leaks in tubular air preheaters. Background Technology

[0002] Tubular air preheaters are heat exchange devices that use the waste heat of flue gas outside the pipes to heat the air inside the pipes used for combustion. They are simple in structure, have high heat transfer efficiency, and good sealing, and are widely used in power plant boilers, especially circulating fluidized bed boilers. However, if a pipe ruptures and leaks, air inside the pipe will flow out under negative pressure, further reducing the oxygen required for combustion, worsening the ignition conditions and process, and affecting combustion efficiency. This also increases auxiliary energy consumption. Therefore, leak detection and sealing of tubular air preheaters are necessary to ensure the unit's load-bearing capacity. Currently, existing detection methods generally involve manual inspection by listening and observation. However, since tubular air preheaters contain hundreds, or even thousands, of pipes, this method is not only inaccurate but also inefficient. Summary of the Invention

[0003] Embodiments of this disclosure provide a leakage detection method, apparatus, and electronic device for a tubular air preheater, aimed at solving one or more of the above-mentioned problems and other potential problems.

[0004] According to a first aspect of this disclosure, a leakage detection method for a tubular air preheater is provided. The method includes acquiring infrared thermal imaging images of each pipe of the tubular air preheater in response to pressure information indicating a negative pressure has been generated in the flue gas region of the tubular air preheater. The method further includes generating a first command in response to the presence of a temperature anomaly region in the infrared thermal imaging image corresponding to a target pipe, and receiving single-end sealing information indicating a blockage at one end of the target pipe, to control the detection target to generate a flame at the other end of the target pipe and acquire a flame image. Furthermore, the method includes generating a first detection result for the target pipe based on the contour shape of the flame in the flame image.

[0005] According to a second aspect of this disclosure, a leakage detection device for a tubular air preheater is provided. The device includes an infrared image acquisition module configured to acquire infrared thermal imaging images of each pipe of the tubular air preheater in response to pressure information indicating a negative pressure has been generated in the flue gas region of the tubular air preheater. The device also includes a flame image acquisition module configured to generate a first command in response to the presence of a temperature anomaly region in the infrared thermal imaging image corresponding to a target pipe, and upon receiving single-end sealing information indicating a blockage at one end of the target pipe, to control the detection target to generate a flame at the other end of the target pipe and acquire a flame image. Furthermore, the device includes a contour detection module configured to generate a first detection result for the target pipe based on the contour shape of the flame in the flame image.

[0006] According to a third aspect of this disclosure, an electronic device is provided, including one or more processors and a memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform a method provided according to a first scheme.

[0007] According to a fourth aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the method provided according to the first aspect.

[0008] It should be understood that the description in the Summary of the Invention section is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0009] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0010] Figure 1 A schematic diagram of an example environment in which several embodiments of the present disclosure may be implemented is shown;

[0011] Figure 2 A schematic flowchart of a leakage detection method for a tubular air preheater according to some embodiments of the present disclosure is shown;

[0012] Figure 3 A flowchart illustrating the overall process of leak detection for a tubular air preheater according to some embodiments of this disclosure is shown.

[0013] Figure 4 A flowchart illustrating the training process of a contour recognition model according to some embodiments of the present disclosure is shown.

[0014] Figure 5A schematic diagram of the structure of a leakage detection device for a tubular air preheater according to some embodiments of the present disclosure is shown;

[0015] Figure 6 A schematic block diagram of an electronic device according to some embodiments of the present disclosure is shown. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0017] The terms “comprising” and “having”, and any variations thereof, in this specification, claims, and the foregoing drawings are intended to cover a non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. Depending on the context, the word “if” as it applies herein may be interpreted as “when”, “in response to determination”, or “in response to detection”.

[0018] As mentioned earlier, to ensure that the tubular air preheater generates sufficient heated air for combustion during normal operation, regular inspections for damage and leaks are necessary. Because the tubular air preheater has numerous, long, and closely spaced pipes for air supply, traditional manual inspection methods involving listening and visual observation are not only labor-intensive and inefficient, but also prone to false positives and missed detections due to the pipe structure. This means that even with regular manual inspections, leaks can still prevent the timely and sufficient supply of air to the boiler, resulting in a discrepancy between the estimated and actual air volume. This affects the boiler load, leading to lower-than-expected combustion efficiency, requiring longer combustion times, and wasting electricity.

[0019] To address this issue, embodiments of this disclosure propose a leakage detection scheme for tubular air preheaters. In these embodiments, a negative pressure is first created in the flue gas area of ​​the tubular air preheater using an induced draft fan, and infrared thermal imaging images of each pipe under this negative pressure state are acquired. Pipes exhibiting abnormal temperature areas in the infrared thermal imaging images are considered potential leaking pipes, and one end of these pipes is sealed for further detection. By generating a flame on the unsealed side of the target pipe and acquiring a flame image, the leakage status of the target pipe can be determined based on the different contour shapes of the flame, thereby generating a corresponding first detection result.

[0020] Using the above method, it is possible to quickly identify potential leaks in various pipelines by analyzing temperature anomalies in infrared thermal imaging images, with little or no manual assistance. Then, targeted single-sided sealing of these pipes allows for further analysis of the flame image's outline to determine if a leak has occurred, generating an initial detection result. This method efficiently and accurately identifies leaking pipelines without requiring excessive manpower, enabling timely intervention and ensuring that the boiler's combustion efficiency meets expectations during normal operation of the tubular air preheater.

[0021] Figure 1 A schematic diagram of an example environment 100 in which various embodiments of this disclosure may be implemented is shown. For example... Figure 1As shown, environment 100 may include terminal 110, detection object 120, and tubular air preheater 130. Terminal 110 may be any device with computing or processing capabilities. For example, terminal 110 may include, but is not limited to, mobile phones, tablets, desktop computers, servers, etc. Detection object 120 may include, but is not limited to, robotic arms, intelligent robots, etc. Detection object 120 may be equipped with at least an ignition unit 121 for generating a flame, an infrared unit 122 for acquiring infrared thermal imaging images, and an acquisition unit 123 for acquiring flame images. One detection object 120 may be set, or multiple detection objects 120 may be set. The above-mentioned units may be integrated on one detection object 120 or set separately on different detection objects 120. After the induced draft fan set in the flue gas area of ​​tubular air preheater 130 is started, a negative pressure is generated in the flue gas area, causing terminal 110 to receive negative pressure information 111. Next, each pipe 112 in the tubular air preheater 130 can be identified through image recognition, and each pipe 112 can be numbered sequentially, or the coordinates of each pipe 112 can be determined. Depending on the detection object 120 and the infrared unit 122, the infrared unit 122 can be controlled to simultaneously emit infrared signals to one or more pipes 112 and acquire infrared thermal imaging images 113 of the pipes 112. After identifying the target pipe 114 based on the temperature anomaly area, one end of the target pipe 114 can be temporarily blocked using a deformable flexible plug, either by a robotic arm or manually. After the blocking is completed, single-end blocking information 115 is fed back to the terminal 110 (single-end blocking information 115 may include the number / coordinate area of ​​the pipe that has been blocked, and text information of the local coordinate area corresponding to the blocked end; it can be generated manually by operating on a designated app page on a mobile terminal, or by a robotic arm performing a single-end blocking operation at a designated coordinate position (i.e., by image recognition). After inserting a flexible plug into a designated coordinate position, the controller of the robotic arm generates a first instruction 116, which in turn controls the detection object 120 to generate flames at the unsealed ports of each target pipe 114, resulting in a flame image 117. Based on the outline shape 118 of the flame in the flame image 117, it is determined whether the flame has shifted towards the inside of the pipe under the influence of the airflow generated by the negative pressure, or even extinguished. This generates a first detection result 119 to characterize whether the target pipe 114 is leaking, assisting personnel in determining whether the target pipe 114 needs to be treated. This method allows for manual assistance in sealing the target pipes, rather than directly using the detection object, robotic arm, or other equipment for intelligent sealing. This reduces the workload of the auxiliary detection personnel and does not affect the overall detection efficiency, thereby reducing the number of devices required or the functional units integrated into the devices, and ultimately reducing the cost of the detection process.

[0022] Figure 2 A schematic flowchart of a leakage detection method 200 for a tubular air preheater according to some embodiments of this disclosure is shown. Method 200 can be executed, for example, by terminal 110. Figure 2 As shown in block 202, method 200 can acquire infrared thermal imaging images of each pipe of the tubular air preheater in response to pressure information indicating that a negative pressure has been generated in the flue gas region of the tubular air preheater. In this embodiment, the induced draft fan already installed in the flue gas region of the tubular air preheater can be controlled to operate, causing a negative pressure to be generated in the flue gas region. After the induced draft fan operates, negative pressure information can be fed back to the terminal to indicate that a negative pressure has been generated in the flue gas region. The terminal can scan each pipe separately using the infrared unit integrated on the detection object or a separately installed infrared thermal imager to obtain an infrared thermal imaging image corresponding to each pipe. The infrared thermal imaging image can use different colors to indicate different temperatures, displaying the heat distribution inside the pipe in the image.

[0023] In block 204, method 200 can respond to the presence of a temperature anomaly region in the infrared thermal imaging image corresponding to the target pipe in each of the pipes, and receive single-end sealing information indicating that one end of the target pipe is sealed, to generate a first instruction to control the detection object to generate a flame at the other end of the target pipe and acquire a flame image. In this embodiment, if there is no leak in the pipe, the negative pressure generated by the flue gas area will not act on the pipe, making the temperature distribution of the air in the pipe relatively uniform and within the normal temperature range. However, if there is a leak in the pipe, under the action of negative pressure, the air in the pipe will flow towards the leak location, causing the leak location to experience a local temperature anomaly due to the accumulation of cold air or hot flue gas. By calculating the average temperature in the pipe and comparing the regional temperature of each area with the average temperature, it is possible to determine whether there is a temperature anomaly region in each infrared thermal imaging image whose regional temperature deviates significantly from the average temperature. If a temperature anomaly region is present in an infrared thermal imaging image, the pipe corresponding to that image will be identified as the target pipe. Considering that the heat distribution in the pipe may not be uniform, and considering the fluctuations in the environment and equipment, relying solely on infrared thermal imaging images cannot accurately determine whether the pipe is actually leaking, and there is a possibility of false detection. Therefore, after identifying the target pipeline, the terminal can generate an instruction to direct the operator or the object being tested to seal one end of the target pipeline using a flexible plug. (If the testing unit is shut down, the end of the pipeline furthest from the object being tested should generally be sealed. If the unit is running, the side closest to the fan should generally be sealed to prevent the positive pressure generated by the fan from affecting the pipeline being tested, ensuring that the air inside the pipeline is relatively still under normal conditions.) After sealing, the operator can send a message to the terminal via mobile phone, or the object being tested can transmit information via a communication connection. The terminal will then generate a first instruction to generate a flame at the unsealed end of the target pipeline and acquire a flame image. The reason for first screening the target pipeline using infrared thermal imaging instead of directly acquiring a flame image is that tubular air preheaters have many pipelines. If a flame image were directly acquired for each pipeline, each pipeline would need to be sealed at one end. Combined with the time required for flame generation and image acquisition, this would result in very low overall testing efficiency. Therefore, by initially screening pipelines based on infrared thermal imaging images, and then acquiring and judging flame images of the selected target pipelines, the detection efficiency can be improved, and the time required to restore the tubular air preheater after the detection is completed can be reduced.

[0024] In box 206, method 200 can generate a first detection result for the target pipe based on the contour shape of the flame in the flame image. In this embodiment, the contour shape of the flame in the flame image can be determined by a pre-trained contour recognition model, or by determining the orange-red region representing the flame in the flame image through image recognition and determining the shape of the orange-red region. When there is no leak, there is no airflow in the pipe due to negative pressure, and therefore no airflow interferes with the flame, so the contour shape of the flame will not deviate significantly. Therefore, the state of the flame (i.e., whether the flame is interfered with by airflow) can be determined by the contour shape, thereby determining whether the target pipe is leaking and generating the corresponding first detection result. As an example, corresponding result information can be pre-set in the database for different contour shapes, and the result information is used to indicate whether there is a leak or not. After determining the actual contour shape of the flame, the leakage status of the target pipe can be determined according to the result information corresponding to the contour shape that best matches the actual contour shape in the database, thereby generating the first detection result.

[0025] In this way, potential leaking pipelines can be quickly screened based on infrared thermal imaging images. Then, the specific leak status of the target pipeline can be determined by analyzing the flame outline in the flame image. This allows for intelligent detection of pipeline leaks without requiring or consuming excessive manpower. Compared to purely manual detection, this method is more efficient and provides more accurate results, ensuring that all pipeline leaks in the tubular air preheater are addressed promptly, and guaranteeing that the boiler's combustion efficiency meets expectations during normal operation.

[0026] Figure 3A flowchart illustrating the overall process 300 for leak detection of a tubular air preheater according to some embodiments of this disclosure is shown. In process 300, after detecting negative pressure information 111, each pipe of the tubular air preheater can be scanned sequentially using a detection object or an infrared imager to obtain an infrared thermal imaging image 113 corresponding to each pipe. For any obtained infrared thermal imaging image 113, the image generally uses different colors (i.e., different pixel values) to represent different temperatures, and the temperature represented by a specific color can be directly obtained according to a pre-set pixel value-temperature mapping curve. Therefore, in block 321, based on the infrared thermal imaging image 113, the temperature at each location within the pipe can be determined, and then the average temperature within the pipe can be calculated. For a target area where the difference between the area temperature and the average temperature is greater than a difference threshold, the temperature of the target area is considered to be significantly abnormal relative to other areas of the pipe. Considering that the pipe will not be in complete thermal equilibrium, the temperature distribution within the pipe will exhibit certain fluctuations, which does not mean that the existence of a target area necessarily indicates a leak. The target area formed temporarily due to fluctuations is generally small, while the target area caused by leakage generally has a stable area of ​​considerable size. Therefore, by comparing the area of ​​each target area with the preset area, the interference caused by fluctuations can be eliminated, the temperature abnormal area 322 in each infrared thermal imaging image can be determined, and the pipe corresponding to the infrared thermal imaging image 113 with the temperature abnormal area 322 can be determined as the target pipe 114.

[0027] As an example, under operating conditions, if a leak occurs, some cold air will enter the flue gas passage directly without preheating, causing a drop in flue gas temperature. To maintain the required combustion temperature, more fuel needs to be consumed, thus increasing heat loss. Simultaneously, due to the mixing of cold air, the heat in the flue gas cannot be effectively transferred to the air entering the boiler, leading to an increase in exhaust gas temperature. Furthermore, the leaked cold air may also lower the air temperature at the air preheater outlet, reducing the temperature of the air entering the boiler and affecting combustion efficiency and the overall performance of the boiler. Therefore, under operating conditions, depending on the pressure difference between the air passage and the flue gas passage, the leaked cold air may accumulate locally through the leak, causing the temperature in the area near the leak to be significantly lower than the surrounding area; alternatively, flue gas may enter the air passage and accumulate in the area near the leak, causing the temperature in that area to be significantly higher than the surrounding area. In the case of unit shutdown, the primary and secondary air fans corresponding to the air duct are already out of service (i.e., the primary and secondary air regulating dampers are closed). A leak in the duct will cause air to flow between the air and flue gas sides, resulting in airflow from the air side to the flue gas side. This airflow will cause the temperature near the leak to differ significantly from the surrounding area. The magnitude of the airflow depends on factors such as the leak area, pressure difference, and fluid properties. For example, a larger leak area and a greater pressure difference will result in a greater airflow velocity and flow rate, accompanied by some noise. In contrast, ducts without leaks remain stationary and do not generate airflow, thus avoiding areas with significant temperature anomalies. Therefore, for both of these operating conditions, areas with abnormal temperatures can be identified.

[0028] After identifying the target pipe 114, a processing instruction is generated for the target pipe 114, instructing the personnel or the object being inspected to seal one end of the target pipe 114 to prevent the small amount of airflow from affecting the flame recognition results. After sealing is completed, single-end sealing information 115 is received. Based on the number / coordinate area in the single-end sealing information 115, the sealed pipe is identified, and a first instruction 116 is generated for that pipe. Under the command control of the first instruction 116, the detection unit moves the end of the ignition unit set to the unsealed end of the pipe, so that the ignition unit generates a flame at that location, and then the flame image 117 is acquired. In the flame image 117, the outline shape 118 of the flame can be extracted, and the leakage of the target pipe 114 can be determined based on the outline shape 118. The extraction of the outline shape 118 can be achieved through a pre-trained outline recognition model, or it can be obtained by directly extracting the edge outline by distinguishing the pixel values ​​of the flame color and the background color, etc. For example, determining whether a leak has occurred could involve pre-setting different profile shapes in a database and pre-labeling each profile shape with different result information based on historical testing experience or human experience. This indicates whether each profile shape represents a leak or not. Since a non-leaking flame is not affected by airflow, it generally does not deviate significantly in any direction. Therefore, profile shapes where the flame deviates, and those with an area smaller than a preset area (corresponding to the flame shrinking or extinguishing under the influence of airflow), represent leaks. The result information for the remaining profile shapes represents no leaks. By determining the result information corresponding to the most matching profile shape in the database, it is possible to determine whether the target pipeline is leaking. One method to determine the best-matching contour shape in the database is to calculate the area overlap rate (the ratio of the intersection area to the union area of ​​the contour shapes), with the contour shape having the highest area overlap rate being the best match. Alternatively, it can be done by calculating the cosine similarity based on the feature vectors of the contour shapes, with the contour shape closest to 1 being the best match. Another method is to calculate the Hu moments of the contour shapes using OpenCV's cv2.HuMoments() function, and the contour shape with the smallest difference in Hu moments being the best match. Another example of determining whether a leak has occurred is to pre-annotate the contour shape samples during the training process of the contour recognition model, using the annotation information to characterize whether the contour shape sample represents a leak. In this way, when the contour shape is obtained from the trained contour recognition model, the corresponding annotation information can be directly added to the contour shape to characterize whether a leak has occurred, and the leak status can be directly read from the annotation information on the contour shape. Based on the judgment result of contour shape 118, a first detection result 343 will be generated, representing the leak detection result of the target pipeline.

[0029] To further improve the accuracy of the detection results, after obtaining the first detection result 343, a second command 344 can be generated to stop the negative pressure in the flue gas area. Next, a pressure sensor is installed at any location within the pipeline, either manually or by the object being tested. To ensure the accuracy of the detection data, this pressure sensor should ideally be located away from any unsealed port. Alternatively, if a pressure sensor is already installed within the pipeline for routine data monitoring, the detection can be performed directly based on that sensor. Then, a booster pump continuously introduces a preset positive pressure into the air area, i.e., the target pipeline, from the unsealed end of the pipeline. With the other end of the pipeline sealed and no leaks, the pressure inside the pipeline should stabilize after a preset time. If a leak exists in the pipeline, the pressure inside will not be well maintained near the preset pressure, and significant pressure changes will still occur after the preset time. Therefore, by comparing the pressure before and after a preset time period, the pressure change ratio 345 of the target pipeline can be determined. Based on the comparison of the pressure change ratio 345 with a ratio threshold, a second detection result 346 is generated to characterize whether the target pipeline is leaking. If the second detection result 346 indicates a leak in the target pipeline, soapy water can be sprayed into the target pipeline manually or by controlling the spraying of the target. Under positive pressure, if a leak exists, the soapy water will accumulate and flow out at the leak point, causing a change in the surface tension of the soapy water, thus forming soap bubbles. If there is no leak, the surface tension of the soapy water will not change significantly, meaning no soap bubbles will form. Therefore, the location of the leak can also be determined by observing the position of soap bubbles in the image inside the pipeline. Finally, the first detection result 343 and the second detection result 346 are combined to determine the final leak detection result 347 of the target pipeline. As an example, if both detection results indicate no leak, the target pipeline is considered to be leak-free; if at least one detection result indicates a leak, the target pipeline is considered to be leaking. To improve the accuracy of the results, if only one detection result indicates a leak, a second round of testing can be performed to obtain new first and second detection results. If at least one of the two detection results still indicates a leak, the target pipeline is considered to be leaking.

[0030] In box 350, the presence or absence of a leak in the target pipe can be determined based on the first detection result 343 or the leak detection result 347. In box 360, if a leak occurs in the target pipe, the leak location 361 can be determined based on the temperature anomaly area in the infrared thermal imaging image corresponding to the target pipe. As an example, the midpoint of the area on the inner wall of the pipe covered by the temperature anomaly area can be identified as the leak location 361. Next, the distance 362 between the leak location 361 and the nearest pipe port will be determined. Furthermore, in box 380, the database can pre-set a standard shape for the flame under conditions without leakage gas flow interference. By calculating the similarity 381 between the contour shape 118 and the standard shape, the actual flame's offset relative to the standard state can be determined. As an example, the similarity 381 can be determined by obtaining the feature vectors corresponding to the contour shape and the standard shape, and then performing Euclidean distance calculation, Pearson correlation coefficient calculation, etc., on the feature vectors. The lower the similarity, the greater the offset, indicating a larger leak gap, and a higher leak level 382 will be assigned.

[0031] In box 370, it will simultaneously determine whether the distance 362 is less than a preset distance and whether the leakage level is lower than a preset level. Considering that the pipeline cannot be removed from the tubular air preheater alone and the pipeline is relatively long, if the distance 362 is less than the preset distance, it is considered that the leakage location 361 is close to the port, making it easy for staff to handle. If the leakage level is also lower than the preset level, it is considered that the leakage gap is small and can be directly repaired. Therefore, a repair reminder message 371 can be generated to remind staff to go to the target pipeline to directly repair the leakage location, so that the repaired target pipeline can continue to be used normally. However, if the distance 362 is not less than the preset distance, it is considered that the leakage location 361 is far from the port. Due to the structural characteristics of the pipeline, staff cannot repair the pipeline. Alternatively, if the leakage level is not lower than the preset level, it indicates that the leakage gap is large and cannot be repaired. In this case, a sealing reminder message 372 will be generated to remind staff to directly seal the target pipeline by plugging both ends with flexible plugs, so that air will not flow into the target pipeline during normal operation. While blocking pipes reduces the amount of air supplied to the boiler, the specific number of blocked pipes can be estimated based on the tubular air preheater's parameters. This allows for adjustments to the combustion rate of the fuel in the boiler, ensuring complete combustion and meeting expected efficiency, thus avoiding unnecessary energy waste. Furthermore, if the number of blocked pipes exceeds a preset threshold, the operator can be alerted to replace the tubular air preheater.

[0032] Upon receiving either repair alert 371 or blocking alert 372, the recipients of the alerts will be adjusted based on the leakage level 382. For example, if the leakage level is below a preset level, the situation is considered relatively minor, and the alert can be sent to the system used by staff, generating a pending task for staff to handle. Conversely, if the leakage level is high, the situation is considered serious and requires immediate attention, and the alert can be sent directly to the pre-stored mobile phone of the relevant administrator, allowing them to be informed and assign appropriate personnel to handle the situation.

[0033] Figure 4 A flowchart illustrating the training process 400 of a contour recognition model according to some embodiments of this disclosure is shown. In process 400, flame feature samples 420-1 are determined by extracting flame features from historical flame images. Flame features may include color features, shape features, and texture features. Color features can be obtained by color space conversion from RGB to HSV or by segmenting the pixel regions of the flame using color thresholding. Shape features can be obtained by identifying flame boundaries using edge detection algorithms or the Canny edge detector. Texture features can be obtained using gray-level co-occurrence matrix, local binary mode, etc. Furthermore, for the historical flame images used for training, contour shapes can be manually labeled on each image beforehand, serving as the contours to be obtained after contour recognition of that image, thus obtaining contour shape samples 420-2. The contour shapes in the contour shape samples need to be manually labeled to distinguish them from the shape features extracted directly by the edge detection algorithm, reducing the decrease in model prediction accuracy caused by errors in the edge detection algorithm. Based on the flame feature samples 420-1 and the contour shape samples 420-2, training samples 420 can be constructed. The initial contour recognition model used for training can be a support vector machine, U-Net, fully convolutional network, or similar model. During the training of the contour recognition model 430 using training samples 420, the model's generator 431 can generate a predicted contour shape 432 based on the flame feature sample 420-1. By comparing the contour shape sample 420-2 with the predicted contour shape 432, a generator loss 433 can be obtained. This generator loss 433 is then used to assess the loss of the predicted contour shape 432, and the loss can be determined using a comparison loss function. The generator loss 433 can be a relatively large value, and it can then be backpropagated to the generator 431 to guide the optimization of the generator 431's parameters, achieving one round of supervised training for the generator 431. This training process can be iterated repeatedly until the generator 431 can generate a more accurate predicted contour shape 432, i.e., until the loss value calculated by the loss function is smaller. After training, the contour recognition model 430 can output a contour shape 118.

[0034] Figure 5 A schematic diagram of the structure of a leakage detection device 500 for a tubular air preheater according to some embodiments of this disclosure is shown. The various embodiments in this specification are described in a progressive manner, and similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions of the method embodiments. Figure 5 As shown, the device 500 includes an infrared image acquisition module 501, configured to acquire infrared thermal imaging images of each pipe of the tubular air preheater in response to pressure information indicating that a negative pressure has been generated in the flue gas region of the tubular air preheater. The device also includes a flame image acquisition module 502, configured to generate a first command in response to the presence of a temperature anomaly region in the infrared thermal imaging image corresponding to a target pipe, and receiving single-end sealing information indicating that one end of the target pipe is blocked, to control the detection object to generate a flame at the other end of the target pipe and acquire a flame image. Furthermore, the device includes a contour detection module 503, configured to generate a first detection result for the target pipe based on the contour shape of the flame in the flame image.

[0035] The infrared image acquisition module 501 is also configured to, for any infrared thermal imaging image, determine a target area where the temperature difference between the area temperature and the average temperature inside the pipe is greater than a threshold; and to determine a target area with an area larger than a preset area as a temperature anomaly area.

[0036] The contour detection module 503 is also configured to generate a second instruction to control the flue gas area to stop generating negative pressure and continuously introduce a preset positive pressure into the target pipe; generate a second detection result of the target pipe based on the pressure change ratio of the target pipe within a preset time period; and determine the leakage detection result of the target pipe based on the first detection result and the second detection result.

[0037] The device 500 also includes a leak location determination module, configured to determine the leak location of the target pipeline based on a temperature anomaly region in response to a first detection result indicating the presence of a leak. The device 500 also includes a first information generation module, configured to generate a repair reminder message for the leak location in response to the distance between the leak location and the nearest port of the target pipeline being less than a preset distance. Furthermore, the device 500 includes a second information generation module, configured to generate a sealing reminder message for the target pipeline in response to the distance between the leak location and the nearest port of the target pipeline being not less than a preset distance.

[0038] The device 500 also includes a leakage level calculation module, configured to determine the leakage level of the target pipeline based on the similarity between its outline shape and a preset standard shape, wherein the leakage level is inversely proportional to the similarity. The device 500 also includes a recipient determination module, configured to determine the recipients of repair or sealing reminder messages based on the leakage level.

[0039] In device 500, the contour shape is obtained based on a contour recognition model. Device 500 also includes a training sample generation module configured to determine training samples based on historical flame images; the training samples include flame feature samples and contour shape samples. Device 500 further includes a sample prediction module configured to generate predicted contour shapes from the initial contour recognition model based on the flame feature samples. Furthermore, device 500 includes a model training module configured to perform at least one round of model training on the initial contour recognition model using the contour shape samples as supervision signals to obtain a contour recognition model.

[0040] The training sample generation module includes a sample acquisition unit, configured to acquire flame feature samples based on historical flame images. It also includes a contour annotation unit, configured to annotate contour shapes in the historical flame images to obtain contour shape samples. Furthermore, the module includes a training sample construction unit, configured to construct training samples based on the flame feature samples and contour shape samples.

[0041] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this specification are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., Digital Versatile Discs (DVDs)), or semiconductor media (e.g., Solid State Disks (SSDs)).

[0042] Figure 6 A block diagram of an electronic device 600 that can implement various embodiments of the present disclosure is shown. For example... Figure 6 As shown, the electronic device 600 includes a processor 610, a disk drive 620, an input / output interface 630, a network interface 640, and a memory 650. The processor 610, disk drive 620, input / output interface 630, network interface 640, and memory 650 can communicate with each other via a communication bus 660.

[0043] The processor 610 can be implemented using a general-purpose CPU, microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits to execute relevant programs and implement the technical solution provided in this application.

[0044] The memory 650 can be implemented in the form of ROM (Read Only Memory), RAM (Read Access Memory), static memory, dynamic storage devices, etc. The memory 650 can store the operating system 651 for controlling the operation of the electronic device 600, and the basic input / output system (BIOS) 652 for controlling the low-level operations of the electronic device 600. Additionally, it can store a web browser 653, a data storage management system 654, etc. In summary, when implementing the technical solution provided in this application through software or firmware, the relevant program code is stored in the memory 650 and is called and executed by the processor 610.

[0045] The input / output interface 630 is used to connect input / output modules to enable information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0046] Network interface 640 is used to connect a communication module (not shown in the figure) to enable communication and interaction between the device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0047] Bus 660 includes a pathway for transmitting information between various components of the device, such as processor 610, disk drive 620, input / input interface 630, network interface 640, and memory 650.

[0048] It should be noted that although the above-described device only shows the processor 610, disk drive 620, input / output interface 630, network interface 640, memory 650, bus 660, etc., in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the method of this application, and does not necessarily include all the components shown in the figures.

[0049] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0050] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. Furthermore, although operations are depicted in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the foregoing discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0051] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A leakage detection method for a tubular air preheater, characterized in that, The method includes: In response to the pressure information used to characterize the negative pressure generated in the flue gas region of the tubular air preheater, infrared thermal imaging images of each pipe of the tubular air preheater are acquired respectively. For any of the infrared thermal imaging images, a target area is identified where the temperature difference between the area temperature and the average temperature inside the pipe is greater than a threshold value; and the target area with an area greater than a preset area is identified as the temperature anomaly area. In response to the presence of a temperature anomaly region in the infrared thermal imaging image corresponding to the target pipe in each of the pipes, and receiving single-end sealing information to characterize the sealing of one end of the target pipe, a first instruction is generated to control the detection object to generate a flame at the other end of the target pipe and to acquire a flame image. Based on the outline shape of the flame in the flame image, a first detection result of the target pipe is generated; A second instruction is generated to control the flue gas area to stop generating negative pressure and to continuously introduce a preset positive pressure into the target pipe; Based on the pressure change ratio of the target pipeline within a preset time period, a second detection result for the target pipeline is generated; and Based on the first detection result and the second detection result, the leakage detection result of the target pipeline is determined; In response to the leakage detection result indicating the presence of a leak, the leakage location of the target pipeline is determined based on the temperature anomaly area, and the leakage level of the target pipeline is determined based on the similarity between the outline shape and a preset standard shape, wherein the leakage level is inversely proportional to the similarity. In response to the fact that the distance between the leak location and the nearest port of the target pipe is less than a preset distance, and the leak level is lower than a preset level, a repair reminder message is generated for the leak location; In response to the fact that the distance between the leak location and the nearest port of the target pipeline is not less than a preset distance, or the leak level is not lower than the preset level, a blocking reminder message for the target pipeline is generated.

2. The leakage detection method for a tubular air preheater according to claim 1, characterized in that, The method further includes: Based on the leakage level, determine the recipient of the repair reminder message or the blocking reminder message.

3. The leakage detection method for a tubular air preheater according to claim 1, characterized in that, The contour shape is obtained based on a contour recognition model; The method further includes: Based on historical flame images, training samples are determined, including flame feature samples and contour shape samples. Based on the flame feature samples, a predicted contour shape is generated by the initial contour recognition model; and Using the contour shape sample as a supervision signal, the initial contour recognition model is trained for at least one round to obtain the contour recognition model.

4. The leakage detection method for a tubular air preheater according to claim 3, characterized in that, The process of determining training samples based on historical flame images includes: Based on historical flame images, obtain flame feature samples; Based on the outline shapes marked on the historical flame images, outline shape samples are obtained; and Training samples are constructed based on the flame feature samples and the contour shape samples.

5. A leakage detection device for a tubular air preheater, characterized in that, The device includes: The infrared image acquisition module is configured to acquire infrared thermal imaging images of each pipe of the tubular air preheater in response to pressure information used to characterize the negative pressure generated in the flue gas region of the tubular air preheater; it is also configured to determine, for any infrared thermal imaging image, a target area where the temperature difference between the area temperature and the average temperature inside the pipe is greater than a difference threshold; and to determine a target area with an area larger than a preset area as a temperature anomaly area. The flame image acquisition module is configured to respond to the presence of a temperature anomaly region in the infrared thermal imaging image corresponding to the target pipe in each of the pipes, and to receive single-end blockage information that characterizes one end of the target pipe being blocked, to generate a first instruction to control the detection object to generate a flame at the other end of the target pipe and to acquire a flame image. The contour detection module is configured to generate a first detection result for the target pipe based on the contour shape of the flame in the flame image; it is also configured to generate a second instruction to control the flue gas area to stop generating negative pressure and continuously introduce a preset positive pressure into the target pipe; it generates a second detection result for the target pipe based on the pressure change ratio of the target pipe within a preset time period; and it determines the leakage detection result of the target pipe based on the first and second detection results. A leak location determination module is configured to determine the leak location of the target pipeline based on the temperature anomaly area in response to the leak detection result indicating the presence of a leak. A leakage level calculation module is configured to determine the leakage level of the target pipeline based on the similarity between the outline shape and a preset standard shape, wherein the leakage level is inversely proportional to the similarity. The first information generation module is configured to generate a repair reminder message for the leak location in response to the fact that the distance between the leak location and the nearest port of the target pipeline is less than a preset distance and the leak level is lower than a preset level. The second information generation module is configured to generate a blocking reminder message for the target pipeline in response to the fact that the distance between the leak location and the nearest port of the target pipeline is not less than a preset distance, or the leak level is not lower than the preset level.

6. An electronic device, comprising: One or more processors, and A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of a leak detection method for a tubular air preheater according to any one of claims 1-4.

7. A computer program product, comprising a computer program that, when executed by a processor, implements a leakage detection method for a tubular air preheater according to any one of claims 1-4.

Images