Safety early warning method of waste heat recovery system based on infrared thermogram

Abstract

The present invention relates to the field of image data processing technology, and specifically to a waste heat recovery system safety warning method based on infrared thermograms, comprising: obtaining historical images and current images of different areas of the waste heat recovery system; calculating the danger threshold of each area; dividing any image in the current image into multiple target images, obtaining the hotspot in any image, the hotspot being the pixel with the largest sum of the absolute value of the temperature difference between any pixel point and the remaining pixel points in any target image; calculating the complexity of each hotspot, the complexity representing the abnormality of the hotspot; if the complexity of the hotspot is greater than or equal to the danger threshold of the area, the area is determined to be abnormal and a warning is issued. The present invention solves the problem of low accuracy of safety warnings for waste heat recovery systems.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and more specifically, to a safety early warning method for a waste heat recovery system based on infrared thermal images. Background Art

[0002] With the continuous development of industrial automation and energy-saving and emission reduction technologies, waste heat recovery systems, which consist of heat exchangers, heat exchange media, steam generators, hot air blowers, etc., have been applied in many industries, especially in high-temperature environments, such as metallurgy, chemical industry, steel and other fields. These systems reduce energy waste and improve energy efficiency by recycling waste heat. However, during long-term operation, the waste heat recovery system may cause equipment abnormalities or even fires and other safety accidents due to changes in environmental factors such as temperature and humidity.

[0003] With the development of infrared thermal imaging technology, monitoring methods based on infrared thermograms have gradually become a new type of safety warning method. Infrared thermograms can accurately reflect the surface temperature distribution of objects, monitor the temperature changes of equipment in real time, and provide early warnings when the temperature is abnormal. Especially in complex industrial environments, the use of infrared thermograms to monitor waste heat recovery systems can not only achieve real-time detection of the operating status of equipment, but also quickly identify potential risk points when the temperature is abnormal. However, when performing infrared thermogram analysis, the existing technology mainly relies on the detection of a single temperature change rate and hot spot areas, and fails to fully consider the dynamic changes of the system and the impact of various environmental factors on safety warnings.

[0004] The patent application document with application publication number CN119090869A discloses an ink tank safety warning method based on infrared thermal images. The patent application document calculates the probability of a local hot spot area causing a fire based on the temperature anomaly of the local hot spot area and the temperature change rate of the solvent accumulation area adjacent to the local hot spot area, and then obtains the fire risk level of the ink tank, so as to facilitate safety warning.

[0005] However, although the above technical solution can achieve early identification and response to potential fire risks to a certain extent, it only makes early warning judgments based on the temperature change rate, and fails to fully consider the applicability and variability in complex environments. The single temperature change rate indicator is not sufficient to fully adapt to various environmental conditions, which leads to the problem of low early warning accuracy. Summary of the invention

[0006] In order to solve the problems of low accuracy and insufficient reliability of detection results raised in the above background technology, the present invention provides the following solutions.

[0007] The present invention provides a waste heat recovery system safety early warning method based on infrared thermogram, comprising: obtaining historical images and current images of different areas of the waste heat recovery system; calculating the first Failure probability in a region , , For the The historical images of the area The entropy value of the infrared heat map of the image, For the The mean entropy value of the infrared thermal map of all images in the historical images of the region, For the The total number of historical images of the region; calculate the Danger threshold for each region , , where To preset the danger threshold, The natural constant An exponential function with as the base; divide any image in the current image into multiple target images, obtain the hot spots in any image, and the hot spots are the pixels with the largest sum of the absolute values ​​of the temperature differences between any pixel point in any target image and the remaining pixels; calculate the The complexity of a hotspot is determined, and the complexity represents the abnormality of the hotspot; if the complexity of the hotspot is greater than or equal to the danger threshold of the area where it is located, the area where it is located is determined to be abnormal and an early warning is issued.

[0008] The above technical solution provides a high-precision and reliable safety warning method, which can not only detect local abnormalities in a timely manner, but also flexibly respond to changes under different working conditions, greatly improving the accuracy and response speed of fault detection, reducing the risk of system downtime and safety accidents, and enhancing overall safety and reliability.

[0009] Furthermore, the complexity is, , where For the The complexity of each hotspot, Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. is the total number of hotspots, is the total number of current images, Hotspot The standard deviation of the magnitude of color gradients in different images relative to the current image.

[0010] The above technical solution can accurately evaluate the detailed features and spatial complexity of each hotspot by considering color gradients and pixel differences. This method can flexibly respond to hotspot changes in different scenarios, improve the ability to identify abnormal areas, enhance the accuracy and robustness of image processing, and ensure effective capture and response to potential problems in complex environments.

[0011] Furthermore, the complexity is, , where For the The complexity of each hotspot, Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. is the total number of hotspots, is the total number of current images, Hotspot The standard deviation of the magnitude of color gradients in different images relative to the current image.

[0012] The above technical solution introduces an improved complexity calculation method, so that the analysis of hot spots not only takes into account the mean difference of color gradients, but also integrates the relative standard deviation between hot spots, and weights the complexity through a nonlinear function. This method can more accurately capture the subtle changes of hot spots in different images, especially when processing images with significant spatial or temperature differences, and can more sensitively reflect the complex characteristics of hot spots. Through this comprehensive analysis, the system can more efficiently identify abnormal hot spots, especially in complex backgrounds or with more interference, and improve the accuracy and robustness of hot spot detection. At the same time, nonlinear weighted processing makes the complexity assessment more in line with the needs of practical applications, and improves the reliability and accuracy of fault prediction and alarm.

[0013] Furthermore, the waste heat recovery system is composed of a heat exchanger, a heat exchange medium, a steam generator and a hot air blower.

[0014] Furthermore, the historical images and the current image are acquired using a CMOS camera or a CCD camera.

[0015] Furthermore, it also includes performing denoising processing on the historical images and the current image.

[0016] The above technical solution helps to significantly improve the image quality and analysis accuracy by denoising each historical image and the current image. By removing the noise in the image, the system can more clearly capture the real thermal information in the image, reduce misjudgment caused by noise interference, and ensure the accuracy of subsequent image analysis, feature extraction and anomaly detection. Furthermore, the early warning includes triggering an alarm signal through an audible and visual alarm.

[0017] The above technical solution triggers the alarm signal through the sound and light alarm, realizing the real-time early warning of abnormal situations. Through the combination of sound and light signals, it can quickly attract the attention of personnel, improve the reaction speed, and deal with potential safety risks in time. This early warning mechanism enhances the reliability and emergency response capabilities of the system, helps to ensure the effectiveness of the security monitoring system at critical moments, and reduces possible losses.

[0018] Furthermore, the preset danger threshold is 0.6.

[0019] Furthermore, dividing the image into a plurality of target images includes equally dividing the image to obtain a plurality of images of the same size.

[0020] Furthermore, dividing the image into a plurality of target images includes dividing the image into a plurality of images of any size.

[0021] The beneficial effects of the present invention are: The present invention provides a high-precision and reliable safety early warning method for waste heat recovery systems by combining multi-dimensional analysis of historical images and current images. By calculating the probability of failure, danger threshold, and hotspot complexity, potential abnormal areas in the system can be accurately identified, thereby achieving early warning of potential safety hazards. In addition, denoising processing is used to improve image quality and ensure the accuracy of image analysis, and by triggering an alarm signal through an audible and visual alarm, operators can be reminded in time at critical moments to effectively avoid accidents. The present invention has the advantages of high efficiency, flexibility, and adaptability, and can continuously monitor and accurately predict failure risks under different working conditions, greatly improving the safety and stability of the waste heat recovery system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a flow chart schematically showing a safety warning method for a waste heat recovery system based on an infrared thermogram according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] An embodiment of a waste heat recovery system safety early warning method based on infrared thermogram.

[0024] like Figure 1As shown, the flowchart of the waste heat recovery system safety early warning method based on infrared thermogram of an embodiment of the present invention includes the following steps: S1: Acquire historical images and current images of different areas of the waste heat recovery system.

[0025] In one embodiment, high-performance CMOS or CCD cameras can be used to accurately capture each historical image and current image. Compared with traditional image acquisition methods, these cameras can provide higher resolution and better dynamic range, thereby ensuring that the acquired image quality is clearer and richer in details. In order to further improve the image quality, advanced denoising processing technology is used to optimize each frame of the image to remove interference information introduced by sensor noise, lighting changes or environmental interference.

[0026] Specifically, the denoising process helps to significantly reduce random noise in the image and improve the signal-to-noise ratio of the image, thereby enhancing the accuracy and robustness of subsequent image analysis algorithms. For example, using denoising methods such as median filtering or Gaussian filtering, image details can be smoothed, so that the accuracy of subsequent steps such as edge detection and texture analysis can be optimized. This denoising process can not only effectively reduce the impact of noise, but also retain key feature information in the image, avoid excessive blurring of image details, and ensure the efficiency and reliability of the analysis results.

[0027] Using CMOS or CCD cameras to acquire high-quality images combined with advanced denoising technology not only effectively improves the quality of the image itself, but also provides a more reliable foundation for subsequent image processing and analysis steps.

[0028] S2: Calculate the failure probability and danger threshold of each area based on each historical image.

[0029] In one embodiment, the calculation Failure probability in a region , , For the The historical images of the area The entropy value of the infrared heat map of the image, For the The mean entropy value of the infrared thermal map of all images in the historical images of the region, For the The total number of historical images of each region; by calculating the failure probability of each region, the entropy change of the infrared thermal map in the historical image is used to quantify the abnormality of the region. Through the standardized calculation of the entropy value of the historical image, the region with a large deviation from the normal situation can be effectively identified, and then the potential failure probability can be inferred. This method can comprehensively evaluate the stability and consistency of each region based on multiple historical image data, thereby improving the accuracy and reliability of fault detection. In addition, through this entropy-based analysis, it is possible to sensitively capture small changes or anomalies and timely warn of potential failures.

[0030] Calculate the Danger threshold for each region , , where To preset the danger threshold, The natural constant An exponential function with a base of . The danger threshold of each area is calculated by combining the failure probability, and the nonlinear relationship of the failure probability is modeled using the exponential function, so as to dynamically adjust the danger threshold. This method can adaptively adjust the alert level according to the failure probability of the area, maintain a lower danger threshold when the failure probability is low, and ensure that the alarm is not excessively triggered during normal operation. When the failure probability is high, the threshold is increased to warn of potential dangers in advance. Through this dynamic adjustment mechanism, it can respond flexibly under different working conditions, effectively improve the accuracy of fault detection and warning, and avoid false alarms and missed alarms.

[0031] S3: Get the hot spots of the current image and calculate the complexity of each hot spot.

[0032] In one embodiment, any one of the current images is divided into multiple target images, and a hot spot in any one of the images is obtained, wherein the hot spot is a pixel point having the largest sum of the absolute values ​​of the temperature differences between any pixel point in any target image and the remaining pixels; the image is divided into multiple target images, including averaging the image to obtain multiple images of the same size.

[0033] Calculate the The complexity of the hotspot is, , where For the The complexity of each hotspot, Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. is the total number of hotspots, is the total number of current images, Hotspot The standard deviation of the magnitude of color gradients in different images relative to the current image.

[0034] By refining the image processing steps and combining hotspot detection and complexity calculation, the hotspots in the image can be accurately analyzed. By dividing the image into multiple target areas and identifying hotspots with significant temperature differences, the most representative abnormal areas in the image can be effectively highlighted. The complexity of calculating the hotspots further enhances the multi-dimensional analysis capability of the hotspots. By considering color gradients and pixel differences, the detailed features and spatial complexity of each hotspot can be accurately evaluated. This method can flexibly respond to hotspot changes in different scenarios, improve the ability to identify abnormal areas, enhance the accuracy and robustness of image processing, ensure the effective capture and response to potential problems in complex environments, and thus improve the accuracy of fault warning and decision support.

[0035] In another embodiment, the image is divided into multiple images of any size. The complexity is, , where For the The complexity of each hotspot, Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. is the total number of hotspots, is the total number of current images, Hotspot The standard deviation of the magnitude of color gradients in different images relative to the current image.

[0036] By introducing an improved complexity calculation method, the analysis of hot spots not only takes into account the mean difference of color gradients, but also integrates the relative standard deviation between hot spots, and weights the complexity through a nonlinear function. This method can more accurately capture the subtle changes of hot spots in different images, especially when processing images with significant spatial or temperature differences, and can more sensitively reflect the complex characteristics of hot spots. Through this comprehensive analysis, the system can more efficiently identify abnormal hot spots, especially in complex backgrounds or with more interference, and improve the accuracy and robustness of hot spot detection. At the same time, nonlinear weighted processing makes the complexity evaluation more in line with the needs of practical applications, and improves the reliability and accuracy of fault prediction and alarm.

[0037] S4: If the complexity of the hotspot is greater than or equal to the danger threshold of the area, the area is determined to be abnormal and an early warning is issued.

[0038] In one embodiment, the preset danger threshold is 0.6. Of course, it can also be determined according to actual conditions. By setting an appropriate danger threshold, the system can evaluate the status of equipment or areas in real time to ensure timely response when an abnormality occurs.

[0039] The early warning includes triggering an alarm signal through an audible and visual alarm, which can not only effectively attract the attention of operators and ensure that the early warning signal can be quickly identified in any environment, but also promptly remind relevant personnel to take necessary safety measures in an emergency. It also provides continuous safety protection for equipment operation under different working conditions, significantly improves the timeliness and accuracy of fault prevention, and reduces the occurrence of potential risks.

[0040] The solution of the present invention realizes accurate safety warning of the waste heat recovery system by combining infrared thermal map analysis, hot spot detection and complexity assessment. Through multi-level analysis of historical images and current images, it is possible to effectively identify possible fault areas, timely calculate fault probabilities and danger thresholds, and provide a scientific basis for the early warning mechanism. The calculation of hot spot complexity further enhances the sensitivity to abnormal changes, and can detect subtle temperature fluctuations and abnormal trends during system operation, and provide real-time early warnings. De-noising improves image quality, avoids noise interference, and ensures data accuracy. In addition, through the early warning method of the sound and light alarm, it is ensured that the attention of the operator can be quickly attracted in abnormal situations, thereby improving the response speed and reliability of the system, reducing the risk of failure, and significantly enhancing the safety and stability of the waste heat recovery system.

[0041] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0042] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A waste heat recovery system safety early warning method based on infrared thermal images, characterized in that: include: Obtain historical images and current images of different areas of the waste heat recovery system; Calculate the Failure probability in a region , , For the The historical images of the area The entropy value of the infrared heat map of the image, For the The mean entropy value of the infrared thermal map of all images in the historical images of the region, For the The total number of historical images of the region; calculate the Danger threshold for each region , , where To preset the danger threshold, The natural constant An exponential function with base ; Divide any image in the current image into multiple target images, obtain the hot spots in any image, and the hot spots are the pixels with the largest sum of the absolute values ​​of the temperature differences between any pixel point and the remaining pixel points in any target image; calculate the The complexity of a hotspot, wherein the complexity represents the abnormality of the hotspot; If the complexity of the hotspot is greater than or equal to the danger threshold of the area where the hotspot is located, the area where the hotspot is located is determined to be abnormal and an early warning is issued.

2. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: The complexity is, , where For the The complexity of each hotspot, Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. is the total number of hotspots, is the total number of current images, Hotspot The standard deviation of the magnitude of color gradients in different images relative to the current image.

3. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: The complexity is, , where For the The complexity of each hotspot, Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. Hotspot In the current image The mean value of the color gradient amplitude of all pixels in the connected domain in the image. is the total number of hotspots, is the total number of current images, Hotspot The standard deviation of the magnitude of color gradients in different images relative to the current image.

4. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: The waste heat recovery system consists of a heat exchanger, a heat exchange medium, a steam generator and a hot air blower.

5. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: The historical images and the current image are acquired by using a CMOS camera or a CCD camera.

6. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: The method also includes performing denoising processing on the historical images and the current image.

7. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: The early warning includes triggering an alarm signal through an audible and visual alarm.

8. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: The preset danger threshold is 0.

6.

9. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: Dividing the image into multiple target images includes equally dividing the image to obtain multiple images of the same size.

10. The waste heat recovery system safety early warning method based on infrared thermogram according to claim 1 is characterized in that: Divide the image into multiple target images, Including dividing the image into multiple images of any size.

Images