Boiler four-tube-wall tube wear prediction method based on big data algorithm

Abstract

The invention relates to the technical field of boiler equipment fault prediction, in particular to a boiler four-tube wall tube wear prediction method based on a big data algorithm, comprising the following steps: constructing a three-dimensional digital model of boiler four tubes, and spatially binding design parameters, historical operation and maintenance records and real-time monitoring data to form a multi-source associated database; constructing a pipe wall temperature prediction model by using a heat conduction mechanism model and a neural network algorithm, and outputting wall temperature distribution data of a full heating surface; inputting the data into a wear, creep and corrosion model, and calculating the residual life and generating a life map in combination with a historical pipe explosion case; generating a three-dimensional risk thermodynamic distribution diagram by adopting a leakage risk probability algorithm and a coupling risk coefficient model; and superposing the thermodynamic diagram into the three-dimensional digital model, and displaying the risk level based on color gradient mapping, so that the method can be widely applied to online state monitoring and intelligent early warning of the four tubes of the coal-fired power plant boiler.

Description

Technical Field

[0001] The present invention relates to the technical field of boiler equipment failure prediction, and in particular to a method for predicting wear of four-tube walls of a boiler based on a big data algorithm. Background Art

[0002] As the capacity of large thermal power generation units' boilers continues to increase, the thermal loads and mechanical stresses borne by their key components—the four-tube system (i.e., water-wall tubes, superheater tubes, economizer tubes, and reheater tubes)—have also increased significantly. These tubes are constantly exposed to extreme conditions such as high temperature, high pressure, flue gas corrosion, and erosion, making them highly susceptible to failures such as wear, creep, and corrosion, becoming a major risk for boiler tube bursts.

[0003] Existing technologies for boiler four-tube failure prediction and early warning still face several key difficulties. First, traditional methods often rely on manual inspections and periodic ultrasonic thickness measurements, which cannot achieve real-time online monitoring of complex structural systems. Second, even with a certain sensor deployment capability, the distribution of measurement points is often discontinuous, resulting in spatial data loss for key variables such as temperature, affecting the integrity of subsequent analysis. Third, existing prediction models are mostly based on a single factor, failing to effectively integrate thermal models and historical data, and lack a comprehensive assessment mechanism for the coupled relationships between multiple failure modes (such as overheating, wear, and corrosion). Summary of the Invention

[0004] Based on the above objectives, the present invention provides a boiler four-tube wall wear prediction method based on big data algorithm. By integrating multi-source data, model prediction and three-dimensional visualization, the boiler four-tube wall wear prediction method improves the risk identification speed and decision response efficiency, and provides comprehensive intelligent decision support for the safe operation of the boiler.

[0005] A method for predicting wear of four-tube wall of boiler based on big data algorithm includes the following steps:

[0006] S1: Build a 3D digital model of the four boiler tubes. Spatially bind the design parameters of the pipe welds, historical operation and maintenance records, and real-time monitoring data to the equipment units in the 3D digital model to generate a multi-source relational database containing equipment attributes and operating status.

[0007] S2: Based on the multi-source correlation database, for pipe sections lacking temperature measurement points, a pipe wall temperature prediction model is established using a heat conduction mechanism model and a neural network algorithm to output wall temperature distribution data of the entire heated surface;

[0008] S3: Inputting the wall temperature distribution data into the wear prediction model, high temperature creep life model and corrosion rate model, and combining with the historical pipe burst case library to calculate the remaining life value of each pipe section and generate a remaining life distribution map;

[0009] S4: Based on the remaining life distribution map and a leakage risk probability algorithm, overlay the coupled risk coefficients of overheating, wear, and corrosion on the areas where the remaining life is below the threshold to generate a three-dimensional risk thermal distribution map;

[0010] S5: Dynamically superimpose the three-dimensional risk heat distribution map and the three-dimensional digital model, and display areas of different risk levels through color gradient mapping technology.

[0011] Optionally, the S1 includes:

[0012] S11: Construct the basic geometric structure information set of the four boiler tubes. By obtaining the boiler design blueprint, 3D scanning data, and piping layout drawings, extract the structural parameters of the water wall, superheater, economizer, and reheater tubes in the four boiler tubes, including tube diameter, tube length, weld location, tube arrangement, and spatial coordinate information. This basic geometric structure information set serves as the structural input basis for the 3D digital model.

[0013] S12: Build a 3D digital model based on the basic geometric structure information set, use 3D modeling software to model the piping system, and generate a 3D digital model with real spatial scale and structural hierarchy according to the spatial structural relationship of the four boiler tubes;

[0014] S13: Obtain the design parameters of the pipeline weld, including welding process, weld type, material grade, design pressure, and design temperature. By consulting boiler manufacturing data, engineering design documents, and welding technical specifications, the weld design parameter set is derived from the design database and indexed and managed using a unique weld identifier.

[0015] S14: Collect historical operation and maintenance records, extract maintenance data of corresponding pipe sections and welds, historical fault records, repair and replacement records, and manual inspection data from the boiler historical operation management system, organize them into a structured historical operation and maintenance record set, and archive them by equipment number and timestamp index;

[0016] S15: Collect real-time monitoring data. Real-time operating status parameters are obtained through temperature sensors, pressure sensors, and acoustic emission monitoring devices deployed at key sections and welds of the four boiler tubes. These parameters are uploaded to the data platform in real time, forming a real-time monitoring data set containing timestamps, spatial locations, and sensor numbers.

[0017] S16: Construct a multi-source data comparison table. Based on the unique identifier of each equipment unit in the 3D digital model, perform data mapping on the design parameter set, historical operation and maintenance record set, and real-time monitoring data set. Establish a unified data field format and standardized data interface to unify various data in terms of space, time, and equipment numbering.

[0018] S17: Spatially bind the multi-source data to the equipment units in the 3D digital model. Through 3D coordinate matching and identifier comparison, the design parameters of the pipeline welds, historical operation and maintenance records, and real-time monitoring data are respectively attached to the corresponding equipment units in the 3D digital model, thus achieving the binding and visualization of multi-source information in the model space.

[0019] S18: Generate a multi-source relational database containing equipment attributes and operating status. With the bound three-dimensional digital model as the core carrier, all related data are aggregated into a structured database. The database fields include equipment unit number, spatial coordinates, structural attributes, design parameters, historical operation and maintenance record entries, and real-time monitoring indicators, realizing comprehensive data fusion from physical structure to operation and maintenance status.

[0020] S19: Verify the integrity of the multi-source associated database through the data consistency verification mechanism, check whether the device number matches, whether the timestamp is continuous, and whether the sensor data is lost, to ensure that the generated multi-source associated database is accurate, timely and traceable.

[0021] Optionally, the S2 includes:

[0022] S21: Based on the generated multi-source relational database, a temperature modeling data subset is extracted from the pipe segments with temperature measurement points, including the wall temperature data, fluid temperature, pipe pressure, flow velocity in the real-time monitoring data, as well as the spatial location, material type, and wall thickness of the pipe segment;

[0023] S22: Construct a heat conduction mechanism model. Based on Fourier's heat conduction law and the typical four-tube wall structure of the boiler, a one-dimensional steady-state heat conduction model is established. The expression is:

[0024] Where q is the heat flux per unit area, λ is the thermal conductivity, is the temperature gradient;

[0025] Using the known measuring point wall temperature and the temperature difference between the inside and outside of the pipe, combined with the thermal conductivity of the material, the actual heat flux variation characteristics are calculated to generate a heat conduction characteristic sample set;

[0026] S23: Construct a neural network temperature mapping model. Based on the heat conduction feature sample set, design a multi-layer feedforward neural network. Use equipment attributes (pipe diameter, wall thickness, material), operating conditions (pressure, flow rate), and heat flux as input layer nodes, and the corresponding wall temperature as output layer nodes. Use the back-propagation algorithm for weight training to form a neural network temperature mapping model.

[0027] S24: The temperature gradient boundary conditions of the heat conduction mechanism model are used as the prior knowledge of the neural network temperature mapping model, and a pipe wall temperature prediction model is established through a model fusion strategy.

[0028] Optionally, the S2 further includes:

[0029] S25: Identify a list of pipe sections that lack temperature measurement points, filter out pipe sections without wall temperature sensor installation records in the multi-source association database, generate a list of pipe sections that lack temperature measurement points based on the equipment number, and extract the equipment attributes and operating condition parameters of each pipe section in the list to form a prediction input set;

[0030] S26: Inputting the prediction input set into the pipe wall temperature prediction model, inputting the prediction input sets of the pipe sections that lack temperature measurement points into the pipe wall temperature prediction model in sequence, and predicting and outputting the wall temperature values ​​of the corresponding pipe sections based on the mapping relationship between heat conduction and operating conditions learned by the model, thereby forming a wall temperature prediction result set;

[0031] S27: The existing measured wall temperature data and the wall temperature prediction result set are spliced ​​and interpolated, and the corresponding wall temperature values ​​are mapped to all pipe sections in spatial order according to the spatial position of the pipe sections in the 3D digital model to generate wall temperature distribution data covering all pipe sections of the four boiler pipes;

[0032] S28: The wall temperature distribution data is structured into a multi-source relational database and passed as input to the wear prediction model, high temperature creep life model and corrosion rate model.

[0033] Optionally, the S3 includes:

[0034] S31: Build a wear prediction model based on wall temperature distribution data. Using the wall temperature distribution data as input, combined with pipe section samples with known wear in a multi-source correlation database, extract feature vectors (including wall temperature gradient, temperature fluctuation frequency, and operating time) to build a wear prediction model based on support vector regression.

[0035] S32: Using the Larson-Miller relation for high-temperature creep of materials, combined with wall temperature distribution data and material parameter tables, a high-temperature creep residual life model is calculated to predict the remaining time to creep failure.

[0036] S33: Using Arrhenius corrosion kinetics, a corrosion rate model is constructed. The temperature in the wall temperature distribution data is used as input, combined with the material and flue gas composition information, to predict the corrosion rate of each pipe segment.

[0037] S34: Extract structured records of all pipe burst events from the power plant's historical operation management system, equipment accident report database, and third-party accident investigation reports, including the burst location, occurrence time, operating conditions, pipe section wall temperature, wear status, creep life value, and corrosion level. This creates an annotated historical pipe burst case library, indexed by pipe section number and event time.

[0038] Optionally, the S3 further includes:

[0039] S35: uniformly inputting the wall temperature distribution data, the equipment attributes of the pipe section in the multi-source association database, and the operating status parameters into the wear prediction model, the high-temperature creep life model, and the corrosion rate model, respectively calculating the wear remaining life, the creep remaining life, and the corrosion remaining life, and recording them as a life indicator set;

[0040] S36: Calculate the comprehensive remaining life value by fusing the life indicator set, and use the weighted fusion algorithm to calculate the comprehensive remaining life value of each pipe segment;

[0041] S37: Correct the remaining life value based on the historical pipe burst case library. Compare the comprehensive remaining life value of each pipe section with the actual failure life under similar working conditions in the historical pipe burst case library. If there is a deviation, correct the remaining life value through the error regression model.

[0042] S38: Map the comprehensive remaining life value of each pipe section to the three-dimensional digital model, draw a remaining life distribution map according to the spatial position, use color gradient to represent the life span, and generate a remaining life distribution map.

[0043] Optionally, the S4 includes:

[0044] S41: According to the boiler operation procedures and safety standards, a lower limit threshold of the remaining life is set, and all pipe sections whose comprehensive remaining life values ​​are less than the set lower limit threshold of the remaining life are screened out from the remaining life distribution map to generate a list of high-risk pipe sections;

[0045] S42: In the list of high-risk pipe sections, the leakage probability is calculated based on the leakage risk probability algorithm, which is calculated as:

[0046]

[0047] Among them, P leak is the leakage probability, λ is the risk growth factor (obtained by fitting based on the historical pipe burst case database), R final The comprehensive remaining life value of each high-risk pipe section;

[0048] S43: Based on the output results of the multi-source correlation database and the pipe wall temperature prediction model, wear amount prediction model and corrosion rate model, a coupling risk coefficient K is constructed. couple The expression is:

[0049] K couple =ω1·K temp +ω2·K wear +ω3·K corr ;

[0050] K tempis the over-temperature risk coefficient, K wear is the wear risk factor, K corr is the corrosion risk coefficient, ω1, ω2, ω3 are the weights of each risk coefficient.

[0051] Optionally, the S4 further includes:

[0052] S44: For each high-risk pipe section, calculate its comprehensive leakage risk index R risk , expressed as:

[0053] R risk =P leak ·K couple ;

[0054] S45: mapping the comprehensive leakage risk index to the three-dimensional digital model. In the three-dimensional digital model, according to the location identifier of each pipe segment, mapping the corresponding comprehensive leakage risk index to the model space to form a structured risk thermal data set, which includes the pipe segment number, spatial coordinates and comprehensive leakage risk index value;

[0055] S46: Use color gradient mapping technology to render the three-dimensional digital model, set the risk level color according to the comprehensive leakage risk index value, and form a three-dimensional risk heat distribution map with spatial continuity and level expression capabilities.

[0056] Optionally, the S5 includes:

[0057] S51: Based on the generated comprehensive leakage risk index, set the risk level classification rules, build a color gradient mapping table, and correspond each level to a color value (RGB or HEX code) one by one to form a risk level-color coding table;

[0058] S52: Dynamically superimpose the 3D risk thermal distribution map onto the 3D digital model. Based on the spatial coordinate system and structural hierarchy of the 3D digital model, load the comprehensive leakage risk index corresponding to each pipe segment in the 3D risk thermal distribution map. Match the equipment units in the model with the equipment numbers to achieve dynamic binding of thermal data and 3D structure, thus forming a 3D risk visualization model.

[0059] S53: Apply color gradient mapping technology to render the three-dimensional digital model, call the color coding table, convert the comprehensive leakage risk index value of each equipment unit into a corresponding color, and display the risk level area in the three-dimensional digital model in a surface coloring manner, forming a three-dimensional digital model layer with the ability to visually express risk classification;

[0060] S54: Dynamically monitor the status of each equipment unit based on the color level in the three-dimensional risk visualization model and set up a three-level early warning signal trigger mechanism.

[0061] Beneficial effects of the present invention:

[0062] This invention constructs a comprehensive three-dimensional digital model, spatially binding the design parameters, historical operation and maintenance records, and real-time monitoring data of the four boiler tubes by equipment unit, generating a multi-source relational database with temporal and spatial consistency. This database not only eliminates data silos but also provides unified, continuously updated data support for subsequent temperature prediction, lifespan assessment, and risk identification, enabling visualization, quantification, and management of the status of the four boiler tubes.

[0063] This invention combines a heat conduction mechanism model with a neural network algorithm to construct a tube wall temperature prediction model. It further integrates a wear prediction model, a high-temperature creep life model, and a corrosion rate model to comprehensively assess the remaining life of each of the four tube sections in the boiler based on wall temperature distribution data. Calibrated with a historical tube burst case database, this model not only accurately identifies high-risk areas but also calculates a spatially integrated risk index through a leakage risk probability algorithm and a coupled risk coefficient model, enabling proactive prediction of localized boiler failures.

[0064] The present invention dynamically superimposes a three-dimensional risk thermal distribution map on a three-dimensional digital model, combines color gradient mapping technology with a three-level early warning signal trigger mechanism, and realizes spatial visualization, hierarchical classification, and dynamic alarm of the four-tube operating status of the boiler, significantly improving the risk identification speed and decision-making response efficiency, and providing comprehensive intelligent decision-making support for the safe operation of the boiler. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0066] Figure 1 Schematic diagram of a method flow in an embodiment of the present invention;

[0067] Figure 2 This is a schematic diagram of the S3 process of an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0069] It should be noted that references in the specification to "one embodiment," "an embodiment," "an exemplary embodiment," "some embodiments," etc. indicate that the described embodiments may include specific features, structures, or characteristics, but not every embodiment necessarily includes such specific features, structures, or characteristics. In addition, when specific features, structures, or characteristics are described in conjunction with an embodiment, it is within the knowledge of persons skilled in the relevant art to implement such features, structures, or characteristics in conjunction with other embodiments (whether or not explicitly described).

[0070] In general, terms can be understood, at least in part, from their use in context. For example, depending at least in part on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending at least in part on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0071] like Figure 1-Figure 2 As shown, a method for predicting wear of four-tube wall of boiler based on big data algorithm includes the following steps:

[0072] S1: Build a 3D digital model of the four boiler tubes. Spatially bind the design parameters of the pipe welds, historical operation and maintenance records, and real-time monitoring data to the equipment units in the 3D digital model to generate a multi-source relational database containing equipment attributes and operating status.

[0073] S2: Based on the multi-source correlation database, for pipe sections lacking temperature measurement points, a pipe wall temperature prediction model is established using the heat conduction mechanism model and neural network algorithm to output the wall temperature distribution data of the entire heated surface;

[0074] S3: Input the wall temperature distribution data into the wear prediction model, high-temperature creep life model, and corrosion rate model. Combined with the historical pipe burst case library, the remaining life value of each pipe section is calculated to generate a remaining life distribution map.

[0075] S4: Based on the remaining life distribution map and the leakage risk probability algorithm, the coupled risk coefficients of overheating, wear, and corrosion are superimposed on the areas where the remaining life is below the threshold to generate a three-dimensional risk thermal distribution map;

[0076] S5: Dynamically superimpose the three-dimensional risk heat distribution map with the three-dimensional digital model, and display areas of different risk levels through color gradient mapping technology.

[0077] S1 includes:

[0078] S11: Construct the basic geometric structure information set of the four boiler tubes. By obtaining the boiler design blueprint, 3D scanning data, and piping layout drawings, extract the structural parameters of the water wall, superheater, economizer, and reheater tubes in the four boiler tubes, including tube diameter, tube length, weld location, tube arrangement, and spatial coordinate information. This basic geometric structure information set serves as the structural input basis for the 3D digital model.

[0079] S12: Construct a 3D digital model based on the basic geometric structure information set. Use 3D modeling software (SolidWorks) to model the piping system. Generate a 3D digital model with realistic spatial scale and structural hierarchy based on the spatial structural relationship of the four boiler tubes. Each pipe segment, weld, and connection node in the model has a unique identifier.

[0080] S13: Obtain the design parameters of the pipeline weld, including welding process, weld type, material grade, design pressure, and design temperature. By consulting boiler manufacturing data, engineering design documents, and welding technical specifications, the weld design parameter set is derived from the design database and indexed and managed using a unique weld identifier.

[0081] S14: Collect historical operation and maintenance records, extract maintenance data of corresponding pipe sections and welds, historical fault records, repair and replacement records, and manual inspection data from the boiler historical operation management system, organize them into a structured historical operation and maintenance record set, and archive them by equipment number and timestamp index;

[0082] S15: Collect real-time monitoring data. Real-time operating status parameters are obtained through temperature sensors, pressure sensors, and acoustic emission monitoring devices deployed at key sections and welds of the four boiler tubes. These parameters are uploaded to the data platform in real time, forming a real-time monitoring data set containing timestamps, spatial locations, and sensor numbers.

[0083] S16: Construct a multi-source data comparison table. Based on the unique identifier of each equipment unit in the 3D digital model, perform data mapping on the design parameter set, historical operation and maintenance record set, and real-time monitoring data set. Establish a unified data field format and standardized data interface to unify various data in terms of space, time, and equipment numbering.

[0084] S17: Spatially bind the multi-source data to the equipment units in the 3D digital model. Through 3D coordinate matching and identifier comparison, the design parameters of the pipeline welds, historical operation and maintenance records, and real-time monitoring data are respectively attached to the corresponding equipment units in the 3D digital model, thus achieving the binding and visualization of multi-source information in the model space.

[0085] S18: Generate a multi-source relational database containing equipment attributes and operating status. With the bound three-dimensional digital model as the core carrier, all related data are aggregated into a structured database. The database fields include equipment unit number, spatial coordinates, structural attributes, design parameters, historical operation and maintenance record entries, and real-time monitoring indicators, realizing comprehensive data fusion from physical structure to operation and maintenance status.

[0086] S19: Verify the integrity of the multi-source associated database through the data consistency verification mechanism, check whether the device number matches, whether the timestamp is continuous, and whether the sensor data is lost, to ensure that the generated multi-source associated database is accurate, timely and traceable, and provide high-quality data support for subsequent temperature prediction modeling and life calculation.

[0087] By constructing a three-dimensional digital model with real structural information and spatially binding design parameters, historical operation and maintenance records, and real-time monitoring data based on equipment units, a multi-source associated database containing equipment attributes and operating status is finally formed. This enables visualization, manageability, and efficient data support for the status of the four tube walls of the boiler, providing a unified data foundation for subsequent wear prediction and life assessment.

[0088] S2 includes:

[0089] S21: Based on the generated multi-source relational database, a temperature modeling data subset is extracted from the pipe segments with temperature measurement points. This includes the wall temperature data, fluid temperature, pipe pressure, flow velocity, as well as the spatial location, material type, and wall thickness of the pipe segments in the real-time monitoring data. This is used as the data foundation for establishing the pipe wall temperature prediction model.

[0090] S22: Construct a heat conduction mechanism model. Based on Fourier's heat conduction law and the typical four-tube wall structure of the boiler, a one-dimensional steady-state heat conduction model is established. The expression is:

[0091] Where q is the heat flux per unit area, λ is the thermal conductivity, is the temperature gradient;

[0092] Using the known measuring point wall temperature and the temperature difference between the inside and outside of the pipe, combined with the thermal conductivity of the material, the actual heat flux variation characteristics are calculated to generate a heat conduction characteristic sample set;

[0093] S23: Construct a neural network temperature mapping model. Based on the heat conduction feature sample set, design a multi-layer feedforward neural network. Use equipment attributes (pipe diameter, wall thickness, material), operating conditions (pressure, flow rate), and heat flux as input layer nodes, and the corresponding wall temperature as output layer nodes. Use the back-propagation algorithm for weight training to form a neural network temperature mapping model.

[0094] S24: The temperature gradient boundary conditions of the heat conduction mechanism model are used as the prior knowledge of the neural network temperature mapping model. A pipe wall temperature prediction model is established through model fusion strategies (such as serial nesting or weighted integration) to achieve data-driven modeling under mechanism constraints and improve prediction accuracy and physical rationality.

[0095] S2 also includes:

[0096] S25: Identify a list of pipe sections that lack temperature measurement points, filter out pipe sections without wall temperature sensor installation records in the multi-source association database, generate a list of pipe sections that lack temperature measurement points based on the equipment number, and extract the equipment attributes and operating condition parameters of each pipe section in the list to form a prediction input set;

[0097] S26: Inputting the prediction input set into the pipe wall temperature prediction model, inputting the prediction input sets of the pipe sections that lack temperature measurement points into the pipe wall temperature prediction model in sequence, and predicting and outputting the wall temperature values ​​of the corresponding pipe sections based on the mapping relationship between heat conduction and operating conditions learned by the model, thereby forming a wall temperature prediction result set;

[0098] S27: The existing measured wall temperature data and the wall temperature prediction result set are spliced ​​and interpolated, and the corresponding wall temperature values ​​are mapped to all pipe sections in spatial order according to the spatial position of the pipe sections in the 3D digital model to generate wall temperature distribution data covering all pipe sections of the four boiler pipes;

[0099] S28: The wall temperature distribution data is structured into a multi-source relational database and passed as input to the wear prediction model, high-temperature creep life model, and corrosion rate model, providing key thermal field basis for remaining life calculation and generation of three-dimensional risk thermal distribution maps;

[0100] By combining the heat conduction mechanism with the neural network algorithm, a pipe wall temperature prediction model with both physical laws and data-driven capabilities was constructed. It can perform high-precision temperature inference on pipe sections that lack temperature measurement points, reconstruct the wall temperature distribution data of the entire heated surface, and provide comprehensive and accurate thermal information support for subsequent life assessment and risk prediction.

[0101] S3 includes:

[0102] S31: A wear prediction model is constructed based on the wall temperature distribution data. The wall temperature distribution data is used as input, and the pipe section samples with known wear in the multi-source correlation database are combined to extract the feature vector (including wall temperature gradient, temperature fluctuation frequency, and running time). The wear prediction model based on support vector regression is constructed. Its expression is:

[0103] W current =f SVR (ΔT,σ T ,t run );

[0104] Among them, W current is the current predicted cumulative wear amount, ΔT is the temperature difference, σ T is the temperature standard deviation, t run is the cumulative running time;

[0105] S32: Using the Larson-Miller relation for high temperature creep of materials, the expression is established as:

[0106]

[0107] Where PL is the Larson-Miller parameter, T is the tube wall temperature (K), t f is the failure time (h), c is the material constant, and the high-temperature creep residual life model is obtained by inverse calculation based on the wall temperature distribution data and the material parameter table to predict the remaining time to creep failure;

[0108] S33: Using Arrhenius corrosion kinetics, a corrosion rate model is constructed, which is expressed as:

[0109]

[0110] Among them, r corr is the corrosion rate, A is the pre-factor, E a is the activation energy, R is the gas constant, and T is the pipe wall temperature (K). The temperature in the wall temperature distribution data is used as input, combined with the material and flue gas composition information, to predict the corrosion rate of each pipe section;

[0111] S34: Extract structured records of all pipe burst events from the power plant's historical operation management system, equipment accident report database, and third-party accident investigation reports, including the burst location, occurrence time, operating conditions, pipe section wall temperature, wear status, creep life value, and corrosion level. This creates an annotated historical pipe burst case library, indexed by pipe section number and event time.

[0112] S3 also includes:

[0113] S35: The wall temperature distribution data, the equipment attributes of the pipe section in the multi-source association database, and the operating status parameters are uniformly input into the wear prediction model, the high-temperature creep life model, and the corrosion rate model. The wear remaining life, creep remaining life, and corrosion remaining life are calculated respectively and recorded as the life index set R, which is expressed as:

[0114] R={R wear ,R creep ,R corr};

[0115] in, R wearis the remaining wear life value calculated based on the wear prediction model, t limit The maximum allowable wear, r wear is the current wear rate;

[0116] R creep is the creep remaining life value calculated based on the high temperature creep life model, and is the corrosion remaining life value calculated based on the corrosion rate model;

[0117] R corr is the corrosion remaining life value calculated based on the corrosion rate model, t remain =t current -t min : Current wall thickness minus the minimum allowable wall thickness;

[0118] S36: Fusion life index set to calculate the comprehensive remaining life value, using weighted fusion algorithm to calculate the comprehensive remaining life value R of each pipe segment final , expressed as:

[0119] R final =αR wear +βR creep +γR corr ;

[0120] Among them, α+β+γ=1, and the weight factor can be adaptively adjusted according to the proportion of each failure mode in the historical pipe burst case library;

[0121] S37: Correct the remaining life value based on the historical pipe burst case library. Compare the comprehensive remaining life value of each pipe section with the actual failure life under similar working conditions in the historical pipe burst case library. If there is a deviation, the remaining life value is corrected through the error regression model to improve the actual reliability of the life prediction.

[0122] S38: Mapping the comprehensive remaining life value of each pipe section to the three-dimensional digital model, drawing a remaining life distribution map according to spatial position, using color gradient to represent the life span, and generating a remaining life distribution map;

[0123] By constructing a wear prediction model, a high-temperature creep life model, and a corrosion rate model, and integrating a historical pipe burst case library to accurately calculate and correct the remaining life, we ultimately generate a remaining life distribution map with spatial expression capabilities, providing highly reliable support for boiler pipe section safety assessment, maintenance plan formulation, and life prediction.

[0124] S4 includes:

[0125] S41: According to the boiler operation procedures and safety standards, set the lower limit threshold of the remaining life R threshold, such as 30 days or 720 hours, filter out all pipe sections (R final <R threshold ), generate a list of high-risk pipe sections for subsequent leakage risk probability calculation;

[0126] S42: In the list of high-risk pipe sections, the leakage probability is calculated based on the leakage risk probability algorithm, which is calculated as:

[0127]

[0128] Among them, P leak is the leakage probability, λ is the risk growth factor (obtained by fitting based on the historical pipe burst case database), R final is the comprehensive remaining life value of each high-risk pipe section. The expression is in the form of exponential decay. The shorter the remaining life, the higher the corresponding leakage probability.

[0129] S43: Based on the output results of the multi-source correlation database and the pipe wall temperature prediction model, wear amount prediction model and corrosion rate model, a coupling risk coefficient K is constructed. couple The expression is:

[0130] K couple =ω1·K temp +ω2·K wear +ω3·K corr ;

[0131] K temp is the over-temperature risk coefficient, K wear is the wear risk factor, K corr is the corrosion risk coefficient, ω1, ω2, ω3 are the weights of each risk coefficient.

[0132] The S4 also includes:

[0133] S44: For each high-risk pipe section, calculate its comprehensive leakage risk index R risk , expressed as:

[0134] R risk =P leak ·K couple ;

[0135] S45: mapping the comprehensive leakage risk index to the three-dimensional digital model. In the three-dimensional digital model, according to the location identifier of each pipe segment, mapping the corresponding comprehensive leakage risk index to the model space to form a structured risk thermal data set, which includes the pipe segment number, spatial coordinates and comprehensive leakage risk index value;

[0136] S46: Use color gradient mapping technology to render the 3D digital model. Set the risk level color according to the comprehensive leakage risk index value. Green indicates low risk, yellow indicates medium risk, and red indicates high risk. This forms a 3D risk heat distribution map with spatial continuity and grade expression capabilities.

[0137] By introducing a leakage risk probability algorithm and a multi-factor coupling risk coefficient model, a detailed assessment of leakage risk can be achieved based on the remaining lifespan. Combined with a three-dimensional digital model, the spatial thermal distribution expression is completed, providing high-resolution, visual support for the identification of high-risk areas and intelligent early warning of boiler pipe sections.

[0138] S5 includes:

[0139] S51: Based on the generated comprehensive leakage risk index, set the risk level classification rules and construct a color gradient mapping table. The rules are as follows:

[0140] R risk <0.3: green, indicating normal;

[0141] 0.3≤R risk <0.5: Yellow warning;

[0142] 0.5≤R risk <0.7: Orange warning;

[0143] R risk ≥0.7: Red shutdown warning;

[0144] Each level is mapped to a color value (RGB or HEX code) to form a risk level-color coding table;

[0145] S52: Dynamically superimpose the 3D risk thermal distribution map onto the 3D digital model. Based on the spatial coordinate system and structural hierarchy of the 3D digital model, load the comprehensive leakage risk index corresponding to each pipe segment in the 3D risk thermal distribution map. Match the equipment units in the model with the equipment numbers to achieve dynamic binding of thermal data and 3D structure, thus forming a 3D risk visualization model.

[0146] S53: Apply color gradient mapping technology to render the three-dimensional digital model, call the color coding table, convert the comprehensive leakage risk index value of each equipment unit into a corresponding color, and display the risk level area in the three-dimensional digital model in a surface coloring manner, forming a three-dimensional digital model layer with the ability to visually express risk classification;

[0147] S54: Dynamically monitor the status of each equipment unit based on the color level in the 3D risk visualization model, set up a three-level early warning signal trigger mechanism, and set the following trigger logic:

[0148] When any pipe section is rendered yellow, a yellow warning signal is triggered, indicating that it needs to be monitored closely;

[0149] When any pipe section is rendered orange, an orange warning signal is triggered and a maintenance plan is recommended;

[0150] When any pipe section is rendered red, a red shutdown warning signal is immediately triggered, and the linkage control system issues a shutdown command.

[0151] By dynamically superimposing the three-dimensional risk thermal distribution map with the three-dimensional digital model, combining color gradient mapping technology to achieve high-resolution spatial display of risk levels, and building a three-level early warning signal mechanism, it effectively supports real-time visual identification, graded early warning and intelligent intervention of boiler pipe section risks, providing digital and automated support to ensure the safe operation of equipment.

[0152] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail below in connection with the preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0153] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for predicting wear of four-tube wall of boiler based on big data algorithm, characterized in that: The following steps are involved: S1: Build a 3D digital model of the four boiler tubes. Spatially bind the design parameters of the pipe welds, historical operation and maintenance records, and real-time monitoring data to the equipment units in the 3D digital model to generate a multi-source relational database containing equipment attributes and operating status. S2: Based on the multi-source correlation database, for pipe sections lacking temperature measurement points, a pipe wall temperature prediction model is established using a heat conduction mechanism model and a neural network algorithm to output wall temperature distribution data of the entire heated surface; S3: Inputting the wall temperature distribution data into the wear prediction model, high temperature creep life model and corrosion rate model, and combining with the historical pipe burst case library to calculate the remaining life value of each pipe section and generate a remaining life distribution map; S4: Based on the remaining life distribution map and a leakage risk probability algorithm, overlay the coupled risk coefficients of overheating, wear, and corrosion on the areas where the remaining life is below the threshold to generate a three-dimensional risk thermal distribution map; S5: Dynamically superimpose the three-dimensional risk heat distribution map and the three-dimensional digital model, and display areas of different risk levels through color gradient mapping technology.

2. The method for predicting wear of four-tube wall of boiler based on big data algorithm according to claim 1 is characterized in that: Said S1 comprises: S11: Construct the basic geometric structure information set of the four boiler tubes. By obtaining the boiler design blueprint, 3D scanning data, and piping layout drawings, extract the structural parameters of the water wall, superheater, economizer, and reheater tubes in the four boiler tubes, including tube diameter, tube length, weld location, tube arrangement, and spatial coordinate information. This basic geometric structure information set serves as the structural input basis for the 3D digital model. S12: Build a 3D digital model based on the basic geometric structure information set, use 3D modeling software to model the piping system, and generate a 3D digital model with real spatial scale and structural hierarchy according to the spatial structural relationship of the four boiler tubes; S13: Obtain the design parameters of the pipeline weld, including welding process, weld type, material grade, design pressure, and design temperature. By consulting boiler manufacturing data, engineering design documents, and welding technical specifications, the weld design parameter set is derived from the design database and indexed and managed using a unique weld identifier. S14: Collect historical operation and maintenance records, extract maintenance data of corresponding pipe sections and welds, historical fault records, repair and replacement records, and manual inspection data from the boiler historical operation management system, organize them into a structured historical operation and maintenance record set, and archive them by equipment number and timestamp index; S15: Collect real-time monitoring data. Real-time operating status parameters are obtained through temperature sensors, pressure sensors, and acoustic emission monitoring devices deployed at key sections and welds of the four boiler tubes. These parameters are uploaded to the data platform in real time, forming a real-time monitoring data set containing timestamps, spatial locations, and sensor numbers. S16: Construct a multi-source data comparison table. Based on the unique identifier of each equipment unit in the 3D digital model, perform data mapping on the design parameter set, historical operation and maintenance record set, and real-time monitoring data set. Establish a unified data field format and standardized data interface to unify various data in terms of space, time, and equipment numbering. S17: Spatially bind the multi-source data to the equipment units in the 3D digital model. Through 3D coordinate matching and identifier comparison, the design parameters of the pipeline welds, historical operation and maintenance records, and real-time monitoring data are respectively attached to the corresponding equipment units in the 3D digital model, thus achieving the binding and visualization of multi-source information in the model space. S18: Generate a multi-source relational database containing equipment attributes and operating status. Using the bound 3D digital model as the core carrier, all related data is aggregated into a structured database. The database fields include equipment unit number, spatial coordinates, structural attributes, design parameters, historical operation and maintenance record entries, and real-time monitoring indicators, achieving comprehensive data integration from physical structure to operation and maintenance status. S19: Verify the integrity of the multi-source associated database through the data consistency verification mechanism, check whether the device number matches, whether the timestamp is continuous, and whether the sensor data is lost, to ensure that the generated multi-source associated database is accurate, timely and traceable.

3. The method for predicting wear of four-tube wall of boiler based on big data algorithm according to claim 2 is characterized in that: The S2 includes: S21: Based on the generated multi-source relational database, a temperature modeling data subset is extracted from the pipe segments with temperature measurement points, including the wall temperature data, fluid temperature, pipe pressure, flow velocity in the real-time monitoring data, as well as the spatial location, material type, and wall thickness of the pipe segment; S22: Construct a heat conduction mechanism model. Based on Fourier's heat conduction law and the typical four-tube wall structure of the boiler, a one-dimensional steady-state heat conduction model is established. The expression is: Where q is the heat flux per unit area, λ is the thermal conductivity, is the temperature gradient; Using the known measuring point wall temperature and the temperature difference between the inside and outside of the pipe, combined with the thermal conductivity of the material, the actual heat flux variation characteristics are calculated to generate a heat conduction characteristic sample set; S23: Construct a neural network temperature mapping model. Based on the heat conduction feature sample set, design a multi-layer feedforward neural network. Use equipment attributes, operating conditions, and heat flux as input layer nodes, and the corresponding wall temperature as output layer nodes. Use the backpropagation algorithm for weight training to form a neural network temperature mapping model. S24: The temperature gradient boundary conditions of the heat conduction mechanism model are used as the prior knowledge of the neural network temperature mapping model, and a pipe wall temperature prediction model is established through a model fusion strategy.

4. The method for predicting wear of four-tube wall of boiler based on big data algorithm according to claim 3 is characterized in that: Said S2 further comprises: S25: Identify a list of pipe sections that lack temperature measurement points, filter out pipe sections without wall temperature sensor installation records in the multi-source association database, generate a list of pipe sections that lack temperature measurement points based on the equipment number, and extract the equipment attributes and operating condition parameters of each pipe section in the list to form a prediction input set; S26: Inputting the prediction input set into the pipe wall temperature prediction model, inputting the prediction input sets of the pipe sections that lack temperature measurement points into the pipe wall temperature prediction model in sequence, and predicting and outputting the wall temperature values ​​of the corresponding pipe sections based on the mapping relationship between heat conduction and operating conditions learned by the model, thereby forming a wall temperature prediction result set; S27: The existing measured wall temperature data and the wall temperature prediction result set are spliced ​​and interpolated, and the corresponding wall temperature values ​​are mapped to all pipe sections in spatial order according to the spatial position of the pipe sections in the 3D digital model to generate wall temperature distribution data covering all pipe sections of the four boiler pipes; S28: The wall temperature distribution data is structured into a multi-source relational database and passed as input to the wear prediction model, high temperature creep life model and corrosion rate model.

5. The method for predicting wear of four-tube wall of boiler based on big data algorithm according to claim 4 is characterized in that: The S3 includes: S31: Construct a wear prediction model based on wall temperature distribution data. Using the wall temperature distribution data as input, combined with pipe section samples with known wear in the multi-source correlation database, extract feature vectors and construct a wear prediction model based on support vector regression. S32: Using the Larson-Miller relation for high-temperature creep of materials, combined with wall temperature distribution data and material parameter tables, a high-temperature creep residual life model is calculated to predict the remaining time to creep failure. S33: Using Arrhenius corrosion kinetics, a corrosion rate model is constructed. The temperature in the wall temperature distribution data is used as input, combined with the material and flue gas composition information, to predict the corrosion rate of each pipe segment. S34: Extract structured records of all pipe burst events from the power plant's historical operation management system, equipment accident report database, and third-party accident investigation reports, including the burst location, occurrence time, operating conditions, pipe section wall temperature, wear status, creep life value, and corrosion level. This creates an annotated historical pipe burst case library, indexed by pipe section number and event time.

6. The method for predicting wear of four-tube wall of boiler based on big data algorithm according to claim 5 is characterized in that: Said S3 further comprises: S35: uniformly inputting the wall temperature distribution data, the equipment attributes of the pipe section in the multi-source association database, and the operating status parameters into the wear prediction model, the high-temperature creep life model, and the corrosion rate model, respectively calculating the wear remaining life, the creep remaining life, and the corrosion remaining life, and recording them as a life indicator set; S36: Calculate the predicted remaining life value by fusing the life indicator set, and use a weighted fusion algorithm to calculate the predicted remaining life value of each pipe segment; S37: Correct the remaining life value based on the historical pipe burst case library. Compare the comprehensive remaining life value of each pipe section with the actual failure life under similar working conditions in the historical pipe burst case library. If there is a deviation, correct the remaining life value through the error regression model. S38: Map the comprehensive remaining life value of each pipe section to the three-dimensional digital model, draw a remaining life distribution map according to the spatial position, use color gradient to represent the life span, and generate a remaining life distribution map.

7. The method for predicting wear of four-tube wall of boiler based on big data algorithm according to claim 6 is characterized in that: The S4 includes: S41: According to the boiler operation procedures and safety standards, a lower limit threshold of the remaining life is set, and all pipe sections whose comprehensive remaining life values ​​are less than the set lower limit threshold of the remaining life are screened out from the remaining life distribution map to generate a list of high-risk pipe sections; S42: In the list of high-risk pipe sections, the leakage probability is calculated based on the leakage risk probability algorithm, which is calculated as: Among them, P leak is the leakage probability, λ is the risk growth factor, R final The comprehensive remaining life value of each high-risk pipe section; S43: Based on the output results of the multi-source correlation database and the pipe wall temperature prediction model, wear amount prediction model and corrosion rate model, a coupling risk coefficient K is constructed. couple The expression is: K couple =ω1·K temp +ω2·K wear +ω3·K corr ; K trmp is the over-temperature risk coefficient, K wear is the wear risk factor, K corr is the corrosion risk coefficient, ω1, ω2, ω3 are the weights of each risk coefficient.

8. The method for predicting wear of four-tube wall of boiler based on big data algorithm according to claim 7 is characterized in that: Said S4 further comprises: S44: For each high-risk pipe section, calculate its comprehensive leakage risk index R risk , expressed as: R risk =P leak ·K couple ; S45: mapping the comprehensive leakage risk index to the three-dimensional digital model. In the three-dimensional digital model, according to the location identifier of each pipe segment, mapping the corresponding comprehensive leakage risk index to the model space to form a structured risk thermal data set, which includes the pipe segment number, spatial coordinates and comprehensive leakage risk index value; S46: Use color gradient mapping technology to render the three-dimensional digital model, set the risk level color according to the comprehensive leakage risk index value, and form a three-dimensional risk heat distribution map with spatial continuity and level expression capabilities.

9. The method for predicting wear of four-tube wall of boiler based on big data algorithm according to claim 8, characterized in that: The S5 includes: S51: Based on the generated comprehensive leakage risk index, risk level classification rules are set, a color gradient mapping table is constructed, and each level is mapped to a color value one by one to form a risk level-color coding table; S52: Dynamically superimpose the 3D risk thermal distribution map onto the 3D digital model. Based on the spatial coordinate system and structural hierarchy of the 3D digital model, load the comprehensive leakage risk index corresponding to each pipe segment in the 3D risk thermal distribution map. Match the equipment units in the model with the equipment numbers to achieve dynamic binding of thermal data and 3D structure, thus forming a 3D risk visualization model. S53: Apply color gradient mapping technology to render the three-dimensional digital model, call the color coding table, convert the comprehensive leakage risk index value of each equipment unit into a corresponding color, and display the risk level area in the three-dimensional digital model in a surface coloring manner, forming a three-dimensional digital model layer with the ability to visually express risk classification; S54: Dynamically monitor the status of each equipment unit based on the color level in the 3D risk visualization model.

Images