Steel structure power room construction quality whole process management and control method based on digital twinning

Abstract

The application discloses a kind of based on digital twinning steel structure power room construction quality whole process management and control method, belong to industrial intelligent manufacturing technical field;The method comprises the following steps: obtaining construction object information and carrying out hierarchical division, constructs and trains three-layer digital twinning model;Acquisition and processing construction data;The layered quality analysis is carried out to the construction data after processing by three-layer digital twinning model, when detecting quality anomaly, start cross-layer association analysis and locate problem root;According to the analysis result of the layered quality analysis, trigger hierarchical early warning, and generate quantitative construction adjustment instruction based on the simulation of three-layer digital twinning model;The application realizes the high-fidelity virtual mapping of construction whole element by constructing basic-structure-construction three-layer digital twinning model and establishing cross-layer parameter association rule.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial intelligent manufacturing, in particular to a steel structure power house construction quality whole process management and control method based on digital twinning. BACKGROUND

[0002] As the core facility of the power system, the construction quality of the steel structure power house is directly related to the safety, stability and service life of the overall structure. With the development of modern engineering construction towards large-scale and complex, the traditional construction quality management method has been difficult to meet the high-standard quality control demand. At present, the following technical limitations exist:

[0003] Firstly, the existing monitoring means has obvious shortcomings, the traditional quality control mainly relies on manual inspection and post-detection, not only the response is lagging, the average discovery time of quality problems is more than 24 hours, and it is greatly affected by personnel experience and has high omission rate; although some researches have tried to use digital twinning technology for monitoring, the monitoring dimension is single, only focuses on limited parameters such as structure displacement and stress, and cannot include influencing factors such as geological conditions and construction operation into a unified analysis framework, so that the root cause of quality problems cannot be accurately traced.

[0004] Secondly, the model and the entity are disconnected, the existing digital twinning model mostly uses static architecture, the parameter threshold is fixed and cannot be changed, which cannot adapt to dynamic factors such as material performance fluctuation and environmental change, at the same time, the model lacks effective precision checking mechanism, and it is difficult to ensure the consistency of virtual model and physical entity, so the monitoring precision is significantly reduced.

[0005] Thirdly, the quality problem rectification lacks quantitative guidance, the existing system is usually limited to abnormal alarm and cannot provide specific rectification scheme guidance. The rectification measures depend on manual experience and have great subjectivity, which leads to low rectification compliance rate, high rework cost and cannot form effective closed-loop management.

[0006] In addition, the data tracing and knowledge reuse ability is insufficient, the traditional method lacks a whole-chain quality data management mechanism, the data is easy to lose and difficult to trace, and the model parameters and analysis algorithms cannot be continuously optimized through historical data, which leads to the repeated occurrence of similar problems in different projects.

[0007] Therefore, an intelligent quality management method capable of realizing whole process, multi-dimensional and self-adaptive is needed to fundamentally solve the problems of one-sided monitoring, rigid model, fuzzy decision and difficult tracing in the existing technology. SUMMARY

[0008] The purpose of the present application is to provide a steel structure power house construction quality whole process management and control method based on digital twinning to solve the problems proposed in the background.

[0009] To solve the above technical problems, the present application provides the following technical solutions:

[0010] The method comprises the following steps:

[0011] Obtain construction object information and perform hierarchical division, construct and train a three-layer digital twin model;

[0012] Collect and process construction data;

[0013] Perform hierarchical quality analysis on the processed construction data through the three-layer digital twin model, and when a quality anomaly is detected, start cross-layer association analysis to locate the problem source;

[0014] According to the analysis result of the hierarchical quality analysis, trigger a hierarchical warning, and based on the simulation of the three-layer digital twin model, generate quantitative construction adjustment instructions;

[0015] Construct a construction quality traceability file and upload it to a blockchain, and dynamically and iteratively optimize the parameter threshold and analysis algorithm of the three-layer digital twin model based on historical construction data.

[0016] Further, the construction of the three-layer digital twin model specifically includes:

[0017] Extract historical qualified data in the construction archives of the same type of steel structure power plant and experimental data built in the laboratory and consistent with the welding environment on site, the historical qualified data including effective data corresponding to welding current, voltage, time, and weld flaw detection qualified rate, the experimental data including weld appearance quality, internal defects, and tensile strength quality results, obtain construction object information, constitute an initial training set of welding parameter-quality result data pairs, and perform de-duplication and outlier rejection processing on the initial training set;

[0018] Divide the construction object information into a basic layer, a structure layer, and a construction layer according to the data type; the basic layer includes geological survey data, the structure layer includes design drawing data, and the construction layer includes construction scheme data;

[0019] Based on BIM technology, construct a three-layer digital twin model corresponding to the basic layer, the structure layer, and the construction layer, and preset parameter association rules in the model; the parameter association rules include linkage rules of component displacement data and foundation settlement data, and mapping rules of welding temperature data and material mechanical property data;

[0020] The three-dimensional coordinate data of the entity structure obtained by laser scanning is compared with the model base layer coordinates, and when the error exceeds the preset threshold, the model parameters are adjusted; the simulation results of the model mechanical properties of the key components are checked with the mechanical test data of the entity components, and when the consistency error of the three-dimensional digital twin model and the entity is less than the error threshold, the trained three-dimensional digital twin model is obtained.

[0021] Further, the mapping rule of the welding temperature data and the material mechanical property data is a mapping formula:

[0022]

[0023] wherein,

[0024] σ is the target mechanical property index of the weld metal, including tensile strength, yield strength or impact toughness;

[0025] I is the welding current; U is the welding voltage; V is the welding speed;

[0026] t is the time coefficient related to the plate thickness or heat conduction; C, Mn, Si,... are the specific chemical element contents of the base material and the welding material.

[0027] Further, the collection and processing of construction data specifically includes:

[0028] Various collection devices are deployed at the construction site to collect construction data; the collection devices include displacement sensors installed at the connection of load-bearing components, infrared temperature sensors installed at the welding operation area, and unmanned aerial vehicles equipped with high-definition cameras;

[0029] Outlier rejection and data standardization processing are performed on the construction data; the outlier rejection adopts the 3σ criterion, and the construction data exceeding the preset normal value interval is automatically marked and triggers secondary collection; the data standardization processing includes converting different dimensions of construction data into standardized numerical values in the range of 0-1;

[0030] A weighted fusion algorithm is used to fuse the same type of construction data from different collection devices to generate fused construction data; the same type of construction data at least includes component displacement data and welding quality data;

[0031] The fused construction data is synchronized in real time to the corresponding level of the three-dimensional digital twin model.

[0032] Further, the component displacement data is calculated using the following weighted fusion formula:

[0033]

[0034] wherein,

[0035] is a displacement value of the component after fusion, is a displacement value of the component collected by the displacement sensor, is a displacement value of the component extracted by the unmanned aerial vehicle image recognition, and α is a displacement weight corresponding to the displacement sensor, and β is a displacement weight corresponding to the unmanned aerial vehicle image recognition;

[0036] For the welding quality data, the following weighted fusion formula is used to generate a welding quality comprehensive evaluation index:

[0037]

[0038] wherein,

[0039] is a welding quality comprehensive evaluation index, is a standardized value of the welding temperature data, is a normalized score of the weld appearance image feature data, is a qualified rate of the ultrasonic flaw detection data, ω1 is a weight of the welding temperature data, ω2 is a weight of the weld appearance image feature data, and ω3 is a weight of the ultrasonic flaw detection data.

[0040] Further, the hierarchical quality analysis and cross-layer correlation analysis specifically include:

[0041] The basic layer analysis inputs the real-time collected ground settlement data into a pre-trained time series prediction model to obtain a ground settlement prediction value in a future period; the ground settlement prediction value is compared with a preset allowed settlement threshold in the model basic layer, and if the ground settlement prediction value exceeds the prediction threshold, it is marked as a basic layer quality hidden danger;

[0042] The structure layer analysis compares the real-time collected component installation coordinate data with a preset three-dimensional coordinate threshold in the model structure layer to calculate a deviation value; a multi-parameter coupling analysis algorithm is used to comprehensively analyze multiple parameters of the welding process to determine a welding quality grade;

[0043] The construction layer analysis analyzes the real-time collected construction operation video stream through a behavior recognition algorithm to identify whether the operation behavior is compliant;

[0044] The cross-layer correlation analysis automatically retrieves and correlates the historical and real-time data of other layers when a quality abnormality is detected in any layer to comprehensively locate the root cause of the quality problem;

[0045] The cross-layer correlation analysis specifically includes:

[0046] When the structural layer detects that the component displacement is out of tolerance, the foundation settlement data at the last time node in the region in the basic layer and the installation process record data of the component in the construction layer are automatically associated and called;

[0047] If the settlement amount in the foundation settlement data is out of limit and the change trend is synchronized with the displacement out of tolerance, it is determined that the problem is caused by foundation settlement;

[0048] If the foundation settlement data is normal, but the installation process record shows that the positioning clamp is not used according to the specification, it is determined that the problem is caused by installation operation error.

[0049] Further, the multi-parameter coupling analysis algorithm calculates the welding quality comprehensive score through the following formula:

[0050]

[0051] wherein,

[0052] is the welding quality comprehensive score, is the normalized value of the welding current, is the normalized value of the welding voltage, is the normalized evaluation value of the welding time and temperature curve characteristics, is the normalized value of the weld height, is the normalized value weight of the welding current, is the normalized value weight of the welding voltage, is the normalized evaluation value weight of the time and temperature curve characteristics, 7 is the normalized value weight of the weld height.

[0053] Further, the triggering of the hierarchical early warning and the generation of the quantified construction adjustment instruction specifically include:

[0054] According to the severity of the quality problem obtained by the layered quality analysis, different levels of early warning are triggered; the early warning levels at least include a first-level early warning requiring immediate stoppage for rectification, a second-level important early warning requiring timely treatment, and a third-level general early warning only for record;

[0055] Based on the simulation analysis result of the three-layer digital twin model, an executable adjustment instruction containing specific operation parameters is generated; the adjustment instruction includes a component correction instruction and a welding parameter adjustment instruction;

[0056] The component correction instruction specifically includes:

[0057] In the three-layer digital twin model, the required correction force is calculated by calling the mechanics simulation module with the target that the displacement deviation after correction is not greater than 2mm; the correction force is calculated through the following mechanics formula:

[0058] F = (ΔL × E × S) / L

[0059] wherein,

[0060] F is the required correction force, ΔL is the amount of displacement to be corrected, E is the elastic modulus of the material of the component, S is the cross-sectional area of the component, and L is the length of the component;

[0061] Based on the required correction force F, the specific steps of applying the correction force in stages, the required tools and their installation positions are planned according to the preset planning standard to form executable correction instructions;

[0062] The welding parameter adjustment instruction specifically includes:

[0063] Retrieve the welding temperature-current-voltage correlation model and historical qualified data stored in the three-layer digital twin model;

[0064] The required current value and voltage value are calculated by backstepping through the correlation model; the correlation model is:

[0065] T = a × I + b × U + c

[0066] wherein,

[0067] T is the welding temperature, I is the welding current, U is the welding voltage, and a, b, and c are model coefficients obtained by training historical data;

[0068] The calculated current value and voltage value, together with their corresponding simulated welding strength results, are generated as adjustment instructions according to the adjustment instruction template.

[0069] Further, the construction quality traceability archive is uploaded to the blockchain, and the parameter thresholds and analysis algorithms of the three-layer digital twin model are dynamically iteratively optimized based on historical construction data, specifically including:

[0070] A construction quality traceability archive containing problem details, rectification instructions, and rectification results is constructed, and the construction quality traceability archive is uploaded to the blockchain;

[0071] Each record of the construction quality traceability archive generates a unique hash value and is stored in association with the hash value of the previous record, forming a chain structure; the construction quality traceability archive supports retrieval by problem ID or component number;

[0072] Based on historical construction data, the parameter thresholds and analysis algorithms in the three-layer digital twin model are optimized and updated using machine learning algorithms;

[0073] The optimization process of the parameter thresholds includes:

[0074] Adopt random forest model, with welding parameters in historical construction process as input characteristics, with quality qualified probability as output;Through random forest model training to find the best temperature threshold, and update the lower limit threshold of welding temperature in the random forest model to the best temperature threshold.

[0075] Compared with the prior art, the beneficial effects achieved by the present application are:

[0076] The present application realizes high-fidelity virtual mapping of construction full elements by constructing a foundation-structure-construction three-layer digital twin model and establishing cross-layer parameter correlation rules;Through multi-source heterogeneous data fusion processing and virtual-real synchronization mechanism, high-frequency synchronization and interaction of physical entities and virtual models are realized;Through hierarchical intelligent analysis, including foundation settlement trend analysis based on time series prediction, multi-parameter coupled welding quality evaluation, and cross-level correlation diagnosis such as structure displacement anomaly and foundation settlement, and collaborative analysis of construction operation, the root cause of quality problems is accurately located;Through quantitative decision support based on mechanical simulation and process parameter optimization, executable precise adjustment instructions are generated;Through the non-tamperable traceability system enabled by blockchain and the model self-optimization mechanism driven by data, the traceability of quality management and the continuous evolution of system performance are realized;The scheme effectively solves the technical pain points of traditional methods, such as single monitoring dimension, one-sided problem diagnosis, experience-dependent decision-making, and static rigidity of models, and realizes important changes in construction quality management, such as from passive inspection to active control, from local isolated analysis to global collaborative diagnosis, from qualitative experience judgment to quantitative scientific decision-making, and from fixed threshold monitoring to adaptive optimization, significantly improving the accuracy, reliability and intelligent level of quality control. BRIEF DESCRIPTION OF DRAWINGS

[0077] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, and are used to explain the present application together with embodiments of the present application, and do not constitute a limitation on the present application.

[0078] Figure 1 is a step schematic diagram of the steel structure power room construction quality whole process management method based on digital twin of the present application. DETAILED DESCRIPTION

[0079] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0080] Please refer to Figure 1In the embodiment, a steel structure power machine room construction quality whole process management method based on digital twinning is provided, the steel structure power machine room is a construction object, construction object information of the construction object includes foundation layer information, structure layer information and construction layer information; the method comprises the following steps:

[0081] S1, construction object information is acquired and hierarchical division is performed, a three-layer digital twinning model is constructed and trained;

[0082] S2, construction data is collected and processed;

[0083] S3, the three-layer digital twinning model is used to perform layered quality analysis on the processed construction data, when quality abnormalities are detected, cross-layer association analysis is started to locate the problem source;

[0084] S4, according to the analysis result of the layered quality analysis, a hierarchical early warning is triggered, and a quantitative construction adjustment instruction is generated based on the simulation of the three-layer digital twinning model;

[0085] S5, a construction quality traceability file is constructed and uploaded to a blockchain, and the parameter threshold and analysis algorithm of the digital twinning model are dynamically iteratively optimized based on historical construction data.

[0086] Through the above steps, the embodiment of the application provides a systematic solution to the three core problems existing in traditional steel structure construction quality management: first, the problem of single monitoring dimension is solved, the three-layer model architecture of foundation-structure-construction is used to realize full-element digital mapping from geological environment to component state to construction operation; second, the technical bottleneck of one-sided problem diagnosis is broken, the root cause of quality abnormalities can be traced from the system level with the help of cross-layer association analysis mechanism; third, the limitation of experience-dependent rectification measures is overcome, and a quantitative and executable precise adjustment scheme is generated based on the simulation deduction ability of the digital twinning model.

[0087] The significant effects realized by the embodiment are reflected in three aspects: at the technical level, a high-frequency interaction channel between physical entities and virtual models is established, realizing real-time synchronization and bidirectional optimization of virtual and real spaces; at the management level, a closed-loop management and control system from quality monitoring, analysis and early warning to decision execution is constructed, transforming traditional passive quality inspection into active preventive quality control; at the system level, the quality data are ensured to be non-tamperable and traceable through blockchain technology, and the system is provided with continuous self-optimization ability relying on machine learning algorithm, significantly improving the accuracy and reliability of quality management.

[0088] In the preferred embodiment, the construction of the three-layer digital twinning model in step S1 specifically comprises:

[0089] S1.1, obtaining an initial training set: extracting historical qualified data of the same type of steel structure power plant construction archives and experimental data built in the laboratory and consistent with the welding environment on site, the historical qualified data including welding current, voltage, time and corresponding effective data of welding flaw detection qualified rate, the experimental data including welding appearance quality, internal defects, tensile strength quality results, obtaining construction object information, constituting an initial training set of welding parameter-quality result data pairs, and performing de-duplication and outlier rejection processing on the initial training set;

[0090] S1.2, hierarchical data acquisition: according to the data type, the construction object information is divided into a basic layer, a structure layer and a construction layer; the basic layer includes geological survey data, the structure layer includes design drawing data, and the construction layer includes construction scheme data; the structure layer design drawing data includes component size deviation allowable value, material parameter and connection mode requirement; the construction layer construction scheme data includes each process construction time window and personnel operation specification standard.

[0091] S1.3, hierarchical model construction: based on BIM technology, three-layer digital twin models corresponding to the basic layer, the structure layer and the construction layer are constructed, and parameter correlation rules are preset in the model; the parameter correlation rules include linkage rules of component displacement data and foundation settlement data, and mapping rules of welding temperature data and material mechanical property data; the basic layer is bound with geological survey data, the structure layer is associated with component three-dimensional coordinates and material performance parameters, and the construction layer is embedded with process flow and quality standard threshold.

[0092] S1.4, model precision verification: compare the three-dimensional coordinate data of the entity structure obtained by laser scanning with the model basic layer coordinates, and adjust the model parameters when the error exceeds the preset threshold 0.5mm; the mechanical property simulation results of the key component model are verified with the mechanical test data of the entity component, and the consistency error between the model and the entity is ensured to be ≤1%.

[0093] The mapping rule of welding temperature data and material mechanical property data preset in the step S1.3 is specifically realized by a multivariate nonlinear regression model, which is used to represent the quantitative relationship between welding heat input, cooling rate and final mechanical property of weld metal; the formula of the model is:

[0094]

[0095] wherein,

[0096] σ is the target mechanical property index of the weld metal, including tensile strength, yield strength or impact toughness;

[0097] I is the welding current (A); U is the welding voltage (V); V is the welding speed (cm / min);

[0098] t is the plate thickness or the time coefficient related to heat conduction (s); C, Mn, Si,... are the specific chemical element contents of the base material and welding material (%);

[0099] The specific function form f of the model and each coefficient are obtained by machine learning training on more than 12000 groups of welding parameter-quality result data pairs in the initial training set, and the model is embedded in the structure layer of the three-layer digital twin model, and is used for real-time prediction of the mechanical properties of the weld under different welding parameters.

[0100] Through the detailed implementation of the above step S1, the embodiment focuses on solving three key problems in the construction of the digital twin model: first, the problem of incomplete data dimension and lack of correlation between process parameters and quality results in the traditional modeling process is solved, and a high-quality training set is constructed by combining historical data and experimental data; second, the limitation of a single model that cannot reflect the multi-dimensional characteristics of construction is overcome, and full-factor digital mapping is achieved through a three-layer architecture of foundation-structure-construction; third, the technical bottleneck of static model parameter solidification and disconnection with the actual construction state is broken through, and the model has dynamic response capability through the establishment of parameter correlation rules.

[0101] The significant effects achieved by the embodiment are as follows: first, the multi-element nonlinear regression model trained by more than 12000 groups of welding parameter-quality result data realizes the precise quantitative mapping of welding process parameters and weld mechanical properties, so that the model has the ability to predict the final quality from the process parameters; second, the consistency error between the model and the entity is controlled within 1% by using the dual verification mechanism of laser scanning and mechanical testing, which ensures the high fidelity of the digital twin model; finally, through the multi-dimensional correlation rules such as pre-set component displacement and foundation settlement, welding temperature and material performance, a cross-layer data linkage mechanism is established, which lays a solid foundation for subsequent real-time monitoring and accurate analysis. These innovations make the digital twin model not a simple three-dimensional visualization model, but an intelligent analysis platform that integrates physical laws and process knowledge.

[0102] Preferably, the step S2 specifically comprises:

[0103] S2.1, Multi-source device deployment: Deploy multiple acquisition devices to collect construction data on the construction site; the acquisition devices include displacement sensors installed at the connection of load-bearing members, infrared temperature sensors installed at the welding operation area, and unmanned aerial vehicles equipped with high-definition cameras; the measurement accuracy of the displacement sensor is ±0.01mm; the temperature measurement range of the infrared temperature sensor is 0-1500℃; the unmanned aerial vehicle is equipped with a high-definition camera with a resolution of 4K, and inspects the overall structure at a frequency of every 30 minutes, collecting member installation flatness data.

[0104] S2.2, Data preprocessing: Perform outlier rejection and data standardization processing on the received acquisition construction data; the outlier rejection uses the 3σ criterion, and automatically marks and triggers secondary collection for construction data that exceeds the preset normal value interval; the data standardization processing converts construction data of different dimensions into standardized values within the range of 0-1; for welding temperature data, the preset normal value interval is 800-1300℃.

[0105] S2.3, Multi-source data fusion: Use a weighted fusion algorithm to fuse the same type of data from different acquisition devices to generate fused data; the weighted fusion algorithm uses different weight configuration strategies for different types of monitoring data; the same type of data at least includes member displacement data and welding quality data;

[0106] S2.4, Virtual-real synchronization: Real-time synchronization of fused data to the corresponding level of the three-layer digital twin model, and ensure that the time difference between the three-layer digital twin model and the physical construction state does not exceed the set threshold.

[0107] In step S2.3, for member displacement data, the following weighted fusion formula is used for calculation:

[0108]

[0109] Wherein,

[0110] is the fused displacement value, is the displacement value collected by the displacement sensor, is the displacement value extracted by the unmanned aerial vehicle image recognition, α is the weight of the displacement sensor data, and the value is 0.7, β is the weight of the unmanned aerial vehicle image recognition displacement data, and the value is 0.3;

[0111] In step S2.3, for welding quality data, the following weighted fusion formula is used to generate a welding quality comprehensive evaluation index:

[0112]

[0113] Wherein,

[0114] is a comprehensive evaluation index of welding quality, is a standardized value of infrared temperature sensor data, is a normalized value of the weld appearance image feature score, is an ultrasonic flaw detection pass rate, ω1 is the weight of temperature data, ω2 is the weight of weld appearance image feature data, and is 0.3, and ω3 is the weight of ultrasonic flaw detection data, and is 0.2;

[0115] The synchronization in step S2 specifically comprises: establishing a real-time transmission and dynamic updating dual-channel synchronization mechanism, and when it is monitored that data transmission delay exceeds a set threshold, automatically switching to a backup communication link to ensure that the time difference between the digital twin model and the entity construction state is not more than 5 minutes.

[0116] Before synchronization, a weighted fusion algorithm is used to fuse the same type of data obtained by multiple source acquisition devices to obtain higher precision fused data.

[0117] Preferably, the step S3 specifically comprises:

[0118] S3.1, base layer analysis: inputting the real-time collected foundation settlement data into a pre-trained time series prediction model to obtain a foundation settlement prediction value in a future period; comparing the foundation settlement prediction value with a preset allowable settlement threshold in the model base layer, and if the foundation settlement prediction value exceeds the threshold, marking as a base layer quality hidden danger;

[0119] The time series prediction model is an ARIMA model; the input of the ARIMA model is a foundation settlement daily data sequence in the past 30 days, and the output is a daily settlement prediction value in the future 7 days; and the allowable settlement threshold is that the daily average settlement amount is not greater than 0.167 mm.

[0120] S3.2, structure layer analysis: comparing the real-time collected component installation coordinate data with a preset three-dimensional coordinate threshold in the model structure layer to calculate a deviation value; and using a multi-parameter coupling analysis algorithm to comprehensively analyze multiple parameters in the welding process to determine a welding quality grade;

[0121] S3.3, construction layer analysis: analyzing the real-time collected construction operation video stream by a behavior recognition algorithm to identify whether the operation behavior is compliant; the behavior recognition algorithm uses a YOLOv8 target detection model combined with a motion classification model; the algorithm first identifies the construction personnel, tools and components in the video stream, then analyzes the motion trajectory and motion features thereof, and compares them with a preset standard motion feature library to determine whether the operation is compliant.

[0122] S3.4, Cross-layer correlation analysis: When quality abnormalities are detected at any level, automatically retrieve and correlate historical and real-time data of other levels to comprehensively locate the root cause of quality problems;

[0123] The cross-layer correlation analysis specifically includes:

[0124] When the component displacement exceeds the limit at the structure level, automatically correlate and retrieve the foundation settlement change data of the area in the foundation level in the past 24 hours, and the installation process record data of the component at the construction level;

[0125] If the foundation settlement data indicates that the settlement exceeds the limit and the change trend is synchronized with the displacement exceeding the limit, it is determined that the root cause of the problem is the foundation settlement;

[0126] If the foundation settlement data is normal, but the installation process record shows that the positioning fixture is not used according to the specification, it is determined that the root cause of the problem is the installation operation error.

[0127] In the step S3.2, the multi-parameter coupling analysis algorithm calculates the welding quality comprehensive score by the following formula:

[0128]

[0129] wherein,

[0130] is the welding quality comprehensive score, is the normalized value of the welding current, is the normalized value of the welding voltage, is the normalized evaluation value of the welding time and temperature curve characteristics, is the normalized value of the weld height, is the normalized value weight of the welding current, is the normalized value weight of the welding voltage, is the normalized evaluation value weight of the time and temperature curve characteristics, 7 is the normalized value weight of the weld height;

[0131] If ≥ 0.8, it is determined that the welding quality is qualified, otherwise it is determined to be unqualified, and according to the deviation of each normalized parameter value, the specific problem type is marked, the problem type includes incomplete fusion and insufficient weld size; the determination condition of incomplete fusion is that the duration of weld temperature curve below 800℃ exceeds 5 seconds.

[0132] Through the implementation of the above steps S2 and S3, the embodiment systematically solves the key technical problems in construction quality monitoring: first, it breaks through the limitations of traditional monitoring methods with single data source and insufficient precision, and realizes the complementation and enhancement of monitoring data in the time and space dimensions through multi-source heterogeneous sensor collaborative collection and weighted fusion processing; second, it overcomes the problem of isolated quality analysis dimension and difficulty in tracing the source, and realizes systematic diagnosis from surface phenomenon to root cause through the establishment of a three-layer linkage analysis mechanism of foundation-structure-construction.

[0133] The innovative effects realized by the embodiment mainly reflect in three aspects: at the data collection level, through the multi-source collaboration of displacement sensors, infrared temperature sensors and unmanned aerial vehicle inspection, combined with the weighted fusion algorithm based on specific weight distribution, the monitoring data precision is improved; at the analysis and early warning level, the ARIMA time series model is used to realize the predictive analysis of ground settlement, the multi-parameter coupling algorithm is used to comprehensively evaluate the welding quality, and the YOLOv8 behavior recognition technology is used to realize the all-round monitoring of the construction process, so that the identification accuracy of quality hidden dangers reaches more than 95%; at the problem diagnosis level, a cross-layer correlation analysis mechanism is innovatively established, when the structure displacement exceeds the tolerance, the ground settlement trend and the construction operation record can be automatically associated and analyzed, and the quality problem root cause such as ground settlement or installation failure can be accurately distinguished, which fundamentally solves the pain point of traditional methods "only reporting exceptions, not finding the root cause", and these technical innovations together constitute a construction quality control system integrating accurate monitoring, intelligent diagnosis and prediction and early warning.

[0134] Preferably, the step S4 specifically comprises:

[0135] S4.1, hierarchical early warning: according to the severity of the quality problem obtained by the hierarchical quality analysis in step S3, triggering different levels of early warning; the early warning levels at least include a first-level early warning requiring immediate stop and rectification, a second-level important early warning requiring timely treatment, and a general third-level early warning only for record;

[0136] The triggering condition of the first-level early warning is that the component displacement exceeds 10mm or the welding temperature is continuously lower than 700℃; after triggering, the system sends out sound and light alarm within ≤10 seconds and pushes the notification containing the accurate position of the problem and the severity to the management terminal of the mobile terminal;

[0137] The triggering condition of the second-level early warning is that the component displacement exceeds 5-10mm or the weld height deviation is 2-3mm; after triggering, the system pushes the early warning information containing the problem details and the preliminary analysis conclusion to the field management terminal within ≤30 seconds;

[0138] The triggering condition of the third-level early warning is that the component displacement exceeds 2-5mm; after triggering, the system only records in the background and generates a daily early warning summary report.

[0139] S4.2, Quantitative adjustment instruction generation: based on the simulation analysis result of the digital twin model, generate an executable adjustment instruction containing specific operation parameters; the adjustment instruction includes component correction instruction and welding parameter adjustment instruction;

[0140] The component correction instruction specifically includes:

[0141] In the digital twin model, the target is to correct the displacement deviation not more than 2mm, and the required correction force is calculated by calling the mechanical simulation module; the correction force is calculated by the following mechanical formula:

[0142] F = (ΔL × E × S) / L

[0143] Wherein,

[0144] F is the required correction force (N), ΔL is the displacement amount that needs to be corrected (m), E is the elastic modulus of the component material (Pa),

[0145] S is the cross-sectional area of the component (m²), and L is the length of the component (m);

[0146] Based on the calculated correction force F, the specific steps of applying the correction force in stages, the required tools and their installation positions are planned to form an executable correction instruction;

[0147] The correction force applied in stages is specifically: the calculated correction force F is divided into 3 times, and the force of each time is F / 3; the instruction specifies the use of two hydraulic jacks with a rated load capacity of 50kN in parallel to implement the correction.

[0148] The welding parameter adjustment instruction specifically includes:

[0149] Retrieve the welding temperature-current-voltage correlation model and historical qualified data stored in the three-layer digital twin model;

[0150] The target is to adjust the welding temperature to the optimal interval, and the required current value (I) and voltage value (U) are calculated by the correlation model; the correlation model is:

[0151] T = a × I + b × U + c

[0152] Wherein,

[0153] T is the welding temperature (℃), I is the welding current (A), U is the welding voltage (V), a, b, c are the model coefficients obtained by training historical data;

[0154] The calculated current value and voltage value, together with the corresponding simulation welding strength result, are generated as adjustment instructions.

[0155] Preferably, the step S5 specifically comprises:

[0156] S5.1, construction of a full-chain traceability archive: constructing a construction quality traceability archive containing problem details, rectification instructions and rectification results, and storing the construction quality traceability archive in a blockchain database to ensure that the data cannot be tampered with;

[0157] Each record of the construction quality traceability archive generates a unique hash value and is stored in association with the hash value of the previous record, forming a chain structure; the construction quality traceability archive supports searching by problem ID or component number, and the search response time is not more than 2 seconds;

[0158] The historical qualified data refers to a set of effective data corresponding to the final quality inspection results, which are selected from past projects of the same type of steel structure power house, such as projects that have been accepted and qualified in the past 5 years, including current, voltage and speed, and the purpose is to let the digital twin model learn the reliable process experience and achievement standard verified by practice in the industry;

[0159] The experimental data refers to the accurate quality result data obtained after systematically adjusting the welding parameters and conducting repeated tests under each set of parameters in a laboratory environment using the same materials as in the field, such as Q355 steel test pieces and equipment, and the purpose is to supplement the missing parameter combinations and results in the historical data in a controllable environment, especially focusing on fault boundary conditions, thereby giving the model a more scientific and comprehensive causal mapping relationship;

[0160] During the construction process of the project, real-time data such as displacement, temperature, image and operation behavior are continuously collected by multi-source collection equipment. After preprocessing and fusion, these data are synchronized to the digital twin model and continuously accumulated over time, forming unique historical construction data for the project. This historical construction data records all states, operations, warnings and rectification results from the start of the project to the current time, and is a digital archive for the entire life cycle of the project;

[0161] S5.2, dynamic iterative optimization of the model: based on the historical construction data, the parameter thresholds and analysis algorithms in the three-dimensional digital twin model are optimized and updated using machine learning algorithms;

[0162] The parameter threshold optimization specifically includes:

[0163] A random forest model is used, with welding parameters in the historical construction process as input features and quality qualification probability as output. Through random forest model training, the optimal temperature threshold is found that makes the qualification probability not less than 99%, and the lower threshold value of welding temperature in the random forest model is updated to the optimal temperature threshold;

[0164] The analysis algorithm is optimized, in particular:

[0165] For an analysis scene with an identification accuracy lower than 90%, the multi-parameter coupling analysis algorithm is retrained by increasing the historical data samples of the scene; the cross-validation method is used to adjust the parameter weight and decision threshold in the algorithm, so that the identification accuracy of the test set is improved to more than 95%.

[0166] Through the implementation of the above steps S4 and S5, the embodiment innovatively solves two key problems in the quality control closed loop: first, it breaks through the limitations of the traditional early warning system, such as response lag, fuzzy classification, and lack of precise execution guidance, and establishes a quantitative decision mechanism based on digital twin simulation; second, it overcomes the defects of traditional quality management, such as poor data traceability, fixed model parameters, and lack of self-optimization ability, and builds an intelligent control system that continuously evolves.

[0167] The significant effects achieved by the embodiment are as follows: in terms of early warning response, the establishment of a three-level early warning mechanism enables accurate control of quality problems in a hierarchical manner, and the response efficiency is significantly improved; in terms of rectification execution, the accurate calculation of the correction force based on mechanical formulas and the three-time application operation strategy, combined with the back-propagation optimization of the welding parameter correlation model, realize a major change from experience-based judgment to quantitative execution, and the rectification first-time pass rate is greatly improved; in terms of quality traceability, the use of blockchain hash chain storage ensures data integrity, supports fast retrieval within 2 seconds, and establishes a full-chain trusted traceability system; in terms of system optimization, the use of a random forest model to optimize the parameter threshold with a target of 99% pass rate, and the use of cross-validation to improve the identification accuracy from 90% to more than 95%, enable the system to have continuous self-evolution capability. These innovations together form a complete quality control closed loop that integrates intelligent early warning, precise execution, trusted traceability, and continuous optimization.

[0168] It should be noted that, in this document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0169] Finally, it should be noted that the above only describes the preferred embodiments of the present application and is not intended to limit the present application, and although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent replacements to some technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for whole-process management and control of construction quality of a steel structure power house based on digital twinning, characterized in that, The method comprises the following steps: acquiring construction object information and performing hierarchical division, constructing and training a three-layer digital twin model; collecting and processing construction data; performing hierarchical quality analysis on the processed construction data through the three-layer digital twin model, and when a quality anomaly is detected, starting cross-layer association analysis to locate the problem source; triggering hierarchical early warning according to the analysis result of the hierarchical quality analysis, and generating quantitative construction adjustment instructions based on the simulation of the three-layer digital twin model; constructing a construction quality traceability file and uploading it to a blockchain, and dynamically and iteratively optimizing the parameter threshold and analysis algorithm of the three-layer digital twin model based on historical construction data; The construction of the three-layer digital twin model specifically comprises: extracting historical qualified data in the construction archives of the same type of steel structure power house and experimental data obtained in a welding environment consistent with the field in the laboratory, the historical qualified data including effective data corresponding to welding current, voltage, time, and weld flaw detection qualified rate, and the experimental data including quality results of weld appearance quality, internal defects, and tensile strength, obtaining construction object information to form an initial training set of welding parameter-quality result data pairs, and performing de-duplication and outlier rejection processing on the initial training set; dividing the construction object information into a basic layer, a structure layer, and a construction layer according to the data type; the basic layer includes geological survey data, the structure layer includes design drawing data, and the construction layer includes construction scheme data; constructing a three-layer digital twin model corresponding to the basic layer, the structure layer, and the construction layer based on BIM technology, and presetting parameter association rules in the model; the parameter association rules include linkage rules of component displacement data and foundation settlement data, and mapping rules of welding temperature data and material mechanical property data; comparing the three-dimensional coordinate data of the entity structure obtained by laser scanning with the model basic layer coordinates, adjusting the model parameters when the error exceeds the preset threshold, and verifying the model mechanical property simulation results of key components with the mechanical test data of the entity components, when the consistency error between the three-layer digital twin model and the entity is less than the error threshold, obtaining the trained three-layer digital twin model.

2. The digital-twin-based construction quality whole-process management and control method for a steel structure power machine room according to claim 1, characterized in that, The mapping formula of the welding temperature data and the material mechanical property data is: ； where σ is the target mechanical property index of the weld metal, including tensile strength, yield strength, or impact toughness; I is the welding current; U is the welding voltage; V is the welding speed; t is the plate thickness or the time coefficient related to heat conduction; C, Mn, Si,... are the specific chemical element contents of the base material and the welding material.

3. The digital-twin-based construction quality whole-process management and control method for a steel structure power machine room according to claim 1, characterized in that, The collection and processing of construction data specifically comprises: deploying various collection devices to collect construction data on the construction site; the collection devices include displacement sensors installed at the connection of load-bearing components, infrared temperature sensors installed at the welding operation area, and unmanned aerial vehicles equipped with high-definition cameras; The construction data is subjected to outlier elimination and data standardization processing; the outlier elimination adopts a 3σ criterion, and construction data exceeding a preset normal value interval is automatically marked and triggers secondary collection; the data standardization processing includes converting construction data of different dimensions into standardized values in the range of 0-1; A weighted fusion algorithm is used to fuse the same type of construction data from different collection devices to generate fused construction data; the same type of construction data at least includes component displacement data and welding quality data; The fused construction data is real-time synchronized to the corresponding level of the three-layer digital twin model.

4. The steel structure power machine room construction quality whole-process management and control method based on digital twinning according to claim 3, characterized in that, The component displacement data is calculated using the following weighted fusion formula: ； wherein, is the displacement value of the component after fusion, is the displacement value of the component collected by the displacement sensor, is the displacement value of the component extracted by the unmanned aerial vehicle image recognition, and α is the displacement weight corresponding to the displacement sensor, and β is the displacement weight corresponding to the unmanned aerial vehicle image recognition. For welding quality data, the following weighted fusion formula is used to generate a welding quality comprehensive evaluation index: ； wherein, is a comprehensive evaluation index of welding quality, is a standardized value of welding temperature data, is a standardized score of weld appearance image feature data, is a qualified rate of ultrasonic flaw detection data, ω1 is the weight of welding temperature data, ω2 is the weight of weld appearance image feature data, and ω3 is the weight of ultrasonic flaw detection data.

5. The digital-twin-based construction quality whole-process management and control method for a steel structure power machine room according to claim 1, characterized in that, The hierarchical quality analysis and cross-layer correlation analysis specifically include: The foundation layer analysis inputs the real-time collected foundation settlement data into a pre-trained time series prediction model to obtain the foundation settlement prediction value in the future period; the foundation settlement prediction value is compared with the preset allowable settlement threshold in the model foundation layer, and if the foundation settlement prediction value exceeds the prediction threshold, it is marked as a foundation layer quality hidden danger; The structure layer analysis compares the real-time collected component installation coordinate data with the preset three-dimensional coordinate threshold in the model structure layer to calculate the deviation value; a multi-parameter coupling analysis algorithm is used to comprehensively analyze multiple parameters of the welding process to determine the welding quality grade; The construction layer analysis analyzes the real-time collected construction operation video stream through a behavior recognition algorithm to identify whether the operation behavior is compliant; The cross-layer correlation analysis automatically retrieves and analyzes the historical and real-time data of other levels when a quality anomaly is detected in any level to comprehensively locate the root cause of the quality problem; The cross-layer correlation analysis specifically includes: When the component displacement exceeds the tolerance in the structure layer, the foundation settlement data in the previous time node in the foundation layer and the installation procedure record data of the component in the construction layer are automatically associated and retrieved; If the settlement amount in the foundation settlement data exceeds the limit and the change trend is synchronized with the displacement overage, the root cause of the problem is determined to be foundation settlement; If the foundation settlement data is normal, but the installation procedure record shows that the positioning clamp is not used according to the specification, the root cause of the problem is determined to be installation operation error.

6. The digital-twin-based construction quality whole-process management and control method for a steel structure power machine room according to claim 5, characterized in that, The multi-parameter coupling analysis algorithm calculates the welding quality comprehensive score through the following formula: ； wherein, is a welding quality composite score, is a normalized value of welding current, is a normalized value of welding voltage, is a normalized evaluation value of a time and temperature curve characteristic, is a normalized value of weld height, is a normalized value weight of welding current, is a normalized value weight of welding voltage, is a normalized evaluation value weight of a time and temperature curve characteristic, 7is a normalized value weight of weld height.

7. The digital-twin-based construction quality whole-process management and control method for a steel structure power machine room according to claim 1, characterized in that, The trigger hierarchical early warning and generation of quantitative construction adjustment instructions specifically include: According to the severity of the quality problem obtained from the hierarchical quality analysis, different levels of early warning are triggered; the early warning levels at least include a first-level early warning requiring immediate stop for rectification, a second-level important early warning requiring timely treatment, and a general third-level early warning only for record; Based on the simulation analysis results of the three-layer digital twin model, executable adjustment instructions containing specific operation parameters are generated; the adjustment instructions include component correction instructions and welding parameter adjustment instructions; The component correction instructions specifically include: In the three-layer digital twin model, the required correction force is calculated by calling the mechanical simulation module with the target of correcting the displacement deviation of no more than 2mm; the correction force is calculated by the following mechanical formula: F = (ΔL × E × S) / L Wherein, F is the required correction force, ΔL is the displacement amount to be corrected, E is the elastic modulus of the component material, S is the cross-sectional area of the component, and L is the length of the component; Based on the required correction force F, the specific steps of applying the correction force in stages, the required tools and their installation positions are planned according to the preset planning standard to form executable correction instructions; The welding parameter adjustment instruction specifically includes: Retrieve the welding temperature-current-voltage correlation model and historical qualified data stored in the three-layer digital twin model; The required current value and voltage value are calculated by backstepping through the correlation model; the correlation model is: T = a × I + b × U + c Wherein, T is the welding temperature, I is the welding current, U is the welding voltage, and a, b, c are model coefficients obtained by training historical data; The calculated current value and voltage value, together with their corresponding simulated welding strength results, are generated as adjustment instructions according to the adjustment instruction template.

8. The digital-twin-based construction quality whole-process management and control method for a steel structure power machine room according to claim 1, characterized in that, The construction quality traceability file is uploaded to the blockchain, and the parameter threshold and analysis algorithm of the three-layer digital twin model are dynamically iteratively optimized based on historical construction data, specifically including: Construct a construction quality traceability file containing problem details, rectification instructions and rectification results, and upload the construction quality traceability file to the blockchain; Each record of the construction quality traceability file generates a unique hash value and is stored in association with the hash value of the previous record, forming a chain structure; the construction quality traceability file supports retrieval by problem ID or component number; Based on historical construction data, the parameter threshold and analysis algorithm in the three-layer digital twin model are optimized and updated using machine learning algorithms; The optimization process of the parameter threshold includes: Using a random forest model, the welding parameters in the historical construction process are used as input features, and the quality qualification probability is used as output; the best temperature threshold is found by training the random forest model, and the lower threshold of the welding temperature in the random forest model is updated to the best temperature threshold.

Images