Boiler economizer coating failure detection method and system based on infrared thermal imaging

Abstract

The application relates to a boiler economizer coating failure detection method and system based on infrared thermal imaging, relates to the technical field of boiler coating detection, and comprises the following steps: based on a coating failure detection instruction, calling infrared thermal imaging time sequence and boiler operation parameter time sequence of a target boiler economizer; dividing and determining the boiler operation condition of the boiler operation parameter time sequence, segmenting the infrared thermal imaging time sequence, and constructing a working condition infrared image sequence; analyzing the time sequence of the working condition infrared image sequence, and obtaining a coating degradation coefficient; obtaining a coating comprehensive degradation coefficient of the target boiler economizer based on the boiler operation condition and the coating degradation coefficient, and carrying out coating failure early warning based on the coating comprehensive degradation coefficient. The application solves the problems that traditional coating failure detection of the external coating of the boiler economizer needs to stop and disassemble the heat preservation layer, the efficiency is low, the best maintenance time is easily missed, quantitative evaluation and predictive maintenance of the heat insulation and corrosion prevention effect of the damaged coating are lacked, and the equipment operation efficiency and service life are affected.

Description

Technical Field

[0001] This application relates to the field of boiler coating inspection, and in particular to a method and system for detecting boiler economizer coating failure based on infrared thermal imaging. Background Technology

[0002] With increasing demands from industrial production for boiler operating efficiency and equipment lifespan, accurate and timely detection of boiler economizer coating failure has become a key technical requirement for ensuring stable equipment operation.

[0003] Currently, detecting failure of the external coating of traditional boiler economizers requires shutting down the machine and disassembling the insulation layer, which is not only inefficient but also makes it easy to miss the best time for maintenance. At the same time, traditional methods are mostly qualitative judgments and lack the ability to quantitatively assess the heat insulation and anti-corrosion effects of the coating after damage and predictive maintenance capabilities. This not only leads to obvious blindness and lag in the maintenance of external coatings, but also reduces heat efficiency due to heat insulation failure and accelerates rust due to anti-corrosion failure, thereby shortening the service life of equipment and increasing industrial operation and maintenance costs and safety risks. Summary of the Invention

[0004] This application provides a method and system for detecting coating failures of boiler economizers based on infrared thermal imaging, which improves the efficiency and accuracy of detecting external coating failures of boiler economizers, and improves the problems of blind maintenance and lag caused by qualitative judgment in traditional detection, effectively ensuring the operating efficiency and service life of boiler economizers.

[0005] The embodiments of this application disclose the following technical solutions:

[0006] In a first aspect, embodiments of this application provide a method for detecting coating failure in boiler economizers based on infrared thermal imaging, the method comprising:

[0007] In response to the coating failure detection command, the infrared thermal imaging timing sequence of the target boiler economizer and the timing sequence of boiler operating parameters are retrieved based on the coating failure detection command.

[0008] The operating conditions are divided according to the boiler operating parameter time sequence to determine multiple boiler operating conditions. The infrared thermal imaging time sequence is then segmented based on the multiple boiler operating conditions to construct multiple operating condition infrared image sequences.

[0009] Time-series analysis was performed on infrared image sequences under multiple operating conditions to obtain the coating degradation coefficient for each boiler operating condition;

[0010] The comprehensive coating degradation coefficient of the target boiler economizer is obtained based on the multiple boiler operating conditions and the coating degradation coefficient of each boiler operating condition, and coating failure early warning is performed based on the comprehensive coating degradation coefficient.

[0011] Secondly, embodiments of this application provide a boiler economizer coating failure detection system based on infrared thermal imaging, the system comprising:

[0012] The instruction response and sequence retrieval module is used to respond to the coating failure detection instruction and retrieve the infrared thermal imaging timing sequence of the target boiler economizer and the boiler operating parameter timing sequence based on the coating failure detection instruction.

[0013] The operating condition division and sequence construction module is used to divide the operating conditions according to the time sequence of the boiler operating parameters, determine multiple boiler operating conditions, and segment the infrared thermal imaging time sequence based on the multiple boiler operating conditions to construct multiple operating condition infrared image sequences.

[0014] The time series analysis and coefficient acquisition module is used to perform time series analysis on infrared image sequences under multiple operating conditions to obtain the coating degradation coefficient for each boiler operating condition.

[0015] The coefficient integration and failure early warning module is used to obtain the overall coating degradation coefficient of the target boiler economizer based on the multiple boiler operating conditions and the coating degradation coefficient of each boiler operating condition, and to perform coating failure early warning based on the overall coating degradation coefficient.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0017] This application proposes a method and system for detecting boiler economizer coating failure based on infrared thermal imaging. By responding to detection commands and retrieving data in a step-by-step manner, dividing operating conditions and constructing infrared image sequences, performing time-series analysis and obtaining degradation coefficients, calculating comprehensive degradation coefficients and failure warnings, the method achieves accurate detection and early warning of boiler economizer coating failure. First, responding to the coating failure detection command, a data extraction time window is constructed by combining the start time of the target boiler economizer's service and the current detection time. Infrared thermal imaging time series and boiler operating parameter time series within the corresponding time range are retrieved. Then, vector transformation and cluster analysis are performed on the boiler operating parameter time series to divide it into multiple boiler operating conditions. Based on the operating condition labeling results, the infrared thermal imaging time series is segmented to construct infrared image sequences corresponding to each operating condition. Subsequently, a preset interval is set to sample the infrared image sequences of each operating condition. A dedicated coating degradation evaluator bound to each operating condition is retrieved, and the coating degradation degree sequence for each operating condition is output through adjacent image group analysis. The coating degradation coefficient for each operating condition is obtained through cumulative decreasing calculation. Next, weighting coefficients are determined based on the duration of each operating condition and the total operating time. The degradation coefficients are weighted and summed with the weights to obtain the comprehensive coating degradation coefficient. Finally, the comprehensive degradation coefficient is compared with a preset coating failure warning threshold. If the coefficient exceeds the threshold, a warning message is generated and output to the user terminal.

[0018] The technical solution of this application solves the problems of traditional boiler economizer external coating failure detection, such as the need for shutdown operation, low detection efficiency, mostly qualitative judgment lacking quantitative evaluation, and difficulty in achieving early warning, as well as the lag in external coating maintenance caused by untimely warnings. It improves the real-time performance, accuracy and reliability of boiler economizer external coating failure detection, and provides strong support for the refined operation and maintenance and safe and stable operation of the equipment. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0020] Figure 1 A schematic flowchart of the boiler economizer coating failure detection method based on infrared thermal imaging provided in this application embodiment;

[0021] Figure 2 A schematic diagram of the structure of a boiler economizer coating failure detection system based on infrared thermal imaging provided in this application embodiment.

[0022] The components represented by each number in the attached diagram are explained below:

[0023] Module 01 for instruction response and sequence retrieval, module 02 for working condition division and sequence construction, module 03 for time series analysis and coefficient acquisition, and module 04 for coefficient synthesis and failure early warning. Detailed Implementation

[0024] This application provides a method and system for detecting coating failures of boiler economizers based on infrared thermal imaging. It addresses the technical problems in the prior art where traditional boiler economizer external coating failure detection requires shutdown operation, has low detection efficiency, is prone to missing the best maintenance time, lacks quantitative assessment of coating degradation degree and predictive maintenance capabilities, and thus leads to blind and delayed external coating maintenance.

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0027] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0028] Example 1, as shown in the appendix Figure 1 As shown, this application provides a method for detecting coating failure in boiler economizers based on infrared thermal imaging. The method includes the following steps:

[0029] S110: In response to the coating failure detection command, retrieve the infrared thermal imaging timing of the target boiler economizer and the timing of boiler operating parameters based on the coating failure detection command;

[0030] In this embodiment of the application, in the scenario where the coating of the boiler economizer is prone to failure due to factors such as high temperature and corrosion during long-term operation, and it is necessary to keep track of the coating status in a timely manner to ensure the stability of the equipment, in order to accurately obtain the basic data required for detection and avoid the impact of data loss or time range deviation on subsequent analysis, it is necessary to first respond to the coating failure detection command and retrieve the corresponding time series data to ensure that subsequent steps such as working condition division and degradation assessment can be carried out based on complete and matching data.

[0031] Specifically, it first receives coating failure detection instructions uploaded by the user. These instructions can be triggered by the user according to maintenance needs, or they can be automatically triggered at preset intervals to adapt to the detection needs in different scenarios.

[0032] Furthermore, the received coating failure detection command is parsed to extract the current detection time. This time is then combined with the start time of the target boiler economizer's commissioning to form a data extraction time window, ensuring that the retrieved data covers the complete change cycle of the coating from its initial use to the current detection node.

[0033] Furthermore, by connecting the thermal imaging database storing image data and the operating parameter database storing operating parameters, and based on the determined data extraction time window, the infrared thermal imaging time series of the target boiler economizer and the boiler operating parameter time series are retrieved from the two databases respectively. Both types of time series data have acquisition timestamps, providing a time dimension basis for subsequent accurate matching of operating conditions and infrared images.

[0034] This step, through standardized command response and data retrieval procedures, provides reliable data support for each subsequent stage of coating failure detection, ensuring the continuity and accuracy of the detection process.

[0035] Step S110 in the method provided in this application embodiment includes:

[0036] Receive coating failure detection instructions uploaded by the user terminal, wherein the coating failure detection instructions are user-triggered instructions or timed trigger instructions;

[0037] The coating failure detection command is parsed to obtain the current detection time, and the data extraction time window is formed by combining it with the service start time of the target boiler economizer.

[0038] The system connects to a thermal imaging database and an operating parameter database, and retrieves the infrared thermal imaging time series and boiler operating parameter time series of the target boiler economizer based on the data extraction time window. Both the infrared thermal imaging time series and the boiler operating parameter time series have acquisition timestamps.

[0039] In this embodiment of the application, in order to provide accurate and complete data support for the steps of working condition division and degradation analysis in the failure detection of boiler economizer coating, it is necessary to first obtain the key timing data of the target equipment through a standardized instruction receiving, parsing and data retrieval process, so as to ensure the continuity, accuracy and reliability of the entire coating failure detection process.

[0040] First, the system receives coating failure detection commands uploaded by the user. These commands can be either user-triggered or timed, allowing for flexible adaptation to different operational and maintenance scenarios.

[0041] Specifically, when maintenance personnel discover abnormal boiler operating parameters or suspect that the coating may be at risk of failure, they can proactively initiate user-triggered commands to achieve immediate detection. For routine periodic coating condition monitoring, timed trigger commands can be automatically generated by preset detection cycles (e.g., once a month or once a quarter), allowing for regular detection without manual intervention and ensuring the continuity of coating condition monitoring.

[0042] Furthermore, the coating failure detection command is parsed to obtain the current detection time, which is then combined with the service start time of the target boiler economizer to form a data extraction time window.

[0043] Among them, the service start time is the time when the boiler economizer is officially put into use. Combining it with the current testing time can ensure that the data extraction range covers the complete service cycle of the coating from its new state to the current testing point, avoiding the inability to capture the long-term degradation trend of the coating due to the data time range being too short.

[0044] For example, if the start date of service for a boiler economizer is January 1, 2024, and the current testing date is January 1, 2025, then the data extraction time window is set to January 1, 2024 to January 1, 2025, which can fully cover the status change data of the coating during its one-year service period.

[0045] Furthermore, after setting the data extraction time window, the system connects the thermal imaging database storing image data and the operating parameter database storing operating parameter data, and retrieves the infrared thermal imaging time sequence of the target boiler economizer and the boiler operating parameter time sequence according to the set time window.

[0046] During the retrieval process, it is necessary to ensure that both types of time-series data have acquisition timestamps to ensure accurate time matching between the extracted infrared thermal imaging data and the boiler operating parameter data. Subsequently, the infrared image at a certain moment can be associated with the corresponding boiler operating parameters such as pressure, temperature, and load through the timestamps, providing a key basis for segmenting the infrared thermal imaging time series according to operating conditions.

[0047] For example, when the acquisition timestamp of an infrared image is 14:00 on June 15, 2024, the operating parameters of the boiler at the same time can be found through the timestamp, and then the operating condition of the boiler corresponding to the infrared image can be determined.

[0048] Meanwhile, during the data retrieval process, it is also necessary to conduct preliminary verification of the completeness and validity of two types of time-series data: if there are missing infrared images in a certain time period in the thermal imaging database, or if the parameter records in a certain moment in the operating parameter database are abnormal, it is necessary to mark and supplement the data with valid data in the adjacent time period in a timely manner, or prompt the operation and maintenance personnel to check the status of the data acquisition equipment, so as to avoid the impact of data incompleteness on subsequent detection and analysis.

[0049] Finally, through the above steps of instruction reception, parsing, and data retrieval, the complete service life of the coating and the effective infrared thermal imaging time sequence and boiler operating parameter time sequence were obtained, laying a scientific data foundation for subsequent coating failure detection.

[0050] S120: Divide the operating conditions according to the boiler operating parameter time sequence, determine multiple boiler operating conditions, and segment the infrared thermal imaging time sequence based on the multiple boiler operating conditions to construct multiple operating condition infrared image sequences.

[0051] In this embodiment of the application, in scenarios where different operating conditions have significantly different effects on coating degradation during boiler operation, in order to avoid coating degradation assessment deviations caused by mixed operating condition data and to accurately match the coating state changes under each operating condition, it is necessary to first divide the operating conditions according to the boiler operating parameters and segment the infrared thermal imaging time sequence to ensure the accuracy and relevance of the subsequent coating degradation coefficient calculation.

[0052] Specifically, the boiler operating parameters are first transformed into vector form, and the scattered parameters such as pressure, temperature, and load are integrated to construct a multi-dimensional parameter feature vector sequence, thus transforming unstructured parameter data into analyzable structured data.

[0053] Furthermore, cluster analysis is performed on the multidimensional parameter feature vector sequence. Vectors with similar features will cluster to form multiple parameter clusters. Each cluster corresponds to a boiler operating condition with similar operating status, thereby determining multiple boiler operating conditions.

[0054] Furthermore, each acquisition timestamp in the boiler operating parameter time series is labeled with its corresponding boiler operating condition to obtain the operating condition labeling result, establishing a correspondence between time and operating condition. Then, based on the operating condition labeling result, all infrared image data belonging to the same boiler operating condition are selected from the infrared thermal imaging time series. These image data are arranged according to the acquisition time order to construct a corresponding infrared image sequence for each boiler operating condition, ultimately resulting in multiple operating condition infrared image sequences.

[0055] This step involves dividing the boiler's operating conditions and constructing a dedicated infrared image sequence for each condition. This allows subsequent coating degradation analysis to be conducted on a single operating condition, eliminating interference between different operating conditions and providing a basis for accurately assessing the degree of damage to the coating under each condition.

[0056] Step S120 in the method provided in this application embodiment includes:

[0057] The boiler operating parameters time series are transformed into vector form to construct a multidimensional parameter feature vector sequence;

[0058] Cluster analysis is performed on the multidimensional parameter feature vector sequence to obtain multiple parameter clusters, and multiple boiler operating conditions are determined based on the multiple parameter clusters.

[0059] Each collection timestamp in the boiler operating parameter time series is marked as the corresponding boiler operating condition to obtain the operating condition marking result;

[0060] Based on the operating condition marking results, all infrared image data belonging to the same boiler operating condition are extracted from the infrared thermal imaging time sequence and arranged in the order of acquisition time to construct the infrared image sequence corresponding to each boiler operating condition, thus obtaining multiple operating condition infrared image sequences.

[0061] In this embodiment of the application, in order to accurately distinguish the impact of different operating conditions on the degradation of the boiler economizer coating and avoid the distortion of coating degradation assessment caused by interference from mixed operating condition data, it is necessary to divide the operating conditions by structuring and clustering analysis of the operating parameters and matching the corresponding infrared image sequence of the operating conditions to ensure that the subsequent coating degradation analysis can be carried out for a single operating condition, thereby improving the accuracy and pertinence of coating failure detection.

[0062] Specifically, the boiler operating parameter time series is first transformed into vector form to construct a multi-dimensional parameter feature vector sequence. The boiler operating parameter time series includes various dynamically changing parameters such as pressure, temperature, and load. The parameter values ​​corresponding to each acquisition timestamp are integrated, and the scattered parameter data is integrated into a multi-dimensional parameter feature vector, providing a structured data foundation for operating condition division.

[0063] Among them, temperature refers to the temperature of the medium inside the boiler tube, that is, the temperature of the working fluid flowing through the pipes of the economizer during boiler operation. This standardizes the temperature parameter benchmark and ensures the consistency and accuracy of temperature indicators when dividing operating conditions.

[0064] For example, if the boiler operating parameters corresponding to a certain acquisition timestamp are "pressure 1.5MPa, temperature 2205℃, load 81%", these three parameters are converted into elements in a vector through numerical extraction and integration, forming a three-dimensional parameter feature vector of [1.5, 205, 81]. If the parameters corresponding to another acquisition timestamp are "pressure 0.6MPa, temperature 172℃, load 42%", they are integrated and converted into a three-dimensional parameter feature vector of [0.6, 172, 42].

[0065] Similarly, by traversing all the collected timestamps using the above transformation method, the entire boiler operating parameter time series can be transformed into a multi-dimensional parameter feature vector sequence composed of multiple structured vectors, allowing the scattered parameter data to form a unified analysis unit, providing clear data support for subsequent clustering and classification of operating conditions.

[0066] Furthermore, cluster analysis is performed on the obtained multidimensional parameter feature vector sequence to obtain multiple parameter clusters, and multiple boiler operating conditions are determined based on these parameter clusters.

[0067] Specifically, an improved K-means clustering algorithm is used for cluster analysis: First, the optimal number of clusters K is determined by the elbow rule, that is, by traversing the range of K=2 to K=8, calculating the sum of squares within each cluster (WCSS) corresponding to each K value and plotting the change curve, and selecting the K value when the curve shows a clear inflection point (such as K=3) as the final number of clusters.

[0068] Next, the key parameters of the algorithm are configured. The initial centroid is selected using K-means++ to avoid clustering bias. After eliminating the difference in the units between parameters through Z-score standardization, Euclidean distance is used to measure vector similarity. At the same time, an upper limit of 500 iterations is set as the termination condition of the algorithm.

[0069] Ultimately, under the above parameter configuration, cluster analysis will group vectors with similar features into the same cluster. Since vectors with similar features correspond to similar operating states of the boiler, a single parameter cluster can represent a boiler operating condition.

[0070] For example, after clustering, clusters may be formed such as "high load stable cluster" (corresponding to pressure 1.2-2.5MPa, temperature 190-230℃, load 75%-100%), "low load cluster" (corresponding to pressure 0.4-0.8MPa, temperature 160-180℃, load 30%-50%), and "start-stop transition cluster" (corresponding to pressure 0.2-1.6MPa, temperature 140-210℃, load 20%-80%). The parameter range corresponding to each cluster represents a specific type of boiler operating condition.

[0071] Furthermore, each data collection timestamp in the boiler operating parameter time series is labeled with the corresponding boiler operating condition, resulting in operating condition labeling results. Since each data collection timestamp corresponds to a set of parameter vectors, and these vectors have been clustered into a certain parameter cluster (i.e., a certain operating condition), the operating condition label can be directly bound to the timestamp. For example, the timestamp "May 10, 2024, 14:00" can be labeled as "high load stable operating condition," and "May 10, 2024, 08:00" can be labeled as "start-stop transition operating condition," thereby establishing a one-to-one correspondence between timestamps and operating conditions.

[0072] Finally, based on the operating condition labeling results, all infrared image data belonging to the same boiler operating condition are extracted from the infrared thermal imaging time series and arranged in the order of acquisition time to construct the infrared image sequence corresponding to each boiler operating condition, thus obtaining multiple operating condition infrared image sequences.

[0073] Each image in the infrared thermal imaging time sequence is accompanied by an acquisition timestamp. By matching the timestamps with the operating condition markings, all infrared images belonging to the same operating condition can be filtered out. For example, by filtering out all infrared images with the timestamp marked "high load stable operating condition" and then sorting them from earliest to latest acquisition time, a "high load stable operating condition infrared image sequence" can be formed. Similarly, "low load operating condition infrared image sequence" and "start-stop transition operating condition infrared image sequence" can be constructed.

[0074] Ultimately, through the matching and sorting steps described above, each infrared image sequence obtained under a specific working condition can fully reflect the coating's change process over time under that condition. This provides dedicated image data support for subsequent calculations of the coating degradation coefficient for a single working condition, effectively eliminating interference between different working conditions and ensuring that the coating degradation assessment results are more consistent with actual operating conditions.

[0075] S130: Perform time-series analysis on infrared image sequences under multiple operating conditions to obtain the coating degradation coefficient for each boiler operating condition;

[0076] In this embodiment of the application, in order to reduce the complexity of data processing and obtain the coating degradation status under various boiler operating conditions efficiently and accurately, it is necessary to first set a preset interval time period to sample the infrared image sequence of the operating conditions, and then analyze it through a dedicated evaluator to obtain the coating degradation degree sequence, and then calculate the coating degradation coefficient to achieve accurate quantitative evaluation of coating degradation under different operating conditions.

[0077] Specifically, the first step is to obtain the preset time interval. Considering that the duration of each working condition is very long and the number of images collected is large, directly evaluating all images together would increase the complexity of data processing. At the same time, the coating degradation coefficient needs to reflect the degree of degradation per unit time. Therefore, the unit time is set to 1 month as the preset time interval, which not only meets the time dimension requirements of the degradation coefficient, but also effectively reduces the number of images to be analyzed.

[0078] Furthermore, based on a preset time interval, multiple infrared image sequences for different working conditions are sampled sequentially. From the infrared image sequence of each working condition, images are extracted at monthly intervals to obtain multiple sampled infrared image sequences, thereby reducing the complexity of subsequent data processing.

[0079] Furthermore, multiple coating degradation evaluators are retrieved based on multiple boiler operating conditions. Each operating condition is matched with a dedicated evaluator to adapt to its unique degradation pattern. Each sampled infrared image sequence is input into the corresponding coating degradation evaluator, and the evaluator analyzes the coating changes between adjacent sampled images to obtain the coating degradation degree sequence corresponding to each boiler operating condition.

[0080] Finally, based on the coating degradation sequence corresponding to each boiler operating condition and combined with a preset one-month unit time, the coating degradation coefficient for each boiler operating condition is calculated. This coefficient can intuitively reflect the monthly degradation degree of the coating under each operating condition.

[0081] This step, through sampling dimensionality reduction and analysis using a dedicated evaluator, accurately quantifies the coating degradation under various working conditions while ensuring data processing efficiency, providing reliable single-condition data support for subsequent comprehensive evaluation of the overall coating condition.

[0082] Step S130 in the method provided in this application embodiment includes:

[0083] Get the preset interval time period;

[0084] Based on the preset interval time period, the multiple working condition infrared image sequences are sampled in time sequence to obtain multiple sampled infrared image sequences;

[0085] Based on the multiple boiler operating conditions, multiple corresponding coating degradation evaluators are retrieved, and each of the sampled infrared image sequences is input into the corresponding coating degradation evaluator to obtain the coating degradation degree sequence corresponding to each boiler operating condition.

[0086] Based on the coating degradation sequence corresponding to each boiler operating condition, the coating degradation coefficient for each boiler operating condition is obtained.

[0087] In this embodiment of the application, in order to reduce the complexity of data processing when the operating time of each boiler is long and the number of corresponding infrared images is large, and at the same time accurately quantify the degree of coating degradation per unit time under different operating conditions, it is necessary to achieve accurate evaluation of the coating degradation state under each operating condition by setting interval sampling, calling a dedicated evaluator, and accumulating and calculating the coating degradation coefficient.

[0088] Specifically, the first step is to obtain the preset time interval. Since the duration of each boiler's operating condition is usually quite long, the corresponding number of infrared thermal imaging images collected will increase significantly. If all images are analyzed directly, the complexity of data processing will be significantly increased, affecting the detection efficiency.

[0089] Meanwhile, the coating degradation coefficient needs to reflect the degree of degradation per unit time. After comprehensive consideration, the unit time is set to 1 month as the preset interval period. This can effectively reduce the number of images to be analyzed and adapt to the time dimension requirements of the degradation coefficient, so as to ensure that the degree of degradation calculated in the subsequent calculation matches the actual time period.

[0090] Furthermore, based on a predetermined preset time interval, multiple working condition infrared image sequences are sampled sequentially to obtain multiple sampled infrared image sequences.

[0091] For example, for the "high-load stable working condition infrared image sequence", if the working condition continues for 6 months and 1 frame of image is collected every hour, and sampling is performed at a preset interval of 1 month, representative frames can be selected from the monthly images to finally obtain a "high-load stable working condition sampling infrared image sequence" composed of 6 sampled images. Through the above sampling method, the amount of data is effectively reduced and the analysis efficiency is improved while preserving the key trend of coating degradation.

[0092] Furthermore, after sampling is completed, multiple coating degradation evaluators are retrieved according to multiple boiler operating conditions. Each sampled infrared image sequence is input into the corresponding coating degradation evaluator to obtain the coating degradation degree sequence corresponding to each boiler operating condition.

[0093] The method provided in this application embodiment, which involves "retrieving multiple corresponding coating degradation evaluators based on the multiple boiler operating conditions, inputting each of the sampled infrared image sequences into the corresponding coating degradation evaluators, and obtaining a coating degradation degree sequence corresponding to each boiler operating condition," includes:

[0094] The first boiler operating condition is obtained from the plurality of boiler operating conditions, the first sampled infrared image sequence corresponding to the first boiler operating condition is extracted, and the first coating degradation evaluator bound to the first boiler operating condition is retrieved.

[0095] The first sampled infrared image sequence is paired up from each pair of adjacent sampled infrared images to obtain multiple sampled infrared image pairs, wherein each sampled infrared image pair includes a preceding infrared image and a following infrared image;

[0096] By iterating through multiple sampled infrared image pairs, the first sampled infrared image pair is obtained;

[0097] The first preceding infrared image and the first following infrared image in the first sampled infrared image pair are respectively input into the preceding image receiver and the following image receiver of the first coating degradation evaluator, and the first coating degradation degree is output.

[0098] Following the method of obtaining the first coating degradation degree of the first sampled infrared image pair, the coating degradation degree of the remaining sampled infrared image pairs is obtained sequentially to obtain the first coating degradation degree sequence;

[0099] Following the method of obtaining the first coating degradation sequence for the first boiler operating condition, the coating degradation sequence for the remaining boiler operating conditions is obtained, thus obtaining the coating degradation sequence for each boiler operating condition.

[0100] In this embodiment of the application, in order to make the coating degradation analysis under different boiler operating conditions more in line with their respective degradation patterns, it is necessary to match a dedicated coating degradation evaluator for each operating condition, and to accurately capture the subtle changes of the coating over time by analyzing adjacent image pairs, so as to generate a degradation degree sequence that can reflect the actual degradation process of the coating under each operating condition.

[0101] Specifically, firstly, a first boiler operating condition (e.g., high-load stable operating condition) is selected from multiple boiler operating conditions, and the corresponding first sampled infrared image sequence is extracted simultaneously. This sequence is based on the infrared image sequence of the high-load stable operating condition sampled at a preset interval (i.e., 1 month) and can reflect the state of the coating at different time points under this operating condition.

[0102] At the same time, the first coating degradation evaluator, which is linked to the operating conditions of the first boiler, is retrieved to accurately analyze the degradation pattern of the coating under this operating condition.

[0103] In the method provided in this application embodiment, the construction steps of the "first coating degradation evaluator" include:

[0104] Obtain the device model of the target boiler economizer, and extract the infrared thermal imaging records of the same model boiler economizer and the same operating conditions based on the device model and the first boiler operating conditions to obtain historical infrared thermal imaging data.

[0105] Based on the preset interval time period and the historical infrared thermal imaging data, a sample infrared image pair set is constructed. Each sample infrared image pair in the sample infrared image pair set includes a sample preceding infrared image and a sample following infrared image.

[0106] The coating degradation degree is labeled for each sample infrared image pair in the sample infrared image pair set to obtain the sample coating degradation degree set;

[0107] A coating degradation evaluator architecture is set up, which includes a preceding image receiver and a following image receiver;

[0108] Based on the set of sample infrared images and the set of sample coating degradation degrees, the coating degradation evaluator architecture is trained to obtain the first coating degradation evaluator;

[0109] The trained first coating degradation evaluator is bound and stored with the first boiler operating conditions.

[0110] In this embodiment of the application, in order for the first coating degradation evaluator to accurately identify the degradation characteristics of the coating under the operating conditions of the first boiler, it is necessary to construct training samples based on historical data of the same type of equipment and the same operating conditions, and through a standardized annotation and training process, enable the evaluator to have targeted degradation identification capabilities, so as to ensure that when the subsequent input sampling infrared image sequence, it can output degradation degree data that accurately reflects the coating degradation under the operating conditions.

[0111] Specifically, the first step is to obtain the equipment model of the target boiler economizer. Different models of boiler economizers may differ in structural design and coating material selection, and these differences will directly affect the degradation pattern of the coating. Therefore, based on the equipment model and the operating conditions of the first boiler, it is necessary to extract the infrared thermal imaging records of the same model of boiler economizer under the same operating conditions from the database to form historical infrared thermal imaging data.

[0112] For example, if the target equipment model is "GL-2000" and the first operating condition is a high-load stable operating condition, then all infrared thermal imaging data of "GL-2000" model boilers under the high-load stable operating condition are extracted to ensure that these historical data are highly consistent with the coating degradation environment of the target equipment.

[0113] Furthermore, a set of sample infrared image pairs is constructed according to a preset interval period (i.e., 1 month). Taking historical data covering 3 years and a preset interval period of 1 month as an example, images are selected from the historical infrared thermal imaging time series of each device of the same model at 1-month intervals. Images from two adjacent intervals are paired, such as images from the 1st month and the 2nd month, images from the 2nd month and the 3rd month, etc. Each pair contains an infrared image preceding the sample representing the early coating state and an infrared image following the sample representing the later coating state. Through the above pairing method, the natural degradation process of the coating over time is simulated to ensure that the samples can reflect the changes in the coating under the working condition within a standard time period.

[0114] Furthermore, the coating degradation degree of each sample infrared image pair in the sample infrared image pair set is labeled to form a sample coating degradation degree set. The labeling process needs to combine the characteristics of infrared thermal imaging images and refer to industry coating degradation assessment standards. Professional technicians determine the degree of coating degradation in each pair of sample images and assign specific values.

[0115] For example, if a sample has no obvious abnormal high temperature area in the preceding infrared image, but an abnormal high temperature area with an area ratio of 1% appears in the subsequent infrared image, the coating degradation degree of the sample pair can be marked as 1.2% based on the high temperature corrosion resistance performance parameters of the coating material, thus ensuring the accuracy and objectivity of the labeling results for each sample.

[0116] Furthermore, a coating degradation evaluator architecture is set up, using an improved U-Net and CNN fusion network: the input layer has two branches, each branch contains 3 convolutional blocks (each convolutional block contains 2 3×3 convolutional layers + BN layer + ReLU activation function) and 2 max pooling layers (2×2 stride); the feature fusion layer uses an attention mechanism to weightedly fuse high-dimensional features from the two branches; the output layer is a 1×1 convolutional layer (Sigmoid activation), outputting a degradation value of 0-100%.

[0117] Simultaneously, the architecture needs to include a preceding image receiver and a following image receiver, used to receive adjacent infrared image pairs to be analyzed. Furthermore, the architecture needs to integrate functional modules such as image feature extraction, difference calculation, and degradation degree mapping to support subsequent training and degradation degree output.

[0118] After completing the architecture setup of the coating degradation evaluator, the architecture is trained based on the set of sample infrared image pairs and the set of sample coating degradation degrees. Specifically, the first coating degradation evaluator is constructed based on an algorithm combining convolutional neural networks (CNN) and temporal difference analysis.

[0119] First, the preceding and following infrared images from the sample infrared image pair set are input into the preceding and following image receivers of the evaluator architecture, respectively. Through convolutional and pooling layers of a CNN, feature extraction is performed on both images. Specifically, basic visual features such as the temperature distribution texture of the coating, the edges of abnormally high-temperature areas, and the integrity of the coating coverage are extracted first. Then, fully connected layers transform these basic features into 64-dimensional high-dimensional feature vectors to accurately capture key information about the coating state.

[0120] Furthermore, a difference calculation module is set up within the architecture to compare the high-dimensional feature vectors of the preceding and subsequent images dimension by dimension, calculate the cosine similarity between the two in the feature space, thereby quantifying the degree of difference in coating features between the two images and obtaining the feature difference value in the preliminary stage.

[0121] Furthermore, a three-layer fully connected network (with 64-32-16 hidden layer nodes respectively) is used to map the feature difference values ​​to specific coating degradation degrees.

[0122] Furthermore, labeled data from the sample coating degradation set are introduced, and feature difference values ​​are correlated with the corresponding sample coating degradation to construct a mapping relationship model.

[0123] During training, mean squared error (MSE) was used as the loss function, and the Adam optimizer was selected (initial learning rate set to 0.001, β1=0.9, β2=0.999). The backpropagation algorithm was used to continuously adjust the convolution kernel parameters, fully connected layer weights, and coefficients of the difference calculation module of the CNN, so that the feature difference value output by the evaluator could be consistent with the actual labeled coating degradation degree.

[0124] In addition, the model parameters are continuously iterated and optimized. In each iteration, a batch of infrared images with a size of 32 samples are selected as input to the architecture. The loss value between the output value and the labeled degradation degree is calculated. If the loss value is greater than the preset threshold (0.001), the parameters are adjusted along the gradient descent direction. The upper limit of the iteration is set to 500 rounds. If the loss value is less than or equal to the preset threshold, it means that the model can accurately judge the degree of coating degradation through the differences in image features.

[0125] Meanwhile, to avoid model overfitting, a dropout mechanism (with a dropout probability of 0.3) is introduced during training to randomly discard some neural network nodes. Cross-validation is used to divide the set of sample infrared image pairs and the set of sample coating degradation into training and validation sets in a 7:3 ratio. After parameter training is completed using the training set, the model's generalization ability is tested through the validation set. An early stopping mechanism is triggered when the validation set loss increases for 10 consecutive rounds to ensure that the model can still output accurate coating degradation when faced with infrared image pairs under the same conditions that were not trained.

[0126] Through the above training process based on sample data, the coating degradation evaluator architecture is finally equipped with the coating degradation identification capability for the first boiler operating condition, forming the first coating degradation evaluator that can accurately output the coating degradation degree under this operating condition.

[0127] Finally, the trained first coating degradation evaluator is bound and stored with the first boiler operating condition. For example, it is associated with the "high load stable condition" tag and stored in the evaluator database. Subsequently, when coating degradation under high load stable condition needs to be analyzed, the evaluator can be directly retrieved without retraining or adaptation, ensuring the efficiency of the detection process.

[0128] Furthermore, the extracted first sampled infrared image sequence is paired with adjacent sampled infrared images to obtain multiple sampled infrared image pairs (including preceding and subsequent infrared images) in order to accurately capture the continuous change process of the coating over time in the sequence.

[0129] In particular, the acquisition time sequence must be strictly followed during the pairing process to ensure that the time relationship between the preceding and subsequent infrared images is accurate, thereby reflecting the natural degradation process of the coating over a continuous time period.

[0130] For example, if the first sampled infrared image sequence contains 6 frames of images arranged chronologically (corresponding to images collected in January, February, March, April, May, and June 2024 respectively), then the image pairs of January and February, February and March, ... May and June will be combined to form 5 sampled infrared image pairs.

[0131] Furthermore, multiple sampled infrared image pairs are traversed, and the sampled infrared image pair with the highest ranking is selected as the first sampled infrared image pair (i.e., the image pair from January and February). During the selection process, the image acquisition timestamps need to be verified again to confirm that the time of the preceding image is earlier than that of the subsequent image, so as to avoid the time order being reversed due to data storage deviations.

[0132] Furthermore, the first preceding infrared image (January 2024 image) and the first following infrared image (February 2024 image) from the first sampled infrared image pair are respectively input into the preceding image receiver and the following image receiver of the first coating degradation evaluator. Based on the recognition logic formed by previous training, the first coating degradation evaluator will automatically extract key information such as the temperature distribution characteristics and abnormal region contours of the coating in the two images, and compare and analyze the feature differences between the two frames of images.

[0133] For example, if the area of ​​the abnormally high temperature region of the coating in the subsequent image increases by 0.8% compared with the previous image, the evaluator will combine the coating degradation law under this operating condition, convert this feature difference into a specific value, and output the first coating degradation degree, such as 1%. This value represents the degree of coating degradation under the first boiler operating condition from January to February.

[0134] Similarly, the remaining sampled infrared image pairs are processed sequentially in the same way as the first coating degradation degree obtained from the first sampled infrared image pair. Taking the image pair from February and March as an example, it is also input into the first coating degradation evaluator to obtain a second coating degradation degree of 0.9%; the image pair from March and April is processed to obtain a third coating degradation degree of 1.1%, until the analysis of all 5 sampled infrared image pairs is completed.

[0135] Furthermore, by arranging the obtained degradation degrees in chronological order according to the corresponding time periods [1.0%, 0.9%, 1.1%, 1.0%, 1.2%], a first coating degradation degree sequence corresponding to the first boiler operating condition can be formed. This sequence fully presents the monthly degradation of the coating under this operating condition over 6 months.

[0136] Similarly, the same operation was performed on the other boiler operating conditions (i.e., low load condition and start-stop transition condition) by obtaining the first coating degradation sequence for the first boiler operating condition. Ultimately, each boiler operating condition generated its own corresponding coating degradation sequence, and each sequence accurately reflected the coating degradation trajectory over time under the corresponding condition, providing basic data matching the operating condition characteristics for subsequent calculations of the coating degradation coefficient for each condition.

[0137] The method provided in this application embodiment, "obtaining the coating degradation coefficient for each boiler operating condition based on the coating degradation degree sequence corresponding to each boiler operating condition," includes:

[0138] Set the initial integrity of the coating;

[0139] The cumulative decrease of the coating degradation degree sequence corresponding to each boiler operating condition is calculated. The current coating integrity degree corresponding to each boiler operating condition is obtained by subtracting each coating degradation degree from the initial integrity degree of the coating.

[0140] Based on the current coating integrity corresponding to each boiler operating condition, the coating degradation coefficient for each boiler operating condition is calculated. The coating degradation coefficient is equal to 1 minus the coating integrity.

[0141] In this embodiment of the application, in order to accurately quantify the cumulative degradation degree of the coating from the time of service under various boiler operating conditions, it is necessary to convert the time-series coating degradation degree into an intuitive coating degradation coefficient index by setting the initial integrity degree of the coating and cumulatively calculating the current integrity degree of the coating, so as to ensure that the degree of damage to the coating under different operating conditions can be clearly judged based on the coefficient.

[0142] Specifically, the initial integrity of the coating is first determined. It is considered that even if a new coating has just been applied, there may be minor defects that are difficult to detect with the naked eye due to the application process and environmental conditions, and it is not in an ideal 100% perfect state.

[0143] Therefore, before the boiler economizer is put into use for the first time, the coating needs to be benchmarked by infrared thermal imaging technology. That is, the temperature distribution image of the coating is captured by infrared thermal imaging equipment, and the integrity of the coating is evaluated by temperature field analysis algorithm.

[0144] If there are no obvious abnormal high-temperature areas in the image and the temperature distribution uniformity meets the coating material standard requirements, the initial integrity may be 98%; if there are extremely small areas of abnormal temperature, the initial integrity may be set to 95% based on the high-temperature corrosion resistance parameters of the coating material. This initial value is not a fixed 100%, but is based on the actual health benchmark detected to ensure that the starting point of subsequent degradation analysis is closer to the true initial state of the coating.

[0145] Furthermore, after determining the initial integrity of the coating, the coating degradation sequence corresponding to each boiler operating condition is cumulatively decreased to obtain the current coating integrity corresponding to each operating condition.

[0146] The coating degradation sequence is time-series data obtained by analyzing adjacent sampled image pairs through an evaluator. For example, the coating degradation sequence for the first boiler operating condition (high load stable condition) is [1.0%, 0.9%, 1.1%, 1.0%, 1.2%], with an initial integrity of 98%.

[0147] In the specific calculation, starting from the initial integrity, the degradation degree of each coating in the sequence is subtracted in turn. That is, first, the first coating degradation degree of 1.0% is subtracted from 98%, and the integrity degree after the first period is 97%; then, the second coating degradation degree of 0.9% is subtracted from 97%, and 96.1% is obtained; the calculation is repeated until the last coating degradation degree of 1.2% in the sequence is subtracted, and finally the current coating integrity degree under this working condition is obtained (i.e., 98%-1.0%-0.9%-1.1%-1.0%-1.2%=92.8%).

[0148] Finally, the coating degradation coefficient is calculated based on the current coating integrity corresponding to each boiler's operating condition.

[0149] The coating integrity represents the current remaining integrity of the coating (i.e., 92.8% means that 92.8% of the coating is still in good condition), while the coating degradation coefficient needs to reflect the degree of degradation. Therefore, it is calculated using the formula "coating degradation coefficient = 1 - coating integrity".

[0150] For example, if the current integrity of the first boiler operating condition is 92.8%, then its coating degradation coefficient is 1-92.8%=7.2%; if the current integrity of another low-load operating condition is 96.5%, then its coating degradation coefficient is 3.5%.

[0151] Ultimately, through the above calculation method, the coating degradation state under different working conditions can be transformed into a coating degradation coefficient value of a unified dimension, which intuitively reflects the difference in damage to the coating under each working condition. The higher the coating degradation coefficient, the more severe the coating degradation under that working condition. For example, the coating degradation coefficient of 7.2% under high load stable working condition is higher than that under low load working condition of 3.5%, indicating that the high load condition causes greater damage to the coating, providing a clear quantitative comparison basis for subsequent comprehensive evaluation of the overall coating degradation.

[0152] S140: Based on the multiple boiler operating conditions and the coating degradation coefficient of each boiler operating condition, the comprehensive coating degradation coefficient of the target boiler economizer is obtained, and coating failure early warning is performed based on the comprehensive coating degradation coefficient.

[0153] In this embodiment of the application, in order to avoid the deviation in judging the overall state of the coating due to relying solely on the coating degradation coefficient under a single working condition, and to ensure that coating failure can be warned in advance to avoid missing maintenance opportunities, it is necessary to calculate the working condition weights and obtain a comprehensive degradation coefficient by weighted summation, and compare it with the warning threshold to provide a coating failure warning, so as to comprehensively and accurately assess the overall degradation state of the coating and ensure the stable operation of the equipment.

[0154] Specifically, firstly, based on the operating condition marking results, the duration of multiple operating conditions for multiple boiler operating conditions is obtained. The operating condition marking results record the operating condition corresponding to each collection timestamp in the boiler operating parameter time series. By statistically analyzing the total duration of all collection timestamps under the same operating condition, the actual running time of each operating condition can be obtained.

[0155] Furthermore, the total operating time of the target boiler economizer is obtained. Based on the duration of multiple operating conditions and the total operating time, the time proportion of each operating condition is obtained as a weighting coefficient corresponding to each boiler operating condition. This weighting coefficient reflects the degree of contribution of different operating conditions to coating degradation; the longer the operating time of an operating condition, the greater its impact on the overall degradation of the coating.

[0156] Furthermore, the coating degradation coefficients for each boiler operating condition are weighted and summed with their corresponding weighting coefficients to obtain the comprehensive coating degradation coefficient of the target boiler economizer. This value comprehensively reflects the overall degree of coating degradation under the combined effect of all operating conditions, avoiding the bias in evaluation caused by ignoring the differences in operating time under different conditions.

[0157] Furthermore, after calculating the overall coating degradation coefficient, a preset coating failure early warning threshold is obtained, and the overall coating degradation coefficient is compared with the coating failure early warning threshold. When the overall coating degradation coefficient is greater than the coating failure early warning threshold, a coating failure early warning message is generated and output to the user terminal to remind maintenance personnel to take timely maintenance measures to prevent further coating failure from affecting equipment operating efficiency and service life.

[0158] This step calculates the comprehensive degradation coefficient by combining the weight of operating conditions and achieves timely early warning based on the comparison of coating failure warning thresholds. It not only comprehensively considers the impact of different operating conditions on the coating, but also triggers maintenance reminders in a timely manner. It effectively solves the problems of difficulty in quantifying the overall state of the coating and untimely early warning in traditional testing, and provides reliable support for the refined management of boiler economizer coatings.

[0159] Step S140 in the method provided in this application embodiment includes:

[0160] Based on the operating condition marking results, the duration of multiple operating conditions for multiple boiler operating conditions is obtained;

[0161] The total operating time of the target boiler economizer is obtained. Based on the duration of the multiple operating conditions and the total operating time, the time ratio of the multiple operating conditions is obtained as the weighting coefficient corresponding to each boiler operating condition.

[0162] The coating degradation coefficient of each boiler operating condition is weighted and summed with the corresponding weighting coefficient to obtain the comprehensive coating degradation coefficient of the target boiler economizer.

[0163] Obtain a preset coating failure early warning threshold, and compare the overall coating degradation coefficient with the coating failure early warning threshold;

[0164] When the overall degradation coefficient of the coating is greater than the coating failure warning threshold, a coating failure warning message is generated and output to the user terminal.

[0165] In this embodiment of the application, in order to comprehensively consider the impact of different operating conditions on the coating of the boiler economizer and avoid relying solely on the coating degradation coefficient of a single operating condition to make a one-sided judgment on the overall state of the coating, it is necessary to calculate the weight of the operating conditions and obtain a comprehensive degradation coefficient by weighting, and combine it with the coating failure early warning threshold to realize coating failure reminder, so as to comprehensively and accurately assess the overall degradation state of the coating and provide a reliable basis for equipment maintenance.

[0166] Specifically, based on the operating condition marking results, the duration of multiple operating conditions for multiple boiler operating conditions is first obtained.

[0167] The operating condition marking result records the specific operating condition corresponding to each collection timestamp in the boiler operating parameter time series. By statistically analyzing the cumulative duration of all collection timestamps under the same operating condition, the actual running time of each operating condition can be obtained.

[0168] For example, the equipment operates for a cumulative total of 200 days under high-load stable conditions, 100 days under low-load conditions, and 50 days under start-stop transition conditions. These duration data directly reflect the duration of the coating's effect under different conditions and are the core basis for subsequent weight calculations.

[0169] Furthermore, the total operating time of the target boiler economizer is obtained. Based on the duration of multiple operating conditions and the total operating time, the time percentage of each operating condition is obtained and used as the weighting coefficient for each boiler operating condition. The total operating time is the sum of the duration of all operating conditions (i.e., 200 + 100 + 50 = 350 days), and the time percentage of each operating condition is the ratio of the duration of a single operating condition to the total duration.

[0170] For example, the time proportion of high-load stable operating conditions is 200 / 350≈0.57, the time proportion of low-load operating conditions is 100 / 350≈0.29, and the time proportion of start-stop transition operating conditions is 50 / 350≈0.14. This time proportion, as a weighting coefficient, can reasonably reflect the contribution of different operating conditions to the overall degradation of the coating. The longer the operating time of an operating condition, the higher its weighting coefficient, and the greater its impact on the overall degradation coefficient.

[0171] Furthermore, the coating degradation coefficients for each boiler operating condition are weighted and summed with their corresponding weighting coefficients to obtain the comprehensive coating degradation coefficient for the target boiler economizer. Assuming the coating degradation coefficient is 8.5% for high-load stable operation, 4.2% for low-load operation, and 6.8% for start-stop transition operation, the comprehensive degradation coefficient is calculated as 8.5%×0.57+4.2%×0.29+6.8%×0.14≈7.015%. This value integrates the degradation effects of all operating conditions and accurately reflects the overall degradation degree of the coating throughout the equipment's operating cycle.

[0172] Furthermore, a preset coating failure warning threshold is obtained, and the overall coating degradation coefficient is compared with the coating failure warning threshold to accurately determine whether the current overall state of the coating has approached or exceeded the safe operating limit.

[0173] The coating failure warning threshold needs to be determined by combining the coating material's tolerance performance, the boiler economizer's operational safety standards, and maintenance experience. For example, for a certain type of high-temperature resistant coating, when the degradation degree exceeds 9%, its heat insulation and corrosion resistance performance will decrease significantly, which may affect the boiler's thermal efficiency or cause equipment failure. Therefore, the coating failure warning threshold is set at 9%.

[0174] Specifically, when the overall degradation coefficient of the coating exceeds the coating failure warning threshold, a coating failure warning message will be generated and output to the user terminal.

[0175] For example, if the calculated overall degradation coefficient of the coating is 9.2%, which exceeds the 9% coating failure warning threshold, an automatic warning message will be generated containing the message "The overall degradation coefficient of the target boiler economizer coating is 9.2%, which exceeds the preset warning threshold of 9%. The coating is at risk of failure. It is recommended to arrange maintenance within 7 days to avoid affecting the equipment's operating efficiency and safety." This message will be pushed through the user's operation and maintenance management platform, SMS, or email to ensure that operation and maintenance personnel can receive timely reminders, plan maintenance work in advance, and avoid missing the best maintenance opportunity, which could lead to further coating failure.

[0176] Meanwhile, during the output of early warning information, the calculation basis of the coating's overall degradation coefficient must also be included, including the duration, weight coefficient, and corresponding coating degradation coefficient of each working condition. This facilitates maintenance personnel to trace the impact of different working conditions on the coating, formulate targeted maintenance plans, and achieve a comprehensive assessment and timely early warning of the overall coating status, effectively ensuring the stable operation and service life of the boiler economizer.

[0177] The embodiments of this application, through the specific implementation methods described above, achieve the following technical effects:

[0178] This application proposes a boiler economizer coating failure detection method based on infrared thermal imaging. First, in response to a coating failure detection command, the method parses the command to obtain the current detection time. A data extraction time window is constructed by combining this with the target boiler economizer's service start time. The method connects to a thermal imaging and operating parameter database, retrieving infrared thermal imaging time series and boiler operating parameter time series with acquisition timestamps within the corresponding time range. Next, the boiler operating parameter time series is vectorized to construct a multi-dimensional parameter feature vector sequence. Multiple boiler operating conditions are then segmented through cluster analysis. Based on the operating condition labeling results, the infrared thermal imaging time series is segmented to construct a dedicated infrared image sequence for each operating condition. Subsequently, based on a preset interval time period... To reduce data processing complexity, the system performs time-series sampling of infrared image sequences under operating conditions. It then retrieves a dedicated coating degradation evaluator bound to the operating conditions, inputs the sampled image sequences into the evaluator to obtain coating degradation degree sequences for each operating condition, and calculates the current integrity by cumulatively decreasing the initial integrity, thereby obtaining the coating degradation coefficient for each operating condition. Finally, based on the operating condition marking results, it obtains the duration of each operating condition and the total operating time of the equipment, calculates the proportion of operating condition time as a weighting coefficient, and weights and sums the coating degradation coefficients for each operating condition with the weighting coefficients to obtain the comprehensive coating degradation coefficient. This coefficient is then compared with a preset coating failure warning threshold. If the coefficient exceeds the threshold, a warning message is generated and output to the user terminal, achieving accurate warning of coating failure.

[0179] The method provided in this application, through the technical solution of "command response and data retrieval - operating condition division and image sequence construction - time-series sampling and degradation assessment - comprehensive coefficient calculation and failure early warning", solves the problems of traditional boiler economizer coating detection, such as the need for shutdown operation, evaluation bias caused by mixed operating condition data, lack of quantitative indicators and early warning capabilities. It avoids untimely coating maintenance caused by incomplete data, operating condition interference or delayed early warning, and improves the real-time performance, accuracy and reliability of coating failure detection, providing technical support for the safe and stable operation and service life extension of boiler economizers.

[0180] Example 2, as shown in the appendix Figure 2 As shown, based on the inventive concept of the boiler economizer coating failure detection method based on infrared thermal imaging provided in Embodiment 1, this application also provides a boiler economizer coating failure detection system based on infrared thermal imaging, specifically including:

[0181] The instruction response and sequence retrieval module 01 is used to respond to the coating failure detection instruction and retrieve the infrared thermal imaging timing sequence of the target boiler economizer and the boiler operating parameter timing sequence based on the coating failure detection instruction.

[0182] The operating condition division and sequence construction module 02 is used to divide the operating conditions according to the boiler operating parameter time sequence, determine multiple boiler operating conditions, and segment the infrared thermal imaging time sequence based on the multiple boiler operating conditions to construct multiple operating condition infrared image sequences.

[0183] The time series analysis and coefficient acquisition module 03 is used to perform time series analysis on infrared image sequences under multiple operating conditions to obtain the coating degradation coefficient for each boiler operating condition.

[0184] The coefficient integration and failure early warning module 04 is used to obtain the overall coating degradation coefficient of the target boiler economizer based on the multiple boiler operating conditions and the coating degradation coefficient of each boiler operating condition, and to perform coating failure early warning based on the overall coating degradation coefficient.

[0185] In one embodiment, the instruction response and sequence retrieval module 01 is further configured to:

[0186] The system receives a coating failure detection command uploaded by the user terminal, which can be a user-triggered command or a timed trigger command; it parses the coating failure detection command to obtain the current detection time, and combines it with the service start time of the target boiler economizer to form a data extraction time window; it connects to the thermal imaging database and the operating parameter database, and retrieves the infrared thermal imaging time sequence and boiler operating parameter time sequence of the target boiler economizer based on the data extraction time window, both of which have acquisition timestamps.

[0187] In one embodiment, the working condition division and sequence construction module 02 is further used for:

[0188] The boiler operating parameter time series is transformed into vector form to construct a multi-dimensional parameter feature vector sequence; cluster analysis is performed on the multi-dimensional parameter feature vector sequence to obtain multiple parameter clusters, and multiple boiler operating conditions are determined based on the multiple parameter clusters; each acquisition timestamp in the boiler operating parameter time series is marked as the corresponding boiler operating condition to obtain the operating condition marking result; based on the operating condition marking result, all infrared image data belonging to the same boiler operating condition are extracted from the infrared thermal imaging time series and arranged in order of acquisition time to construct the infrared image sequence corresponding to each boiler operating condition, thus obtaining multiple operating condition infrared image sequences.

[0189] In one embodiment, the time series analysis and coefficient acquisition module 03 is further used for:

[0190] A preset time interval is obtained; based on the preset time interval, the infrared image sequences of the multiple operating conditions are sampled in time to obtain multiple sampled infrared image sequences; multiple coating degradation evaluators are retrieved according to the multiple boiler operating conditions, and each sampled infrared image sequence is input into the corresponding coating degradation evaluator to obtain the coating degradation degree sequence corresponding to each boiler operating condition; based on the coating degradation degree sequence corresponding to each boiler operating condition, the coating degradation coefficient of each boiler operating condition is obtained.

[0191] Furthermore, the time series analysis and coefficient acquisition module 03 also includes:

[0192] A first boiler operating condition is obtained from the multiple boiler operating conditions. A first sampled infrared image sequence corresponding to the first boiler operating condition is extracted, and a first coating degradation evaluator bound to the first boiler operating condition is retrieved. Adjacent sampled infrared images in the first sampled infrared image sequence are paired to obtain multiple sampled infrared image pairs, each of which includes a preceding infrared image and a following infrared image. Multiple sampled infrared image pairs are traversed to obtain a first sampled infrared image pair. The first preceding infrared image and the first following infrared image in the first sampled infrared image pair are respectively input to the preceding image receiver and the following image receiver of the first coating degradation evaluator, and a first coating degradation degree is output. Following the method of obtaining the first coating degradation degree of the first sampled infrared image pair, the coating degradation degree of the remaining sampled infrared image pairs is obtained sequentially to obtain a first coating degradation degree sequence. Following the method of obtaining the first coating degradation degree sequence of the first boiler operating condition, the coating degradation degree sequence corresponding to the remaining boiler operating conditions is obtained to obtain a coating degradation degree sequence corresponding to each boiler operating condition.

[0193] Furthermore, the time series analysis and coefficient acquisition module 03 also includes:

[0194] Obtain the device model of the target boiler economizer; extract infrared thermal imaging records of the same model boiler economizer under the same operating conditions based on the device model and the first boiler operating conditions to obtain historical infrared thermal imaging data; construct a sample infrared image pair set according to the preset interval time period and the historical infrared thermal imaging data, each sample infrared image pair in the sample infrared image pair set includes a preceding infrared image and a following infrared image; label the coating degradation degree of each sample infrared image pair in the sample infrared image pair set to obtain a sample coating degradation degree set; set a coating degradation evaluator architecture, which includes a preceding image receiver and a following image receiver; train the coating degradation evaluator architecture based on the sample infrared image pair set and the sample coating degradation degree set to obtain the first coating degradation evaluator; bind and store the trained first coating degradation evaluator with the first boiler operating conditions.

[0195] Furthermore, the time series analysis and coefficient acquisition module 03 also includes:

[0196] Set the initial integrity of the coating; perform cumulative decreasing calculation on the coating degradation sequence corresponding to each boiler operating condition, subtract each coating degradation from the initial integrity of the coating in sequence to obtain the current coating integrity corresponding to each boiler operating condition; calculate the coating degradation coefficient for each boiler operating condition based on the current coating integrity corresponding to each boiler operating condition, where the coating degradation coefficient is equal to 1 minus the coating integrity.

[0197] In one embodiment, the coefficient synthesis and failure early warning module 04 is also used for:

[0198] Based on the operating condition marking results, the duration of multiple operating conditions for multiple boiler operating conditions is obtained; the total operating time of the target boiler economizer is obtained, and based on the duration of multiple operating conditions and the total operating time, the time ratio of multiple operating conditions is obtained as the weight coefficient corresponding to each boiler operating condition; the coating degradation coefficient of each boiler operating condition is weighted and summed with the corresponding weight coefficient to obtain the comprehensive coating degradation coefficient of the target boiler economizer. A preset coating failure warning threshold is obtained, and the comprehensive coating degradation coefficient is compared with the coating failure warning threshold; when the comprehensive coating degradation coefficient is greater than the coating failure warning threshold, coating failure warning information is generated and output to the user terminal.

[0199] It should be noted that the order of the embodiments described above 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 implementations, multitasking and parallel processing are possible or may be advantageous.

[0200] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0201] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for detecting boiler economizer coating failure based on infrared thermal imaging, characterized in that, The method includes: In response to the coating failure detection command, the infrared thermal imaging timing sequence of the target boiler economizer and the timing sequence of boiler operating parameters are retrieved based on the coating failure detection command. The operating conditions are divided according to the boiler operating parameter time sequence to determine multiple boiler operating conditions. The infrared thermal imaging time sequence is then segmented based on the multiple boiler operating conditions to construct multiple operating condition infrared image sequences. Time-series analysis was performed on infrared image sequences under multiple operating conditions to obtain the coating degradation coefficient for each boiler operating condition; The comprehensive coating degradation coefficient of the target boiler economizer is obtained based on the multiple boiler operating conditions and the coating degradation coefficient of each boiler operating condition, and coating failure early warning is performed based on the comprehensive coating degradation coefficient.

2. The method according to claim 1, characterized in that, In response to a coating failure detection command, the system retrieves the infrared thermal imaging timing sequence of the target boiler economizer and the timing sequence of boiler operating parameters based on the coating failure detection command, including: Receive coating failure detection instructions uploaded by the user terminal, wherein the coating failure detection instructions are user-triggered instructions or timed trigger instructions; The coating failure detection command is parsed to obtain the current detection time, and the data extraction time window is formed by combining it with the service start time of the target boiler economizer. The system connects to a thermal imaging database and an operating parameter database, and retrieves the infrared thermal imaging time series and boiler operating parameter time series of the target boiler economizer based on the data extraction time window. Both the infrared thermal imaging time series and the boiler operating parameter time series have acquisition timestamps.

3. The method according to claim 2, characterized in that, Based on the boiler operating parameter time series, operating conditions are divided to determine multiple boiler operating conditions. Then, based on these multiple boiler operating conditions, the infrared thermal imaging time series is segmented to construct multiple operating condition infrared image sequences, including: The boiler operating parameters time series are transformed into vector form to construct a multidimensional parameter feature vector sequence; Cluster analysis is performed on the multidimensional parameter feature vector sequence to obtain multiple parameter clusters, and multiple boiler operating conditions are determined based on the multiple parameter clusters. Each collection timestamp in the boiler operating parameter time series is marked as the corresponding boiler operating condition to obtain the operating condition marking result; Based on the operating condition marking results, all infrared image data belonging to the same boiler operating condition are extracted from the infrared thermal imaging time sequence and arranged in the order of acquisition time to construct the infrared image sequence corresponding to each boiler operating condition, thus obtaining multiple operating condition infrared image sequences.

4. The method according to claim 1, characterized in that, Time-series analysis was performed on infrared image sequences under multiple operating conditions to obtain the coating degradation coefficient for each boiler operating condition, including: Get the preset interval time period; Based on the preset interval time period, the multiple working condition infrared image sequences are sampled in time sequence to obtain multiple sampled infrared image sequences; Based on the multiple boiler operating conditions, multiple corresponding coating degradation evaluators are retrieved, and each of the sampled infrared image sequences is input into the corresponding coating degradation evaluator to obtain the coating degradation degree sequence corresponding to each boiler operating condition. Based on the coating degradation sequence corresponding to each boiler operating condition, the coating degradation coefficient for each boiler operating condition is obtained.

5. The method according to claim 4, characterized in that, Based on the multiple boiler operating conditions, multiple corresponding coating degradation evaluators are retrieved, and each of the sampled infrared image sequences is input into the corresponding coating degradation evaluator to obtain the coating degradation degree sequence corresponding to each boiler operating condition, including: The first boiler operating condition is obtained from the plurality of boiler operating conditions, the first sampled infrared image sequence corresponding to the first boiler operating condition is extracted, and the first coating degradation evaluator bound to the first boiler operating condition is retrieved. The first sampled infrared image sequence is paired up from each pair of adjacent sampled infrared images to obtain multiple sampled infrared image pairs, wherein each sampled infrared image pair includes a preceding infrared image and a following infrared image; By iterating through multiple sampled infrared image pairs, the first sampled infrared image pair is obtained; The first preceding infrared image and the first following infrared image in the first sampled infrared image pair are respectively input into the preceding image receiver and the following image receiver of the first coating degradation evaluator, and the first coating degradation degree is output. Following the method of obtaining the first coating degradation degree of the first sampled infrared image pair, the coating degradation degree of the remaining sampled infrared image pairs is obtained sequentially to obtain the first coating degradation degree sequence; Following the method of obtaining the first coating degradation sequence for the first boiler operating condition, the coating degradation sequence for the remaining boiler operating conditions is obtained, thus obtaining the coating degradation sequence for each boiler operating condition.

6. The method according to claim 5, characterized in that, The construction steps of the first coating degradation evaluator include: Obtain the device model of the target boiler economizer, and extract the infrared thermal imaging records of the same model boiler economizer and the same operating conditions based on the device model and the first boiler operating conditions to obtain historical infrared thermal imaging data. Based on the preset interval time period and the historical infrared thermal imaging data, a sample infrared image pair set is constructed. Each sample infrared image pair in the sample infrared image pair set includes a sample preceding infrared image and a sample following infrared image. The coating degradation degree is labeled for each sample infrared image pair in the sample infrared image pair set to obtain the sample coating degradation degree set; A coating degradation evaluator architecture is set up, which includes a preceding image receiver and a following image receiver; Based on the set of sample infrared images and the set of sample coating degradation degrees, the coating degradation evaluator architecture is trained to obtain the first coating degradation evaluator; The trained first coating degradation evaluator is bound and stored with the first boiler operating conditions.

7. The method according to claim 4, characterized in that, Based on the coating degradation degree sequence corresponding to each boiler operating condition, the coating degradation coefficient for each boiler operating condition is obtained, including: Set the initial integrity of the coating; The cumulative decrease of the coating degradation degree sequence corresponding to each boiler operating condition is calculated. The current coating integrity degree corresponding to each boiler operating condition is obtained by subtracting each coating degradation degree from the initial integrity degree of the coating. Based on the current coating integrity corresponding to each boiler operating condition, the coating degradation coefficient for each boiler operating condition is calculated. The coating degradation coefficient is equal to 1 minus the coating integrity.

8. The method according to claim 3, characterized in that, The overall coating degradation coefficient of the target boiler economizer is obtained based on the multiple boiler operating conditions and the coating degradation coefficient for each boiler operating condition, including: Based on the operating condition marking results, the duration of multiple operating conditions for multiple boiler operating conditions is obtained; The total operating time of the target boiler economizer is obtained. Based on the duration of the multiple operating conditions and the total operating time, the time ratio of the multiple operating conditions is obtained as the weighting coefficient corresponding to each boiler operating condition. The coating degradation coefficient of each boiler operating condition is weighted and summed with the corresponding weighting coefficient to obtain the comprehensive coating degradation coefficient of the target boiler economizer.

9. The method according to claim 1, characterized in that, Based on the comprehensive degradation coefficient of the coating, a coating failure early warning system is implemented, including: Obtain a preset coating failure early warning threshold, and compare the overall coating degradation coefficient with the coating failure early warning threshold; When the overall degradation coefficient of the coating is greater than the coating failure warning threshold, a coating failure warning message is generated and output to the user terminal.

10. A boiler economizer coating failure detection system based on infrared thermal imaging, characterized in that, The system is used to perform the boiler economizer coating failure detection method based on infrared thermal imaging as described in any one of claims 1-9, and the system comprises: The instruction response and sequence retrieval module is used to respond to the coating failure detection instruction and retrieve the infrared thermal imaging timing sequence of the target boiler economizer and the boiler operating parameter timing sequence based on the coating failure detection instruction. The operating condition division and sequence construction module is used to divide the operating conditions according to the time sequence of the boiler operating parameters, determine multiple boiler operating conditions, and segment the infrared thermal imaging time sequence based on the multiple boiler operating conditions to construct multiple operating condition infrared image sequences. The time series analysis and coefficient acquisition module is used to perform time series analysis on infrared image sequences under multiple operating conditions to obtain the coating degradation coefficient for each boiler operating condition. The coefficient integration and failure early warning module is used to obtain the overall coating degradation coefficient of the target boiler economizer based on the multiple boiler operating conditions and the coating degradation coefficient of each boiler operating condition, and to perform coating failure early warning based on the overall coating degradation coefficient.

Images