Method for cleaning the inner wall of a boiler based on a drum robot

Abstract

The present application relates to the technical field of boiler inner wall coke cleaning, and particularly relates to a method for cleaning coke on the inner wall of a boiler based on a drum robot. The method comprises the following steps: using the drum robot to preliminarily clean the coke area; then judging whether the subsequent photographed image is disturbed according to the flue gas, if disturbed, obtaining a thermal imaging image sequence of the coke area; obtaining the temperature value stability and temperature value consistency of the pixel points at the same position respectively, and obtaining the stability degree of the pixel points at the position, then obtaining the stable pixel points and unstable pixel points; obtaining the flue gas concentration relative value of the unstable pixel points, obtaining the weighted stability degree of each flue gas concentration relative value, and obtaining a fitting curve, then obtaining the corrected stability degree of the unstable pixel points and stable pixel points, and finally obtaining a corrected image; evaluating the cleaning condition of the coke area based on the corrected image until the coke cleaning is completed. The present application can improve the efficiency of coke cleaning.

Description

Technical Field

[0001] This invention relates to the field of boiler inner wall coke cleaning technology, specifically to a boiler inner wall coke cleaning method based on a drum robot. Background Technology

[0002] Coke buildup on the boiler's inner wall is a solid deposit formed during boiler operation, caused by the adhesion, carbonization, and gradual accumulation of combustible materials such as fuel residue on the boiler's inner wall due to high temperatures. Coke buildup has significant thermal resistance, affecting the boiler's heat transfer efficiency, leading to localized overheating, increased fuel consumption and emissions, and posing certain safety hazards. Therefore, timely cleaning of coke buildup on the boiler's inner wall is necessary to maintain normal boiler operation. Traditionally, coke removal relies on manual labor; however, due to the high temperature and enclosed space inside the boiler, manual cleaning is highly dangerous, slow, and inefficient. With technological advancements, robots are gradually replacing manual coke removal.

[0003] Because coke has poor thermal conductivity, the temperature at the coke-covered area is significantly higher than that of the clean inner wall. Therefore, robots often use thermal imaging to identify the location of coke on the boiler inner wall. However, during the coke removal process by the drum robot, a large amount of flue gas is often generated. Flue gas has properties such as absorption, scattering, and blocking of thermal infrared waves, resulting in blurry and inaccurate thermal imaging images, affecting the assessment of the coke condition. Furthermore, in poorly ventilated areas within the boiler, flue gas often lingers for a longer period, impacting the efficiency of the robot's coke removal work. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide a method for cleaning coke buildup on the inner wall of a boiler based on a drum robot. The specific technical solution adopted is as follows:

[0005] One embodiment of the present invention provides a method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, the method comprising:

[0006] The initial thermal imaging image inside the boiler is captured using a drum robot, and the coke accumulation area is determined based on the initial thermal imaging image; the drum robot is then used to perform preliminary cleaning of the coke accumulation area.

[0007] Based on the concentrations of steam, carbon dioxide, and particulate matter in the boiler after initial cleaning, it is determined whether subsequent images will be interfered with. If subsequent images are interfered with, a series of thermal imaging images will be continuously captured on a coked area within a set time period.

[0008] Based on the temperature values ​​of pixels at the same location in each image of the thermal imaging image sequence, the stability and consistency of temperature value changes of the pixels at that location are obtained, and the stability of the pixels at that location is obtained; stable pixels and unstable pixels are obtained based on the stability of the pixels at each location.

[0009] The relative value of the smoke concentration of an unstable pixel is obtained based on the distance between the unstable pixel and all pixels contained in the focal region; the weighted average of the stability of each unstable pixel under a certain relative value of smoke concentration is used to obtain the weighted stability of that relative value of smoke concentration, and a fitting curve is obtained.

[0010] The corrected stability of each unstable pixel is obtained based on the fitted curve; the corrected stability of the stable pixels is obtained, and the temperature values ​​of each unstable pixel in the last image of the thermal imaging image sequence are corrected to obtain a corrected image; the cleaning of the focused area is evaluated based on the corrected image.

[0011] Preferably, determining whether subsequent images are interfered with based on the concentrations of steam, carbon dioxide, and particulate matter in the boiler after preliminary cleaning includes:

[0012] The differences in steam concentration, carbon dioxide concentration, and particulate matter concentration in the boiler before and after preliminary cleaning are normalized and averaged to obtain the probability of interference with the thermal imaging image. If the probability of interference with the thermal imaging image is greater than or equal to the first threshold, then subsequent images are interfered with.

[0013] Preferably, the stability of the temperature value and the consistency of temperature value change of the pixel at the same location are obtained based on the temperature value of the pixel at the same location in each image of the thermal imaging image sequence, including:

[0014] The standard deviation of the temperature values ​​of pixels at the same location in each image of the thermal imaging image sequence is added to a first preset value and inverted to obtain the temperature value stability of the pixel at that location. The temperature values ​​of pixels at the same location are arranged in chronological order to obtain a time-series temperature value sequence. The difference between the next and previous temperature values ​​in every two adjacent temperature values ​​in the time-series temperature value sequence is calculated, and the obtained differences are sorted in chronological order to obtain a difference value sequence. The consistency of temperature value change of the pixel at that location is obtained based on the difference value sequence.

[0015] Preferably, obtaining the consistency of temperature value changes of pixels at that location based on the difference sequence includes:

[0016] Obtain the absolute value of the difference between the next and previous differences in every two adjacent differences in the difference sequence, and calculate the reciprocal of the average of all absolute differences. Then normalize to obtain the consistency of temperature value changes.

[0017] Preferably, the method for obtaining the stability level is as follows:

[0018] The minimum and maximum standard deviations of temperature values ​​for each pixel at each location in the thermal imaging image sequence are obtained and denoted as minimum standard deviation and maximum standard deviation, respectively. The weights corresponding to temperature value stability and temperature value change consistency are obtained based on the minimum and maximum standard deviations, respectively. The stability of the pixel at a location is obtained by weighted summation of the temperature value stability and temperature value change consistency based on the weights corresponding to temperature value stability and temperature value change consistency.

[0019] Preferably, the weights corresponding to temperature value stability and temperature value change consistency are obtained based on the minimum standard deviation and the maximum standard deviation, respectively, including:

[0020] The weight corresponding to the consistency of temperature value changes is the ratio of the minimum standard deviation to the maximum standard deviation; the weight corresponding to the stability of temperature values ​​is the difference between the first preset value and the weight corresponding to the consistency of temperature value changes.

[0021] Preferably, stable pixels and unstable pixels are obtained based on the stability of pixels at each location, including:

[0022] Specifically, if the stability of a pixel at a certain position in each image of a thermal imaging image sequence is less than the second threshold, then all pixels at that position are unstable pixels; if the stability is greater than or equal to the second threshold, then all pixels at that position are stable pixels.

[0023] Preferably, obtaining the relative value of the smoke concentration of an unstable pixel based on the distance between the unstable pixel and all pixels contained in the coking region includes:

[0024] Based on the coordinates of an unstable pixel in the image and the coordinates of each pixel in the focal region, the distance between the unstable pixel and each pixel in the focal region is obtained using the Euclidean distance calculation method, and the sum of the distances is obtained. The sum of the distances is then inverted and normalized to obtain the relative flue gas concentration value of the unstable pixel.

[0025] Preferably, the weighted stability of the relative smoke concentration is obtained by weighted averaging the stability of unstable pixels under a certain relative smoke concentration, including:

[0026] The absolute values ​​of the differences between the stability of an unstable pixel at a given relative flue gas concentration value and the stability of other unstable pixels at the same relative flue gas concentration value are calculated and summed to obtain the summation result. The summation result is then inverted and normalized to obtain the probability of the stability of the unstable pixel at the given relative flue gas concentration value being affected by flue gas.

[0027] The stability of an unstable pixel at a given relative smoke concentration value is determined by comparing the probability of its stability being affected by smoke with the sum of the probabilities of all unstable pixels at that relative smoke concentration value. The weight of the unstable pixel at that relative smoke concentration value is then calculated by weighting the stability of each unstable pixel at that relative smoke concentration value.

[0028] Preferably, obtaining the fitted curve includes:

[0029] The weighted stability of each relative flue gas concentration value and the weighted stability of each relative flue gas concentration value are fitted to obtain a fitting curve; the horizontal axis of the fitting curve is each relative flue gas concentration value, and the vertical axis is the weighted stability of each relative flue gas concentration value.

[0030] Preferably, the correction stability of each unstable pixel is obtained based on the fitted curve, including:

[0031] The relative flue gas concentration value of unstable pixels is substituted into the fitting curve to obtain the corrected stability of unstable pixels.

[0032] Preferably, the temperature values ​​of unstable pixels in the last image of the thermal imaging image sequence are corrected to obtain a corrected image, including:

[0033] Obtain the neighboring pixels of an unstable pixel in the last image; use the ratio of the corrected stability of a neighboring pixel to the sum of the corrected stability of all neighboring pixels of the unstable pixel as the weight of the neighboring pixel; use the weights corresponding to the neighboring pixels of the unstable pixel to perform a weighted summation of the temperature values ​​of each neighboring pixel to obtain the corrected temperature value of the unstable pixel; obtain the corrected temperature values ​​of all unstable pixels in the last image to obtain the corrected image.

[0034] Preferably, obtaining the corrected stability of stable pixels includes:

[0035] Set the first preset value to the degree of stability correction for stable pixels.

[0036] The embodiments of the present invention have at least the following beneficial effects: This application first determines the coking area in the boiler through the initial thermal imaging image, performs preliminary cleaning of the coking area, and then judges whether the subsequent images will be interfered with when analyzing the coking situation based on the water vapor concentration, carbon dioxide concentration and soot particle concentration in the boiler after preliminary cleaning. If it is interfered with, the thermal imaging image sequence is obtained by continuously shooting a coking area within a set time for subsequent analysis, so as to obtain a more accurate coking judgment result.

[0037] Furthermore, the stability of temperature values ​​at the same location on each image in the thermal imaging image sequence is analyzed to obtain the stability of each pixel, thereby identifying stable and unstable pixels. Then, the relative values ​​of flue gas concentration and the weighted stability of each relative value of flue gas concentration for unstable pixels are obtained, and a fitting curve is obtained. Based on the fitting curve, the corrected stability of unstable pixels is obtained to correct the stability of unstable pixels and improve the accuracy of the analysis. Finally, the temperature values ​​of each unstable pixel in the last image are corrected to obtain a corrected image. This can effectively avoid the problem of over-cleaning or incomplete cleaning caused by the interference of flue gas generated during the decoking process on the judgment of the coke accumulation on the inner wall of the boiler, thus improving the effectiveness of the robot decoking work. Attached Figure Description

[0038] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 A flowchart illustrating a method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, provided as an embodiment of the present invention. Detailed Implementation

[0040] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a boiler inner wall coke cleaning method based on a drum robot proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0042] The following description, in conjunction with the accompanying drawings, details a specific scheme for a boiler inner wall coke cleaning method based on a drum robot provided by the present invention.

[0043] Example:

[0044] The main application scenario of this invention is as follows: When the drum robot is performing coke removal work on the inner wall of a boiler, it may generate a large amount of flue gas, which will affect the clarity and accuracy of the thermal imaging image, resulting in inaccurate assessment results of the coke accumulation on the inner wall of the boiler and affecting the efficiency of the robot's coke removal work. This application corrects the affected thermal imaging image to facilitate the cleaning of the coke accumulation.

[0045] Please see Figure 1 The diagram illustrates a method flowchart for cleaning coke buildup on the inner wall of a boiler based on a drum robot, according to an embodiment of the present invention. The method includes the following steps:

[0046] Step S1: Use a drum robot to capture an initial thermal imaging image of the boiler, and determine the coke accumulation area based on the initial thermal imaging image; use the drum robot to perform preliminary cleaning of the coke accumulation area.

[0047] First, an infrared thermal imager needs to be installed at the front of the drum robot to capture thermal images of the boiler interior. This initial thermal image of the boiler's inner wall is then captured using the infrared thermal imager installed at the front of the drum robot. Because the inner wall of a normal boiler has good thermal conductivity and a smooth surface, the temperature transition is natural. However, the temperature difference between the coking area and the non-coking area is significant. Therefore, edge detection is performed on the initial thermal image to identify the coking area. Then, the position of the drum robot is adjusted so that it faces the center of the coking area and performs initial cleaning of the coking area.

[0048] Since a single cleaning may not completely remove the scorched area, it is necessary to assess the condition of the area after the initial cleaning to determine if it has been effectively cleaned. Therefore, after the initial cleaning, a thermal image of the scorched area is taken again to evaluate the cleaning progress.

[0049] However, the cleaning process may generate a significant amount of smoke (including water vapor, carbon dioxide, and particulate matter). Taking a thermal image of the coked area again after the initial cleaning may be affected by interference, reducing its accuracy. To better assess the degree of smoke interference, a smoke analyzer and a light scattering dust collector (laser) need to be installed on the robot. These will be used to monitor the current environmental concentrations of water vapor (H), carbon dioxide (CO), and particulate matter (PM), respectively, and to collect relevant smoke concentration data while taking thermal images before and after cleaning.

[0050] Step S2: Based on the concentration of water vapor, carbon dioxide and dust particles in the boiler after preliminary cleaning, determine whether subsequent images will be interfered with; if subsequent images are interfered with, continuously capture thermal imaging image sequences of a coking area within a set time period.

[0051] Water vapor, carbon dioxide, and particulate matter in the flue gas are the main factors interfering with thermal imaging images. Therefore, the higher the concentration of water vapor, carbon dioxide, and particulate matter in the environment when taking thermal imaging images, the more likely the images will be interfered with. Since the concentration of water vapor, carbon dioxide, and particulate matter in the environment is mainly generated during the cleaning process, the concentration of the relevant flue gas before the initial cleaning is taken as the normal concentration inside the boiler. Based on the difference between the relevant flue gas concentration after the initial cleaning and the normal concentration, the probability of interference with the current thermal imaging image is calculated.

[0052] The concentrations of steam, carbon dioxide, and particulate matter inside the boiler after initial cleaning are used to determine whether subsequent images will be affected.

[0053] Specifically, the differences in steam concentration, carbon dioxide concentration, and particulate matter concentration in the boiler before and after preliminary cleaning are normalized and averaged to obtain the probability of interference in the thermal imaging image.

[0054] The specific calculation model for the probability of interference in thermal imaging images is as follows:

[0055] ,

[0056] in, Indicates the possibility of interference with thermal imaging images. , and These represent the concentrations of steam, carbon dioxide, and particulate matter in the boiler after preliminary cleaning. , and These represent the concentrations of steam, carbon dioxide, and particulate matter in the boiler before the initial cleaning. This indicates the increase in water vapor concentration in the environment after cleaning compared to before. The larger the difference, the greater the likelihood of interference with subsequent thermal imaging images. Used to normalize the increase in water vapor concentration. This indicates the possibility that subsequent thermal imaging images may be affected by water vapor. This represents the maximum value of historical steam concentration data inside the boiler. This represents the maximum historical carbon dioxide concentration data inside the boiler. This represents the maximum historical concentration of particulate matter in the boiler. This indicates the possibility that subsequent thermal imaging images may be affected by carbon dioxide interference. This indicates the possibility that subsequent thermal imaging images may be interfered with by smoke particles. The likelihood of interference with the captured thermal image is represented by the average probability of subsequent thermal imaging images being affected by water vapor, carbon dioxide, and soot particles.

[0057] Furthermore, a first threshold is set. Preferably, the value of the first threshold is 0.4 (a reference value, which the implementer can adjust according to the actual situation; if higher sensitivity is required, the value can be increased appropriately). If the possibility of interference with the thermal imaging image is greater than or equal to the first threshold, the subsequent captured images will be interfered with. In this case, it is necessary to continuously capture a focused area within a set time to obtain the corresponding thermal imaging image sequence for subsequent analysis, so as to perform image correction and obtain a more accurate assessment of the focused situation. Preferably, the set time is 1 minute, and the shooting interval during continuous shooting is 10 seconds. The implementer can adjust it according to the actual situation.

[0058] If the probability of interference in the thermal imaging image is less than the first threshold, it indicates that the accuracy of the evaluation image of the in-focus area is high. In this case, the in-focus area can be directly evaluated using the thermal imaging images of the in-focus area taken later. It should also be noted that if, after initial cleaning, it is determined that subsequent images are not interfered with, the in-focus area can be evaluated using the thermal imaging images of the in-focus area taken later. After evaluation, if further in-focus cleaning of the in-focus area is required, the image should be evaluated again for interference after cleaning. This will determine whether to directly use the thermal imaging images of the in-focus area taken later for in-focus assessment, or whether to take a series of images to form a thermal imaging image sequence for subsequent analysis before evaluating the in-focus area. This process continues until the evaluation result indicates that the in-focus area is completely cleaned. At this point, the next in-focus area is evaluated and cleaned according to the plan, following the same procedure as the first in-focus area, until the in-focus area is completely cleared.

[0059] Step S3: Based on the temperature value of the pixel at the same position in each image of the thermal imaging image sequence, obtain the temperature value stability and temperature value change consistency of the pixel at that position, and obtain the stability of the pixel at that position; obtain stable pixels and unstable pixels according to the stability of the pixel at each position.

[0060] When subsequent image capture of a fouled area is affected by water vapor and smoke concentration, it is necessary to analyze the thermal imaging image sequence of the fouled area to obtain the final image that can be used to assess the fouling situation.

[0061] Water vapor, carbon dioxide, and particulate matter affect thermal imaging images by absorbing, reflecting, and blocking infrared radiation. Although the cleaning process increases the concentration of these gases and dust, they gradually settle and recover after cleaning. This means that the concentration of water vapor, carbon dioxide, and particulate matter in the environment is dynamic after cleaning. Therefore, pixels affected by interference in the thermal imaging image will show unstable changes. However, the actual boiler inner wall and coke buildup do not change over time. Therefore, by comparing the changes in multiple consecutive thermal imaging frames, inaccurate thermal imaging areas (areas affected by flue gas interference) can be identified.

[0062] First, since the temperature values ​​of undisturbed pixels in a thermal imaging image sequence corresponding to a focal area do not change significantly, it is necessary to obtain and analyze the temperature values ​​at the same location in each image of the thermal image sequence. For example, for a pixel with coordinates (a, b) in the first image, we need to find pixels with coordinates (a, b) in the remaining images. These pixels are the pixels at the same location.

[0063] Based on the temperature values ​​of pixels at the same location in each image of a thermal imaging image sequence, the stability of the temperature value and the consistency of temperature value changes of the pixels at that location are obtained.

[0064] Specifically, the standard deviation of the temperature values ​​of pixels at the same location in each image of the thermal imaging image sequence is added to a first preset value and inverted to obtain the temperature value stability of the pixel at that location; the temperature values ​​of pixels at the same location are arranged in chronological order to obtain a temporal temperature value sequence; the difference between the latter and former temperature values ​​in every two adjacent temperature values ​​in the temporal temperature value sequence is calculated, and the obtained differences are sorted in chronological order to obtain a difference value sequence; the consistency of temperature value change of the pixel at that location is obtained based on the difference value sequence.

[0065] The specific calculation model for temperature stability is as follows:

[0066] ,

[0067] in, This represents the stability of the temperature value of the pixel at position i in each image of a thermal imaging image sequence. This represents the standard deviation of the temperature value of the pixel at position i in each image of a thermal imaging image sequence. A smaller standard deviation indicates less temperature variation at that pixel location, meaning a more stable temperature value. Used to ensure that the fraction is meaningful and The first preset value is 1.

[0068] Furthermore, since the temperature changes of pixels at the same location in each undisturbed image should be regular without reheating, i.e. the temperature may gradually decrease over time without repeated fluctuations, the consistency of pixel temperature changes in multiple thermal imaging images is crucial.

[0069] The consistency of temperature value changes at a given pixel location is obtained by analyzing the difference sequence. Specifically, the absolute value of the difference between the second and first adjacent differences in the difference sequence is calculated, and the reciprocal of the average of all absolute differences is taken. This average is then normalized to obtain the consistency of temperature value changes.

[0070] The specific calculation model for the consistency of temperature value changes is as follows:

[0071] ,

[0072] in, This indicates the consistency of temperature value changes at the i-th position in each image of a thermal imaging image sequence; norm represents the normalization operation, used for normalization. This indicates the number of images in a thermal imaging image sequence. This represents the difference in temperature between the (t+1)th image and the tth image at the i-th position in the thermal imaging image sequence, which is also the difference between the (t+1)th temperature value and the t-th temperature value in the temperature value sequence. This represents the absolute value of the difference in temperature change of the pixel at position i between three adjacent thermal imaging images. The smaller the absolute value, the more stable the temperature change trend of the pixel at position i in the continuous image sequence.

[0073] Although the boiler has been shut down and cooled when the drum robot enters to clean it, the boiler's inner wall may still continue to cool during the robot's operation. Generally, the temperature change of the boiler's inner wall is slow at this time, but when the external environment is cold, it may exacerbate the cooling inside the boiler. Therefore, if the normal cooling rate of the boiler's inner wall is relatively fast, the temperature change at undisturbed pixel locations will also be more obvious. That is, the stability of the temperature value of each pixel cannot well distinguish between undisturbed and disturbed pixels, so it is necessary to judge based on the stability of the temperature change trend. Therefore, the minimum standard deviation of the temperature value of each pixel in multiple frames of thermal imaging images represents the severity of the current normal cooling of the boiler's inner wall, and this is used as the influence weight of the temperature value stability.

[0074] Therefore, the minimum and maximum values ​​of the standard deviation of the temperature value of each pixel at each position in each image of the thermal imaging image sequence are obtained and denoted as the minimum standard deviation and the maximum standard deviation, respectively. The weights corresponding to the temperature value stability and the temperature value change consistency are obtained based on the minimum standard deviation and the maximum standard deviation, respectively. The stability of the pixel at a position is obtained by weighted summing of the temperature value stability and the temperature value change consistency of the pixel at that position based on the weights corresponding to the temperature value stability and the temperature value change consistency.

[0075] If the boiler's inner wall cools down relatively quickly, the temperature value in the normal area will be unstable. Therefore, temperature stability alone cannot effectively distinguish between disturbed and undisturbed pixels. Thus, the weight of temperature change consistency should be increased, while the weight of temperature value stability should be decreased. This is achieved by obtaining the weights corresponding to temperature value stability and temperature change consistency based on the minimum and maximum standard deviations, respectively. Specifically, the weight corresponding to temperature value change consistency is the ratio of the minimum and maximum standard deviations; the weight corresponding to temperature value stability is the difference between the first preset value and the weight corresponding to temperature value change consistency.

[0076] The specific model for calculating stability is as follows:

[0077] ,

[0078] in, This represents the stability of the pixel at position i in each image of a thermal imaging image sequence. This indicates the consistency of temperature value changes at the i-th position in each image of a thermal imaging image sequence. This represents the stability of the temperature value of the pixel at position i in each image of a thermal imaging image sequence. This represents the minimum standard deviation of the temperature values ​​of a pixel across multiple frames of thermal imaging images; it is also known as the minimum standard deviation. This represents the maximum standard deviation of the temperature values ​​of a pixel across multiple frames of thermal imaging images; it is also known as the maximum standard deviation. One standard deviation represents the standard deviation of the temperature values ​​of a pixel at the same location across multiple frames of thermal imaging images. This indicates the severity of the normal cooling of the boiler's inner wall. Used for Normalization is performed.

[0079] A second threshold is set, preferably 0.7 (based on empirical values; the implementer can adjust the value of the second threshold based on multiple experiments or statistical distribution of stability data). Further, stable and unstable pixels are identified based on the stability of pixels at each location. Specifically, if the stability of a pixel at a given location in the thermal imaging image sequence is less than the second threshold, then all pixels at that location are considered unstable; if the stability is greater than or equal to the second threshold, then all pixels at that location are considered stable. The unstable pixels in the last image of the thermal imaging image sequence need to be analyzed and corrected subsequently.

[0080] Step S4: Obtain the relative value of the smoke concentration of an unstable pixel based on the distance between the unstable pixel and all pixels contained in the coking region; calculate the weighted average of the stability of each unstable pixel under a certain relative value of smoke concentration to obtain the weighted stability of that relative value of smoke concentration, and obtain the fitting curve.

[0081] While the stability of the pixels obtained from the above steps can reflect the interference of smoke on the thermal imaging image of the pixel location to some extent, the temperature may also change significantly due to the use of heating and other cleaning methods during the defocusing process. Therefore, the stability of a pixel at a single location cannot accurately reflect the degree of interference at that location, i.e., the accuracy of the pixel location temperature value.

[0082] Because water vapor, carbon dioxide, and soot particles that interfere with the realism of thermal imaging images are generated during the robot's defocusing process, the smoke concentration at the defocusing location is relatively high, while the surrounding area experiences changes in smoke concentration due to variations at the defocusing location, resulting in a relatively lower smoke concentration. Therefore, by calculating the distance between pixels at different locations and the current defocusing area, the difference in smoke concentration between each pixel and the center point can be estimated. Based on the smoke concentration at different locations and its stability across multiple frames, the influence of smoke concentration on pixel stability in the image is obtained. Furthermore, inaccurate pixels are corrected based on the temperature values ​​of less affected pixels.

[0083] The relative smoke concentration of an unstable pixel is obtained based on the distance between that unstable pixel and all pixels within the focal region. For pixels at the same location, if a pixel at that location is unstable, since its position is the same across all images, only one unstable pixel needs to be considered for calculation.

[0084] Specifically, based on the coordinates of an unstable pixel in the image and the coordinates of each pixel in the focal region, the distance between the unstable pixel and each pixel in the focal region is obtained using the Euclidean distance calculation method, and the sum of the distances is obtained. The sum of the distances is then inverted and normalized to obtain the relative flue gas concentration value of the unstable pixel.

[0085] The specific calculation model for relative flue gas concentration values ​​is as follows:

[0086] ,

[0087] in, This represents the relative smoke concentration value of the i-th unstable pixel. This indicates the number of pixels within the focal region, which is the focal region initially detected. This represents the coordinates of the i-th unstable pixel in the image. This represents the coordinates of the j-th pixel in the focal region. This represents the distance between the coordinates of the i-th unstable pixel and the coordinates of the j-th pixel in the focal region, calculated using the Euclidean distance method. This indicates that the smaller the distance between the i-th unstable pixel and the pixels in the focus region, the greater its relative smoke concentration. `norm` represents the normalization operation. The relative smoke concentration value refers to the relative smoke concentration value at the location of the unstable pixel.

[0088] Next, the stability level and relative smoke concentration values ​​of unstable pixels at all locations are projected onto a two-dimensional Cartesian coordinate system. The horizontal axis represents the relative smoke concentration value, and the vertical axis represents the stability level. The same relative smoke concentration value may correspond to multiple stability levels. Some values ​​may be unstable due to other factors rather than smoke interference, differing significantly from data values ​​solely affected by smoke and lacking regularity. Therefore, under the same relative smoke concentration value, the probability that each stability level is affected by smoke is calculated based on the density of each stability level compared to other stability levels. It should be noted that for each image in the thermal imaging image sequence, "the same location" means the same position on each image, with identical unstable and stable pixels, stability levels, and relative smoke concentration values. Here, we only analyze the unstable and stable pixels on a single image to analyze the characteristics of each location on the image.

[0089] Specifically, the absolute values ​​of the differences between the stability of an unstable pixel at a given relative flue gas concentration value and the stability of other unstable pixels at the same relative flue gas concentration value are calculated and summed to obtain the summation result. The summation result is then inverted and normalized to obtain the probability of the stability of the unstable pixel at that relative flue gas concentration value being affected by flue gas.

[0090] Its specific calculation model is as follows:

[0091] ,

[0092] in, This represents the probability that the stability of the v-th unstable pixel is affected by the smoke under the r-th relative smoke concentration value. This represents the number of unstable pixels under the r-th relative flue gas concentration value. This represents the stability of the v-th unstable pixel under the r-th relative flue gas concentration value. This represents the stability of the a-th unstable pixel under the r-th relative flue gas concentration value. This represents the sum of the differences between the stability of the v-th unstable pixel and the stability of other unstable pixels under the same relative flue gas concentration. The smaller this sum, the greater the density of the stability of the v-th unstable pixel relative to the stability of other unstable pixels, meaning it is more likely to be affected by flue gas.

[0093] Next, the probability of the stability of an unstable pixel being affected by smoke under a certain relative smoke concentration value is compared with the sum of the probabilities of the stability of all unstable pixels under the same relative smoke concentration value to obtain the weight of the stability of the unstable pixel under that relative smoke concentration value. The stability of each unstable pixel is then weighted and averaged using the weights of the stability of each unstable pixel under that relative smoke concentration value to obtain the weighted stability of that relative smoke concentration value.

[0094] Its specific calculation model is as follows:

[0095] ,

[0096] in, This represents the weighted stability of the r-th relative flue gas concentration value. This represents the number of weighted stability values ​​for the r-th relative flue gas concentration. This represents the probability that the stability of the v-th unstable pixel is affected by the smoke under the r-th relative smoke concentration value. This represents the probability that the stability of the a-th unstable pixel is affected by the flue gas under the r-th relative flue gas concentration value. Indicates the sum; The weight representing the stability of the v-th unstable pixel under the r-th relative flue gas concentration value. This represents the stability of the v-th unstable pixel under the r-th relative flue gas concentration value.

[0097] Finally, a fitting curve was obtained by fitting the weighted stability of each relative flue gas concentration value and the weighted stability of each relative flue gas concentration value. The fitting method used was the least squares method. This fitting curve represents the influence of flue gas concentration on the temperature of pixels in the thermal imaging image, and can be used to further correct the stability of unstable pixels.

[0098] Step S5: Obtain the corrected stability of each unstable pixel based on the fitted curve; obtain the corrected stability of the stable pixel, and correct the temperature value of each unstable pixel in the last image of the thermal imaging image sequence to obtain a corrected image; evaluate the cleaning of the coking area based on the corrected image.

[0099] In step S5, a fitted curve is obtained. Further, the stability of each unstable pixel can be corrected based on the fitted curve. Specifically, the relative smoke concentration value of the unstable pixel is substituted into the fitted curve to obtain the corrected stability of the unstable pixel, representing the degree to which the location of the unstable pixel is affected by smoke. A larger value indicates a smaller degree of influence from smoke. For stable pixels, their corresponding temperature values ​​are considered relatively realistic; therefore, a first preset value of 1 is set as their corrected stability.

[0100] Furthermore, using the image from the thermal imaging image sequence closest to the current moment as the image to be corrected, the neighboring pixels of an unstable pixel in the last image are obtained. The ratio of the correction stability of a neighboring pixel to the sum of the correction stability of all neighboring pixels of the unstable pixel is used as the weight corresponding to that neighboring pixel. The temperature values ​​of all neighboring pixels are weighted and summed using the weights corresponding to the unstable pixel to obtain the corrected temperature value of the unstable pixel. The corrected temperature values ​​of all unstable pixels in the last image are obtained to obtain the corrected image. The neighboring pixels of an unstable pixel refer to the pixels within its eight-neighborhood.

[0101] The specific calculation model for the temperature correction value of an unstable pixel in the last image is as follows:

[0102] ,

[0103] in, This represents the correction value for the temperature of the b-th unstable pixel in the last image. This represents the number of neighboring pixels of the b-th unstable pixel. This represents the degree of stability correction of the i-th neighboring pixel among the neighboring pixels of the b-th unstable pixel. This represents the temperature value of the i-th neighboring pixel among the neighboring pixels of the b-th unstable pixel. This represents the sum of the corrected stability levels of all neighboring pixels of the b-th unstable pixel. This represents the weight of the i-th neighboring pixel among the neighboring pixels of the b-th unstable pixel. This process corrects the temperature values ​​of all unstable pixels in the last image, resulting in the corrected image.

[0104] Furthermore, the cleaning status of the defocused area is evaluated based on the corrected image. Edge detection is performed on the corrected image to identify areas with higher temperature values, which are the areas of defocusing that have not been completely cleaned and need to be cleaned again. Then, another image is captured for evaluation until the defocused area is completely cleaned. If all areas in the corrected image show a stable and natural temperature transition during the evaluation, it indicates that the defocused area has been cleaned, and the cleaning process can proceed to the next defocused area. The acquisition of the defocused area is a prior art technique that can be achieved through edge detection, and will not be described in detail here.

[0105] In summary, this application analyzes the stability of different pixels in multiple thermal imaging images to determine the degree of temperature interference of each pixel in the image by smoke. By combining the change characteristics of the interfered pixel and its neighboring pixels in multiple images, the temperature values ​​of the interfered pixels in the image are corrected. Then, the coking situation after decoking is evaluated based on the corrected image, which can effectively improve the efficiency of the drum robot in cleaning coking.

[0106] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0107] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0108] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, characterized in that, The method includes: The initial thermal imaging image inside the boiler is captured using a drum robot, and the coke accumulation area is determined based on the initial thermal imaging image; the drum robot is then used to perform preliminary cleaning of the coke accumulation area. Based on the concentrations of steam, carbon dioxide, and particulate matter in the boiler after initial cleaning, it is determined whether subsequent images will be interfered with. If subsequent images are interfered with, a series of thermal imaging images will be continuously captured on a coked area within a set time period. Based on the temperature values ​​of pixels at the same location in each image of the thermal imaging image sequence, the stability and consistency of temperature value changes of the pixels at that location are obtained, and the stability of the pixels at that location is obtained; stable pixels and unstable pixels are obtained based on the stability of the pixels at each location. The relative value of the smoke concentration of an unstable pixel is obtained based on the distance between the unstable pixel and all pixels contained in the focal region; the weighted average of the stability of each unstable pixel under a certain relative value of smoke concentration is used to obtain the weighted stability of that relative value of smoke concentration, and a fitting curve is obtained. The corrected stability of each unstable pixel is obtained based on the fitted curve; the corrected stability of the stable pixels is obtained, and the temperature values ​​of each unstable pixel in the last image of the thermal imaging image sequence are corrected to obtain a corrected image; the cleaning of the focused area is evaluated based on the corrected image.

2. The method for cleaning coke buildup on the inner wall of a boiler based on a drum robot according to claim 1, characterized in that, The method of determining whether subsequent images are interfered with based on the concentrations of steam, carbon dioxide, and particulate matter in the boiler after preliminary cleaning includes: The differences in steam concentration, carbon dioxide concentration, and particulate matter concentration in the boiler before and after preliminary cleaning are normalized and averaged to obtain the probability of interference with the thermal imaging image. If the probability of interference with the thermal imaging image is greater than or equal to the first threshold, then subsequent images are interfered with.

3. The method for cleaning coke buildup on the inner wall of a boiler based on a drum robot according to claim 1, characterized in that, The method of obtaining the stability and consistency of temperature value changes of pixels at the same location in each image of a thermal imaging image sequence includes: The standard deviation of the temperature values ​​of pixels at the same location in each image of the thermal imaging image sequence is added to a first preset value and inverted to obtain the temperature value stability of the pixel at that location. The temperature values ​​of pixels at the same location are arranged in chronological order to obtain a time-series temperature value sequence. The difference between the next and previous temperature values ​​in every two adjacent temperature values ​​in the time-series temperature value sequence is calculated, and the obtained differences are sorted in chronological order to obtain a difference value sequence. The consistency of temperature value change of the pixel at that location is obtained based on the difference value sequence.

4. The method for cleaning coke buildup on the inner wall of a boiler based on a drum robot according to claim 3, characterized in that, The step of obtaining the consistency of temperature value changes of pixels at a given location based on the difference sequence includes: Obtain the absolute value of the difference between the next and previous differences in every two adjacent differences in the difference sequence, and calculate the reciprocal of the average of all absolute differences. Then normalize to obtain the consistency of temperature value changes.

5. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, as described in claim 1, is characterized in that... The method for obtaining the stability level is as follows: The minimum and maximum standard deviations of temperature values ​​at each location in each image of the thermal imaging image sequence are obtained and denoted as minimum standard deviation and maximum standard deviation, respectively. The weights corresponding to temperature value stability and temperature value change consistency are obtained based on the minimum and maximum standard deviations, respectively. The stability of a pixel at a given location is obtained by weighting the temperature stability and temperature change consistency based on the weights corresponding to temperature stability and temperature change consistency.

6. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, as described in claim 5, is characterized in that... The weights for temperature value stability and temperature value variation consistency obtained based on the minimum and maximum standard deviations, respectively, include: The weight corresponding to the consistency of temperature value changes is the ratio of the minimum standard deviation to the maximum standard deviation; the weight corresponding to the stability of temperature values ​​is the difference between the first preset value and the weight corresponding to the consistency of temperature value changes.

7. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, as described in claim 1, is characterized in that... The process of obtaining stable and unstable pixels based on the stability of pixels at each location includes: Specifically, if the stability of a pixel at a certain position in each image of a thermal imaging image sequence is less than the second threshold, then all pixels at that position are unstable pixels; if the stability is greater than or equal to the second threshold, then all pixels at that position are stable pixels.

8. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, as described in claim 1, is characterized in that... The step of obtaining the relative value of the smoke concentration of an unstable pixel based on the distance between the unstable pixel and all pixels contained in the coking region includes: Based on the coordinates of an unstable pixel in the image and the coordinates of each pixel in the focal region, the distance between the unstable pixel and each pixel in the focal region is obtained using the Euclidean distance calculation method, and the sum of the distances is obtained. The sum of the distances is then inverted and normalized to obtain the relative flue gas concentration value of the unstable pixel.

9. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot according to claim 1, characterized in that, The weighted average of the stability of unstable pixels under a certain relative value of smoke concentration to obtain the weighted stability of that relative value of smoke concentration includes: The absolute values ​​of the differences between the stability of an unstable pixel at a given relative flue gas concentration value and the stability of other unstable pixels at the same relative flue gas concentration value are calculated and summed to obtain the summation result. The summation result is then inverted and normalized to obtain the probability of the stability of the unstable pixel at the given relative flue gas concentration value being affected by flue gas. The stability of an unstable pixel at a given relative smoke concentration value is determined by comparing the probability of its stability being affected by smoke with the sum of the probabilities of all unstable pixels at that relative smoke concentration value. The weight of the unstable pixel at that relative smoke concentration value is then calculated by weighting the stability of each unstable pixel at that relative smoke concentration value.

10. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, as described in claim 1, is characterized in that... The process of obtaining the fitted curve includes: The weighted stability of each relative flue gas concentration value and the weighted stability of each relative flue gas concentration value are fitted to obtain a fitting curve; the horizontal axis of the fitting curve is each relative flue gas concentration value, and the vertical axis is the weighted stability of each relative flue gas concentration value.

11. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, as described in claim 1, is characterized in that... The step of obtaining the corrected stability of each unstable pixel based on the fitted curve includes: The relative flue gas concentration value of unstable pixels is substituted into the fitting curve to obtain the corrected stability of unstable pixels.

12. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, as described in claim 1, is characterized in that... The step of correcting the temperature values ​​of unstable pixels in the last image of the thermal imaging image sequence to obtain a corrected image includes: Obtain the neighboring pixels of an unstable pixel in the last image; use the ratio of the corrected stability of a neighboring pixel to the sum of the corrected stability of all neighboring pixels of the unstable pixel as the weight of the neighboring pixel; use the weights corresponding to the neighboring pixels of the unstable pixel to perform a weighted summation of the temperature values ​​of each neighboring pixel to obtain the corrected temperature value of the unstable pixel; obtain the corrected temperature values ​​of all unstable pixels in the last image to obtain the corrected image.

13. A method for cleaning coke buildup on the inner wall of a boiler based on a drum robot, as described in claim 1, is characterized in that... The process of obtaining the corrected stability of stable pixels includes: Set the first preset value to the degree of stability correction for stable pixels.

Images