Boiler four-tube full life cycle information tracing method and system based on BIM

Abstract

The invention discloses a BIM-based boiler four-tube full life cycle information tracing method and system, and relates to the related field of data processing, and the method comprises the steps: carrying out BIM modeling according to boiler four tubes to obtain a four-tube BIM model, and carrying out the welding line full life cycle monitoring of the boiler four tubes based on the four-tube BIM model to obtain a four-tube welding line dynamic model; performing multi-degree-of-freedom risk feature focusing on the four-pipe weld joint dynamic model to obtain a four-pipe risk feature focusing map; according to the four-tube risk feature focusing atlas, deducing and tracing welding seam cracks, bump fractures and corrosion failure accidents of the four tubes of the boiler to obtain a first deducing and tracing atlas, a second deducing and tracing atlas and a third deducing and tracing atlas; and performing multi-dimensional coupling accident deduction tracing according to the first deduction tracing atlas, the second deduction tracing atlas and the third deduction tracing atlas to obtain a four-tube deduction tracing atlas. The technical problem that existing boiler four-tube information tracing is insufficient in accuracy and efficiency is solved, and the technical effect of accurately and efficiently conducting boiler four-tube full-life-cycle information tracing is achieved.

Description

Technical Field

[0001] This application relates to the field of data processing, and in particular to a method and system for tracing the entire life cycle information of boiler four-tubes based on BIM. Background Technology

[0002] As a critical component of thermal power plants, the operating status of the boiler's four tubes directly affects the safety and economy of the unit. Due to prolonged exposure to harsh conditions such as high temperature, high pressure, and corrosion, the boiler's four tubes are prone to various types of failures. Once a failure occurs, it can not only cause equipment downtime and economic losses, but may even lead to serious safety accidents. Therefore, effective monitoring and accident tracing of the boiler's four tubes are crucial. Currently, the status monitoring and accident tracing of boiler four tubes mainly rely on traditional detection technologies combined with manual recording and analysis. Various detection devices are used periodically to inspect the boiler's four tubes, obtaining current status data. This data is then manually processed and analyzed to determine if there are any hidden dangers and to trace the cause of accidents. This method suffers from long detection cycles and limited detection range, making it difficult to achieve continuous and comprehensive monitoring of the boiler's four tubes throughout their entire lifecycle, resulting in some potential hazards not being detected in a timely manner.

[0003] Currently, the relevant technologies for tracing information on the four tubes of boilers suffer from technical problems of insufficient accuracy and efficiency. Summary of the Invention

[0004] This application provides a BIM-based method and system for tracing the entire lifecycle information of four boiler tubes. It employs BIM modeling of the four boiler tubes to obtain a BIM model of each tube, and then uses this model to monitor the weld seams throughout their entire lifecycle, obtaining a dynamic model of the weld seams. The dynamic model of the weld seams is then subjected to multi-degree-of-freedom risk feature focusing to generate a risk feature focusing map of the four tubes. Based on this map, accident simulations and tracing are performed on the four boiler tubes for weld cracks, weld bulging fractures, and weld corrosion failures, resulting in first, second, and third simulation and tracing maps. These maps are then combined to perform multi-dimensional coupled accident simulation and tracing, forming a four-tube simulation and tracing map. These technical means solve the technical problems of insufficient accuracy and efficiency in existing boiler four-tube information tracing, achieving a precise and efficient technical effect for tracing the entire lifecycle information of four boiler tubes.

[0005] This application provides a BIM-based method for tracing the entire lifecycle information of four boiler tubes, comprising: performing BIM modeling of the four boiler tubes to obtain a four-tube BIM model; performing full lifecycle monitoring of the welds of the four boiler tubes based on the four-tube BIM model to obtain a dynamic model of the four-tube welds; focusing multi-degree-of-freedom risk features on the dynamic model of the four-tube welds to obtain a four-tube risk feature focusing map; performing weld crack accident simulation and tracing of the four boiler tubes based on the four-tube risk feature focusing map to obtain a first simulation and tracing map; performing weld bulge fracture accident simulation and tracing of the four boiler tubes based on the four-tube risk feature focusing map to obtain a second simulation and tracing map; performing weld corrosion failure accident simulation and tracing of the four boiler tubes based on the four-tube risk feature focusing map to obtain a third simulation and tracing map; and performing multi-dimensional coupled accident simulation and tracing based on the first simulation and tracing map, the second simulation and tracing map, and the third simulation and tracing map to obtain a four-tube simulation and tracing map.

[0006] In a possible implementation, the dynamic model of the four-pipe weld is subjected to multi-degree-of-freedom risk feature focusing to obtain a four-pipe risk feature focusing map. The following processing is then performed: weld crack risk feature focusing is performed based on the dynamic model of the four-pipe weld to obtain a first risk feature focusing map; weld bulge fracture risk feature focusing is performed based on the dynamic model of the four-pipe weld to obtain a second risk feature focusing map; weld corrosion failure risk feature focusing is performed based on the dynamic model of the four-pipe weld to obtain a third risk feature focusing map; the first risk feature focusing map, the second risk feature focusing map, and the third risk feature focusing map are integrated to generate the four-pipe risk feature focusing map.

[0007] In a possible implementation, weld crack risk features are focused based on the dynamic model of the four-tube weld to obtain a first risk feature focus map, and the following processing is performed: Based on the dynamic model of the four-tube weld, the first monitoring feature sequence corresponding to the water-cooled wall tube weld is read; a coordinate system is fitted based on the historical set of weld crack risks to construct an analytical coordinate system for crack risk contribution; the crack risk contribution of each monitoring feature in the first monitoring feature sequence is evaluated based on the analytical coordinate system for crack risk contribution to obtain a first crack risk contribution sequence; attention enhancement is applied to the first monitoring feature sequence based on the first crack risk contribution sequence to obtain a first crack risk feature focus vector; based on the first crack risk feature focus vector, weld crack risk features are further focused on the four boiler tubes according to the dynamic model of the four-tube weld to generate the first risk feature focus map.

[0008] In a possible implementation, the risk features of weld bulge fracture are focused based on the dynamic model of the four-tube weld to obtain a second risk feature focus map. The following processing is then performed: a coordinate system is fitted based on the historical set of weld bulge fracture risks to construct an analytical coordinate system for the contribution of bulge fracture risk; the contribution of the first monitoring feature sequence to bulge fracture risk is evaluated based on the analytical coordinate system to obtain a first bulge fracture risk contribution sequence; attention enhancement is applied to the first monitoring feature sequence based on the first bulge fracture risk contribution sequence to obtain a first bulge fracture risk feature focus vector; based on the first bulge fracture risk feature focus vector, the weld bulge fracture risk features of the four boiler tubes are further focused based on the dynamic model of the four-tube weld to generate the second risk feature focus map.

[0009] In a possible implementation, the boiler's four tubes are subjected to weld crack accident simulation and tracing based on the four-tube risk feature focus map to obtain a first simulation and tracing map. The following processing is then performed: weld crack accident events are retrieved for the water-cooled wall tubes to obtain a first weld crack accident event group; the first weld crack accident event group is randomly perturbed using an adversarial example generator to obtain a second weld crack accident event group; multiple learners are adversarially trained based on the second weld crack accident event group to obtain multiple crack accident simulation models and multiple adversarial training loss datasets; the multiple crack accident simulation models are fused and trained based on the multiple adversarial training loss datasets to obtain a water-cooled wall tube crack accident simulation channel; the first crack risk feature focus vector is input into the water-cooled wall tube crack accident simulation channel to obtain a first crack accident simulation result; based on the first crack accident simulation result, the boiler's four tubes are further simulated and traced for weld crack accidents based on the four-tube risk feature focus map to generate the first simulation and tracing map.

[0010] In a possible implementation, the boiler's four tubes are retrospectively analyzed for weld bulge fracture accidents based on the four-tube risk feature focus map to obtain a second retrospective map. The following processing is then performed: weld bulge fracture accident events are retrieved for the water-cooled wall tubes to obtain a first weld bulge fracture accident event group; the first weld bulge fracture accident event group is randomly perturbed to obtain a second weld bulge fracture accident event group; multiple learners are adversarially trained based on the second weld bulge fracture accident event group to obtain multiple bulge fracture accident simulation models and multiple training loss datasets; the multiple bulge fracture accident simulation models are fused and trained based on the multiple training loss datasets to obtain a water-cooled wall tube bulge fracture accident simulation channel; the first bulge fracture risk feature focus vector is input into the water-cooled wall tube bulge fracture accident simulation channel to obtain a first bulge fracture accident simulation result; based on the first bulge fracture accident simulation result, the boiler's four tubes are further retrospectively analyzed for weld bulge fracture accidents based on the four-tube risk feature focus map to generate the second retrospective map.

[0011] In a possible implementation, multi-dimensional coupled accident simulation and tracing is performed based on the first, second, and third simulation and tracing maps to obtain a four-tube simulation and tracing map. The following processing is then performed: Weld crack coupled accident simulation and tracing is conducted on the four boiler tubes based on the first, second, and third simulation and tracing maps to obtain the first coupled accident simulation and tracing result; weld bulge fracture is then investigated on the four boiler tubes based on the first, second, and third simulation and tracing maps. Coupled accident simulation and tracing are performed to obtain a second coupled accident simulation and tracing result; based on the first, second, and third simulation and tracing maps, the weld corrosion failure coupled accident simulation and tracing of the four boiler tubes is performed to obtain a third coupled accident simulation and tracing result; the first, second, and third simulation and tracing maps, the first coupled accident simulation and tracing result, the second coupled accident simulation and tracing result, and the third coupled accident simulation and tracing result are correlated and sorted to generate the four tube simulation and tracing map.

[0012] In a possible implementation, based on the four-tube BIM model, the full life cycle monitoring of the welds of the four boiler tubes is performed to obtain a dynamic model of the four-tube welds. The following processes are then performed: the structured data of the welds of the four boiler tubes is imported into the four-tube BIM model to obtain a digital model of the four tubes; the weld monitoring dataset of the four boiler tubes is cleaned to obtain a weld monitoring data stream; and the weld monitoring data stream is synchronized to the digital model of the four tubes to generate the dynamic model of the four-tube welds.

[0013] In a possible implementation, the following process is performed: the boiler four tubes include water-cooled wall tubes, economizer tubes, superheater tubes, and reheater tubes.

[0014] This application also provides a BIM-based boiler four-tube full lifecycle information traceability system, including: a weld full lifecycle monitoring module, used to perform BIM modeling based on the boiler four tubes to obtain a four-tube BIM model, and to perform full lifecycle monitoring of the welds of the boiler four tubes based on the four-tube BIM model to obtain a four-tube weld dynamic model; a multi-degree-of-freedom risk feature focusing module, used to perform multi-degree-of-freedom risk feature focusing on the four-tube weld dynamic model to obtain a four-tube risk feature focusing map; and a weld crack accident inference and traceability module, used to perform weld crack accident inference and traceability on the boiler four tubes based on the four-tube risk feature focusing map. The system obtains a first retrospective map; a weld bulge fracture accident retrospective module is used to perform weld bulge fracture accident retrospective on the four boiler tubes based on the risk feature focusing map of the four tubes, and obtain a second retrospective map; a weld corrosion failure accident retrospective module is used to perform weld corrosion failure accident retrospective on the four boiler tubes based on the risk feature focusing map of the four tubes, and obtain a third retrospective map; a multidimensional coupled accident retrospective module is used to perform multidimensional coupled accident retrospective based on the first retrospective map, the second retrospective map, and the third retrospective map, and obtain a four-tube retrospective map.

[0015] The proposed method and system for tracing the entire lifecycle information of four boiler tubes based on BIM involves: first, creating a BIM model of the four boiler tubes; then, performing full lifecycle monitoring of the welds of the four boiler tubes based on the BIM model to obtain a dynamic model of the welds; next, focusing multi-degree-of-freedom risk features on the dynamic model of the welds to obtain a risk feature focus map; then, using the risk feature focus map to simulate and trace weld crack accidents in the four boiler tubes to obtain a first simulation and trace map; then, using the risk feature focus map to simulate and trace weld bulge fracture accidents in the four boiler tubes to obtain a second simulation and trace map; then, using the risk feature focus map to simulate and trace weld corrosion failure accidents in the four boiler tubes to obtain a third simulation and trace map; finally, using the first, second, and third simulation and trace maps, performing multi-dimensional coupled accident simulation and trace to obtain a four-tube simulation and trace map. It has achieved the technical effect of accurately and efficiently tracing the information of the four tubes of the boiler throughout its entire life cycle. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 A flowchart illustrating the BIM-based boiler four-tube full lifecycle information traceability method provided in this application embodiment.

[0018] Figure 2 A schematic diagram of the structure of the BIM-based boiler four-tube full life cycle information traceability system provided in the embodiments of this application.

[0019] Figure labeling: 10 Weld life cycle monitoring module, 20 Multi-degree-of-freedom risk feature focusing module, 30 Weld crack accident simulation and tracing module, 40 Weld bulge fracture accident simulation and tracing module, 50 Weld corrosion failure accident simulation and tracing module, 60 Multi-dimensional coupled accident simulation and tracing module. Detailed Implementation

[0020] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application.

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0023] This application provides a BIM-based method for tracing the entire lifecycle information of four boiler tubes, such as... Figure 1 As shown, the method includes: Step S100: Perform BIM modeling based on the four boiler tubes to obtain the four-tube BIM model, and perform full life cycle monitoring of the welds of the four boiler tubes based on the four-tube BIM model to obtain the dynamic model of the welds of the four boiler tubes. The four boiler tubes include water-cooled wall tubes, economizer tubes, superheater tubes and reheater tubes.

[0024] Specifically, BIM modeling, short for Building Information Modeling, is a building design and construction management method based on three-dimensional digital technology. It integrates various information about a building into a single model to facilitate collaborative work and information sharing among stakeholders. In this application, BIM software is used to create a three-dimensional model of the boiler's four tubes based on the actual design drawings, dimensional parameters, and material information. By defining the attributes of different components, such as the pipe wall thickness, diameter, and thermal expansion coefficient of the material, a four-tube BIM model is constructed. The four boiler tubes are water-cooled wall tubes, economizer tubes, superheater tubes, and reheater tubes.

[0025] Based on the four-pipe BIM model, sensor technology is integrated. For example, strain sensors and temperature sensors are installed at key locations along the weld. Strain sensors monitor stress changes in the weld in real time during operation, while temperature sensors monitor weld temperature to determine if overheating or other abnormalities are present. Data collected by the sensors is transmitted to the computer system in real time via a data acquisition system, such as a programmable logic controller (PLC) or data acquisition card. The four-pipe BIM model is updated based on the collected data to obtain a dynamic model of the four-pipe weld, which reflects the changes in the weld's state at different points in time.

[0026] In one possible implementation, based on the four-tube BIM model, full lifecycle monitoring of the welds of the four boiler tubes is performed to obtain a dynamic model of the four-tube welds. Step S100 further includes step S110, importing the structured data of the welds of the four boiler tubes into the four-tube BIM model to obtain a digital model of the four tubes. Specifically, structured data of the welds of the four boiler tubes is collected, including weld location information, i.e., the specific coordinate position of the weld on the four boiler tubes; material information, i.e., the type of welding material used in the weld; and specification information, such as the size and shape of the weld. The collected weld structured data is imported into the established four-tube BIM model. The four-tube BIM model integrates the weld data and associates it with the geometric model of the four tubes to form a digital model of the four tubes. In this model, each weld has its corresponding detailed information.

[0027] Step S120 involves cleaning the weld monitoring dataset of the four boiler tubes to obtain a weld monitoring data stream. Specifically, monitoring data of the boiler tube welds, including various parameters such as temperature, stress, and deformation, are collected in real time using various monitoring devices to form a weld monitoring dataset. Due to interference from various factors during monitoring, such as sensor malfunctions and environmental noise, the weld monitoring dataset contains outliers, missing values, and noisy data. Data cleaning removes these defective data, ensuring data quality and accuracy. Specific operations include identifying and removing outliers, filling in missing data, and using filtering algorithms to remove noisy data. After data cleaning, a clean and accurate weld monitoring data stream is obtained.

[0028] Step S130: The weld monitoring data stream is synchronized to the four-pipe digital model to generate the four-pipe weld dynamic model. Specifically, the cleaned weld monitoring data stream is synchronized to the four-pipe digital model in real time. For example, temperature monitoring data is assigned to the corresponding weld locations based on the weld location information. As the monitoring data is continuously updated, the various weld parameters in the four-pipe digital model also change dynamically, thereby generating the four-pipe weld dynamic model.

[0029] Step S200: Perform multi-degree-of-freedom risk feature focusing on the dynamic model of the four-pipe weld to obtain a four-pipe risk feature focusing map.

[0030] Multi-degree-of-freedom risk feature focusing refers to the analysis and focusing of the risk characteristics of boiler four-tube welds from multiple dimensions by integrating various risk factors and monitoring data throughout the entire life cycle monitoring process. Specifically, data mining and machine learning algorithms are used to analyze the data in the dynamic model of the four-tube welds. For example, principal component analysis is used to reduce the dimensionality of the monitoring data and extract the main characteristic parameters affecting weld risk, such as stress concentration, temperature fluctuation range, and corrosion rate. Then, clustering analysis algorithms are used to classify weld areas with similar risk characteristics, generating a four-tube risk feature focusing map. This map shows the risk level and risk feature distribution of different areas of the boiler four-tube welds.

[0031] In one possible implementation, multi-degree-of-freedom risk feature focusing is performed on the dynamic model of the four-tube weld to obtain a four-tube risk feature focusing map. Step S200 further includes step S210, performing weld crack risk feature focusing based on the dynamic model of the four-tube weld to obtain a first risk feature focusing map. Specifically, weld cracks are one of the common and serious defects in boiler four-tube systems. The dynamic model of the four-tube weld includes various real-time monitoring data of the weld, such as stress, temperature, and deformation. Crack formation is related to abnormal changes in these parameters. For example, when the stress on the weld exceeds the strength limit of its material, cracks may occur; rapid temperature changes may also cause thermal stress in the weld, thereby triggering cracks.

[0032] Specifically, by analyzing relevant data from the dynamic model of the four-pipe weld, and using stress-strain analysis and fracture mechanics models, risk factors that may lead to weld cracking were identified. These risk factors include areas of localized stress concentration and areas with large temperature fluctuations. These risk factors were then marked and visualized within the dynamic model of the four-pipe weld, forming a primary risk feature focus map. This map shows which areas of the weld are at risk of cracking and their distribution.

[0033] Step S220: Based on the dynamic model of the four-pipe weld, focus the risk features of weld bulge fracture to obtain a second risk feature focus map. Specifically, weld bulge fracture is caused by defects inside the weld, such as porosity and slag inclusions. When subjected to external loads, stress concentration occurs at these defects, leading to local bulging of the weld and ultimately fracture.

[0034] Specifically, deformation monitoring data from a dynamic model of a four-pipe weld is used, combined with principles of materials mechanics and structural mechanics, to analyze the risk of weld bulge fracture. By calculating the strain distribution of the weld and identifying abnormal deformation areas, potential locations for bulge fracture are determined. These risk locations are then marked and visualized in the model, generating a second risk feature focus map. This map illustrates the distribution of weld bulge fracture risk.

[0035] Step S230: Based on the dynamic model of the four-tube weld, focus the risk characteristics of weld corrosion failure to obtain a third risk characteristic focus map. Specifically, the environment in which the boiler's four tubes are located is relatively harsh, containing corrosive media such as water and impurities in steam. Under long-term exposure to corrosive media, the weld will undergo corrosion failure. The corrosion failure process is related to the material properties of the weld, environmental conditions, corrosion potential, and corrosion rate.

[0036] Specifically, corrosion-related monitoring data is extracted from the dynamic model of the four-pipe weld, and the corrosion failure risk of the weld is assessed by combining corrosion science theories and models. Risk points for weld corrosion failure are identified by analyzing corrosion data trends and identifying severely corroded areas. These risk points are then visualized in the model, forming a third risk feature focus map. This map illustrates the distribution of weld corrosion failure risk.

[0037] Step S240: Integrate the first risk feature focus map, the second risk feature focus map, and the third risk feature focus map to generate the four-pipe risk feature focus map. Specifically, the first, second, and third risk feature focus maps are superimposed and fused. During the fusion process, a weight allocation method can be used to assign corresponding weights to each risk feature based on the importance and impact of different risk types. Based on the weights, a comprehensive calculation and analysis of each risk feature is performed to generate the four-pipe risk feature focus map. This map shows the comprehensive distribution of various risks on the weld.

[0038] In one possible implementation, weld crack risk features are focused based on the four-tube weld dynamic model to obtain a first risk feature focus map. Step S210 further includes step S211, which reads the first monitoring feature sequence corresponding to the water-cooled wall tube weld based on the four-tube weld dynamic model. Specifically, the four-tube weld dynamic model integrates various real-time monitoring data of the welds of the water-cooled wall tubes, economizer tubes, superheater tubes, and reheater tubes. Each weld has its specific monitoring parameters, which reflect the health status of the weld. For the water-cooled wall tube weld, its monitoring feature sequence includes multiple key parameters, such as stress, temperature, and vibration frequency. From the four-tube weld dynamic model, all monitoring data related to the water-cooled wall tube weld are extracted through a data interface or query method and arranged in a certain time order or logical order to form the first monitoring feature sequence. For example, stress, temperature, and vibration frequency data can be recorded once per minute, and the data over a period of time can be organized into a sequence form.

[0039] Step S212 involves fitting a coordinate system based on the historical weld crack risk set to construct an analytical coordinate system for crack risk contribution. Specifically, the historical weld crack risk set contains various case data of cracks occurring in the boiler's four-tube welds over a past period. This data records information such as monitoring parameter values, environmental conditions, and operating conditions at the time of crack occurrence. First, the data in the historical weld crack risk set is preprocessed, including data cleaning and feature extraction. Then, statistical analysis methods, such as principal component analysis and correlation analysis, are used to determine the main monitoring features affecting crack risk. Next, a multi-dimensional coordinate system is constructed based on these main monitoring features, with each coordinate axis representing a monitoring feature. By fitting the crack occurrence data from the historical data, the weight and range of each monitoring feature in the coordinate system are determined, thereby constructing the analytical coordinate system for crack risk contribution.

[0040] Step S213: Evaluate the crack risk contribution of each monitoring feature in the first monitoring feature sequence according to the crack risk contribution analysis coordinate system to obtain the first crack risk contribution sequence. Specifically, substitute each monitoring feature value in the first monitoring feature sequence into the crack risk contribution analysis coordinate system, and calculate the crack risk contribution of each monitoring feature according to the weight and range of the coordinate system. For example, if the stress monitoring feature has a large weight in the coordinate system and the current stress value is high, then the stress monitoring feature will have a large contribution to the crack risk. Arrange the crack risk contributions of each monitoring feature in order to form the first crack risk contribution sequence.

[0041] Step S214: Attention enhancement is applied to the first monitoring feature sequence based on the first crack risk contribution sequence to obtain a first crack risk feature focusing vector. Specifically, attention enhancement is a data processing method used to highlight monitoring features with a significant impact on crack risk and weaken features with a smaller impact. Specifically, each monitoring feature in the first monitoring feature sequence is assigned a weight, which is proportional to its corresponding contribution in the first crack risk contribution sequence. Then, each monitoring feature is multiplied by its corresponding weight to obtain a weighted monitoring feature value. These weighted monitoring feature values ​​are combined into a vector, i.e., the first crack risk feature focusing vector. This vector can more accurately reflect the crack risk characteristics of the water-cooled wall pipe weld.

[0042] Step S215: Based on the first crack risk feature focusing vector, the weld crack risk features of the four boiler tubes are further focused according to the dynamic model of the four-tube weld, generating the first risk feature focusing map. Specifically, the process of focusing the weld crack risk features of the economizer tube, superheater tube, and reheater tube is similar to that of the water-cooled wall tube. The monitoring feature sequences corresponding to the welds of the economizer tube, superheater tube, and reheater tube are read from the dynamic model of the four-tube weld, respectively, and are denoted as the second monitoring feature sequence, the third monitoring feature sequence, and the fourth monitoring feature sequence. A crack risk contribution analysis coordinate system suitable for each tube is constructed based on the weld crack risk history set. According to the constructed crack risk contribution analysis coordinate system, the crack risk contribution of each monitoring feature in the second, third, and fourth monitoring feature sequences is evaluated, and the second crack risk contribution sequence, the third crack risk contribution sequence, and the fourth crack risk contribution sequence are obtained respectively. Based on the crack risk contribution sequence obtained above, attention enhancement is performed on the second, third, and fourth monitoring feature sequences respectively to obtain the second crack risk feature focus vector, the third crack risk feature focus vector, and the fourth crack risk feature focus vector.

[0043] Based on the first crack risk feature focusing vector, and the second, third, and fourth crack risk feature focusing vectors obtained in the same manner, a comprehensive crack risk feature focusing of the boiler's four tubes is performed according to the dynamic model of the four-tube welds. The crack risk information reflected by these four vectors is uniformly mapped to the corresponding weld locations of the tubes in the dynamic model of the four-tube welds. Visualization techniques, such as color coding and contour plotting, are used to display the crack risk level of different tube welds in the model. For example, different colors are used to represent different crack risk levels, with red indicating high risk and green indicating low risk, ultimately generating the first risk feature focusing map.

[0044] In one possible implementation, the risk features of weld bulge fracture are focused based on the dynamic model of the four-tube weld to obtain a second risk feature focus map. Step S220 further includes step S221, which involves fitting a coordinate system based on the historical set of weld bulge fracture risks to construct an analytical coordinate system for the contribution of bulge fracture risk. Specifically, the historical set of weld bulge fracture risks records various case data of bulge fractures in the four-tube welds of the boiler in the past, including information such as various monitoring parameters, operating conditions, and environmental conditions at the time of bulge fracture. First, the data in the historical set of weld bulge fracture risks is preprocessed, including removing outliers, filling missing values, and standardizing the data format. Then, multivariate statistical analysis methods, such as factor analysis and regression analysis, are used to determine the main monitoring features affecting the risk of bulge fracture. Then, a multidimensional coordinate system is constructed based on these main monitoring features, with each coordinate axis corresponding to a monitoring feature. By fitting the occurrence of bulge fractures in the historical data, the weight and value range of each monitoring feature in the coordinate system are determined, thereby constructing an analytical coordinate system for the contribution of bulge fracture risk.

[0045] Step S222: Evaluate the contribution of the first monitoring feature sequence to the bulge fracture risk according to the analytical coordinate system for bulge fracture risk contribution, thereby obtaining a first bulge fracture risk contribution sequence. Specifically, each monitoring feature value in the first monitoring feature sequence is sequentially substituted into the analytical coordinate system for bulge fracture risk contribution. Following the calculation method specified by the coordinate system and combining the weights of each monitoring feature, the bulge fracture risk contribution of each monitoring feature is obtained. These contributions are then arranged in the order of the monitoring features to form the first bulge fracture risk contribution sequence.

[0046] Step S223: Attention enhancement is applied to the first monitoring feature sequence based on the first bulge fracture risk contribution sequence to obtain a first bulge fracture risk feature focusing vector. Specifically, a weight is assigned to each monitoring feature in the first monitoring feature sequence, and this weight is proportional to the corresponding contribution in the first bulge fracture risk contribution sequence. Each monitoring feature is multiplied by its corresponding weight to obtain a weighted monitoring feature value. These weighted monitoring feature values ​​are combined into a vector, namely the first bulge fracture risk feature focusing vector, which reflects the bulge fracture risk characteristics of the water-cooled wall pipe weld.

[0047] Step S224: Based on the first bulge fracture risk feature focusing vector, the boiler's four tubes are further focused on the weld bulge fracture risk features according to the four-tube weld dynamic model, generating the second risk feature focusing map. Specifically, the information in the first bulge fracture risk feature focusing vector is mapped to the weld position of the water-cooled wall tube in the four-tube weld dynamic model. Simultaneously, following the same method as steps S221-S223, the economizer tube, superheater tube, and reheater tube are processed respectively to obtain the corresponding second, third, and fourth bulge fracture risk feature focusing vectors, which are then mapped to the weld positions of the corresponding tubes in the four-tube weld dynamic model. Through visualization technology, such as using different colors to represent different bulge fracture risk levels, the degree of bulge fracture risk of the four tube welds is displayed in the model, ultimately generating the second risk feature focusing map.

[0048] The method for focusing on the risk features of weld corrosion failure based on the dynamic model of the four-pipe weld to obtain the third risk feature focus spectrum is similar to the method for obtaining the first risk feature focus spectrum and the second risk feature focus spectrum, and will not be repeated here.

[0049] Step S300: Based on the risk feature focus map of the four tubes, perform weld crack accident simulation and tracing on the four tubes of the boiler to obtain the first simulation and tracing map.

[0050] Specifically, a weld crack accident simulation model is established by combining machine learning models and historical accident databases. Based on the risk characteristic focus map of the four pipes, weld crack accidents are simulated and traced back to trace the causes, development process, and consequences of weld crack accidents, generating a first simulation and tracing map that shows the possible development path, key nodes, and scope of impact of weld crack accidents.

[0051] In one possible implementation, the boiler's four tubes are used to extrapolate and trace weld crack accidents based on the risk feature focus map of the four tubes to obtain a first extrapolation and tracing map. Step S300 further includes step S310, which involves retrieving weld crack accident events for the water-cooled wall tubes to obtain a first weld crack accident event group. Specifically, a data retrieval system is established to filter water-cooled wall tube weld crack accident events that meet the requirements from multiple data sources such as the boiler operation database, equipment maintenance records, and accident report documents, according to predetermined search conditions, such as the time range of the accident, the specific location of the water-cooled wall tube, and the severity of the crack. These events are then organized and classified to form the first weld crack accident event group.

[0052] Step S320: Randomly perturb the first weld crack accident event group using an adversarial example generator to obtain a second weld crack accident event group. Specifically, an adversarial example generator is a tool capable of generating new data similar to the original data but with minor perturbations. Introducing adversarial examples increases the diversity and complexity of the data, simulating various uncertainties and anomalies that may occur in actual operation. By randomly perturbing the original event group, some weld crack accident events that may occur in reality but have not yet been recorded or observed are generated, thereby improving the model's generalization ability and adaptability to complex situations.

[0053] Using adversarial example generation algorithms, such as the fast gradient sign method and the projected gradient descent method, each event in the first weld crack accident event group is taken as input and randomly perturbed by the adversarial example generator. The perturbation methods include modifying some parameter values ​​in the event, adding or deleting some relevant information, etc. After processing, the second weld crack accident event group is obtained.

[0054] Step S330: Adversarial training is performed on multiple learners based on the second weld crack accident event group to obtain multiple crack accident inference models and multiple adversarial training loss datasets. Specifically, the second weld crack accident event group is divided into a training set and a validation set. Multiple different learners, such as decision trees, support vector machines, and neural networks, are selected and adversarially trained using the training set. During training, the parameters of the learners are continuously adjusted based on adversarial examples generated by the adversarial example generator, enabling the learners to accurately infer and trace weld crack accidents. Simultaneously, the loss value of each learner during training is recorded, forming multiple adversarial training loss datasets. These datasets are used to evaluate the training effect and performance of the models. After training, multiple crack accident inference models are obtained.

[0055] Step S340: The multiple crack accident simulation models are fused and trained based on the multiple adversarial training loss datasets to obtain a water-cooled wall tube crack accident simulation channel. Specifically, different learners have different characteristics and advantages. Fusing and training multiple crack accident simulation models is used to combine the strengths of each model and improve the overall model's performance and accuracy. A model fusion method is selected, such as weighted average, voting, or stacking. Based on the multiple adversarial training loss datasets, the weights or fusion methods of each crack accident simulation model are determined. Multiple models are fused and trained, and by continuously adjusting the fusion parameters, the fused model achieves optimal performance on the validation set, ultimately obtaining the water-cooled wall tube crack accident simulation channel.

[0056] Step S350: Input the first crack risk feature focusing vector into the water-cooled wall tube crack accident prediction channel to obtain the first crack accident prediction result. Specifically, the first crack risk feature focusing vector contains the current crack risk feature information of the water-cooled wall tube weld, such as the length, width, depth, location of the crack, and surrounding environmental parameters. Input this feature information into the water-cooled wall tube crack accident prediction channel. The integrated model in the channel performs comprehensive analysis and judgment based on this information to predict the probability of occurrence, development trend, and possible consequences of water-cooled wall tube weld crack accidents, thereby obtaining the first crack accident prediction result. This result can be presented in the form of probability value, risk level, development trend curve, etc.

[0057] Step S360: Based on the first crack accident simulation results, the boiler's four tubes are further simulated and traced for weld crack accidents according to the four-tube risk characteristic focusing map, generating the first simulation and tracing map. Specifically, based on the first crack accident simulation results and combined with the four-tube risk characteristic focusing map, the probability and development process of weld crack accidents in the economizer tube, superheater tube, and reheater tube are simulated and traced using a method similar to steps S310-S350. The simulation results of each tube are integrated and visualized to generate the first simulation and tracing map, which shows information such as the occurrence sequence, propagation path, and impact range of weld crack accidents in the boiler's four tubes.

[0058] Step S400: Based on the risk feature focus map of the four tubes, perform a retrospective analysis of the weld bulge fracture accident of the four tubes of the boiler to obtain a second retrospective analysis map.

[0059] Specifically, a model for predicting weld bulge fracture accidents is established by combining machine learning models and historical accident databases. Based on the risk characteristic focus map of the four pipes, the model for predicting and tracing weld bulge fracture accidents is conducted to trace the causes, development process, and consequences of weld bulge fracture accidents, generating a second prediction and tracing map that shows the possible development path, key nodes, and scope of impact of weld bulge fracture accidents.

[0060] In one possible implementation, based on the risk characteristic focus map of the four tubes, the boiler's four tubes are used to extrapolate and trace weld bulge fracture accidents to obtain a second extrapolation and tracing map. Step S400 further includes step S410, which involves retrieving weld bulge fracture accident events for the water-cooled wall tubes to obtain a first weld bulge fracture accident event group. Specifically, similar to step S310, a data retrieval system is established to filter out water-cooled wall tube weld bulge fracture accident events that meet the requirements from multiple data sources such as the boiler operation database, equipment maintenance records, and accident report documents, according to predetermined retrieval conditions, such as the time range of the accident, the specific location of the water-cooled wall tube, and the severity of the bulge fracture. These events are then organized and classified to form the first weld bulge fracture accident event group.

[0061] Step S420: Randomly perturb the first weld bulge fracture accident event group to obtain the second weld bulge fracture accident event group. Specifically, similar to step S320, an adversarial example generation algorithm, such as the fast gradient sign method or projective gradient descent method, is used. Each event in the first weld bulge fracture accident event group is taken as input, and the adversarial example generator randomly perturbs it. The perturbation methods include modifying certain parameter values ​​in the event, adding or deleting some relevant information, etc. After processing, the second weld bulge fracture accident event group is obtained.

[0062] Step S430: Adversarial training is performed on multiple learners based on the second weld bulge fracture accident event group to obtain multiple bulge fracture accident inference models and multiple training loss datasets. Specifically, similar to step S330, the second weld bulge fracture accident event group is divided into a training set and a validation set. Multiple different learners are selected, and adversarial training is performed on each using the training set. During training, the parameters of the learners are continuously adjusted based on adversarial examples generated by the adversarial example generator, enabling the learners to accurately infer and trace weld bulge fracture accidents. Simultaneously, the loss value of each learner during training is recorded, forming multiple adversarial training loss datasets. These datasets are used to evaluate the training effect and performance of the models. After training, multiple bulge fracture accident inference models are obtained.

[0063] Step S440: The multiple bulge fracture accident simulation models are fused and trained based on the multiple training loss datasets to obtain a simulation channel for water-cooled wall tube bulge fracture accidents. Specifically, similar to step S340, a model fusion method is selected, and the weights or fusion method of each bulge fracture accident simulation model are determined based on multiple adversarial training loss datasets. Multiple models are fused and trained, and the fusion parameters are continuously adjusted to optimize the performance of the fused model on the validation set, ultimately obtaining the simulation channel for water-cooled wall tube bulge fracture accidents.

[0064] Step S450: Input the first bulge fracture risk feature focusing vector into the water-cooled wall tube bulge fracture accident simulation channel to obtain the first bulge fracture accident simulation result. Specifically, similar to step S350, the first bulge fracture risk feature focusing vector contains the current bulge fracture risk feature information of the water-cooled wall tube weld, such as the length, width, depth, location of the bulge fracture, and surrounding environmental parameters. Input this feature information into the water-cooled wall tube bulge fracture accident simulation channel. The integrated model in the channel performs comprehensive analysis and judgment based on this information to predict the probability of occurrence, development trend, and possible consequences of the water-cooled wall tube weld bulge fracture accident, thereby obtaining the first bulge fracture accident simulation result. This result can be presented in the form of probability value, risk level, development trend curve, etc.

[0065] Step S460: Based on the simulation results of the first bulge fracture accident, the boiler's four tubes are further simulated and traced for weld bulge fracture accidents according to the risk feature focus map of the four tubes, generating the second simulation and tracing map. Specifically, similar to step S360, based on the simulation results of the first bulge fracture accident and combined with the risk feature focus map of the four tubes, the probability and development process of weld bulge fracture accidents of economizer tubes, superheater tubes, and reheater tubes are simulated and traced using methods similar to steps S410-S450. The simulation results of each tube are integrated and visualized to generate the second simulation and tracing map, which shows the occurrence sequence, propagation path, and impact range of weld bulge fracture accidents of the boiler's four tubes.

[0066] Step S500: Based on the risk characteristic focus map of the four tubes, perform weld corrosion failure accident simulation and tracing on the four tubes of the boiler to obtain the third simulation and tracing map.

[0067] Specifically, similar to steps S300 and S400, weld corrosion failure accident simulation and tracing are performed. Each of the four pipes is analyzed for its weld corrosion failure accidents. First, relevant accident event data is collected to construct an initial event cluster. Then, the event cluster is randomly perturbed to enhance model adaptability. Next, the perturbed event cluster is used to perform adversarial training on multiple learners, obtaining multiple corrosion failure accident simulation models for that pipe and corresponding training loss datasets. Then, based on these training loss datasets, the multiple simulation models are fused and trained to obtain the weld corrosion failure accident simulation channel for that pipe. Finally, the corrosion risk feature focus vector of that pipe is input into the simulation channel to obtain the weld corrosion failure accident simulation result for that pipe. After completing the weld corrosion failure accident simulation for each of the four pipes, the simulation results for the four pipes are integrated to generate a third simulation and tracing map.

[0068] Step S600: Perform multi-dimensional coupled accident simulation and tracing based on the first simulation and tracing map, the second simulation and tracing map, and the third simulation and tracing map to obtain a four-tube simulation and tracing map.

[0069] Specifically, the first, second, and third simulation traceability maps are integrated and analyzed. System integration technology and multi-factor coupling analysis methods are used to analyze the mutual influence and correlation between different accident types. For example, weld cracks may lead to stress concentration, thereby accelerating the corrosion process; while corrosion reduces the strength of the pipeline, increasing the risk of bulging fracture. By comprehensively considering these factors, a comprehensive simulation traceability of various accidents that may occur in the four boiler tubes under different operating conditions is performed. Using visualization technology, the multi-dimensional coupled accident simulation results are used to generate a four-tube simulation traceability map, demonstrating the various accident combinations that may occur in the four boiler tubes under complex operating conditions and their development process.

[0070] In one possible implementation, multi-dimensional coupled accident simulation and tracing are performed based on the first, second, and third simulation and tracing maps to obtain a four-tube simulation and tracing map. Step S600 further includes step S610, performing weld crack coupled accident simulation and tracing on the four boiler tubes based on the first, second, and third simulation and tracing maps to obtain a first coupled accident simulation and tracing result. Specifically, during the operation of the four boiler tubes, different types of weld accidents do not occur in isolation but are interconnected and influence each other. For example, weld corrosion can lead to a decline in material properties, thereby increasing the risk of weld crack formation; stress changes generated when a weld bulge fractures can also trigger cracks in nearby welds. The first simulation and tracing map performs separate simulation and tracing for weld crack accidents, but the occurrence and development of weld cracks may be affected by other accident types such as weld bulge fractures and weld corrosion failures.

[0071] Information directly related to weld crack accidents, such as the initial location, morphology, size, propagation direction, and time sequence of crack occurrence, is extracted from the first retrospective analysis map. Simultaneously, key data on weld bulge fracture accidents, including the location, size, formation time, and stress state at fracture, are obtained from the second retrospective analysis map. Information on weld corrosion failure accidents, such as the location, extent, rate, and distribution of corrosive media, is collected from the third retrospective analysis map. This data from different maps is then integrated to establish a unified dataset.

[0072] By applying theoretical knowledge from engineering mechanics, materials science, and corrosion science, and combining it with practical operational experience, a coupling relationship model between weld cracks and other accident types is established. For example, based on stress-strength interference theory, the model analyzes how the stress generated by weld bulge fracture affects the weld's crack resistance; using a corrosion kinetic model, it analyzes the promoting effect of weld corrosion on crack initiation and propagation. These models describe the interaction mechanisms between different accident types. Using the integrated dataset as input, the established coupling relationship models are used for extrapolation and retrospection, starting from the initial state, simulating how weld bulge fracture and weld corrosion failure affect the occurrence and development of weld cracks under different operating conditions. The results of the extrapolation and retrospection are analyzed to summarize the key influencing factors, development patterns, and possible consequences of weld crack coupling accidents. For example, it determines under what corrosion levels and bulge fracture conditions weld cracks are most likely to propagate and become unstable. The analysis results are compiled into a first coupling accident extrapolation and retrospection report.

[0073] Step S620: Based on the first, second, and third simulation traceability maps, a simulation traceability of the weld bulge fracture coupled accident is performed on the four boiler tubes to obtain the second coupled accident simulation traceability result. Specifically, the weld bulge fracture accident is also potentially affected by other accident types. For example, the presence of weld cracks can disrupt the structural continuity of the weld, leading to local stress concentration and increasing the likelihood of bulge formation; weld corrosion can thin the weld material, reduce its strength, and decrease its resistance to bulge deformation.

[0074] Based on the principles of plasticity and fracture mechanics, a coupling relationship model between weld bulge fracture and other accident types is constructed. For example, based on the influence of stress concentration caused by weld cracks on the initiation and development of bulges, a relationship model between the stress concentration factor and the bulge size growth is established. Combining the influence of corrosion on material properties, the critical stress criterion for bulge fracture is modified to reflect the actual strength changes of materials under corrosive environments. Using the integrated dataset from step S610 as input, the established coupling relationship model is used for simulation and retrospective analysis, simulating how weld cracks and weld corrosion interact under different operating conditions, affecting the formation, development, and fracture process of weld bulges. The simulation and retrospective results are analyzed to summarize the key influencing factors, development patterns, and potential risks of coupled accidents involving weld bulge fracture. For example, it is determined at what crack size and corrosion level weld bulges are most prone to fracture. The analysis results are compiled into a second coupled accident simulation and retrospective report.

[0075] Step S630: Based on the first, second, and third simulation traceability maps, perform a simulation traceability of the coupled accident of weld corrosion failure in the four boiler tubes to obtain the simulation traceability result of the third coupled accident. Specifically, similar to steps S610 and S620, the weld corrosion failure process is not independent. Weld cracks provide an intrusion channel for corrosive media, accelerating the corrosion process; weld bulge fracture leads to changes in local environmental conditions, such as changes in temperature, pressure, and media flow state, thereby affecting the corrosion rate and distribution.

[0076] Using electrochemical corrosion theory and fluid mechanics principles, a coupling model between weld corrosion failure and other accident types is constructed. For example, a model is established to relate crack size to corrosion current density, analyzing the impact of the additional surface area provided by the crack on the corrosion rate; the influence of changes in medium flow caused by bulging fracture on the transport and deposition of corrosion products is analyzed, establishing a correlation model between fluid dynamics and corrosion product distribution. Using the integrated dataset from step S610 as input, the established coupling model is used for extrapolation and tracing, simulating how weld cracks and weld bulging fractures interact under different operating conditions, affecting the initiation, development, and failure process of weld corrosion. The extrapolation and tracing results are analyzed to summarize the key influencing factors, development patterns, and potential hazards of coupled weld corrosion failure accidents. For example, it is determined under which crack and bulging conditions the weld corrosion rate is fastest. The analysis results are compiled into a third coupled accident extrapolation and tracing report.

[0077] Step S640 involves correlating and analyzing the first, second, and third inductive tracing maps, the first coupled accident inductive tracing result, the second coupled accident inductive tracing result, and the third coupled accident inductive tracing result to generate the four-pipe inductive tracing map. Specifically, the inherent relationships between the first, second, and third inductive tracing maps and the three coupled accident inductive tracing results are analyzed to determine the causal relationships, temporal sequence, and spatial distribution among different accident types, and to identify key nodes and influencing factors between accidents. For example, it analyzes how weld cracks lead to weld bulge fracture, which in turn accelerates weld corrosion; or how weld corrosion weakens the weld structure, increasing the risk of crack and bulge fracture.

[0078] Based on the results of data correlation analysis, a structure for the four-tube retrospective map is designed. This can be implemented using hierarchical, network, or flowchart formats, with the four boiler tubes as the core nodes and different types of accidents and coupling relationships as child nodes, represented by arrows or lines. For example, with the four boiler tubes as the center, three main branches can be derived: weld cracks, weld bulging fractures, and weld corrosion failures. Each branch can be further subdivided into detailed information on individual accidents and coupled accidents, demonstrating the hierarchical relationships and interactions between accidents.

[0079] Key information from various simulation and tracing results is extracted and integrated, and then displayed in a suitable way in the graph. For example, different colors, shapes, or symbols can be used to distinguish different types of accidents and coupling relationships, and necessary text descriptions and annotations can be added to make the graph clearer and easier to understand. For example, red can be used to represent serious accidents, and yellow can represent potential risk accidents; solid lines can represent direct causal relationships, and dashed lines can represent indirect influence relationships, etc.

[0080] This application embodiment employs BIM modeling of the four boiler tubes to obtain a four-tube BIM model, and uses this model to perform full life-cycle monitoring of the welds, obtaining a dynamic model of the four-tube welds. Multi-degree-of-freedom risk feature focusing is then performed on the dynamic model of the four-tube welds to generate a four-tube risk feature focusing map. Based on the four-tube risk feature focusing map, weld crack, weld bulge fracture, and weld corrosion failure accidents are simulated and traced for the four boiler tubes, resulting in first, second, and third simulation and tracing maps. Combining the first, second, and third simulation and tracing maps, multi-dimensional coupled accident simulation and tracing are performed to form a four-tube simulation and tracing map. These technical means solve the technical problems of insufficient accuracy and efficiency in existing boiler four-tube information tracing, achieving the technical effect of accurate and efficient full life-cycle information tracing for boiler four tubes.

[0081] In the above text, refer to Figure 1 This paper describes in detail a BIM-based method for tracing the entire lifecycle information of four tubes in a boiler according to an embodiment of the present invention. Next, we will refer to... Figure 2 This invention describes a BIM-based boiler four-tube full lifecycle information traceability system according to an embodiment of the present invention.

[0082] The BIM-based boiler four-tube full lifecycle information traceability system according to embodiments of the present invention addresses the technical problems of insufficient accuracy and efficiency in existing boiler four-tube information traceability, achieving precise and efficient technical effects in tracing the full lifecycle information of boiler four-tubes. The BIM-based boiler four-tube full lifecycle information traceability system includes: a weld full lifecycle monitoring module 10, a multi-degree-of-freedom risk feature focusing module 20, a weld crack accident simulation and traceability module 30, a weld bulge fracture accident simulation and traceability module 40, a weld corrosion failure accident simulation and traceability module 50, and a multi-dimensional coupled accident simulation and traceability module 60.

[0083] The weld lifecycle monitoring module 10 is used to perform BIM modeling on the four boiler tubes to obtain a four-tube BIM model, and to perform full lifecycle monitoring of the welds on the four boiler tubes based on the four-tube BIM model to obtain a dynamic model of the four-tube welds; the multi-degree-of-freedom risk feature focusing module 20 is used to perform multi-degree-of-freedom risk feature focusing on the dynamic model of the four-tube welds to obtain a risk feature focusing map of the four tubes; the weld crack accident inference and tracing module 30 is used to perform weld crack accident inference and tracing on the four boiler tubes based on the risk feature focusing map of the four tubes to obtain a first inference and tracing map; weld bulge The fracture accident simulation and tracing module 40 is used to perform weld bulge fracture accident simulation and tracing on the four boiler tubes based on the risk feature focusing map of the four tubes, and obtain a second simulation and tracing map; the weld corrosion failure accident simulation and tracing module 50 is used to perform weld corrosion failure accident simulation and tracing on the four boiler tubes based on the risk feature focusing map of the four tubes, and obtain a third simulation and tracing map; the multidimensional coupled accident simulation and tracing module 60 is used to perform multidimensional coupled accident simulation and tracing based on the first simulation and tracing map, the second simulation and tracing map and the third simulation and tracing map, and obtain a four-tube simulation and tracing map.

[0084] The detailed description of the specific configuration of the multi-degree-of-freedom risk feature focusing module 20 is explained as follows: As mentioned above, multi-degree-of-freedom risk feature focusing is performed on the dynamic model of the four-pipe weld to obtain a four-pipe risk feature focusing map. The multi-degree-of-freedom risk feature focusing module 20 may further include: a weld crack risk feature focusing unit for performing weld crack risk feature focusing based on the dynamic model of the four-pipe weld to obtain a first risk feature focusing map; a weld bulge fracture risk feature focusing unit for performing weld bulge fracture risk feature focusing based on the dynamic model of the four-pipe weld to obtain a second risk feature focusing map; a weld corrosion failure risk feature focusing unit for performing weld corrosion failure risk feature focusing based on the dynamic model of the four-pipe weld to obtain a third risk feature focusing map; and a map integration unit for integrating the first risk feature focusing map, the second risk feature focusing map, and the third risk feature focusing map to generate the four-pipe risk feature focusing map.

[0085] The process involves focusing weld crack risk features based on the dynamic model of the four-tube weld to obtain a first risk feature focus map. The weld crack risk feature focus unit may further include: a first monitoring feature sequence reading subunit for reading the first monitoring feature sequence corresponding to the water-cooled wall tube weld based on the dynamic model of the four-tube weld; a coordinate system fitting subunit for fitting a coordinate system based on the historical set of weld crack risks to construct an analytical coordinate system for crack risk contribution; a crack risk contribution evaluation subunit for evaluating the crack risk contribution of each monitoring feature in the first monitoring feature sequence based on the analytical coordinate system for crack risk contribution to obtain a first crack risk contribution sequence; an attention enhancement subunit for enhancing the attention of the first monitoring feature sequence based on the first crack risk contribution sequence to obtain a first crack risk feature focus vector; and a first risk feature focus map generation subunit for continuing to focus weld crack risk features on the four boiler tubes based on the first crack risk feature focus vector and the dynamic model of the four-tube weld to generate the first risk feature focus map.

[0086] Specifically, the weld bulge fracture risk feature focusing unit can further include: a coordinate system fitting subunit for fitting a coordinate system based on the historical set of weld bulge fracture risks to construct an analytical coordinate system for bulge fracture risk contribution; a bulge fracture risk contribution evaluation subunit for evaluating the contribution of the first monitoring feature sequence to bulge fracture risks based on the analytical coordinate system to obtain a first bulge fracture risk contribution sequence; an attention enhancement subunit for enhancing the attention of the first monitoring feature sequence based on the first bulge fracture risk contribution sequence to obtain a first bulge fracture risk feature focusing vector; and a second risk feature focusing map generation subunit for continuing to focus on the weld bulge fracture risk features of the four boiler tubes based on the first bulge fracture risk feature focusing vector and the four-tube weld dynamic model to generate the second risk feature focusing map.

[0087] The detailed description of the specific configuration of the weld crack accident inference and tracing module 30 is explained as follows: As mentioned above, the weld crack accident inference and tracing of the four boiler tubes is performed based on the risk feature focusing map of the four tubes to obtain a first inference and tracing map. The weld crack accident inference and tracing module 30 may further include: a weld crack accident event retrieval unit for retrieving weld crack accident events for the water-cooled wall tubes to obtain a first weld crack accident event group; a random perturbation unit for randomly perturbing the first weld crack accident event group through an adversarial sample generator to obtain a second weld crack accident event group; and an adversarial training unit for training the second weld crack accident event group. Multiple learners undergo adversarial training to obtain multiple crack accident simulation models and multiple adversarial training loss datasets. A fusion training unit is used to perform fusion training on the multiple crack accident simulation models based on the multiple adversarial training loss datasets to obtain a water-cooled wall tube crack accident simulation channel. A first crack accident simulation unit is used to input a first crack risk feature focusing vector into the water-cooled wall tube crack accident simulation channel to obtain a first crack accident simulation result. A first simulation tracing map generation unit is used to continue to perform weld crack accident simulation tracing on the four boiler tubes based on the first crack accident simulation result and the four tube risk feature focusing map to generate the first simulation tracing map.

[0088] The detailed description of the specific configuration of the weld bulge fracture accident simulation and tracing module 40 is explained as follows: As mentioned above, the weld bulge fracture accident simulation and tracing of the four boiler tubes is performed based on the risk feature focusing map of the four tubes to obtain a second simulation and tracing map. The weld bulge fracture accident simulation and tracing module 40 may further include: a weld bulge fracture accident event retrieval unit for retrieving weld bulge fracture accident events for water-cooled wall tubes to obtain a first weld bulge fracture accident event group; a random perturbation unit for randomly perturbing the first weld bulge fracture accident event group to obtain a second weld bulge fracture accident event group; and an adversarial training unit for training multiple learning systems based on the second weld bulge fracture accident event group. The learning instrument performs adversarial training to obtain multiple bulge fracture accident simulation models and multiple training loss datasets; the fusion training unit is used to perform fusion training on the multiple bulge fracture accident simulation models based on the multiple training loss datasets to obtain a water-cooled wall tube bulge fracture accident simulation channel; the first bulge fracture accident simulation unit is used to input the first bulge fracture risk feature focusing vector into the water-cooled wall tube bulge fracture accident simulation channel to obtain the first bulge fracture accident simulation result; the second simulation tracing map generation unit is used to continue to simulate and trace the weld bulge fracture accident of the four boiler tubes based on the first bulge fracture accident simulation result and the four tube risk feature focusing map, and generate the second simulation tracing map.

[0089] The detailed configuration of the multidimensional coupled accident simulation and tracing module 60 is explained below: As mentioned above, multidimensional coupled accident simulation and tracing is performed based on the first, second, and third simulation and tracing maps to obtain a four-tube simulation and tracing map. The multidimensional coupled accident simulation and tracing module 60 may further include: a weld crack coupled accident simulation and tracing unit used to perform weld crack coupled accident simulation and tracing on the four boiler tubes based on the first, second, and third simulation and tracing maps to obtain a first coupled accident simulation and tracing result; and a weld bulge fracture coupled accident simulation and tracing unit used to perform weld crack coupled accident simulation and tracing based on the first and second simulation and tracing maps. The third inference and tracing diagram is used to perform weld bulging fracture coupled accident inference and tracing on the four boiler tubes to obtain the second coupled accident inference and tracing result; the weld corrosion failure coupled accident inference and tracing unit is used to perform weld corrosion failure coupled accident inference and tracing on the four boiler tubes according to the first inference and tracing diagram, the second inference and tracing diagram and the third inference and tracing diagram to obtain the third coupled accident inference and tracing result; the association and sorting unit is used to perform association and sorting on the first inference and tracing diagram, the second inference and tracing diagram, the third inference and tracing diagram, the first coupled accident inference and tracing result, the second coupled accident inference and tracing result and the third coupled accident inference and tracing result to generate the four tubes inference and tracing diagram.

[0090] The detailed description of the specific configuration of the weld life cycle monitoring module 10 is as follows: As mentioned above, the full life cycle monitoring of the welds of the four boiler tubes is performed based on the four-tube BIM model to obtain a dynamic model of the four-tube welds. The weld life cycle monitoring module 10 may further include: a weld structured data import unit for importing the weld structured data of the four boiler tubes into the four-tube BIM model to obtain a digital model of the four tubes; a data cleaning unit for cleaning the weld monitoring dataset of the four boiler tubes to obtain a weld monitoring data stream; and a weld monitoring data stream synchronization unit for synchronizing the weld monitoring data stream to the digital model of the four tubes to generate the dynamic model of the four-tube welds.

[0091] The weld life cycle monitoring module 10 may further include: the boiler four tubes include water-cooled wall tubes, economizer tubes, superheater tubes and reheater tubes.

[0092] The BIM-based boiler four-tube full life cycle information traceability system provided in this embodiment of the invention can execute the BIM-based boiler four-tube full life cycle information traceability method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0093] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0094] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A BIM-based method for tracing the entire lifecycle information of four boiler tubes, characterized in that, The method includes: Based on the four tubes of the boiler, BIM modeling is performed to obtain the four tubes BIM model, and the full life cycle monitoring of the welds of the four tubes of the boiler is performed based on the four tubes BIM model to obtain the four tubes weld dynamic model. Multi-degree-of-freedom risk feature focusing is performed on the dynamic model of the four-pipe weld to obtain the risk feature focusing map of the four-pipe weld. Based on the risk characteristic focus map of the four tubes, the weld crack accident of the four tubes of the boiler is simulated and traced to obtain the first simulation and traceability map; Based on the risk characteristic focus map of the four tubes, the boiler four tubes are used to simulate and trace the weld bulge fracture accident to obtain the second simulation and tracing map. Based on the risk characteristic focus map of the four tubes, the weld corrosion failure accident of the four tubes of the boiler is simulated and traced to obtain the third simulation and traceability map. Based on the first, second, and third inference and tracing maps, multidimensional coupled accident inference and tracing are performed to obtain a four-tube inference and tracing map.

2. The BIM-based boiler four-tube full lifecycle information traceability method as described in claim 1, characterized in that, Multi-degree-of-freedom risk feature focusing was performed on the dynamic model of the four-pipe weld to obtain a risk feature focusing map of the four pipes, including: Based on the dynamic model of the four-pipe weld, the risk features of weld cracks are focused to obtain the first risk feature focus map. Based on the dynamic model of the four-pipe weld, the risk features of weld bulge fracture are focused to obtain a second risk feature focus map. Based on the dynamic model of the four-pipe weld, the risk characteristics of weld corrosion failure are focused to obtain the third risk characteristic focus map. The first risk feature focus map, the second risk feature focus map, and the third risk feature focus map are integrated to generate the four-tube risk feature focus map.

3. The BIM-based boiler four-tube full lifecycle information traceability method as described in claim 2, characterized in that, Based on the dynamic model of the four-pipe weld, weld crack risk features are focused to obtain a first risk feature focus map, including: Based on the dynamic model of the four-pipe weld, the first monitoring feature sequence corresponding to the water-cooled wall pipe weld is read; Based on the historical set of weld crack risk, a coordinate system is fitted to construct an analytical coordinate system for crack risk contribution. The crack risk contribution of each monitoring feature in the first monitoring feature sequence is evaluated according to the crack risk contribution analytical coordinate system to obtain the first crack risk contribution sequence. Attention enhancement is performed on the first monitoring feature sequence based on the first crack risk contribution sequence to obtain the first crack risk feature focusing vector; Based on the first crack risk feature focusing vector, the weld crack risk features of the four boiler tubes are further focused according to the dynamic model of the four tube welds to generate the first risk feature focusing map.

4. The BIM-based boiler four-tube full lifecycle information traceability method as described in claim 3, characterized in that, Based on the dynamic model of the four-pipe weld, the risk characteristics of weld bulge fracture are focused to obtain a second risk characteristic focus map, including: Based on the historical set of weld bulge fracture risk, a coordinate system is fitted to construct an analytical coordinate system for the contribution of bulge fracture risk. The contribution of the bulge fracture risk to the first monitoring feature sequence is evaluated according to the analytical coordinate system of the bulge fracture risk contribution, and the first bulge fracture risk contribution sequence is obtained. Attention enhancement is performed on the first monitoring feature sequence based on the first bulge fracture risk contribution sequence to obtain the first bulge fracture risk feature focus vector; Based on the first bulge fracture risk feature focusing vector, the boiler four tubes are further focused on the weld bulge fracture risk features according to the dynamic model of the four-tube weld, generating the second risk feature focusing map.

5. The method for tracing the entire lifecycle information of boiler four-tube system based on BIM as described in claim 1, characterized in that, Based on the risk characteristic focus map of the four tubes, a weld crack accident simulation and tracing was performed on the four tubes of the boiler to obtain a first simulation and tracing map, including: We conducted a weld crack accident event retrieval on the water-cooled wall tubes to obtain the first weld crack accident event group; The first weld crack accident event group is randomly perturbed by an adversarial sample generator to obtain the second weld crack accident event group. Based on the second weld crack accident event group, multiple learners are subjected to adversarial training to obtain multiple crack accident inference models and multiple adversarial training loss datasets. The multiple crack accident simulation models are fused and trained based on the multiple adversarial training loss datasets to obtain a water-cooled wall tube crack accident simulation channel. The first crack risk feature focusing vector is input into the water-cooled wall tube crack accident simulation channel to obtain the first crack accident simulation result; Based on the first crack accident simulation results, the weld crack accident simulation and tracing of the four boiler tubes is continued according to the risk feature focus map of the four tubes, and the first simulation and tracing map is generated.

6. The method for tracing the entire lifecycle information of a boiler's four tubes based on BIM as described in claim 1, characterized in that, Based on the risk characteristic focus map of the four tubes, a second simulation and tracing map of weld bulge fracture accidents was obtained for the four tubes of the boiler, including: A search was conducted on weld bulge fracture accident events of water-cooled wall tubes to obtain the first weld bulge fracture accident event group. Randomly perturb the first weld bulge fracture accident event group to obtain the second weld bulge fracture accident event group. Based on the second weld bulge fracture accident event group, multiple learners are subjected to adversarial training to obtain multiple bulge fracture accident inference models and multiple training loss datasets. Based on the multiple training loss datasets, the multiple bulge fracture accident simulation models are fused and trained to obtain a water-cooled wall tube bulge fracture accident simulation channel. Input the first bulge fracture risk feature focusing vector into the water-cooled wall tube bulge fracture accident simulation channel to obtain the first bulge fracture accident simulation result; Based on the first bulge fracture accident simulation results, the boiler's four tubes are further simulated and traced for weld bulge fracture accidents according to the four tube risk characteristic focus map, generating the second simulation and tracing map.

7. The method for tracing the entire lifecycle information of a boiler's four tubes based on BIM as described in claim 1, characterized in that, Based on the first, second, and third inference and tracing maps, a multi-dimensional coupled accident inference and tracing is performed to obtain a four-pipe inference and tracing map, including: Based on the first, second, and third simulation and tracing diagrams, the weld crack coupling accident simulation and tracing of the four boiler tubes is performed to obtain the first coupling accident simulation and tracing result. Based on the first, second, and third simulation and tracing diagrams, the boiler's four tubes were simulated and traced for a coupled accident involving weld bulging fracture, and the simulation and tracing results for the second coupled accident were obtained. Based on the first, second, and third inference and tracing diagrams, the weld corrosion failure coupled accident of the four tubes of the boiler is simulated and traced to obtain the third coupled accident simulation and tracing result. The first inference and tracing map, the second inference and tracing map, the third inference and tracing map, the first coupled accident inference and tracing result, the second coupled accident inference and tracing result, and the third coupled accident inference and tracing result are correlated and sorted to generate the four-pipe inference and tracing map.

8. The BIM-based boiler four-tube full lifecycle information traceability method as described in claim 1, characterized in that, Based on the four-tube BIM model, perform full lifecycle monitoring of the weld seams of the four boiler tubes to obtain a dynamic model of the four-tube weld seams, including: The weld structure data of the four boiler tubes are imported into the four tubes BIM model to obtain the digital model of the four tubes; Data cleaning was performed on the weld monitoring dataset of the four boiler tubes to obtain the weld monitoring data stream; The weld monitoring data stream is synchronized to the four-pipe digital model to generate the four-pipe weld dynamic model.

9. The BIM-based boiler four-tube full lifecycle information traceability method as described in claim 1, characterized in that, The boiler has four tubes: water-cooled wall tubes, economizer tubes, superheater tubes, and reheater tubes.

10. A BIM-based boiler four-tube full lifecycle information traceability system, characterized in that, The system is used to implement the BIM-based boiler four-tube full life cycle information traceability method according to any one of claims 1-9, and the system includes: The weld life cycle monitoring module is used to perform BIM modeling based on the four tubes of the boiler, obtain the four tubes BIM model, and perform full life cycle monitoring of the welds of the four tubes of the boiler based on the four tubes BIM model to obtain the four tubes weld dynamic model. The multi-degree-of-freedom risk feature focusing module is used to perform multi-degree-of-freedom risk feature focusing on the dynamic model of the four-pipe weld seam to obtain the four-pipe risk feature focusing map. The weld crack accident simulation and tracing module is used to perform weld crack accident simulation and tracing on the four tubes of the boiler based on the risk feature focusing map of the four tubes, and obtain the first simulation and tracing map; The weld bulge fracture accident simulation and tracing module is used to simulate and trace weld bulge fracture accidents of the four boiler tubes based on the risk characteristic focusing map of the four tubes, and obtain a second simulation and tracing map; The weld corrosion failure accident simulation and tracing module is used to perform weld corrosion failure accident simulation and tracing on the four boiler tubes based on the risk characteristic focusing map of the four tubes, and obtain a third simulation and tracing map. The multidimensional coupled accident simulation and tracing module is used to perform multidimensional coupled accident simulation and tracing based on the first simulation and tracing map, the second simulation and tracing map and the third simulation and tracing map to obtain a four-tube simulation and tracing map.

Images