Urban construction and traffic engineering investigation and design enterprise product quality management system and method

Abstract

The invention provides an urban construction and traffic engineering investigation and design enterprise product quality management system and method. The system comprises a cross-project type quality data fusion module, a dynamic defect evolution prediction module, an intelligent layered spot check engine, a three-dimensional quality situation deduction module, a conference knowledge graph construction module, a result influence evaluation module and a self-adaptive file management system. All the modules are connected through a unified data bus to form a closed-loop management process of quality data acquisition, analysis, decision and feedback. According to the system, through the functions of multi-modal data fusion, defect prediction, intelligent spot check, situation deduction, knowledge precipitation, achievement evaluation, file management and the like, the whole-process intelligent management and control of the project quality are realized, and the management efficiency and the decision scientificity are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of urban construction and traffic engineering survey and design, and more particularly to a product quality management system and method for urban construction and traffic engineering survey and design enterprises. Background Art

[0002] In the field of urban construction and transportation engineering survey and design, quality management faces numerous challenges as projects continue to expand in scale and complexity. Traditional quality management methods, relying primarily on manual inspection and empirical judgment, struggle to cope with complex projects involving multiple engineering types and interdisciplinary expertise. While existing technologies have introduced information technology, data fusion capabilities are limited, preventing them from effectively integrating survey and design data from diverse fields such as construction, municipal engineering, and transportation engineering, leading to a serious phenomenon of information silos. Regarding defect detection and prediction, traditional methods struggle to accurately simulate the spatiotemporal evolution of defects, preventing early warning of potential risks. Furthermore, quality verification plans are often fixed and lack dynamic adjustment mechanisms, making them difficult to adapt to the dynamic changes of engineering projects. Regarding knowledge management, technical meetings lack decision-making relevance and knowledge accumulation, hindering effective knowledge transfer and decision-making support. Furthermore, existing technologies lack quantitative methods for assessing the impact of engineering survey and design results, making it difficult to fully measure their value contribution. Regarding document management, traditional methods are unable to dynamically restructure document structures based on project needs, resulting in inefficient document management.

[0003] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the existing technology: the existing technology cannot effectively integrate quality data of multiple engineering types, lacks the ability to accurately predict the temporal and spatial evolution of quality defects, the quality verification plan lacks a dynamic adjustment mechanism, the knowledge management and decision-making support capabilities are insufficient, the results evaluation lacks quantitative means, and the document management efficiency is low, which makes it difficult to meet the high-precision, high-efficiency and intelligent quality management requirements of urban construction and transportation engineering survey and design companies. Summary of the Invention

[0004] The present invention provides a product quality management system and method for urban construction and traffic engineering survey and design enterprises.

[0005] In a first aspect of the present invention, a product quality management system for an urban construction and transportation engineering survey and design enterprise is provided, comprising:

[0006] A cross-project quality data fusion module, configured to integrate survey and design data for construction, municipal, transportation, landscape, water supply and drainage, HVAC, electrical, and geotechnical engineering projects through a multimodal feature alignment algorithm;

[0007] A dynamic defect evolution prediction module is configured to build a quality defect propagation model based on a spatiotemporal correlation analysis algorithm;

[0008] An intelligent layered spot check engine configured to generate multi-level quality verification plans through a hybrid optimization model;

[0009] A three-dimensional quality situation deduction module is configured to use a parameterized driving engine to build an interactive spatiotemporal topological map of quality data;

[0010] The conference knowledge graph construction module is configured to generate a decision-making association network for technical conferences based on an event chain mining algorithm;

[0011] The impact assessment module is configured to quantify the value contribution of scientific and technological achievements through the technology diffusion model;

[0012] An adaptive document management system configured to achieve dynamic reorganization of quality management documents based on a semantic trajectory tracking algorithm;

[0013] The system connects each module through a unified data bus to form a closed loop of quality data collection-analysis-decision-feedback.

[0014] Furthermore, the dynamic defect evolution prediction module includes:

[0015] A multi-stage defect assessment unit extracts the spatiotemporal features of design defects based on a convolutional long short-term memory network. These features include the defect location coordinate set L(x, y, z), the defect area growth rate ΔA / Δt, and the associated component stress value σ.

[0016] The propagation path modeling unit uses an improved cellular automaton algorithm to simulate the defect diffusion process and defines the evolution rules as follows:

[0017] P d =λ×(1+e -(α×S+β×T) ) -1

[0018] Among them, P d is the probability of defect diffusion, λ is the material attenuation coefficient, S is the stress concentration, T is the duration of environmental action, α and β are empirical correction factors;

[0019] Risk warning unit, when the condition ∑(P d ×V d ) / C m >γ triggers a graded alarm, where V d is the defect impact volume, C m is the component safety threshold, and γ is the risk tolerance coefficient.

[0020] Furthermore, the intelligent layered spot check engine includes:

[0021] Verification strategy generation unit, based on the engineering complexity index E c=ω1×N c +ω2×D g +ω3×T s Construct a project classification model, where B c is the number of cross-disciplinary majors, D g is the geological risk level, T s is the technical novelty, ω i is the weight coefficient;

[0022] The multi-objective optimization unit uses a hybrid strategy of taboo search and genetic algorithm to solve the optimal spot check combination. The objective function is:

[0023] min∑(1-R k )×I k +max(T j )-min(T j )

[0024] Constraint: ∑C m ≤B,R k ≥θ

[0025] where R k is the risk coefficient of project k, I k is the impact factor, T j is the workload of inspector j, C m is the inspection cost, B is the total budget, and θ is the minimum coverage threshold;

[0026] Dynamic adjustment unit, updates the verification priority based on real-time quality data flow and defines the adjustment factor Where ΔQ is the change in quality score, σ is the historical volatility, and Δt is the monitoring period.

[0027] Furthermore, the three-dimensional quality situation deduction module includes:

[0028] The topological relationship reconstruction unit generates a three-dimensional grid structure with mechanical properties based on the BIM model and point cloud data, and defines the grid node attribute set including material strength σ m , deformation δ, defect density ρ d ;

[0029] The multi-parameter coupling unit maps the quality score Q, rectification efficiency η, and risk level R into a color gradient value C = α × Q + β × log (R) + γ × η through a field theory model, where α, β, and γ are visualization weight coefficients;

[0030] The interactive deduction unit responds to user operation instructions to dynamically simulate the quality evolution process, including:

[0031] Spatial section analysis based on virtual reality equipment

[0032] Chain traceability of associated component quality data

[0033] Parallel comparison of the rectification effects of multiple plans.

[0034] Furthermore, the conference affairs knowledge graph construction module includes:

[0035] The meeting decision chain extraction unit uses a multi-head attention mechanism to identify technical decision points and their relationships from meeting minutes, and defines the relationship strength as:

[0036]

[0037] where Q i is the query vector of the i-th decision point, K j is the key vector of the j-th associated point, and d is the feature dimension;

[0038] The knowledge accumulation unit builds a four-dimensional graph containing topics, conclusions, responsible parties, and execution status based on a dynamic graph convolutional network;

[0039] The intelligent reminder unit automatically triggers a supervision notification when it detects that the decision execution progress deviation ΔD>ε, where ΔD=|actual progress D a -Plan Progress D p |, ε is the tolerance threshold.

[0040] Furthermore, the achievement impact assessment module includes:

[0041] Technology diffusion calculation unit, based on the improved PageRank algorithm to quantify the impact of the results:

[0042] PR(u)=(1-d)+d×∑(PR(v)×W vu / (v))

[0043] Where d is the damping coefficient, W vu is the technical association weight of achievement v to u, C(v) is the total number of outbound links of achievement v;

[0044] Value contribution analysis unit, building a c , technology improvement degree T i 、Industry recognition I a The evaluation matrix:

[0045] V=ω1×E c +ω2×arctan(T i )+ω3×log(I a +1)

[0046] where ω i Adjust weights for sectors;

[0047] The association recommendation unit generates an optimization plan for the combination of results based on the spectral clustering algorithm.

[0048] Furthermore, the adaptive file management system includes:

[0049] The semantic trajectory analysis unit uses graph embedding technology to construct the file usage behavior feature space and defines the relevance A of user u to file f. uf =∑δ(t)×sim(q u ,c f ), where δ(t) is the time decay function, sim is the semantic similarity, and q u Query features for users, c f is the file content feature;

[0050] Dynamic recombination unit, when A is detected uf Automatically increase the index priority of file f when it exceeds the threshold θ for n consecutive times;

[0051] The intelligent snapshot unit generates a file version evolution graph based on the differential hash algorithm and visualizes the distribution of revision points.

[0052] Furthermore, it also includes:

[0053] The quality data lineage tracking module is configured as follows:

[0054] Use blockchain technology to record the entire process of quality data generation, modification, and circulation

[0055] Build a verifiable chain containing data source HASH value, operator ID, and timestamp to achieve privacy protection for data auditing through zero-knowledge proof

[0056] The data lineage integrity verification conditions are:

[0057] Verify(Sig i ,H(H i |H i-1 ))=True

[0058] Represents the signature Sig of the i-th block i With the current hash H i , pre-order hash H i-1 The connection verification passed.

[0059] Furthermore, the cross-project quality data fusion module includes:

[0060] The heterogeneous data alignment unit uses a dual attention mechanism to address the temporal and spatial basis differences of multi-source data and defines the spatial alignment weight as:

[0061]

[0062] The time alignment weight is:

[0063] W t =exp(-|t i -t j |σ t )

[0064] where σ s , σ t is a learnable parameter;

[0065] Feature decoupling unit, which separates the domain features and common features of engineering data through adversarial generative networks;

[0066] Incremental learning unit, when a new engineering type is added, the feature extractor is updated based on the knowledge distillation algorithm without compromising the performance of the existing model.

[0067] In a second aspect of the present invention, a product quality management method for an urban construction and transportation engineering survey and design enterprise is provided, comprising:

[0068] Establish a multimodal quality data fusion channel across engineering types and achieve spatiotemporal benchmark alignment through a dual attention mechanism;

[0069] Construct a defect spatiotemporal evolution prediction model and use a cellular automation and deep learning fusion algorithm to simulate the defect propagation path;

[0070] Generate a multi-objective optimized intelligent stratified inspection plan, and dynamically adjust the inspection strategy by combining taboo search and genetic algorithm;

[0071] Create a parametric three-dimensional mass deduction system to achieve multi-field coupling visualization of mass data and mechanical properties;

[0072] Implement a technical conference knowledge accumulation mechanism based on decision chain mining to build a four-dimensional correlation map;

[0073] Develop a technology diffusion-driven achievement value assessment system to quantify the multi-dimensional contributions of scientific and technological achievements;

[0074] Deploy a semantic trajectory-aware file management system to dynamically reconstruct the file organization structure;

[0075] The method achieves full-process trusted traceability of quality data through blockchain evidence storage and zero-knowledge proof.

[0076] The above-described embodiments of the present invention have at least the following beneficial effects: First, the present invention integrates survey and design data from multiple fields, such as construction, municipal engineering, and transportation engineering, through a cross-project quality data fusion module, breaking down information silos and enabling efficient data sharing and collaborative analysis. The dynamic defect evolution prediction module constructs a quality defect propagation model based on a spatiotemporal correlation analysis algorithm, accurately simulating the spatiotemporal evolution of defects, providing early warning of potential risks, and providing a scientific basis for quality control. The intelligent layered spot check engine generates multi-level quality verification plans and dynamically adjusts verification strategies using tabu search and genetic algorithms, improving verification efficiency and accuracy while reducing labor costs. The three-dimensional quality situation deduction module utilizes a parameterized drive engine to construct an interactive spatiotemporal topology map of quality data, enabling visualization of the multi-field coupling of quality data and mechanical properties, intuitively displaying the quality situation evolution process, and providing visual support for decision-making. The conference knowledge graph construction module generates a decision-making correlation network for technical conferences based on an event chain mining algorithm, accumulating conference knowledge and forming an effective knowledge transfer and decision-making support system. The achievement impact assessment module quantifies the value contribution of scientific and technological achievements, providing a scientific evaluation basis for their promotion and application. The adaptive document management system can dynamically reorganize quality management documents based on semantic trajectory tracking algorithms, improve document management efficiency and accuracy, and ensure the timeliness and availability of documents.

[0077] In addition, the present invention connects each module through a unified data bus to form a closed-loop management process of quality data collection-analysis-decision-feedback, which can realize intelligent control of the entire quality management process, improve management efficiency and scientific decision-making. The quality data lineage tracking module uses blockchain technology to record the entire process of data generation, modification and circulation, and realizes privacy protection of data auditing through zero-knowledge proof to ensure the reliable traceability and security of data. Overall, the present invention can significantly improve the product quality management level of urban construction and transportation engineering survey and design companies, enhance the quality and efficiency of engineering projects, and promote the intelligent development of the industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily apparent by reading the following detailed description with reference to the accompanying drawings, in which several embodiments of the present invention are shown by way of example and not limitation, in which:

[0079] Figure 1 A schematic diagram of the structure of a product quality management system for urban construction and transportation engineering survey and design enterprises provided by one embodiment of the present invention;

[0080] Figure 2 A schematic diagram of a process for product quality management of an urban construction and transportation engineering survey and design enterprise provided by one embodiment of the present invention;

[0081] Figure 3 The figure schematically shows the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0082] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided solely to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. Rather, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0083] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software.

[0084] It should be noted that any number of elements in the drawings is for illustration only and not for limitation, and any naming is only for distinction and does not have any limiting meaning.

[0085] Reference below Figure 1 , Figure 1 This is a schematic diagram of the structure of a product quality management system for urban construction and transportation engineering survey and design enterprises provided by one embodiment of the present invention. Figure 1 As shown, a product quality management system 100 for urban construction and transportation engineering survey and design enterprises includes:

[0086] A cross-project quality data fusion module 101 is configured to integrate the survey and design data of construction projects, municipal projects, transportation projects, landscape projects, water supply and drainage projects, heating and ventilation projects, electrical projects, and geotechnical projects through a multimodal feature alignment algorithm;

[0087] A dynamic defect evolution prediction module 102 is configured to construct a quality defect propagation model based on a spatiotemporal correlation analysis algorithm;

[0088] an intelligent layered spot check engine 103 configured to generate a multi-level quality check plan through a hybrid optimization model;

[0089] A three-dimensional quality situation deduction module 104 is configured to construct an interactive spatiotemporal topological map of quality data using a parameterized driving engine;

[0090] The conference knowledge graph construction module 105 is configured to generate a decision association network of technical conferences based on an event chain mining algorithm;

[0091] The achievement impact assessment module 106 is configured to quantify the value contribution of scientific and technological achievements through a technology diffusion model;

[0092] an adaptive document management system 107 configured to implement dynamic reorganization of quality management documents based on a semantic trajectory tracking algorithm;

[0093] The system connects each module through a unified data bus to form a closed loop of quality data collection-analysis-decision-feedback.

[0094] It should be noted that the core of the present invention is to build a product quality management system for urban construction and transportation engineering survey and design enterprises, integrate the survey and design data of multi-field projects through a cross-engineering type quality data fusion module, and break down data barriers. Among them, the multimodal feature alignment algorithm can map the data features of different engineering types to a unified space to achieve data integration. The dynamic defect evolution prediction module predicts the defect propagation path through a spatiotemporal correlation analysis algorithm, the intelligent layered spot check engine uses a hybrid optimization model to generate an efficient verification plan, and the three-dimensional quality situation deduction module uses a parameterized driving engine to realize the visual deduction of quality data. These modules work together to form a closed-loop management process from data collection to analysis, decision-making to feedback, thereby improving quality management efficiency.

[0095] Specifically, the multimodal feature alignment algorithm integrates data by aligning its temporal and spatial features. For example, the temporal features of construction project data are construction progress timestamps, and the spatial features are the three-dimensional coordinates of the building structure; the temporal features of traffic project data are traffic flow monitoring times, and the spatial features are the road geographic coordinates. Through a dual attention mechanism, the system addresses the issue of temporal and spatial basis differences in multi-source data. Spatial alignment weights are determined by calculating spatial feature similarity, and temporal alignment weights are determined by calculating timestamp differences. The dynamic defect evolution prediction module uses a convolutional long short-term memory network to extract spatiotemporal features of defects, such as the defect location coordinate set, area growth rate, and associated component stress values. The propagation path modeling unit uses an improved cellular automaton algorithm to simulate the defect diffusion process. The risk warning unit triggers hierarchical alarms based on defect impact volume, component safety threshold, and risk tolerance coefficient. The intelligent hierarchical spot check engine's inspection strategy generation unit constructs a project tiering model based on engineering complexity indicators. The multi-objective optimization unit uses a hybrid strategy of tabu search and genetic algorithm to determine the optimal spot check combination. The dynamic adjustment unit updates inspection priorities based on real-time quality data streams. The topological relationship reconstruction unit of the three-dimensional quality situation deduction module generates a three-dimensional grid structure with mechanical properties based on the BIM model and point cloud data. The multi-parameter coupling unit maps parameters such as quality score, rectification efficiency, and risk level into color gradient values through the field theory model. The interactive deduction unit can dynamically simulate the quality evolution process in response to user instructions, including spatial section analysis, chain tracing, and comparative display of rectification effects of multiple schemes.

[0096] Optimally, the heterogeneous data alignment unit of the cross-project quality data fusion module employs a dual attention mechanism to address spatial and temporal benchmark discrepancies. The feature decoupling unit separates domain features from common features using a generative adversarial network, and the incremental learning unit updates the feature extractor based on a knowledge distillation algorithm. The multi-stage defect assessment unit of the dynamic defect evolution prediction module extracts spatial features through convolution operations and temporal features using a long short-term memory network. The verification strategy generation unit of the intelligent layered spot check engine calculates a project complexity score by weightedly summing the number of cross-disciplinary areas, geological risk level, and technological novelty. The topological relationship reconstruction unit of the three-dimensional quality situation deduction module combines BIM model geometry and point cloud data spatial information to generate a three-dimensional mesh structure. The multi-parameter coupling unit displays parameters in the form of color gradients using a field theory model and visualization weight coefficients. Through these optimized and refined technical approaches, the system can more efficiently and comprehensively monitor and manage project quality.

[0097] In some embodiments, the dynamic defect evolution prediction module includes:

[0098] A multi-stage defect assessment unit extracts the spatiotemporal features of design defects based on a convolutional long short-term memory network. These features include the defect location coordinate set L(x, y, z), the defect area growth rate ΔA / Δt, and the associated component stress value σ.

[0099] The propagation path modeling unit uses an improved cellular automaton algorithm to simulate the defect diffusion process and defines the evolution rules as follows:

[0100] P d =λ×(1+e -(α×S+β×T) ) -1

[0101] Among them, P d is the probability of defect diffusion, λ is the material attenuation coefficient, S is the stress concentration, T is the duration of environmental action, α and β are empirical correction factors;

[0102] Risk warning unit, when the condition ∑(P d ×V d ) / C m >γ triggers a graded alarm, where V d is the defect impact volume, C m is the component safety threshold, and γ is the risk tolerance coefficient.

[0103] It should be noted that the core function of the dynamic defect evolution prediction module is to extract the spatiotemporal features of engineering design defects, simulate the diffusion process, and issue risk warnings through a multi-stage defect assessment unit, a propagation path modeling unit, and a risk warning unit. The convolutional long short-term memory network is a deep learning model that combines the spatial feature extraction capabilities of convolutional neural networks (CNNs) with the time series modeling capabilities of long short-term memory networks (LSTMs). It is used to extract the spatiotemporal features of design defects, including the defect location coordinate set, defect area growth rate, and associated component stress values. The propagation path modeling unit uses an improved cellular automaton algorithm to simulate the defect diffusion process by defining evolutionary rules. The risk warning unit triggers graded alarms based on parameters such as the defect impact volume, component safety threshold, and risk tolerance coefficient to provide timely warnings of potential risks.

[0104] Specifically, the multi-stage defect assessment unit uses convolution operations to extract spatial features of defects, such as the defect's location coordinate set. Local features are calculated by sliding the convolution kernel to capture the spatial distribution of defects. Simultaneously, a long-short-term memory network is used to extract temporal features, such as the defect area growth rate and the temporal trend of the stress values of associated components. The propagation path modeling unit uses an improved cellular automaton algorithm to simulate the defect diffusion process by defining evolutionary rules. These rules take into account factors such as the material attenuation coefficient, stress concentration, and the duration of environmental exposure, and predict the defect propagation path by calculating the probability of defect diffusion. The risk warning unit determines whether to trigger an alarm based on the ratio of the defect's impact volume to the component's safety threshold and the risk tolerance factor. When the ratio exceeds the set risk tolerance factor, the system issues a graded alarm, prompting relevant personnel to take timely action. This mechanism can effectively prevent the further expansion of defects and reduce engineering risks.

[0105] Preferably, the multi-stage defect assessment unit can optimize the accuracy of spatial feature extraction by adjusting the size and depth of the convolution kernel, and at the same time set the number of hidden layer units of the long short-term memory network to better capture time series characteristics. For example, for complex engineering structures, larger convolution kernels and deeper network structures can be used to extract finer spatial features. In the propagation path modeling unit, the parameters of the evolution rules can be adjusted according to the specific engineering materials and environmental conditions. For example, for high stress concentration areas, the weight of the stress concentration can be appropriately increased to more accurately simulate the defect diffusion process. In the risk warning unit, the risk tolerance factor can be set according to the importance and safety level of the project. For example, in high-risk projects, a lower risk tolerance factor can be set to provide early warning. In addition, machine learning algorithms can be introduced to optimize the evolution rules, train the model through historical data, and automatically adjust the parameters to improve prediction accuracy.

[0106] In some embodiments, the intelligent layered spot check engine includes:

[0107] Verification strategy generation unit, based on the engineering complexity index E c =ω1×N c +ω2×D g +ω3×T s Construct a project classification model, where B c is the number of cross-disciplinary majors, D g is the geological risk level, T s is the technical novelty, ω i is the weight coefficient;

[0108] The multi-objective optimization unit uses a hybrid strategy of taboo search and genetic algorithm to solve the optimal spot check combination. The objective function is:

[0109] min∑(1-R k )×I k +max(T j )-min(T j )

[0110] Constraint: ∑C m ≤B,R k ≥θ

[0111] where R k is the risk coefficient of project k, I k is the impact factor, T j is the workload of inspector j, C m is the inspection cost, B is the total budget, and θ is the minimum coverage threshold;

[0112] Dynamic adjustment unit, updates the verification priority based on real-time quality data flow and defines the adjustment factor Where ΔQ is the change in quality score, σ is the historical volatility, and Δt is the monitoring period.

[0113] It should be noted that the core function of the intelligent hierarchical inspection engine is to achieve efficient generation and dynamic adjustment of project-level inspection plans through the inspection strategy generation unit, multi-objective optimization unit, and dynamic adjustment unit. Among them, the engineering complexity index is a quantitative indicator used to evaluate the complexity of the project. It comprehensively considers factors such as the number of cross-disciplinary disciplines, geological risk level, and technological novelty, and is obtained by weighted summation of weight coefficients. The inspection strategy generation unit constructs a project classification model based on this index to determine the inspection priority and strategy for different projects. The multi-objective optimization unit adopts a hybrid strategy of taboo search and genetic algorithm to solve the optimal inspection combination. The objective function comprehensively considers factors such as project risk factor, inspector workload, and inspection cost to achieve a balance between inspection efficiency and cost. The dynamic adjustment unit updates the inspection priority based on the real-time quality data stream to ensure the dynamic adaptability of the inspection strategy.

[0114] Specifically, setting a project complexity index requires comprehensive consideration of multiple factors. The number of interdisciplinary areas refers to the number of different disciplines involved in a project. For example, a construction project may involve multiple disciplines, such as structure, electrical engineering, and HVAC. The greater the number of interdisciplinary areas, the higher the project complexity. The geological risk level is an assessment of project risk based on geological conditions. For example, projects constructed on soft soil or in earthquake-prone areas have a higher geological risk level. Technological novelty reflects the extent to which new technologies are adopted in a project; the application of new technologies may introduce greater risk and complexity. Weighting coefficients can be adjusted to reflect the importance of different factors based on project type and management requirements. In the multi-objective optimization unit, the taboo search algorithm is used to avoid local optimal solutions by dynamically adjusting the search direction through the establishment of a taboo table and a set of candidate solutions. The genetic algorithm uses crossover, mutation, and selection operations to rapidly converge to the global optimal solution or a near-optimal solution. In the objective function, the project risk coefficient reflects the magnitude of the project's potential risk, the impact factor is used to adjust the risk weights between different projects, and the workload and cost of inspection personnel are practical constraints during the verification implementation process. In the dynamic adjustment unit, the adjustment factor is calculated based on the change in quality score, historical volatility and monitoring period, and is used to dynamically adjust the verification priority to ensure that the verification strategy can respond to changes in project quality in a timely manner.

[0115] Preferably, the weight coefficient of the engineering complexity index can be refined and adjusted according to the project type and management objectives. For example, in technology-intensive projects, the weight of technological novelty can be increased; in projects with complex geological conditions, the weight of geological risk level can be increased. In the multi-objective optimization unit, more optimization algorithms, such as simulated annealing algorithm, can be introduced to further improve the optimization effect. The constraints in the objective function can be adjusted according to the actual project situation. For example, for projects with limited budgets, the weight of inspection costs can be increased; for time-sensitive projects, time constraints can be introduced. In the dynamic adjustment unit, the calculation of the adjustment factor can introduce more quality-related parameters, such as defect density or rectification efficiency, to more comprehensively reflect the quality status of the project. In addition, it can also be combined with machine learning algorithms to automatically adjust the weight coefficients and optimization strategies through learning and analysis of historical data, thereby further improving the intelligence level of the intelligent layered spot check engine.

[0116] In some embodiments, the three-dimensional quality situation deduction module includes:

[0117] The topological relationship reconstruction unit generates a three-dimensional grid structure with mechanical properties based on the BIM model and point cloud data, and defines the grid node attribute set including material strength σ m , deformation δ, defect density ρ d ;

[0118] The multi-parameter coupling unit maps the quality score Q, rectification efficiency η, and risk level R into a color gradient value C = α × Q + β × log (R) + γ × η through a field theory model, where α, β, and γ are visualization weight coefficients;

[0119] The interactive deduction unit responds to user operation instructions to dynamically simulate the quality evolution process, including:

[0120] Spatial section analysis based on virtual reality equipment

[0121] Chain traceability of associated component quality data

[0122] Parallel comparison of the rectification effects of multiple plans.

[0123] It should be noted that the core function of the three-dimensional quality situation deduction module is to realize the visual deduction and dynamic analysis of quality data through the topological relationship reconstruction unit, the multi-parameter coupling unit and the interactive deduction unit. Among them, the topological relationship reconstruction unit generates a three-dimensional grid structure with mechanical properties based on BIM (Building Information Model) and point cloud data, which is used to simulate the distribution and evolution of engineering quality data in space. The multi-parameter coupling unit maps parameters such as quality score, rectification efficiency, and risk level to color gradient values through the field theory model to realize the visual display of multi-dimensional data. The interactive deduction unit responds to user operation instructions, dynamically simulates the quality evolution process, supports functions such as spatial section analysis, chain tracing, and comparative display of rectification effects of multiple schemes, and provides intuitive decision support for engineering quality management.

[0124] Specifically, in the topological relationship reconstruction unit, the BIM model is a digital model that contains multi-dimensional data such as building geometry information, material properties, and component relationships, while point cloud data is high-precision spatial data obtained through technologies such as laser scanning. By fusing the BIM model with point cloud data, a three-dimensional grid structure with mechanical properties can be generated. The grid node properties include parameters such as material strength, deformation, and defect density. The multi-parameter coupling unit uses a field theory model to visualize parameters such as quality score, rectification efficiency, and risk level through color gradient values. The visualization weight coefficient can be adjusted according to actual needs to highlight the importance of different parameters. The interactive deduction unit supports multiple user interaction methods. For example, through virtual reality equipment, spatial section analysis can be performed, and users can view the quality data of different sections in real time. The chain traceability function can link component quality data to display the propagation path of defects. The multi-scheme rectification effect comparison display allows users to input different rectification plans and display their possible effects through simulation.

[0125] Preferably, more advanced modeling algorithms, such as deep learning-based mesh generation algorithms, can be introduced into the topological relationship reconstruction unit to improve the accuracy and complexity adaptability of the three-dimensional mesh structure. In the multi-parameter coupling unit, the visualization weight coefficient can be dynamically adjusted according to the project type and management needs. For example, in high-risk projects, the weight of the risk level can be increased; in the quality optimization stage, the weight of the rectification efficiency can be increased. In the interactive deduction unit, more interactive functions can be added, such as user-defined dynamic simulation scenarios and real-time data feedback mechanisms. In addition, artificial intelligence algorithms, such as reinforcement learning, can be introduced to optimize the recommendation logic of the rectification plan, and dynamically adjust the recommendation strategy according to real-time quality data to further enhance the intelligence level of the system and user experience.

[0126] In some embodiments, the conference affairs knowledge graph construction module includes:

[0127] The meeting decision chain extraction unit uses a multi-head attention mechanism to identify technical decision points and their relationships from meeting minutes, and defines the relationship strength as:

[0128]

[0129] where Q i is the query vector of the i-th decision point, K j is the key vector of the j-th associated point, and d is the feature dimension;

[0130] The knowledge accumulation unit builds a four-dimensional graph containing topics, conclusions, responsible parties, and execution status based on a dynamic graph convolutional network;

[0131] The intelligent reminder unit automatically triggers a supervision notification when it detects that the decision execution progress deviation ΔD>ε, where ΔD=|actual progress D a -Plan Progress D p |, ε is the tolerance threshold.

[0132] It should be noted that the core function of the conference knowledge graph construction module is to achieve efficient extraction, knowledge precipitation and dynamic monitoring of technical conference decision information through the conference decision chain extraction unit, knowledge precipitation unit and intelligent reminder unit. Among them, the conference decision chain extraction unit adopts a multi-head attention mechanism to identify technical decision points and their correlation relationships from the meeting minutes, and determines the connection between decision points by calculating the relationship strength. The knowledge precipitation unit constructs a four-dimensional graph containing topics, conclusions, responsible parties, and execution status based on a dynamic graph convolutional network to achieve structured storage of conference knowledge. The intelligent reminder unit automatically triggers supervision notifications by detecting deviations in the decision execution progress to ensure the effective implementation of meeting decisions.

[0133] Specifically, in the meeting decision chain extraction unit, the multi-head attention mechanism is a deep learning technology that can simultaneously focus on multiple key information points in the text, and identify decision points and their correlations by calculating the similarity between the query vector and the key vector. In the formula for calculating the relationship strength, the square root of the feature dimension is used to adjust the magnitude of the similarity to ensure the rationality of the calculation results. In the knowledge precipitation unit, the dynamic graph convolutional network is a neural network model for processing graph-structured data. It can dynamically construct a graph based on the information in the meeting minutes. The nodes in the graph include key information such as topics, conclusions, responsible parties, and execution status, and the edges represent the correlation between this information. In the intelligent reminder unit, the decision execution progress deviation is determined by calculating the difference between the actual progress and the planned progress. When the deviation exceeds the set tolerance threshold, the system will automatically trigger a supervision notification to remind relevant personnel to deal with it in a timely manner.

[0134] Preferably, more contextual information can be introduced into the meeting decision chain extraction unit to optimize the recognition effect of decision points. For example, the accuracy of decision point extraction can be further improved by combining the meeting agenda and the speech records of the participants. In the knowledge precipitation unit, a dynamic update mechanism of the graph can be added to adjust the graph content in real time according to the subsequent progress of the meeting to ensure the timeliness of knowledge precipitation. In the intelligent reminder unit, more complex progress monitoring algorithms can be introduced, such as a progress prediction model based on machine learning, to provide early warning of possible progress deviations and provide corresponding adjustment suggestions. In addition, intelligent reminders can be integrated with the company's instant messaging tools or project management software to achieve real-time push and feedback of reminder information, further improving the practicality of the system.

[0135] In some embodiments, the outcome impact assessment module includes:

[0136] Technology diffusion calculation unit, based on the improved PageRank algorithm to quantify the impact of the results:

[0137] PR(u)=(1-d)+d×∑(PR(v)×W vu / (v))

[0138] Where d is the damping coefficient, W vu is the technical association weight of achievement v to u, C(v) is the total number of outbound links of achievement v;

[0139] Value contribution analysis unit, building a c , technology improvement degree T i 、Industry recognition I a The evaluation matrix:

[0140] V=ω1×E c +ω2×arctan(T i)+ω3×log(I a +1)

[0141] where ω i Adjust weights for sectors;

[0142] The association recommendation unit generates an optimization plan for the combination of results based on the spectral clustering algorithm.

[0143] It should be noted that the core function of the achievement impact assessment module is to achieve quantitative evaluation and optimized recommendation of the value of scientific and technological achievements through the technology diffusion calculation unit, value contribution analysis unit and association recommendation unit. Among them, the technology diffusion calculation unit quantifies the influence of the achievements based on the improved PageRank algorithm, and evaluates the degree of dissemination and application of the achievements within the industry by calculating the technical association weights and the total number of outbound links between the achievements. The value contribution analysis unit constructs an evaluation matrix that includes economic benefits, technological advancement and industry recognition, and comprehensively evaluates the multi-dimensional value of scientific and technological achievements. The association recommendation unit generates an achievement combination optimization plan based on the spectral clustering algorithm to provide strategic support for the promotion and application of the achievements.

[0144] Specifically, in the technology diffusion calculation unit, the improved PageRank algorithm is an influence assessment method based on graph structure. It calculates the influence value of each achievement by analyzing the technical association relationship between the achievements. The technical association weight reflects the correlation and dependency between the achievements, and the total number of outbound links indicates the scope of influence of one achievement on other achievements. In the value contribution analysis unit, the evaluation matrix comprehensively considers the three dimensions of economic benefits, technological advancement, and industry recognition by setting different weight coefficients. Economic benefits can be measured by market returns or cost savings; technological advancement can be evaluated by the degree of technological improvement or innovation; and industry recognition can be determined by awards, citations, or expert evaluations. In the association recommendation unit, the spectral clustering algorithm clusters the achievements into different combinations by analyzing the similarities between the achievements, providing a basis for optimizing the promotion plan.

[0145] Preferably, more dimensional association weights can be introduced into the technology diffusion calculation unit, such as considering the time correlation or geographical distribution of the results, so as to more comprehensively reflect the dynamic process of technology diffusion. In the value contribution analysis unit, industry adjustment weights can be introduced to calibrate the evaluation results according to the characteristics of different industries to make them more in line with actual needs. For example, in emerging industries, the weight of technological advancement can be appropriately increased; in traditional industries, the weight of economic benefits can be more important. In the association recommendation unit, machine learning algorithms such as collaborative filtering can be combined to generate personalized results recommendation plans based on the user's historical behavior and preferences. In addition, the evaluation results can be combined with the company's strategic goals to provide decision support for the prioritization of results and resource allocation.

[0146] In some embodiments, the adaptive file management system includes:

[0147] The semantic trajectory analysis unit uses graph embedding technology to construct the file usage behavior feature space and defines the relevance A of user u to file f. uf =∑δ(t)×sim(q u ,c f ), where δ(t) is the time decay function, sim is the semantic similarity, and q u Query features for users, c f is the file content feature;

[0148] Dynamic recombination unit, when A is detected uf Automatically increase the index priority of file f when it exceeds the threshold θ for n consecutive times;

[0149] The intelligent snapshot unit generates a file version evolution graph based on the differential hash algorithm and visualizes the distribution of revision points.

[0150] It should be noted that the core function of the adaptive file management system is to achieve dynamic management and optimization of quality management files through the semantic trajectory analysis unit, dynamic reorganization unit and intelligent snapshot unit. Among them, the semantic trajectory analysis unit uses graph embedding technology to construct the file usage behavior feature space, and dynamically adjusts the index priority of the file by analyzing the semantic correlation between the user and the file to improve the flexibility and efficiency of file management. The dynamic reorganization unit automatically adjusts the organizational structure of the file according to the user's behavior pattern and the frequency of file use to ensure the dynamic adaptability of file management. The intelligent snapshot unit generates a file version evolution map based on the differential hash algorithm, visually displays the distribution of the file's revision points, and helps users quickly understand the changes in the file.

[0151] Specifically, in the semantic trajectory analysis unit, graph embedding technology is a technology that maps graph structured data to a low-dimensional space, which can effectively integrate user behavioral characteristics and file content characteristics. By calculating the semantic similarity between users and files, combined with the time decay function, the user's association with the file can be obtained. The time decay function is used to reflect the changes in user interest in files over time, and the semantic similarity is calculated by analyzing user query features and file content features. The dynamic reorganization unit automatically adjusts the index priority of the file according to the continuous change of the association. For example, when a file is frequently used in multiple consecutive operations, its index priority will be increased so that users can access it faster. In the intelligent snapshot unit, the differential hash algorithm is used to quickly detect changes in file content, and the generated file version evolution map can intuitively display the revision history and key change points of the file.

[0152] Preferably, in the semantic trajectory analysis unit, the calculation accuracy of semantic similarity can be further improved by introducing a more complex semantic analysis model, such as a Transformer-based model. The time decay function can be personalized according to different types of files and user behavior patterns. For example, for technical documents, the decay rate can be relatively slow to reflect their long-term value. In the dynamic reorganization unit, the file reorganization strategy can be optimized in combination with the user's role and permission information. For example, a more convenient file access path can be provided for project managers. In the intelligent snapshot unit, a version comparison function can be introduced to allow users to directly view the differences between different versions and improve the transparency of file management. In addition, file management can be combined with project progress management to automatically adjust the classification and priority of files according to the project stage, further improving the intelligence level of file management.

[0153] In some embodiments, further comprising:

[0154] The quality data lineage tracking module is configured as follows:

[0155] Use blockchain technology to record the entire process of quality data generation, modification, and circulation

[0156] Build a verifiable chain containing data source HASH value, operator ID, and timestamp to achieve privacy protection for data auditing through zero-knowledge proof

[0157] The data lineage integrity verification conditions are:

[0158] Verify(Sig i ,H(H i |H i-1 ))=True

[0159] Represents the signature Sig of the i-th block i With the current hash H i , pre-order hash H i-1 The connection verification passed.

[0160] It's important to note that the core function of the quality data lineage tracking module is to record the entire process of quality data generation, modification, and transfer through blockchain technology, constructing a verifiable chain consisting of data source hash values, operator IDs, and timestamps. This chain also uses zero-knowledge proofs to protect privacy during data audits. Data lineage integrity verification ensures data integrity and immutability through blockchain signature mechanisms, providing technical support for the trusted traceability of quality data. This mechanism not only ensures data authenticity but also protects the privacy of stakeholders during the data audit process.

[0161] Specifically, blockchain technology is a distributed ledger technology that ensures data immutability and transparency through encryption algorithms and consensus mechanisms. In the quality data lineage tracking module, the data source HASH value is a unique identifier obtained by hashing the original data and is used to verify the integrity and source of the data. The operator ID is used to record the identity of the person who operated on the data, and the timestamp records the specific time of the data operation. The verifiable chain constructed through this information can clearly track the flow of data. Zero-knowledge proof is a cryptographic technology that allows the authenticity of data to be verified without revealing the specific content of the data, thereby protecting privacy during the data audit process. Data lineage integrity verification is achieved through the blockchain's signature mechanism, ensuring that the signature of each block is verified in conjunction with the current hash and the previous hash, thereby ensuring the integrity and credibility of the data.

[0162] Preferably, more efficient encryption algorithms can be introduced to optimize blockchain performance, such as using elliptic curve cryptography (ECC) to increase the speed of signing and verification. Data source hash values can be generated using a combination of multiple hash algorithms, such as SHA-256 and SHA-3, to enhance data tamper resistance. Operator ID management can be integrated with the company's identity authentication system to ensure the accuracy and traceability of operator identities. The accuracy of timestamps can be adjusted based on actual needs. For example, in scenarios requiring high precision, millisecond or microsecond timestamps can be used. Furthermore, smart contract technology can be introduced to automatically enforce rules and permission verification in data flows, further improving the level of automation in data management.

[0163] In some embodiments, the cross-engineering type quality data fusion module includes:

[0164] The heterogeneous data alignment unit uses a dual attention mechanism to address the temporal and spatial basis differences of multi-source data and defines the spatial alignment weight as:

[0165]

[0166] The time alignment weight is:

[0167] W t =exp(-|t i -t j |σ t )

[0168] where σ s , σ t is a learnable parameter;

[0169] Feature decoupling unit, which separates the domain features and common features of engineering data through adversarial generative networks;

[0170] Incremental learning unit, when a new engineering type is added, the feature extractor is updated based on the knowledge distillation algorithm without compromising the performance of the existing model.

[0171] It should be noted that the core function of the cross-engineering type quality data fusion module is to achieve efficient integration and dynamic updating of multi-source heterogeneous data through the heterogeneous data alignment unit, feature decoupling unit and incremental learning unit. Among them, the heterogeneous data alignment unit adopts a dual attention mechanism to solve the problem of differences in multi-source data in spatiotemporal benchmarks. By calculating the spatial alignment weight and the temporal alignment weight, the data of different engineering types are aligned to the same benchmark. The feature decoupling unit uses an adversarial generative network to separate the domain features and public features of the engineering data, ensuring that the features after data fusion are clear and independent. The incremental learning unit is based on the knowledge distillation algorithm. When a new engineering type is added, it updates the feature extractor without compromising the performance of the existing model, ensuring the dynamic adaptability of the system.

[0172] Specifically, the dual attention mechanism in the heterogeneous data alignment unit is a deep learning-based technology that adjusts the spatiotemporal basis of the data through spatial alignment weights and temporal alignment weights. The spatial alignment weights are determined by calculating the differences in spatial features of different data sources, while the temporal alignment weights are adjusted based on the differences in timestamps. The adversarial generative network in the feature decoupling unit consists of a generator and a discriminator. The generator is responsible for generating features that are close to the real data, while the discriminator is used to distinguish the generated features from the real features, separating the domain features from the common features through adversarial training. The knowledge distillation algorithm in the incremental learning unit is a model compression technology that supports new engineering types by migrating the knowledge of the existing model to the new model while maintaining the performance of the original model. In terms of parameter setting, the spatial alignment weights and temporal alignment weights can be automatically adjusted through the learning algorithm to adapt to different data sources and engineering types.

[0173] Preferably, a more advanced attention mechanism, such as the multi-head attention mechanism in the Transformer architecture, can be introduced into the heterogeneous data alignment unit to improve alignment accuracy. In the feature decoupling unit, the generator and discriminator of the adversarial generative network can be designed as a deeper network structure to better separate the features of complex engineering data. In the incremental learning unit, the knowledge distillation algorithm can adjust the smoothness of knowledge transfer by introducing a temperature parameter, thereby more effectively updating the feature extractor when new engineering types are added. In addition, meta-learning technology can be combined to enable the system to quickly adapt to new engineering types and data patterns, further improving the system's generalization ability and dynamic adaptability.

[0174] The above-mentioned embodiments of the present invention have the following beneficial effects: Through the cross-project quality data fusion module, the present invention can integrate survey and design data from multiple engineering fields, breaking down information silos and enabling efficient data sharing and collaborative analysis. The dynamic defect evolution prediction module accurately simulates the spatiotemporal evolution of defects, providing early warning of potential risks and providing a scientific basis for quality control. The intelligent layered spot check engine generates multi-level quality verification plans and dynamically adjusts verification strategies, improving verification efficiency and accuracy while reducing labor costs. The three-dimensional quality situation deduction module constructs an interactive spatiotemporal topological map of quality data, enabling multi-field coupling visualization of quality data and mechanical properties, intuitively displaying the quality situation evolution process and providing visual support for decision-making. The conference knowledge graph construction module generates a decision-making association network for technical meetings, accumulating meeting knowledge and forming an effective knowledge transfer and decision-making support system. The achievement impact assessment module quantifies the value contribution of scientific and technological achievements, providing a scientific evaluation basis for their promotion and application. The adaptive document management system dynamically reorganizes quality management documents, improving document management efficiency and accuracy and ensuring document timeliness and availability.

[0175] Furthermore, the quality data lineage tracking module records the entire process of data generation, modification, and transfer. Through blockchain technology and zero-knowledge proofs, it enables trusted data traceability and privacy protection, ensuring data integrity and security. Overall, this invention can significantly improve the product quality management level of urban construction and transportation engineering survey and design companies, enhance the quality and efficiency of engineering projects, and promote the intelligent development of the industry.

[0176] like Figure 2 As shown, in some embodiments, a product quality management method 200 for an urban construction and transportation engineering survey and design enterprise is provided. The method 200 includes:

[0177] Step 201: Establish a multimodal quality data fusion channel across engineering types and achieve spatiotemporal reference alignment through a dual attention mechanism;

[0178] Step 202: construct a defect spatiotemporal evolution prediction model, and use a cellular automation and deep learning fusion algorithm to simulate the defect propagation path;

[0179] Step 203: Generate a multi-objective optimized intelligent layered inspection plan, and dynamically adjust the inspection strategy by combining tabu search and genetic algorithm;

[0180] Step 204 , creating a parameterized three-dimensional quality deduction system to realize multi-field coupling visualization of quality data and mechanical properties;

[0181] Step 205: Implement a technical conference knowledge sedimentation mechanism based on decision chain mining to construct a four-dimensional association graph;

[0182] Step 206: Develop a technology diffusion-driven achievement value assessment system to quantify the multi-dimensional contributions of scientific and technological achievements;

[0183] Step 207: deploy a semantic trajectory-aware file management system to dynamically reconstruct the file organization structure;

[0184] The method achieves full-process trusted traceability of quality data through blockchain evidence storage and zero-knowledge proof.

[0185] It is understandable that the steps and references in the Urban Construction and Transportation Engineering Survey and Design Enterprise Product Quality Management Method 200 Figure 1 The modules, features, and beneficial effects described above for the product quality management system for urban construction and transportation engineering survey and design enterprises also apply to the product quality management method 200 for urban construction and transportation engineering survey and design enterprises and the operations contained therein, and will not be further described here.

[0186] Reference below Figure 3 , which shows a schematic structural diagram of an electronic device 300 suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 3 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0187] like Figure 3 As shown, the electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. Various programs and data required for the operation of the electronic device 300 are also stored in the RAM 303. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0188] Typically, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as needed.

[0189] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or a terminal device such as a computer, a server, a mobile phone, or a tablet.

[0190] The above descriptions are merely some preferred embodiments of the present invention and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also encompass other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned inventive concept. For example, a technical solution formed by mutually replacing the above-mentioned features with (but not limited to) technical features having similar functions disclosed in the embodiments of the present invention.

Claims

1. A product quality management system for urban construction and transportation engineering survey and design enterprises, characterized by: include: A cross-project quality data fusion module, configured to integrate survey and design data for construction, municipal, transportation, landscape, water supply and drainage, HVAC, electrical, and geotechnical engineering projects through a multimodal feature alignment algorithm; A dynamic defect evolution prediction module is configured to build a quality defect propagation model based on a spatiotemporal correlation analysis algorithm; An intelligent layered spot check engine configured to generate multi-level quality verification plans through a hybrid optimization model; A three-dimensional quality situation deduction module is configured to use a parameterized driving engine to build an interactive spatiotemporal topological map of quality data; The conference knowledge graph construction module is configured to generate a decision-making association network for technical conferences based on an event chain mining algorithm; The impact assessment module is configured to quantify the value contribution of scientific and technological achievements through the technology diffusion model; An adaptive document management system configured to achieve dynamic reorganization of quality management documents based on a semantic trajectory tracking algorithm; The system connects each module through a unified data bus to form a closed loop of quality data collection-analysis-decision-feedback.

2. The system according to claim 1, wherein: The dynamic defect evolution prediction module includes: A multi-stage defect assessment unit extracts the spatiotemporal features of design defects based on a convolutional long short-term memory network. These spatiotemporal features include the defect location coordinate set L(x, y, z), the defect area growth rate ΔA / Δt, and the associated component stress value σ. The propagation path modeling unit uses an improved cellular automaton algorithm to simulate the defect diffusion process and defines the evolution rules as follows: P d =λ×(1+e -(α×；+β×T) ) -1 Among them, P d is the probability of defect diffusion, λ is the material attenuation coefficient, S is the stress concentration, T is the duration of environmental action, α and β are empirical correction factors; Risk warning unit, when the condition ∑(P d ×V d ) / C m >γ triggers a graded alarm, where V d is the defect impact volume, C m is the component safety threshold, and γ is the risk tolerance coefficient.

3. The system according to claim 1, wherein: The intelligent layered spot check engine includes: The verification strategy generation unit builds a project classification model based on the engineering complexity index. The engineering complexity index is shown in the following formula: E c =ω1×N c +ω2×D g +ω3×T s , where N c is the number of cross-disciplinary majors, D g is the geological risk level, T s is the technical novelty, ω i is the weight coefficient, E c is the engineering complexity index; The multi-objective optimization unit uses a hybrid strategy of taboo search and genetic algorithm to solve the optimal spot check combination. The objective function is: min∑(1-R k )×I k +max(T j )-min(T j ) The constraint condition is: ∑C m ≤B,R k ≥θ Among them, R k is the risk coefficient of project k, I k is the impact factor, T j is the workload of inspector j, C m is the inspection cost, B is the total budget, and θ is the minimum coverage threshold; Dynamic adjustment unit, updates the verification priority based on real-time quality data flow and defines the adjustment factor Where ΔQ is the change in quality score, σ is the historical volatility, and Δt is the monitoring period.

4. The system according to claim 1, wherein: The three-dimensional quality situation deduction module includes: The topological relationship reconstruction unit generates a three-dimensional grid structure with mechanical properties based on the BIM model and point cloud data, and defines the grid node attribute set including material strength σ m , deformation δ, defect density ρ d ; The multi-parameter coupling unit maps the quality score Q, rectification efficiency η, and risk level R into color gradient values through the field theory model, as shown in the following formula; C = α × Q + β × log (R) + γ × η, where α, β, and γ are visualization weight coefficients; The interactive deduction unit responds to user operation instructions to dynamically simulate the quality evolution process, including: Spatial section analysis based on virtual reality equipment; Chain traceability of associated component quality data; Parallel comparison of the rectification effects of multiple plans.

5. The system according to claim 1, wherein: The conference knowledge graph construction module includes: The meeting decision chain extraction unit uses a multi-head attention mechanism to identify technical decision points and their relationships from meeting minutes, and defines the relationship strength as: Among them, Q i is the query vector of the i-th decision point, K j is the key vector of the j-th associated point, d is the feature dimension, S ij is the strength of the relationship; The knowledge accumulation unit builds a four-dimensional graph containing topics, conclusions, responsible parties, and execution status based on a dynamic graph convolutional network; The intelligent reminder unit automatically triggers a supervision notification when it detects that the decision execution progress deviation ΔD>ε, where the decision execution progress deviation is as shown in the following formula; ΔD=|Actual progress D a -Plan Progress D p |, ε is the tolerance threshold.

6. The system according to claim 1, wherein: The outcome impact assessment module includes: The technology diffusion calculation unit quantifies the impact of the results based on the improved PageRank algorithm, as shown in the following formula; PR(u)=(1-d)+d×∑(PR(v)×W vu / (v)) Where d is the damping coefficient, W vu is the technical association weight from achievement v to u, C(v) is the total number of outbound links of achievement v, and PR(u) is the impact value of the achievement; Value contribution analysis unit, building a c , technology improvement degree T i 、Industry recognition I a The evaluation matrix is shown in the following formula; V=ω1×E c +ω2×arctan(T i )+ω3×log(I a +1) Among them, ω i Adjust weights for sectors; The association recommendation unit generates an optimization plan for the combination of results based on the spectral clustering algorithm.

7. The system according to claim 1, wherein: The adaptive file management system includes: The semantic trajectory analysis unit uses graph embedding technology to construct the file usage behavior feature space and defines the relevance of user u to file f, as shown in the following formula; A uf =∑δ(t)×sim(q u ,c f ), where δ(t) is the time decay function, sim is the semantic similarity, and q u Query features for users, c f is the file content feature, A uf is the relevance of user u to file f; Dynamic recombination unit, when A is detected uf Automatically increase the index priority of file f when it exceeds the threshold θ for n consecutive times; The intelligent snapshot unit generates a file version evolution graph based on the differential hash algorithm and visualizes the distribution of revision points.

8. The system according to claim 1, wherein: Also includes: The quality data lineage tracking module is configured as follows: Use blockchain technology to record the entire process of quality data generation, modification, and circulation; Build a verifiable chain containing data source HASH value, operator ID, and timestamp; Privacy protection for data auditing is achieved through zero-knowledge proof; The data lineage integrity verification condition is the signature Sig of the i-th block i With the current hash H i , pre-order hash H i-1 The connection verification passed.

9. The system according to claim 1, wherein: The cross-project quality data fusion module includes: The heterogeneous data alignment unit uses a dual attention mechanism to address the spatiotemporal basis differences of multi-source data and defines spatial alignment weights and temporal alignment weights. Feature decoupling unit, which separates the domain features and common features of engineering data through adversarial generative networks; Incremental learning unit, when a new engineering type is added, the feature extractor is updated based on the knowledge distillation algorithm without compromising the performance of the existing model.

10. A product quality management method for urban construction and transportation engineering survey and design enterprises, characterized in that: include: Establish a multimodal quality data fusion channel across engineering types and achieve spatiotemporal benchmark alignment through a dual attention mechanism; Construct a defect spatiotemporal evolution prediction model and use a cellular automation and deep learning fusion algorithm to simulate the defect propagation path; Generate a multi-objective optimized intelligent stratified inspection plan, and dynamically adjust the inspection strategy by combining taboo search and genetic algorithm; Create a parametric three-dimensional mass deduction system to achieve multi-field coupling visualization of mass data and mechanical properties; Implement a technical conference knowledge accumulation mechanism based on decision chain mining to build a four-dimensional correlation map; Develop a technology diffusion-driven achievement value assessment system to quantify the multi-dimensional contributions of scientific and technological achievements; Deploy a semantic trajectory-aware file management system to dynamically reconstruct the file organization structure; The method achieves full-process trusted traceability of quality data through blockchain evidence storage and zero-knowledge proof.

Images