Fault prediction and diagnosis method and system based on operating thermal parameters

Abstract

The application discloses a boiler fault prediction and diagnosis method based on operation thermal parameters, and belongs to the technical field of boiler monitoring; the method comprises the following steps: establishing a boiler fault simulation model, containing a virtual boiler structure, and constructing a three-dimensional positioning model with the center coinciding with the virtual boiler structure, which is used for expressing internal abnormal parts and forming abnormal expression blocks; constructing various boiler operation abnormal characteristics, driving the simulation model to simulate under various load conditions to obtain abnormal expressions; analyzing the abnormal expressions, extracting dynamic expressions of the abnormal expression blocks and corresponding operation thermal parameter groups, and establishing and saving the correlation between the two and the abnormal characteristics; and constructing a block dynamic expression analysis model based on the above, analyzing real-time operation thermal parameters, matching corresponding block dynamic expressions, and determining the boiler operation abnormal characteristics. The application generates abnormal samples through simulation and realizes accurate mapping of parameters and spatial block dynamics, thereby improving the accuracy and positioning precision of fault prediction.

Description

Technical Field

[0001] This invention relates to the field of boiler monitoring technology, and in particular to a method and system for fault prediction and diagnosis based on operating thermal parameters. Background Technology

[0002] In the operation of large thermal equipment such as power plant boilers, to ensure safety, stability, efficiency, and economy, it is usually necessary to monitor operating thermal parameters (such as temperature, pressure, flow rate, and heat flux density) in real time and implement fault prediction and diagnosis based on this information. Existing technologies mainly employ methods such as threshold alarms, statistical analysis, expert systems, or black-box models based on machine learning to identify anomalies by analyzing deviations in operating thermal parameters. However, these methods generally suffer from insufficient diagnostic accuracy, ambiguous fault location, and difficulty in handling multi-load variable operating conditions and complex fault scenarios. In particular, they cannot effectively correlate dynamic changes in parameters with abnormal evolution in specific spatial locations within the boiler, leading to difficulties in identifying fault types, locations, and severity, and easily causing false alarms, missed alarms, or delayed responses, affecting equipment lifespan and operational economy. Summary of the Invention

[0003] The purpose of this invention is to provide a relatively accurate method and system for predicting and diagnosing boiler faults.

[0004] This invention discloses a fault prediction and diagnosis method based on operating thermodynamic parameters, including: Step S100: Establish a boiler fault simulation model, including a virtual boiler structure, and establish a three-dimensional positioning model for the virtual boiler structure. Align the three-dimensional positioning model with the virtual boiler structure and use the three-dimensional positioning model to represent the abnormal parts inside the virtual boiler structure to obtain the abnormal performance block. Step S200: Construct several abnormal boiler operation characteristics, and based on these abnormal boiler operation characteristics, drive the boiler fault simulation model to simulate several load requirements to obtain several abnormal boiler operation manifestations. Step S300: Analyze the abnormal operation of the boiler, determine the dynamic performance of the corresponding abnormal performance block, and the dynamic performance of the corresponding operating thermodynamic parameter group, wherein the correlation between the dynamic performance of the block and the abnormal operation characteristics of the boiler is saved. Step S400: Using the correspondence between the dynamic performance of the parameter group and the dynamic performance of the block as the basis for model construction, construct the block dynamic performance analysis model, and use the dynamic performance analysis model to analyze the real-time operating thermodynamic parameters, determine the corresponding block dynamic performance, and then determine the abnormal characteristics of boiler operation.

[0005] In some embodiments disclosed in this invention, the stereo positioning model includes: The three-dimensional positioning coordinate system is defined by delineating the boundaries of the coordinate origin of the three-dimensional positioning coordinate system at a preset distance in the positive and negative directions of the x, y, and z axes, and obtaining the x-axis scale line, y-axis scale line, and z-axis scale line based on the x, y, and z axes, and the space mapped by all the scale lines is identified as the three-dimensional positioning space. The three-dimensional positioning space is gridded, and positioning coordinates are configured at the center of each positioning grid based on the three-dimensional positioning coordinate system.

[0006] In some embodiments disclosed in this invention, the method for driving a boiler fault simulation model to perform several required simulations based on abnormal boiler operation characteristics includes: Step S201: Several types of abnormal boiler operation characteristics are predefined. Each type of abnormal characteristic includes an abnormality type, an abnormality location parameter, and an abnormality severity parameter. The abnormality type is selected from the common boiler fault category group, which includes at least one of the following: slagging on the heating surface, pipeline leakage, superheater overheating, incomplete combustion, water-cooled wall tube rupture, and air preheater blockage. The abnormality location parameter specifies the abnormality occurrence block based on the positioning coordinates of the three-dimensional positioning model. The abnormality severity parameter is expressed by quantitative indicators, including an abnormality degree percentage or damage coefficient. Step S202: For each type of abnormal boiler operation characteristics, generate multiple abnormal feature instances, and form a set of abnormal feature variants by combining different abnormal location parameters and abnormal severity parameters; Step S203: Set several load requirements. The load requirements include a set of boiler operating condition parameters, including full load condition, partial load condition, start-up and shutdown condition and variable load condition. The load requirements are controlled by parameters such as boiler output percentage, fuel supply, feedwater flow rate, and air volume ratio. Step S204: Inject the abnormal feature instance into the boiler fault simulation model, including applying the physical disturbance model corresponding to the abnormal location block of the virtual boiler structure, and adjusting the disturbance intensity according to the abnormality severity parameter. Step S205: Drive the boiler fault simulation model after the injection of anomalies to perform simulations under each load requirement, and record the dynamic performance of the abnormal performance blocks during the simulation process. The block dynamic performance includes: during the operation of the boiler fault simulation model, real-time acquisition of thermodynamic parameter values ​​of each positioning grid within the abnormal performance block, including at least one of temperature, pressure, flow rate, and heat flux density; based on preset abnormal threshold rules, marking each positioning grid as an abnormal grid or a normal grid, the abnormal threshold rules including absolute parameter deviation threshold, relative deviation threshold, or statistical significance threshold compared with the normal operating condition benchmark; spatial clustering processing of the marked abnormal positioning grids to form one or more abnormal grid clusters, identifying the abnormal grid clusters as abnormal performance blocks, and recording the performance of abnormal performance blocks in consecutive frames as block dynamic performance.

[0007] In some embodiments disclosed in this invention, the method for constructing a block dynamic performance analysis model includes: Step S401: Construct a parameter interval mapping template for each operating thermal parameter. The parameter interval mapping template includes several continuously arranged parameter interval display units. Determine the parameter interval to which the parameter value of each frame in the dynamic display of the operating thermal parameter belongs, and drive the highlighting display of the corresponding parameter interval display unit to obtain the dynamic highlighting display of the parameter interval display unit on the parameter interval mapping template. Step S402: Arrange the parameter interval mapping templates in a queue to form a parameter interval mapping template queue, classify the visual features of the overall dynamic highlighting display of the parameter interval mapping template queue, and establish an index association relationship between the overall dynamic highlighting display and the dynamic performance of blocks to obtain the block dynamic performance analysis model.

[0008] In some embodiments disclosed in this invention, the method for visual feature classification of the overall dynamic highlighting display of the parameter interval mapping template queue includes: Step S4021: Perform frame skipping and classification on all overall dynamic highlighting displays. The frame skipping and classification method includes randomly selecting several frames as a group, denoted as a random frame group, comparing the equivalence of the overall dynamic highlighting displays on the corresponding frames in the random frame group, and if they are all equal, classifying the corresponding overall dynamic highlighting displays into one category to obtain the overall dynamic highlighting display group of the random frame group, and counting the display quantity in the overall dynamic highlighting display group. If the display quantity is less than or equal to a preset value, the random frame group is retained and denoted as the primary effective frame group. Step S4022: Compare and deduplicate the overall dynamic highlighting display groups corresponding to all primary valid frame groups, and count whether the overall dynamic highlighting display corresponding to the remaining primary valid frame groups can completely map all overall dynamic highlighting displays. If so, use the highlighting display sequence mapped on the overall dynamic highlighting display group corresponding to the remaining primary valid frame groups as index labels.

[0009] In some embodiments disclosed in this invention, the method for retaining random frame groups further includes: Step S40221: Compare the overall dynamic highlighting display group of the currently generated random frame group with the previously generated overall dynamic highlighting display group. If there are duplicate overall dynamic highlighting display groups and the number is greater than or equal to a preset value, then delete the corresponding random frame group.

[0010] In some embodiments disclosed in this invention, the method for analyzing real-time operating thermodynamic parameters using a dynamic performance analysis model includes: Step S403: For each operating thermal parameter in the operating thermal parameter group, call the corresponding parameter interval mapping template, determine the interval affiliation of the parameter value of each frame in the real-time parameter dynamic display, determine the parameter interval display unit to which the parameter value of each frame belongs, and drive the corresponding parameter interval display unit to highlight the display, so as to obtain the real-time dynamic highlight display sequence on the parameter interval mapping template corresponding to each operating thermal parameter. Step S404: Arrange all parameter interval mapping templates corresponding to the operating thermal parameters in a queue to form a real-time parameter interval mapping template queue. Based on the real-time dynamic highlighting display sequence, generate the overall real-time dynamic highlighting display of the real-time parameter interval mapping template queue. The overall real-time dynamic highlighting display presents the spatiotemporal evolution pattern of all highlighting display units in the queue in the form of time series frames. Step S405: Match the overall real-time dynamic highlighting display by utilizing the consistency between the pre-stored index tags in the block dynamic performance analysis model and the overall real-time dynamic highlighting display.

[0011] In some embodiments disclosed in this invention, the method for analyzing real-time operating thermodynamic parameters using a dynamic performance analysis model further includes: Step S406: Randomly combine and compare the dynamic performance of the preliminarily determined blocks to determine the degree of equivalent expression of the dynamic performance of the blocks within the combination. If the degree of equivalent expression is lower than the preset value, the dynamic performance analysis model is driven to perform a precise analysis of the real-time operating thermal parameters. Among them, methods for accurate analysis of real-time operating thermodynamic parameters include: Step S4061: Compare the previously matched overall dynamic highlighting display and overall real-time dynamic highlighting display frame by frame to determine the most matching overall dynamic highlighting display, and identify its corresponding block dynamic performance as the reference block dynamic performance for interpreting abnormal boiler operation characteristics.

[0012] In some embodiments disclosed in this invention, the method for determining the equivalent representation level of the dynamic performance of blocks within a combination includes: Step S4062: Compare the abnormal performance blocks of each frame and calculate the overlap ratio of the overlapping volume between the two to the total volume of the two. If there are abnormal performance blocks that do not overlap between the two in the comparison of the first preset number of frames, the equivalent expression degree is determined to be the lowest value. Step S4063: Calculate the average overlap ratio between abnormal performance blocks in each frame, determine the lowest overlap ratio, analyze the preset ratio range to which the lowest overlap ratio belongs, and then determine the single-frame equivalent correction coefficient. Based on the average overlap ratio and the equivalent correction coefficient, determine the single-frame equivalent expression parameter. Step S4064: Analyze the number of equivalent frames in the second preset frame number whose single-frame equivalent expression parameter is greater than or equal to the preset value, and based on the number of equivalent frames, correct the single-frame equivalent expression parameter to obtain the degree of equivalent expression between the dynamic performance of blocks. The expression for calculating the degree of equivalent expression is as follows: ; Where T represents the degree of equivalent expression, This represents the average overlap ratio. The adjustment factor is used to account for the influence of the average overlap ratio. The lowest overlap ratio, This is the output function for the single-frame equivalence correction coefficient. Based on the preset score ratio range to which the lowest overlap ratio belongs, it outputs the corresponding single-frame equivalence correction coefficient. B is the adjustment constant for the influence of the average overlap ratio. For the same number of frames, Adjust the coefficient to be equivalent to the impact of the number of frames. Adjust constants to account for the impact of the same number of frames.

[0013] In some embodiments disclosed in this invention, a fault prediction and diagnosis system based on operating thermodynamic parameters is also disclosed, including: The first module is used to establish a boiler fault simulation model, including a virtual boiler structure, and to establish a three-dimensional positioning model for the virtual boiler structure. The three-dimensional positioning model and the virtual boiler structure are aligned at the center, and the abnormal parts inside the virtual boiler structure are represented by the three-dimensional positioning model to obtain the abnormal performance block. The second module is used to construct several abnormal boiler operation characteristics, and based on these abnormal boiler operation characteristics, drive the boiler fault simulation model to simulate several load requirements, thereby obtaining several abnormal boiler operation manifestations. The third module is used to analyze abnormal boiler operation, determine the dynamic performance of the corresponding abnormal performance block, and the dynamic performance of the corresponding operating thermodynamic parameter group. It saves the correlation between the dynamic performance of the block and the abnormal boiler operation characteristics. The fourth module is used to build a block dynamic performance analysis model based on the correspondence between the dynamic performance of parameter groups and the dynamic performance of blocks. The dynamic performance analysis model is then used to analyze the real-time operating thermodynamic parameters, determine the corresponding block dynamic performance, and thus identify abnormal boiler operation characteristics.

[0014] This invention discloses a boiler fault prediction and diagnosis method based on operating thermodynamic parameters, belonging to the field of boiler monitoring technology. It includes establishing a boiler fault simulation model, comprising a virtual boiler structure, and constructing a three-dimensional positioning model coinciding with the virtual boiler's center to represent internal abnormal parts and form abnormal behavior blocks. Multiple abnormal boiler operation characteristics are constructed, driving the simulation model to simulate under various load conditions to obtain abnormal behaviors. The abnormal behaviors are analyzed, extracting the dynamic behaviors of the abnormal behavior blocks and the corresponding dynamic behaviors of the operating thermodynamic parameter groups, establishing and saving the correlation between these and the abnormal characteristics. Based on this, a block dynamic behavior analysis model is constructed to analyze real-time operating thermodynamic parameters, match the corresponding block dynamic behaviors, and thus determine the abnormal boiler operation characteristics. This invention improves the accuracy of fault prediction and location precision by generating abnormal samples through simulation and achieving precise mapping between parameters and spatial block dynamics.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the method steps of the fault prediction and diagnosis method based on operating thermal parameters disclosed in an embodiment of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only for illustration and explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Those skilled in the art can make some non-essential improvements and adjustments based on the following content of the present invention. In the present invention, unless otherwise expressly specified and limited, the technical terms used in the present invention should have the ordinary meaning understood by those skilled in the art.

[0019] Example: The purpose of this invention is to provide a relatively accurate method and system for predicting and diagnosing boiler faults.

[0020] This invention discloses a fault prediction and diagnosis method based on operating thermodynamic parameters. (See reference...) Figure 1 ,include: Step S100: Establish a boiler fault simulation model, including a virtual boiler structure, and establish a three-dimensional positioning model for the virtual boiler structure. Align the three-dimensional positioning model with the virtual boiler structure and use the three-dimensional positioning model to represent the abnormal parts inside the virtual boiler structure to obtain abnormal performance blocks.

[0021] The core principle of step S100 is to construct a boiler fault simulation model and introduce a stereoscopic positioning model to achieve a spatial and visual representation of anomalies within the virtual boiler, thereby providing a refined spatial reference basis for subsequent fault diagnosis. This step first creates a virtual boiler structure as a digital twin, simulating the physical geometry and thermodynamic processes of a real boiler. Then, an independent stereoscopic positioning model (based on a three-dimensional coordinate system and mesh partitioning) is established, and the centers of the two models are aligned to ensure that the positioning coordinates accurately correspond to the actual parts of the boiler. When an anomaly occurs in the simulation, the stereoscopic positioning model is used to mark and cluster the abnormal parameter deviations in the corresponding mesh, forming intuitive "anomaly manifestation blocks." This principle is similar to adding "spatial coordinate labels" to complex equipment in a virtual reality environment, allowing the originally abstract thermal anomalies to be transformed into specific location blocks. For example, when a water-cooled wall tube ruptures, the abnormal high temperature or pressure deviation will be highlighted as a red high-brightness block in the corresponding mesh cluster of the water-cooled wall in the virtual boiler. This transforms the fault from a simple parameter anomaly into a locatable and quantifiable spatial dynamic entity, laying the foundation for subsequent multi-scenario fault sample generation and feature association.

[0022] Step S200: Construct several abnormal boiler operation characteristics, and based on these abnormal boiler operation characteristics, drive the boiler fault simulation model to simulate several load requirements, thereby obtaining several abnormal boiler operation behaviors.

[0023] The core principle of step S100 is to construct a highly realistic boiler fault simulation model and introduce a stereo positioning model to achieve a spatial and visual representation of anomalies within the virtual boiler, thereby providing a refined spatial reference basis for subsequent fault diagnosis. This step first creates a virtual boiler structure as a digital twin, simulating the physical geometry and thermodynamic processes of a real boiler. Then, an independent stereo positioning model (based on a three-dimensional coordinate system and mesh partitioning) is established, ensuring that the centers of the two models coincide and that the positioning coordinates accurately correspond to the actual parts of the boiler. When an anomaly occurs in the simulation, the stereo positioning model is used to mark and cluster the abnormal parameter deviations in the corresponding mesh, forming intuitive "anomaly manifestation blocks." This principle is similar to adding "spatial coordinate labels" to complex equipment in a virtual reality environment, allowing the originally abstract thermal anomalies to be transformed into specific location blocks. For example, when a water-cooled wall tube ruptures, the abnormal high temperature or pressure deviation will be highlighted as a red high-brightness block in the corresponding mesh cluster of the water-cooled wall in the virtual boiler. This transforms the fault from a simple parameter anomaly into a locatable and quantifiable spatial dynamic entity, laying the foundation for subsequent multi-scenario fault sample generation and feature association.

[0024] Step S300: Analyze the abnormal operation of the boiler, determine the dynamic performance of the corresponding abnormal performance block, and the dynamic performance of the corresponding operating thermodynamic parameter group, and save the correlation between the dynamic performance of the block and the abnormal operation characteristics of the boiler.

[0025] Step S300 works by deeply analyzing the simulated abnormal boiler operation, extracting the dynamic evolution characteristics of the abnormality in spatial blocks and the temporal variation characteristics of the global operating thermodynamic parameters, and establishing a close correlation between these and the original abnormal characteristics. This constructs a multi-level mapping knowledge base from "parameter dynamics" to "spatial fault dynamics" and then to "fault essence." During simulation, this step collects thermodynamic parameters (such as temperature and pressure) of the grid within the abnormal block in real time, labels and clusters the abnormal grids based on threshold rules to form dynamic blocks, and records the temporal performance of the overall parameter group. These dynamic characteristics are then matched one-to-one with the injected abnormal characteristics (such as type, location, and severity). For example, in an air preheater blockage simulation, the block dynamics might show a gradual expansion of the bottom grid cluster and a continuous decrease in heat flux density, while the corresponding parameter group dynamics show a slow decrease in the main reheat steam temperature. By saving this correlation, reverse diagnosis can be achieved subsequently.

[0026] Step S400: Using the correspondence between the dynamic performance of the parameter group and the dynamic performance of the block as the basis for model construction, construct the block dynamic performance analysis model, and use the dynamic performance analysis model to analyze the real-time operating thermodynamic parameters, determine the corresponding block dynamic performance, and then determine the abnormal characteristics of boiler operation.

[0027] Step S400 is based on the correspondence between the offline accumulated parameter group dynamic performance and the block dynamic performance. A dedicated block dynamic performance analysis model is constructed, and this model is used in real-time operation to quickly match and analyze the actually collected thermal parameters. This achieves efficient mapping from "real-time parameter stream" to "historical fault dynamic mode," ultimately accurately determining the current boiler's abnormal operating characteristics (type, location, severity). Essentially, this model is a pattern recognition system based on visual spatiotemporal features. It transforms multi-parameter time series into a queued "dynamic highlighting display" pattern and achieves highly robust matching through pre-stored indexes and combined comparisons. When real-time parameters are input, the model first coarsely matches the overall dynamic pattern. If ambiguity exists, it performs fine-grained frame-by-frame matching and overlap volume quantification analysis. For example, when a parameter dynamic mode similar to the "slagging on the heating surface" case in the simulation library is detected during actual boiler operation, the model can quickly lock the corresponding block dynamic performance and output a diagnostic result indicating that slagging occurs in the middle of the furnace with a severity of approximately 50%. This principle enables the interpretability and high precision of the diagnostic process, avoids the unreliability of traditional black-box models, and shifts fault prediction from passive alarm to active and precise location.

[0028] The operating thermal parameters include furnace outlet flue gas temperature, superheater wall temperature, reheater wall temperature, main steam pressure, reheat steam pressure, feedwater temperature, feedwater flow rate, fuel quantity, forced draft volume, induced draft volume, furnace negative pressure, flue gas oxygen content, exhaust gas temperature, desuperheating water flow rate, boiler evaporation rate, and excess air coefficient.

[0029] In some embodiments disclosed in this invention, the stereo positioning model includes: The three-dimensional positioning coordinate system is used to define the boundaries of the coordinate system at a preset distance from the origin of the three-dimensional positioning coordinate system to the positive and negative directions of the x, y, and z axes, thereby obtaining the x-axis scale line, y-axis scale line, and z-axis scale line based on the x, y, and z axes, and the space mapped by all the scale lines is identified as the three-dimensional positioning space.

[0030] The three-dimensional positioning space is gridded, and positioning coordinates are configured at the center of each positioning grid based on the three-dimensional positioning coordinate system.

[0031] In some embodiments disclosed in this invention, the method for driving a boiler fault simulation model to perform several required simulations based on abnormal boiler operation characteristics includes: Step S201: Several types of abnormal boiler operation characteristics are predefined. Each type of abnormal characteristic includes an abnormality type, an abnormality location parameter, and an abnormality severity parameter. The abnormality type is selected from the common boiler fault category group, which includes at least one of the following: slagging on the heating surface, pipeline leakage, superheater overheating, incomplete combustion, water-cooled wall tube rupture, and air preheater blockage. The abnormality location parameter specifies the abnormality occurrence block based on the positioning coordinates of the three-dimensional positioning model. The abnormality severity parameter is expressed by quantitative indicators, including an abnormality degree percentage or damage coefficient.

[0032] Step S202: For each type of abnormal boiler operation characteristic, generate multiple abnormal characteristic instances, and form a set of abnormal characteristic variants by combining different abnormal location parameters and abnormal severity parameters.

[0033] Step S203: Set several load requirements. The load requirements include a set of boiler operating condition parameters, including full load condition, partial load condition, start-up and shutdown condition, and variable load condition. The load requirements are controlled by parameters such as boiler output percentage, fuel supply, feedwater flow rate, and air volume ratio.

[0034] Step S204: Injecting abnormal feature instances into the boiler fault simulation model, including applying physical disturbance models corresponding to the abnormal location blocks of the virtual boiler structure, and adjusting the disturbance intensity according to the abnormality severity parameter.

[0035] Step S205: Drive the boiler fault simulation model after the injection of anomalies to perform simulations under each load requirement, and record the dynamic performance of the abnormal block during the simulation process.

[0036] The block dynamic performance includes: during the operation of the boiler fault simulation model, real-time acquisition of thermodynamic parameter values ​​of each positioning grid within the abnormal performance block, including at least one of temperature, pressure, flow rate, and heat flux density; based on preset abnormal threshold rules, marking each positioning grid as an abnormal grid or a normal grid, the abnormal threshold rules including absolute parameter deviation threshold, relative deviation threshold, or statistical significance threshold compared with the normal operating condition benchmark; spatial clustering processing of the marked abnormal positioning grids to form one or more abnormal grid clusters, identifying the abnormal grid clusters as abnormal performance blocks, and recording the performance of abnormal performance blocks in consecutive frames as block dynamic performance.

[0037] In some embodiments disclosed in this invention, the method for constructing a block dynamic performance analysis model includes: Step S401: Construct a parameter interval mapping template for each operating thermal parameter. The parameter interval mapping template includes several consecutively arranged parameter interval display units. Determine the parameter interval to which the parameter value of each frame in the dynamic display of the operating thermal parameter belongs, and drive the highlighting display of the corresponding parameter interval display unit to obtain the dynamic highlighting display of the parameter interval display unit on the parameter interval mapping template.

[0038] The core principle of step S401 is to transform the continuous time-series dynamics of a single operating thermodynamic parameter into a visualized, discretized dynamic highlighting mode. This is achieved through a parameter interval mapping template, converting the numerical sequence into a "visual signal," thus facilitating subsequent multi-parameter integration and pattern recognition. This step pre-constructs a parameter interval mapping template for each thermodynamic parameter (such as main steam temperature and furnace pressure). This template is essentially a one-dimensional continuous strip divided into multiple parameter interval representation units, similar to a segmented progress bar or heatmap bar. When processing each frame of dynamic parameter data, the system determines in real time which interval the current parameter value falls into and drives the corresponding unit to highlight it (e.g., by deepening the color, increasing the brightness, or filling the pattern), thereby generating a dynamic highlighting sequence for that parameter over the entire time series.

[0039] Step S402: Arrange the parameter interval mapping templates in a queue to form a parameter interval mapping template queue, classify the visual features of the overall dynamic highlighting display of the parameter interval mapping template queue, and establish an index association relationship between the overall dynamic highlighting display and the dynamic performance of blocks to obtain the block dynamic performance analysis model.

[0040] The principle of step S402 lies in forming a multi-dimensional "overall dynamic highlighting display" spatiotemporal pattern (similar to a multi-line heatmap or animation frame sequence) by dynamically highlighting multiple single-parameter parameters in a queue-like arrangement. This pattern is then visually categorized and indexed to construct an efficient block dynamic performance analysis model, enabling rapid retrieval and matching from parameter dynamics to spatial block fault dynamics. In some embodiments disclosed in this invention, the method for visual feature classification of the overall dynamic highlighting display of the parameter interval mapping template queue includes: Step S4021: Perform frame skipping and classification on all overall dynamic highlighting displays. The frame skipping and classification method includes randomly selecting several frames as a group, denoted as a random frame group, comparing the equivalence of the overall dynamic highlighting displays on the corresponding frames in the random frame group, and if they are all equal, classifying the corresponding overall dynamic highlighting displays into one category to obtain the overall dynamic highlighting display group of the random frame group, and counting the display quantity in the overall dynamic highlighting display group. If the display quantity is less than or equal to a preset value, the random frame group is retained and denoted as the primary valid frame group.

[0041] Step S4022: Compare and deduplicate the overall dynamic highlighting display groups corresponding to all primary valid frame groups, and count whether the overall dynamic highlighting display corresponding to the remaining primary valid frame groups can completely map all overall dynamic highlighting displays. If so, use the highlighting display sequence mapped on the overall dynamic highlighting display group corresponding to the remaining primary valid frame groups as index labels.

[0042] In some embodiments disclosed in this invention, the method for retaining random frame groups further includes: Step S40221: Compare the overall dynamic highlighting display group of the currently generated random frame group with the previously generated overall dynamic highlighting display group. If there are duplicate overall dynamic highlighting display groups and the number is greater than or equal to a preset value, then delete the corresponding random frame group.

[0043] In some embodiments disclosed in this invention, the method for analyzing real-time operating thermodynamic parameters using a dynamic performance analysis model includes: Step S403: For each operating thermal parameter in the operating thermal parameter group, call the corresponding parameter interval mapping template, determine the interval affiliation of the parameter value of each frame in the real-time parameter dynamic display, determine the parameter interval display unit to which the parameter value of each frame belongs, and drive the corresponding parameter interval display unit to highlight the display, so as to obtain the real-time dynamic highlight display sequence on the parameter interval mapping template corresponding to each operating thermal parameter.

[0044] Step S404: Arrange all parameter interval mapping templates corresponding to the operating thermal parameters in a queue to form a real-time parameter interval mapping template queue. Based on the real-time dynamic highlighting display sequence, generate the overall real-time dynamic highlighting display of the real-time parameter interval mapping template queue. The overall real-time dynamic highlighting display presents the spatiotemporal evolution pattern of all highlighting display units in the queue in the form of time series frames.

[0045] Step S405: Match the overall real-time dynamic highlighting display by utilizing the consistency between the pre-stored index tags in the block dynamic performance analysis model and the overall real-time dynamic highlighting display.

[0046] In some embodiments disclosed in this invention, the method for analyzing real-time operating thermodynamic parameters using a dynamic performance analysis model further includes: Step S406: Randomly combine and compare the dynamic performance of the preliminarily determined blocks to determine the degree of equivalent expression of the dynamic performance of the blocks within the combination. If the degree of equivalent expression is lower than the preset value, the dynamic performance analysis model is driven to perform accurate analysis of the real-time operating thermal parameters.

[0047] Among them, methods for accurate analysis of real-time operating thermodynamic parameters include: Step S4061: Compare the previously matched overall dynamic highlighting display and overall real-time dynamic highlighting display frame by frame to determine the most matching overall dynamic highlighting display, and identify its corresponding block dynamic performance as the reference block dynamic performance for interpreting abnormal boiler operation characteristics.

[0048] In some embodiments disclosed in this invention, the method for determining the equivalent representation level of the dynamic performance of blocks within a combination includes: Step S4062: Compare the abnormal performance blocks of each frame and calculate the overlap ratio of the overlapping volume between the two to the total volume of the two. If there are abnormal performance blocks that do not overlap between the two in the comparison of the first preset number of frames, the equivalent expression level is determined to be the lowest value.

[0049] Step S4063: Calculate the average overlap ratio between abnormal performance blocks in each frame, determine the lowest overlap ratio, analyze the preset ratio range to which the lowest overlap ratio belongs, and then determine the single-frame equivalent correction coefficient. Based on the average overlap ratio and the equivalent correction coefficient, determine the single-frame equivalent expression parameter.

[0050] Step S4064: Analyze the number of equivalent frames in the second preset frame number whose single-frame equivalent expression parameter is greater than or equal to the preset value, and based on the number of equivalent frames, correct the single-frame equivalent expression parameter to obtain the degree of equivalent expression between the dynamic performance of blocks.

[0051] The expression for calculating the degree of equivalent expression is as follows: .

[0052] Where T represents the degree of equivalent expression, This represents the average overlap ratio. The adjustment factor is used to account for the influence of the average overlap ratio. The lowest overlap ratio, This is the output function for the single-frame equivalence correction coefficient. Based on the preset score ratio range to which the lowest overlap ratio belongs, it outputs the corresponding single-frame equivalence correction coefficient. B is the adjustment constant for the influence of the average overlap ratio. For the same number of frames, Adjust the coefficient to be equivalent to the impact of the number of frames. Adjust constants to account for the impact of the same number of frames.

[0053] L: Adjustment coefficient for the average overlap ratio; an adjustable coefficient used to amplify or reduce the weighting effect of the average overlap ratio. Adjusting L allows the average overlap to play a greater role in the exponential component, reflecting an emphasis on overall similarity.

[0054] B: The average overlap ratio affects the adjustment constant; a constant term, added in parentheses to the exponent part, is used to adjust the average overlap for translation or baseline adjustment, avoiding extreme T values ​​caused by an exponent that is too small or too large.

[0055] R: Equivalent Frame Count Influence Adjustment Factor; an adjustable factor used to adjust the weight of the equivalent frame count (H) to make its influence on the logarithmic part stronger or weaker.

[0056] C: Equivalent frame rate impact adjustment constant; a constant term, added within the logarithmic brackets, to ensure the logarithmic parameter is not zero or negative, and to provide baseline bonus.

[0057] In some embodiments disclosed in this invention, a fault prediction and diagnosis system based on operating thermodynamic parameters is also disclosed, including: The first module is used to establish a boiler fault simulation model, including a virtual boiler structure. A three-dimensional positioning model is established for the virtual boiler structure. The three-dimensional positioning model and the virtual boiler structure are aligned at the center. The three-dimensional positioning model is used to represent the abnormal parts inside the virtual boiler structure to obtain the abnormal performance blocks.

[0058] The second module is used to construct several abnormal boiler operation characteristics, and based on these abnormal boiler operation characteristics, drive the boiler fault simulation model to simulate several load requirements, thereby obtaining several abnormal boiler operation manifestations. The third module is used to analyze abnormal boiler operation, determine the dynamic performance of the corresponding abnormal performance block, and the dynamic performance of the corresponding operating thermodynamic parameter group. It saves the correlation between the dynamic performance of the block and the abnormal boiler operation characteristics. The fourth module is used to build a block dynamic performance analysis model based on the correspondence between the dynamic performance of parameter groups and the dynamic performance of blocks. The dynamic performance analysis model is then used to analyze the real-time operating thermodynamic parameters, determine the corresponding block dynamic performance, and thus identify abnormal boiler operation characteristics.

[0059] This invention discloses a boiler fault prediction and diagnosis method based on operating thermodynamic parameters, belonging to the field of boiler monitoring technology. It includes establishing a boiler fault simulation model, comprising a virtual boiler structure, and constructing a three-dimensional positioning model coinciding with the virtual boiler's center to represent internal abnormal parts and form abnormal behavior blocks. Multiple abnormal boiler operation characteristics are constructed, driving the simulation model to simulate under various load conditions to obtain abnormal behaviors. The abnormal behaviors are analyzed, extracting the dynamic behaviors of the abnormal behavior blocks and the corresponding dynamic behaviors of the operating thermodynamic parameter groups, establishing and saving the correlation between these and the abnormal characteristics. Based on this, a block dynamic behavior analysis model is constructed to analyze real-time operating thermodynamic parameters, match the corresponding block dynamic behaviors, and thus determine the abnormal boiler operation characteristics. This invention improves the accuracy of fault prediction and location precision by generating abnormal samples through simulation and achieving precise mapping between parameters and spatial block dynamics.

[0060] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0061] This invention discloses a boiler fault prediction and diagnosis method based on operating thermodynamic parameters, belonging to the field of boiler monitoring technology. It includes establishing a boiler fault simulation model, comprising a virtual boiler structure, and constructing a three-dimensional positioning model coinciding with the virtual boiler's center to represent internal abnormal parts and form abnormal behavior blocks. Multiple abnormal boiler operation characteristics are constructed, driving the simulation model to simulate under various load conditions to obtain abnormal behaviors. The abnormal behaviors are analyzed, extracting the dynamic behaviors of the abnormal behavior blocks and the corresponding dynamic behaviors of the operating thermodynamic parameter groups, establishing and saving the correlation between these and the abnormal characteristics. Based on this, a block dynamic behavior analysis model is constructed to analyze real-time operating thermodynamic parameters, match the corresponding block dynamic behaviors, and thus determine the abnormal boiler operation characteristics. This invention improves the accuracy of fault prediction and location precision by generating abnormal samples through simulation and achieving precise mapping between parameters and spatial block dynamics.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A fault prediction and diagnosis method based on operating thermodynamic parameters, characterized in that, include: Step S100: Establish a boiler fault simulation model, including a virtual boiler structure, and establish a three-dimensional positioning model for the virtual boiler structure. Align the three-dimensional positioning model with the virtual boiler structure and use the three-dimensional positioning model to represent the abnormal parts inside the virtual boiler structure to obtain the abnormal performance block. Step S200: Construct several abnormal boiler operation characteristics, and based on these abnormal boiler operation characteristics, drive the boiler fault simulation model to simulate several load requirements to obtain several abnormal boiler operation manifestations. Step S300: Analyze the abnormal operation of the boiler, determine the dynamic performance of the corresponding abnormal performance block, and the dynamic performance of the corresponding operating thermodynamic parameter group, wherein the correlation between the dynamic performance of the block and the abnormal operation characteristics of the boiler is saved. Step S400: Using the correspondence between the dynamic performance of the parameter group and the dynamic performance of the block as the basis for model construction, construct the block dynamic performance analysis model, and use the dynamic performance analysis model to analyze the real-time operating thermodynamic parameters, determine the corresponding block dynamic performance, and then determine the abnormal characteristics of boiler operation.

2. The fault prediction and diagnosis method based on operating thermodynamic parameters according to claim 1, characterized in that, The 3D positioning model includes: The three-dimensional positioning coordinate system is defined by delineating the boundaries of the coordinate origin of the three-dimensional positioning coordinate system at a preset distance in the positive and negative directions of the x, y, and z axes, and obtaining the x-axis scale line, y-axis scale line, and z-axis scale line based on the x, y, and z axes, and the space mapped by all the scale lines is identified as the three-dimensional positioning space. The three-dimensional positioning space is gridded, and positioning coordinates are configured at the center of each positioning grid based on the three-dimensional positioning coordinate system.

3. The fault prediction and diagnosis method based on operating thermodynamic parameters according to claim 2, characterized in that, Based on the abnormal characteristics of boiler operation, several methods for driving boiler fault simulation models to perform simulations that meet the requirements include: Step S201: Several types of abnormal boiler operation characteristics are predefined. Each type of abnormal characteristic includes an abnormality type, an abnormality location parameter, and an abnormality severity parameter. The abnormality type is selected from the common boiler fault category group, which includes at least one of the following: slagging on the heating surface, pipeline leakage, superheater overheating, incomplete combustion, water-cooled wall tube rupture, and air preheater blockage. The abnormality location parameter specifies the abnormality occurrence block based on the positioning coordinates of the three-dimensional positioning model. The abnormality severity parameter is expressed by quantitative indicators, including an abnormality degree percentage or damage coefficient. Step S202: For each type of abnormal boiler operation characteristics, generate multiple abnormal feature instances, and form a set of abnormal feature variants by combining different abnormal location parameters and abnormal severity parameters; Step S203: Set several load requirements. The load requirements include a set of boiler operating condition parameters, including full load condition, partial load condition, start-up and shutdown condition and variable load condition. The load requirements are controlled by parameters such as boiler output percentage, fuel supply, feedwater flow rate, and air volume ratio. Step S204: Inject the abnormal feature instance into the boiler fault simulation model, including applying the physical disturbance model corresponding to the abnormal location block of the virtual boiler structure, and adjusting the disturbance intensity according to the abnormality severity parameter. Step S205: Drive the boiler fault simulation model after the injection of anomalies to perform simulations under each load requirement, and record the dynamic performance of the abnormal performance blocks during the simulation process. The block dynamic performance includes: during the operation of the boiler fault simulation model, real-time acquisition of thermodynamic parameter values ​​of each positioning grid within the abnormal performance block, including at least one of temperature, pressure, flow rate, and heat flux density; based on preset abnormal threshold rules, marking each positioning grid as an abnormal grid or a normal grid, the abnormal threshold rules including absolute parameter deviation threshold, relative deviation threshold, or statistical significance threshold compared with the normal operating condition benchmark; spatial clustering processing of the marked abnormal positioning grids to form one or more abnormal grid clusters, identifying the abnormal grid clusters as abnormal performance blocks, and recording the performance of abnormal performance blocks in consecutive frames as block dynamic performance.

4. The fault prediction and diagnosis method based on operating thermodynamic parameters according to claim 2, characterized in that, Methods for constructing block dynamic performance analysis models include: Step S401: Construct a parameter interval mapping template for each operating thermal parameter. The parameter interval mapping template includes several continuously arranged parameter interval display units. Determine the parameter interval to which the parameter value of each frame in the dynamic display of the operating thermal parameter belongs, and drive the highlighting display of the corresponding parameter interval display unit to obtain the dynamic highlighting display of the parameter interval display unit on the parameter interval mapping template. Step S402: Arrange the parameter interval mapping templates in a queue to form a parameter interval mapping template queue, classify the visual features of the overall dynamic highlighting display of the parameter interval mapping template queue, and establish an index association relationship between the overall dynamic highlighting display and the dynamic performance of blocks to obtain the block dynamic performance analysis model.

5. The fault prediction and diagnosis method based on operating thermodynamic parameters according to claim 4, characterized in that, Methods for visual feature classification of the overall dynamic highlighting display of the parameter range mapping template queue include: Step S4021: Perform frame skipping and classification on all overall dynamic highlighting displays. The frame skipping and classification method includes randomly selecting several frames as a group, denoted as a random frame group, comparing the equivalence of the overall dynamic highlighting displays on the corresponding frames in the random frame group, and if they are all equal, classifying the corresponding overall dynamic highlighting displays into one category to obtain the overall dynamic highlighting display group of the random frame group, and counting the display quantity in the overall dynamic highlighting display group. If the display quantity is less than or equal to a preset value, the random frame group is retained and denoted as the primary effective frame group. Step S4022: Compare and deduplicate the overall dynamic highlighting display groups corresponding to all primary valid frame groups, and count whether the overall dynamic highlighting display corresponding to the remaining primary valid frame groups can completely map all overall dynamic highlighting displays. If so, use the highlighting display sequence mapped on the overall dynamic highlighting display group corresponding to the remaining primary valid frame groups as index labels.

6. The fault prediction and diagnosis method based on operating thermodynamic parameters according to claim 5, characterized in that, Methods for preserving random frame groups also include: Step S40221: Compare the overall dynamic highlighting display group of the currently generated random frame group with the previously generated overall dynamic highlighting display group. If there are duplicate overall dynamic highlighting display groups and the number is greater than or equal to a preset value, then delete the corresponding random frame group.

7. The fault prediction and diagnosis method based on operating thermodynamic parameters according to claim 5, characterized in that, Methods for analyzing real-time operating thermodynamic parameters using dynamic performance analysis models include: Step S403: For each operating thermal parameter in the operating thermal parameter group, call the corresponding parameter interval mapping template, determine the interval affiliation of the parameter value of each frame in the real-time parameter dynamic display, determine the parameter interval display unit to which the parameter value of each frame belongs, and drive the corresponding parameter interval display unit to highlight the display, so as to obtain the real-time dynamic highlight display sequence on the parameter interval mapping template corresponding to each operating thermal parameter. Step S404: Arrange all parameter interval mapping templates corresponding to the operating thermal parameters in a queue to form a real-time parameter interval mapping template queue. Based on the real-time dynamic highlighting display sequence, generate the overall real-time dynamic highlighting display of the real-time parameter interval mapping template queue. The overall real-time dynamic highlighting display presents the spatiotemporal evolution pattern of all highlighting display units in the queue in the form of time series frames. Step S405: Match the overall real-time dynamic highlighting display by utilizing the consistency between the pre-stored index tags in the block dynamic performance analysis model and the overall real-time dynamic highlighting display.

8. The fault prediction and diagnosis method based on operating thermodynamic parameters according to claim 7, characterized in that, Methods for analyzing real-time operating thermodynamic parameters using dynamic performance analysis models also include: Step S406: Randomly combine and compare the dynamic performance of the preliminarily determined blocks to determine the degree of equivalent expression of the dynamic performance of the blocks within the combination. If the degree of equivalent expression is lower than the preset value, the dynamic performance analysis model is driven to perform a precise analysis of the real-time operating thermal parameters. Among them, methods for accurate analysis of real-time operating thermodynamic parameters include: Step S4061: Compare the previously matched overall dynamic highlighting display and overall real-time dynamic highlighting display frame by frame to determine the most matching overall dynamic highlighting display, and identify its corresponding block dynamic performance as the reference block dynamic performance for interpreting abnormal boiler operation characteristics.

9. The fault prediction and diagnosis method based on operating thermodynamic parameters according to claim 8, characterized in that, Methods for determining the degree of equivalent representation of the dynamic behavior of blocks within a combination include: Step S4062: Compare the abnormal performance blocks of each frame and calculate the overlap ratio of the overlapping volume between the two to the total volume of the two. If there are abnormal performance blocks that do not overlap between the two in the comparison of the first preset number of frames, the equivalent expression degree is determined to be the lowest value. Step S4063: Calculate the average overlap ratio between abnormal performance blocks in each frame, determine the lowest overlap ratio, analyze the preset ratio range to which the lowest overlap ratio belongs, and then determine the single-frame equivalent correction coefficient. Based on the average overlap ratio and the equivalent correction coefficient, determine the single-frame equivalent expression parameter. Step S4064: Analyze the number of equivalent frames in the second preset frame number whose single-frame equivalent expression parameter is greater than or equal to the preset value, and based on the number of equivalent frames, correct the single-frame equivalent expression parameter to obtain the degree of equivalent expression between the dynamic performance of blocks. The expression for calculating the degree of equivalent expression is as follows: ； Where T represents the degree of equivalent expression, This represents the average overlap ratio. The adjustment factor is used to account for the influence of the average overlap ratio. The lowest overlap ratio, This is the output function for the single-frame equivalence correction coefficient. Based on the preset score ratio range to which the lowest overlap ratio belongs, it outputs the corresponding single-frame equivalence correction coefficient. B is the adjustment constant for the influence of the average overlap ratio. For the same number of frames, Adjust the coefficient to be equivalent to the impact of the number of frames. Adjust constants to account for the impact of the same number of frames.

10. A fault prediction and diagnosis system based on operating thermodynamic parameters, characterized in that, include: The first module is used to establish a boiler fault simulation model, including a virtual boiler structure, and to establish a three-dimensional positioning model for the virtual boiler structure. The three-dimensional positioning model and the virtual boiler structure are aligned at the center, and the abnormal parts inside the virtual boiler structure are represented by the three-dimensional positioning model to obtain the abnormal performance block. The second module is used to construct several abnormal boiler operation characteristics, and based on these abnormal boiler operation characteristics, drive the boiler fault simulation model to simulate several load requirements, thereby obtaining several abnormal boiler operation manifestations. The third module is used to analyze abnormal boiler operation, determine the dynamic performance of the corresponding abnormal performance block, and the dynamic performance of the corresponding operating thermodynamic parameter group. It saves the correlation between the dynamic performance of the block and the abnormal boiler operation characteristics. The fourth module is used to build a block dynamic performance analysis model based on the correspondence between the dynamic performance of parameter groups and the dynamic performance of blocks. The dynamic performance analysis model is then used to analyze the real-time operating thermodynamic parameters, determine the corresponding block dynamic performance, and thus identify abnormal boiler operation characteristics.

Images