Curtain wall construction stage management method and system

Abstract

The invention relates to the technical field of building construction, in particular to a curtain wall construction stage management method and system, and the method comprises the steps: generating a design scheme through a generative adversarial network, and carrying out the multi-objective optimization through reinforcement learning and a genetic algorithm; integrating construction site data in real time by adopting a digital twinborn technology, generating a construction twinborn model, and dynamically adjusting a construction strategy through a self-adaptive control algorithm; deploying a multi-scale sensor network for real-time monitoring, performing anomaly detection by using deep learning, and providing a construction strategy adjustment basis in combination with a causal analysis model; by integrating the advanced digital technology, comprehensive optimization from design to construction is achieved, the responsiveness and safety of construction are enhanced, and the construction quality and efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of building construction, and in particular to a curtain wall construction phase management method and system. Background Art

[0002] Curtain wall construction phase management is crucial to the successful implementation of construction projects, especially in modern buildings, where curtain wall structures not only play an aesthetic decorative role, but must also meet functionality and safety requirements. Since curtain walls are usually composed of complex multi-curved surfaces, their construction quality directly affects the overall performance and safety of the building. Effective curtain wall construction management can ensure the quality of the curtain wall system, shorten the construction period, reduce costs, and improve overall construction efficiency.

[0003] The existing multi-curved curtain wall construction technology (China invention patent, publication number: CN117494563A, name: a multi-curved curtain wall system full process construction method) mainly relies on two methods: mechanical bending and prefabricated molds. However, these technologies have significant shortcomings:

[0004] Mechanical bending technology: This method is relatively complicated in operation, and the residual stress inside the glass is poorly controlled, which can easily lead to unpredictable stress concentration and damage to the glass during use;

[0005] Prefabricated mold method: This method requires the production of a large number of molds to adapt to changes in the curtain wall structure. The mold production cost is high and it is not suitable for precise molding of complex curved surfaces;

[0006] Stress levels are difficult to predict: Since the stress levels of multi-curved glass panels are difficult to accurately predict using traditional finite element calculation methods, it is impossible to quickly and simply detect and evaluate the stress state of the glass panels during construction, which limits the ability to control construction quality.

[0007] These defects make it difficult for existing technologies to achieve rapid response and efficient management in curtain wall construction, resulting in long construction periods and difficult quality control. Summary of the invention

[0008] In view of the many problems existing in the above-mentioned prior arts, the present invention provides a curtain wall construction stage management method and system. The present invention combines digital technology with intelligent algorithms, and realizes efficient management and optimization of curtain wall construction through real-time monitoring and dynamic adjustment. Specifically, the present invention generates a design scheme through a generative adversarial network, and optimizes it using reinforcement learning and genetic algorithms. Subsequently, digital twin technology is used for dynamic simulation and adjustment of the construction process, and the sensor network provides real-time data support, and performs anomaly detection through deep learning. Finally, a causal analysis model is used to optimize the construction strategy. The present invention realizes comprehensive optimization from design to construction, and significantly improves the construction quality and efficiency.

[0009] A curtain wall construction phase management method comprises the following steps:

[0010] Acquire multi-source data of the construction site through drones, robots and IoT sensor networks, generate a comprehensive data set after data fusion and correction, and use the comprehensive data set to build and optimize a dynamic three-dimensional model to reflect the actual conditions of the construction site;

[0011] Based on the dynamic three-dimensional model, a preliminary design scheme is generated by using a generative adversarial network, and multi-objective optimization is performed by using a reinforcement learning algorithm and a genetic algorithm. Fuzzy logic reasoning is used to comprehensively evaluate the optimized design scheme, and a final design scheme is generated and updated in real time.

[0012] According to the final design plan, the digital twin technology is used to integrate the construction site data in real time to generate a construction twin model, and the construction strategy and resource allocation are dynamically adjusted through adaptive control algorithms and sensor networks;

[0013] Deploy a multi-scale sensor network to monitor the construction environment in real time, integrate monitoring data and detect anomalies through deep learning, combine causal analysis models to provide a basis for adjusting construction strategies, and achieve dynamic feedback control and adaptive optimization of construction strategies;

[0014] Define quality standards according to project requirements, develop and implement quality control plans, use sensor monitoring systems for quality assessment and automated acceptance, generate acceptance reports, and conduct long-term structural monitoring and maintenance plan development through health monitoring systems.

[0015] Preferably, the comprehensive data set generated after data fusion and correction of the multi-source data includes point cloud data, image data and environmental parameter data fused by a weighted averaging algorithm, and the weight coefficient in the weighted averaging algorithm is dynamically adjusted according to the reliability and signal-to-noise ratio of each data type.

[0016] Preferably, the generative adversarial network consists of a generator and a discriminator. The generator maximizes the misjudgment probability of the discriminator by generating potential design solutions, while the discriminator compares the generated design solutions with actual construction data to minimize the misjudgment probability, and generates a preliminary design solution through adversarial training optimization.

[0017] Preferably, the process of performing multi-objective optimization by using a reinforcement learning algorithm and a genetic algorithm comprises the following steps: defining structural stability, material cost and construction efficiency as optimization objectives, searching for optimal design parameters in the design space using a reinforcement learning algorithm, performing iterative optimization in the population using selection, crossover and mutation operations of a genetic algorithm, and generating an optimized design parameter combination to achieve the effect of multi-objective optimization.

[0018] Preferably, the specific calculation process of the multi-objective function is to calculate the contribution of each optimization target by linear combination, wherein the calculation formula of the multi-objective function is:

[0019] F (x) =α·S(x)+β·C(x)+γ·E(x)

[0020] Among them, F (x) is the multi-objective function value; S(x) is the structural stability function value; C(x) is the material cost function value; E(x) is the construction efficiency function value; α, β and γ are the weight coefficients of structural stability, material cost and construction efficiency respectively; and the sum of these weight coefficients is one.

[0021] Preferably, the fuzzy logic reasoning conducts a comprehensive evaluation of the optimization design scheme by establishing a fuzzy rule base, the fuzzy rule base conducts reasoning based on the fuzzy indicator matrix of at least two optimization design schemes, calculates the membership of each design indicator through a fuzzy membership function, and utilizes a fuzzy reasoning mechanism to comprehensively evaluate all candidate design schemes, and selects the scheme with the highest membership as the final design scheme.

[0022] Preferably, the digital twin technology constructs a construction twin model by real-time collection and integration of construction site data, simulates and predicts the construction process, and dynamically adjusts construction strategies and resource allocation through adaptive control algorithms to respond to site changes and optimize the construction process.

[0023] Preferably, the multi-scale sensor network includes temperature, humidity and vibration sensors deployed at the construction site, which monitor the construction environment in real time by integrating data from various sensors, and use deep learning models to extract features and detect anomalies from the monitoring data to identify potential risks in the construction process.

[0024] Preferably, the causal analysis model identifies the causal chain existing in the construction process by performing causal reasoning on the anomaly detection data, and generates construction strategy adjustment suggestions based on the causal analysis results. The construction strategy adjustment suggestions include: resource reallocation and process optimization.

[0025] A system for implementing the curtain wall construction phase management method, comprising:

[0026] A data acquisition module, configured to acquire multi-source data of the construction site through drones, robots and IoT sensor networks, and the module is used to perform data fusion and correction on the multi-source data to generate a comprehensive data set;

[0027] a modeling module configured to construct and optimize a dynamic three-dimensional model using the comprehensive data set to reflect the actual conditions of the construction site;

[0028] The design and optimization module is configured to generate a preliminary design scheme based on the dynamic three-dimensional model using a generative adversarial network, perform multi-objective optimization using a reinforcement learning algorithm and a genetic algorithm, perform a comprehensive evaluation of the optimized design scheme using fuzzy logic reasoning, and generate and update the final design scheme in real time;

[0029] A digital twin module is configured to integrate construction site data in real time according to the final design plan, generate a construction twin model, and dynamically adjust construction strategies and resource allocation through adaptive control algorithms and sensor networks;

[0030] The monitoring and feedback module is configured to deploy a multi-scale sensor network to monitor the construction environment in real time, integrate monitoring data and perform anomaly detection through deep learning, and provide a basis for adjusting the construction strategy in combination with a causal analysis model to achieve dynamic feedback control and adaptive optimization of the construction strategy;

[0031] The quality control and maintenance module is configured to define quality standards according to project requirements, develop and execute quality control plans, use sensor monitoring systems to perform quality assessment and automated acceptance, generate acceptance reports, and perform long-term structural monitoring and maintenance plan development through health monitoring systems.

[0032] Compared with the prior art, the advantages and beneficial effects of the present invention are:

[0033] The present invention uses generative adversarial network (GAN) technology to achieve rapid generation and optimization of design solutions, thereby enhancing the diversity and innovation of the design.

[0034] The present invention realizes multi-objective optimization of the design scheme by combining reinforcement learning and genetic algorithm, ensuring the best balance between structural stability, material cost and construction efficiency;

[0035] The present invention integrates construction site data in real time through digital twin technology, realizes dynamic simulation and strategy adjustment of the construction process, and improves the responsiveness and accuracy of the construction;

[0036] The present invention realizes real-time monitoring and anomaly detection of construction sites through the combination of multi-scale sensor networks and deep learning technology, thus enhancing construction safety.

[0037] The present invention identifies the cause-effect relationship chain in the construction process through a cause-effect analysis model, realizes adaptive optimization of the construction strategy, and improves the construction efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the process of the present invention;

[0039] Figure 2A schematic diagram of data collection and initial modeling in the present invention;

[0040] Figure 3 Schematic diagrams for generating design solutions, optimizations and updates in the present invention;

[0041] Figure 4 It is a schematic diagram of the construction process management and feedback in the present invention;

[0042] Figure 5 It is a structural block diagram of the system of the present invention. DETAILED DESCRIPTION

[0043] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0044] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0045] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0046] like Figure 1-Figure 2 As shown, a curtain wall construction phase management method includes the following steps:

[0047] Acquire multi-source data of the construction site through drones, robots and IoT sensor networks, generate a comprehensive data set after data fusion and correction, and use the comprehensive data set to build and optimize a dynamic three-dimensional model to reflect the actual conditions of the construction site;

[0048] Preferably, the comprehensive data set generated after data fusion and correction of the multi-source data includes point cloud data, image data and environmental parameter data fused by a weighted averaging algorithm, and the weight coefficient in the weighted averaging algorithm is dynamically adjusted according to the reliability and signal-to-noise ratio of each data type.

[0049] Drones and robots are used to collect comprehensive data on construction sites. Drones can provide a wide-area perspective from the air, collect panoramic image data of the building, and use the onboard laser radar system to obtain high-precision point cloud data. Robots perform close-range data collection tasks on the ground, collecting more detailed structural information and environmental parameters such as temperature, humidity, and wind speed. These data sources are different but complementary, allowing the entire construction site to be perceived in all directions.

[0050] The acquired multi-source data is fused and corrected to generate a comprehensive data set. At this stage, data fusion is a key step. The principle is to merge data from different sources into a unified comprehensive data set through a weighted average algorithm. The core of the weighted average algorithm in this application is to assign different weights to each data type according to its reliability and signal-to-noise ratio. This means that higher quality and more reliable data account for a larger proportion in the fusion process, thereby improving the accuracy and consistency of the comprehensive data set. For example, in bad weather conditions, the image data of the drone may be affected by insufficient light. At this time, the algorithm will automatically reduce the weight of the image data and increase the weight of the point cloud data and environmental parameter data to ensure the reliability of the generated data set.

[0051] The integrated data set after fusion is used to build and optimize the dynamic 3D model. The dynamic 3D model is a digital representation of the construction site. Its role is to accurately reflect the current situation of the construction site and provide a data basis for subsequent design and construction. By optimizing the dynamic 3D model, potential problems on the site can be identified and pre-adjusted. Model optimization includes eliminating redundant information, correcting errors, and updating the model based on the latest data to reflect changes on the site. The optimized dynamic 3D model can not only be used to visualize the construction progress, but also as a tool to simulate and verify the construction plan.

[0052] In this process, real-time dynamic updates and accurate modeling provide strong support for the management of the entire construction phase. This method can effectively improve the execution efficiency of the construction plan and reduce construction errors and rework rates caused by information asymmetry. For example, when installing curtain walls, the dynamic 3D model can provide accurate positioning information of curtain wall units in real time to ensure the accuracy and safety of installation.

[0053] In summary, using drones, robots and IoT sensor networks to acquire and fuse multi-source data, generate and optimize dynamic 3D models is an important technical means in modern curtain wall construction management. Through efficient data collection and processing, it provides real-time and accurate information support for the construction site, significantly improving the accuracy and efficiency of construction.

[0054] like Figure 3As shown, based on the dynamic three-dimensional model, a generative adversarial network is used to generate a preliminary design scheme, and a reinforcement learning algorithm and a genetic algorithm are used to perform multi-objective optimization, and fuzzy logic reasoning is used to perform a comprehensive evaluation of the optimized design scheme, and the final design scheme is generated and updated in real time;

[0055] Improve the adaptability and efficiency of the design through advanced algorithms and evaluation methods. This process can be achieved through the following technical means: Generative Adversarial Networks (GAN), reinforcement learning algorithms, genetic algorithms, and fuzzy logic reasoning.

[0056] Generative adversarial networks (GANs) are used to generate preliminary design solutions. GANs consist of a generator and a discriminator. Through adversarial training between the two, the generator can produce solutions that are very close to the real design data. Specifically, the generator is responsible for generating new design solutions, while the discriminator determines whether these generated solutions match the existing design standards. As the training progresses, the generator gradually generates design solutions that are more in line with actual needs. The effect of this process is to increase the diversity and innovation of design solutions while maintaining a high degree of match with actual needs. For example, in curtain wall design, GAN can generate curtain wall appearance solutions with a variety of design styles, which not only meet aesthetic requirements, but also meet engineering and technical requirements.

[0057] Multi-objective optimization is performed through reinforcement learning algorithms and genetic algorithms. In this application, the reinforcement learning algorithm simulates various possibilities on the construction site and dynamically adjusts the design parameters to optimize construction efficiency and material usage. This algorithm uses a reward mechanism to evaluate the pros and cons of different design schemes and select a better design path. At the same time, the genetic algorithm uses selection, crossover, and mutation operations to perform population iterative optimization of the design scheme by simulating the biological evolution process. This multi-objective optimization not only considers structural stability and material costs, but also includes factors such as construction efficiency. For example, when optimizing curtain wall materials, genetic algorithms can help find a combination with the least material usage but the best strength and aesthetics.

[0058] The optimized design scheme needs to be comprehensively evaluated, and fuzzy logic reasoning is introduced here. The fuzzy logic system can handle the uncertainty and ambiguity in the design scheme, and comprehensively evaluate multiple indicators of different design schemes through the fuzzy reasoning mechanism. This process will generate a fuzzy evaluation matrix, quantify the indicators of each design scheme according to the membership function, and select the best design scheme as the final scheme. This evaluation mechanism ensures that the design scheme is not only optimal from a technical perspective, but also meets the requirements of actual construction conditions.

[0059] The final design generated and updated in real time by these technical means greatly improves the flexibility and accuracy of the design process. It can dynamically respond to actual changes on the construction site, provide real-time optimized design solutions, and ensure that every step in the construction process is carried out in the best state. For example, when the conditions on the construction site change, the system can quickly adjust the design plan and regenerate the optimal plan to adapt to the new site conditions, thereby reducing construction delays and material waste.

[0060] In short, by combining the multi-objective optimization of generative adversarial networks, reinforcement learning and genetic algorithms, and the comprehensive evaluation of fuzzy logic reasoning, the entire design optimization process has been made intelligent and efficient. This not only makes the management of the curtain wall construction phase more accurate and efficient, but also provides construction personnel with a flexible and real-time design update system, greatly improving the quality and efficiency of construction.

[0061] Preferably, the generative adversarial network consists of a generator and a discriminator. The generator maximizes the misjudgment probability of the discriminator by generating potential design solutions, while the discriminator compares the generated design solutions with actual construction data to minimize the misjudgment probability, and generates a preliminary design solution through adversarial training optimization.

[0062] In the present invention, the application of generative adversarial networks (GANs) can significantly improve the generation and optimization process of design solutions. GANs consist of two main components: a generator and a discriminator, which work together through adversarial training to generate high-quality design solutions.

[0063] The generator is responsible for creating potential design solutions, with the goal of generating designs that are extremely similar to real construction data. These solutions should be as close to actual needs as possible in terms of form, function, and construction feasibility. The generator generates potential curtain wall design solutions through a series of neural network operations, starting with noisy data or random inputs, and gradually adjusting the output. In this process, the generator constantly tries to "cheat" the discriminator so that it cannot easily distinguish the generated design from the real data. The success of the generator lies in maximizing the probability of misjudgment of the discriminator, that is, the design solution that the generator can generate is mistakenly judged by the discriminator as real data.

[0064] The role of the discriminator is to distinguish between real design solutions and those generated by the generator. It analyzes the characteristics of the design solutions to determine whether these solutions meet the standards of actual construction. The discriminator uses supervised learning techniques and is trained on known real data sets to improve its discrimination ability. The goal of the discriminator is to minimize the probability of misjudgment and ensure that it can correctly identify the difference between the generated solution and the real data.

[0065] Through repeated adversarial training, the generator and the discriminator evolve each other. In each training cycle, the generator tries to generate a design that is closer to the real data, while the discriminator adjusts its judgment criteria through feedback to enhance its ability to distinguish. Ultimately, this adversarial training mechanism not only improves the design generation ability of the generator, but also improves the discrimination ability of the discriminator, optimizing the quality of the design.

[0066] The application of GAN is of great significance in curtain wall construction management. By generating adversarial networks, diverse design solutions can be quickly generated, which can effectively integrate innovation and practicality. For example, GAN can generate a variety of curtain wall appearance and structural designs in a short period of time for engineers to choose and further optimize. This method not only saves design time, but also fully considers the actual construction conditions and restrictions in the design stage, ensuring that the generated design solutions are feasible and innovative.

[0067] In summary, the generative adversarial network of the present invention can effectively generate and optimize preliminary design solutions through adversarial training of the generator and the discriminator. This process not only improves the diversity and quality of design solutions, but also provides a more innovative and operable design basis for the construction process.

[0068] Preferably, the process of performing multi-objective optimization by using a reinforcement learning algorithm and a genetic algorithm comprises the following steps: defining structural stability, material cost and construction efficiency as optimization objectives, searching for optimal design parameters in the design space using a reinforcement learning algorithm, performing iterative optimization in the population using selection, crossover and mutation operations of a genetic algorithm, and generating an optimized design parameter combination to achieve the effect of multi-objective optimization.

[0069] In the present invention, the multi-objective optimization process optimizes the design parameters by combining the reinforcement learning algorithm and the genetic algorithm to improve the practicality and efficiency of the design solution. This process involves multiple key steps, including goal definition, reinforcement learning search and genetic algorithm optimization.

[0070] In multi-objective optimization, clear optimization goals are the basis for achieving the design process. In curtain wall design, common optimization goals include structural stability, material cost and construction efficiency. Structural stability ensures the design requirements in terms of safety and durability, material cost controls the budget to improve economic benefits, and construction efficiency is related to the project schedule and resource utilization. Taking these three as optimization goals, all important aspects of the design solution can be fully balanced in the design stage.

[0071] Reinforcement learning algorithms are used to search for optimal design parameters in the design space. Reinforcement learning is a machine learning method that finds the optimal strategy through a trial-and-error process. In curtain wall design, reinforcement learning algorithms can dynamically explore the impact of different combinations of design parameters, and identify and select design parameters that meet the optimization goals through a reward mechanism. For example, when optimizing material usage, reinforcement learning algorithms can identify which design parameter combinations can reduce material usage while ensuring strength, thereby reducing costs. The advantage of reinforcement learning is that it is highly adaptable and can continuously adjust and tend to a better parameter configuration during the design process.

[0072] Genetic algorithms play a key role in multi-objective optimization. They imitate the natural selection process and perform iterative optimization in the population through selection, crossover and mutation operations. The selection operation refers to selecting excellent design solutions for reproduction based on the fitness of individuals; the crossover operation generates a new design solution by combining the characteristics of two design solutions; and the mutation operation introduces random changes to explore new design spaces. The iterative process of genetic algorithms can quickly optimize the combination of design parameters and find a set of design solutions that meet multi-objective requirements. For example, when optimizing the appearance and structural design of curtain walls, genetic algorithms can quickly screen out a combination of solutions that are both beautiful and practical from many possible design solutions.

[0073] Through the combination of reinforcement learning and genetic algorithms, the multi-objective optimization process can effectively generate optimized design parameter combinations. Ultimately, this combined approach not only improves the quality and practicality of the design solution, but also provides more accurate and efficient design solution support for the construction phase. Through this optimization, curtain wall construction can achieve reasonable allocation of resources, effective cost control, and significant improvement in construction efficiency, thereby improving the overall quality and economic benefits of the project.

[0074] In summary, the combination of reinforcement learning algorithm and genetic algorithm in the multi-objective optimization of curtain wall construction stage shows its powerful computing power and adaptive characteristics, making the design scheme more innovative and practical. This method provides an efficient solution for multi-dimensional optimization in complex construction projects.

[0075] Preferably, the specific calculation process of the multi-objective function is to calculate the contribution of each optimization target by linear combination, wherein the calculation formula of the multi-objective function is:

[0076] F (x) =α·S(x)+β·C(x)+γ·E(x)

[0077] Among them, F (x)is the multi-objective function value; S(x) is the structural stability function value; C(x) is the material cost function value; E(x) is the construction efficiency function value; α, β and γ are the weight coefficients of structural stability, material cost and construction efficiency respectively; and the sum of these weight coefficients is one.

[0078] In the present invention, the multi-objective function is used to evaluate and optimize multiple aspects of the design solution to achieve the optimization of comprehensive performance. The function measures the contribution of each optimization target in a linear combination manner, thereby providing a basis for design decisions.

[0079] The multi-objective function linearly combines each objective through weight coefficients to balance the influence of each objective in the overall optimization. The setting of weight coefficients can reflect the priorities of different stages of the project. For example, in the initial design, more attention may be paid to the control of material costs, while in the construction stage, more attention may be paid to structural stability and construction efficiency. By adjusting the weight coefficients, it can be flexibly adapted to the specific needs and constraints of the project.

[0080] A significant effect of using multi-objective functions for optimization is that it can consider multiple interrelated design goals at the same time, rather than focusing on a single goal. In this way, a balance point can be found in the plan to achieve the coordinated development of various goals. Especially in complex construction environments, this multi-objective optimization strategy can effectively reduce costs, improve efficiency, and ensure the safety and stability of the structure.

[0081] In practical applications, the optimization of multi-objective functions can be achieved through the following steps:

[0082] (1) Define each objective function: First, clarify the definition and calculation method of each objective. For example, the structural stability function S(x) can be evaluated based on the finite element analysis model, the material cost function C(x) can be calculated by estimating the material usage and price, and the construction efficiency function E(x) can be measured by time scheduling and resource utilization efficiency.

[0083] (2) Setting weight coefficients: Set weight coefficients α, β, and γ according to project requirements. For example, α = 0.4, β = 0.3, and γ = 0.3 can be selected to indicate priority attention to structural stability.

[0084] (3) Calculate the multi-objective function value: Use the given design parameters to calculate the multi-objective function value F of each design solution. (k) , thereby providing a comprehensive evaluation of different design options.

[0085] (4) Select and optimize design schemes: Based on the calculated multi-objective function values, select the design scheme with the highest function value, and use the algorithm to iteratively optimize to improve the overall performance of the design scheme.

[0086] Through the optimization of multi-objective functions, the management of the curtain wall construction phase can better adapt to different construction conditions and requirements, achieve comprehensive optimization of the design scheme, and provide guarantees for the successful implementation of the project. This method not only improves the scientificity and rationality of the design, but also lays the foundation for the efficient advancement of the construction process.

[0087] Preferably, the fuzzy logic reasoning conducts a comprehensive evaluation of the optimization design scheme by establishing a fuzzy rule base, the fuzzy rule base conducts reasoning based on the fuzzy indicator matrix of at least two optimization design schemes, calculates the membership of each design indicator through a fuzzy membership function, and utilizes a fuzzy reasoning mechanism to comprehensively evaluate all candidate design schemes, and selects the scheme with the highest membership as the final design scheme.

[0088] In the present invention, fuzzy logic reasoning is used to comprehensively evaluate the optimization design scheme to cope with the challenges of uncertainty and ambiguity in design decision-making. By establishing a fuzzy rule base, the system can select from multiple candidate design schemes to ensure that the final selected scheme has the best comprehensive performance.

[0089] The core of fuzzy logic reasoning is to use fuzzy set theory to deal with uncertainty and fuzzy information. In the process of optimizing the evaluation of design solutions, it is necessary to first establish a fuzzy rule base, which contains a large number of rules related to design indicators. Each design solution can be described by its corresponding fuzzy indicator matrix, which may include structural strength, aesthetics, construction feasibility, etc. The fuzzy membership function is used to calculate the membership of each design indicator, that is, to what extent each indicator meets a specific standard or expectation.

[0090] The fuzzy reasoning mechanism comprehensively evaluates these memberships through the fuzzy rule base and generates an overall membership value for each candidate design solution. The result of the comprehensive evaluation assigns a membership value to each solution, which reflects the degree to which the solution meets all design goals. Ultimately, the system selects the solution with the highest membership as the final design solution.

[0091] Through fuzzy logic reasoning, curtain wall construction phase management can effectively handle the complexity and ambiguity in the design scheme and provide more flexible and reliable decision support. This method can help the design team quickly identify and select the scheme that best meets the project needs when faced with multiple design options, thereby improving design quality and efficiency.

[0092] In practical applications, fuzzy logic reasoning can be achieved through the following steps:

[0093] (1) Define fuzzy indicators: Determine the various indicators that need to be evaluated in the design scheme, such as structural strength, construction cost, and appearance aesthetics, and establish fuzzy sets for each indicator.

[0094] (2) Establish a fuzzy rule base: Based on the design standards and empirical knowledge, a fuzzy rule base is established. For example, a rule can be defined as “If the structural strength is high and the construction cost is low, then the design is good.”

[0095] (3) Calculate membership: Use fuzzy membership function to calculate the membership of each design indicator. For example, use trapezoidal or triangular membership function to represent the membership of a certain design scheme in “low cost”.

[0096] (4) Reasoning and evaluation: Through the fuzzy reasoning mechanism, the membership of each candidate design scheme is comprehensively calculated to obtain an overall membership value.

[0097] (5) Select the final solution: Compare the overall membership values ​​of all candidate solutions and select the solution with the highest membership as the final design solution.

[0098] Through these steps, fuzzy logic reasoning provides a powerful tool for the curtain wall construction stage, helping the design team to make optimal decisions under uncertain and complex conditions. This method not only improves the scientificity and rationality of the design plan, but also enhances the responsiveness and flexibility of the design stage.

[0099] like Figure 4 As shown, according to the final design scheme, the digital twin technology is used to integrate the construction site data in real time to generate a construction twin model, and the construction strategy and resource allocation are dynamically adjusted through adaptive control algorithms and sensor networks;

[0100] Preferably, the digital twin technology constructs a construction twin model by real-time collection and integration of construction site data, simulates and predicts the construction process, and dynamically adjusts construction strategies and resource allocation through adaptive control algorithms to respond to site changes and optimize the construction process.

[0101] In this invention, the application of digital twin technology has brought the monitoring and optimization of the construction process to a new level. By integrating construction site data in real time and generating a construction twin model, the construction team can accurately simulate and predict construction progress in a virtual environment, thereby making more accurate and dynamic construction decisions.

[0102] Digital twin technology is a cutting-edge technology that integrates physical and digital space. It collects various data on the construction site in real time (such as structural deformation, environmental parameters, construction progress, etc.) to build a virtual model consistent with the actual construction environment, namely the construction twin model. This model not only faithfully reflects the current construction status, but also simulates future construction scenarios by integrating historical data and prediction algorithms.

[0103] The construction twin model provides a comprehensive view for construction management. Combined with sensor networks and adaptive control algorithms, construction strategies and resource allocation can be adjusted in real time. The adaptive control algorithm dynamically optimizes the construction process and adjusts resource input by analyzing changes in field data reflected in the construction twin model. For example, when sensors detect abnormal structural stress in a certain area, the digital twin system can quickly assess the potential risks in the area and automatically recommend adjusting the construction sequence or adding support structures.

[0104] The application of digital twin technology has significantly improved the efficiency and accuracy of construction management. Through real-time monitoring and dynamic adjustment, the construction team can quickly respond to changes on site, optimize resource use, and reduce construction delays and material waste. In addition, the construction twin model also supports the simulation and optimization of complex construction processes, helping the construction team identify and solve potential problems before actual execution, thereby improving construction quality and safety.

[0105] In practical applications, digital twin technology realizes its functions through the following steps:

[0106] (1) Data collection and integration: A sensor network is deployed at the construction site to collect various data in real time. These data include but are not limited to structural monitoring data, environmental conditions, equipment status, etc., and are integrated into the construction twin model.

[0107] (2) Construction twin model construction: Use the collected data to build a construction twin model. This model not only reflects the current status of the construction site, but also combines historical data and predictive analysis to generate future construction scenarios.

[0108] (3) Simulation and prediction: Simulation is performed in the construction twin model to predict problems that may arise in future construction, such as material shortages and structural instability.

[0109] (4) Adaptive control and optimization: Adaptive control algorithms are used to analyze data changes in the construction twin model and adjust construction strategies in real time. For example, if the model predicts that construction will be delayed at a certain stage, resource allocation can be adjusted to ensure construction progress.

[0110] (5) Feedback and adjustment: Monitor construction progress in real time and input feedback information into the construction twin model for re-evaluation and adjustment to ensure dynamic optimization of the construction process.

[0111] Through digital twin technology, the management of curtain wall construction phase has achieved a transformation from passive response to active control. This not only improves the efficiency and accuracy of construction, but also provides an intelligent solution that can flexibly respond to complex changes on site.

[0112] Deploy a multi-scale sensor network to monitor the construction environment in real time, integrate monitoring data and detect anomalies through deep learning, combine causal analysis models to provide a basis for adjusting construction strategies, and achieve dynamic feedback control and adaptive optimization of construction strategies;

[0113] In this invention, the deployment of multi-scale sensor networks provides a comprehensive solution for real-time monitoring of construction sites. By integrating monitoring data and using deep learning for anomaly detection, the construction team can promptly discover and respond to potential problems, and combine causal analysis models to achieve adaptive optimization of construction strategies.

[0114] A multiscale sensor network is a system that integrates different types of sensors to comprehensively monitor various parameters of a construction site. These sensors include temperature sensors, humidity sensors, vibration sensors, pressure sensors, etc., which can collect multi-dimensional data of the construction environment at different scales. After integration, these data form a comprehensive monitoring data set, providing a detailed information basis for construction management.

[0115] Deep learning algorithms play an important role in this process, analyzing integrated monitoring data for anomaly detection. Deep learning algorithms can automatically identify complex patterns and features in data and quickly discover abnormal behaviors or states. For example, when a sensor detects abnormal vibration, the deep learning model can combine historical data and current data analysis to determine whether this is a potential structural problem.

[0116] Combined with the causal analysis model, the construction management system can further analyze the cause-effect relationship in the anomaly detection data. The causal analysis model helps the construction team understand the root cause of the anomaly by identifying the cause-effect chain between variables. For example, by analyzing the sensor data, the system may find that the increase in pressure in a certain stage of construction is caused by improper material selection in the early stage, thus providing a scientific basis for the adjustment of subsequent construction strategies.

[0117] The combination of multi-scale sensor networks and deep learning makes monitoring of construction sites more accurate and real-time. Through anomaly detection and causal analysis, construction teams can achieve dynamic feedback control and strategy adjustment. In this way, construction management can not only respond quickly to environmental changes, but also implement intelligent optimization and adjustment during the construction process, reduce construction risks, and improve construction quality.

[0118] In practical applications, multiscale sensor networks achieve their functions through the following steps:

[0119] (1) Sensor deployment: Various types of sensors are deployed in key areas of the construction site to achieve comprehensive monitoring of parameters such as temperature, humidity, vibration, and pressure.

[0120] (2) Data integration and analysis: The sensor data collected in real time is summarized through the data integration system to form a monitoring data set, and then analyzed using deep learning algorithms to detect abnormal situations.

[0121] (3) Anomaly detection: Deep learning models can quickly detect abnormal data points by learning and identifying patterns and trends in the data. For example, detecting a sudden increase in vibration values ​​at a certain location may indicate a structural problem.

[0122] (4) Causal analysis: Use causal analysis models to conduct in-depth analysis of anomaly detection results, identify possible causes and consequences, and provide a basis for adjusting construction strategies.

[0123] (5) Feedback control and optimization: Based on the results of causal analysis, construction strategies are dynamically adjusted, such as optimizing resource allocation and adjusting the construction sequence, to improve construction safety and efficiency.

[0124] Through these steps, the application of multi-scale sensor networks and deep learning in the curtain wall construction stage has greatly improved the intelligent level of construction management. It provides strong support for real-time monitoring, rapid response and strategy optimization during the construction process, making the construction process more flexible and safer.

[0125] Preferably, the multi-scale sensor network includes temperature, humidity and vibration sensors deployed at the construction site, which monitor the construction environment in real time by integrating data from various sensors, and use deep learning models to extract features and detect anomalies from the monitoring data to identify potential risks in the construction process.

[0126] In the present invention, the application of a multi-scale sensor network provides an effective solution for real-time monitoring of the construction environment. The network includes temperature, humidity and vibration sensors deployed at the construction site, which can fully and meticulously perceive changes in the construction environment and provide key data support for construction management.

[0127] Multiscale sensor networks capture dynamic changes in environmental parameters by deploying multiple sensors at different locations and heights on the construction site. Temperature sensors are used to monitor temperature fluctuations in the construction area, which have a significant impact on material properties and construction processes in different seasons and weather conditions. Humidity sensors can detect the moisture content in the air to help assess the adaptability of construction materials and the suitability of the construction environment. Vibration sensors are used to monitor the mechanical vibrations generated during the construction process and the impact of the external environment on the structure to ensure construction safety and structural stability.

[0128] Integrating these sensor data can form a comprehensive construction environment monitoring dataset. The deep learning model can automatically identify complex patterns and potential anomalies in the data by performing feature extraction and anomaly detection on this dataset. Feature extraction involves converting raw data into a form that is easier to analyze and understand. These features can be used to describe the current state of the construction environment. Anomaly detection refers to identifying abnormal events or conditions that are different from the expected data pattern. These anomalies may indicate potential construction risks.

[0129] Through the combination of multi-scale sensor networks and deep learning, construction management can understand the specific conditions of the site in real time, detect and respond to potential risks in advance. This approach improves the safety and efficiency of construction and reduces the risk of construction delays and accidents caused by environmental factors. In addition, through in-depth analysis of monitoring data, the construction team can optimize material selection and construction methods in the design stage, thereby improving the overall project quality.

[0130] In practical applications, multiscale sensor networks achieve their functions through the following steps:

[0131] (1) Sensor deployment and data collection: Temperature, humidity, and vibration sensors are deployed at key points and risk areas on the construction site. The sensors collect environmental parameters in real time, forming a multi-dimensional data stream.

[0132] (2) Data integration and feature extraction: Data from different sensors are aggregated through the data integration platform. The deep learning model extracts features from the integrated data and identifies important indicators that reflect the environmental status.

[0133] (3) Anomaly detection: The deep learning model is trained using historical data to establish a normal operating pattern. It identifies abnormal situations that deviate from the normal pattern in real-time data. For example, an abnormal increase in vibration values ​​at a certain location may indicate a construction equipment failure or structural damage.

[0134] (4) Risk assessment and response: Based on the detected abnormalities, the construction management system automatically generates a risk assessment report and recommends corresponding response measures, such as adjusting the construction plan or strengthening on-site monitoring.

[0135] Through the application of multi-scale sensor networks and deep learning technology, curtain wall construction phase management can achieve more refined and intelligent on-site monitoring and risk management. This not only improves the safety and reliability of construction, but also provides important technical support for efficient construction decision-making.

[0136] Preferably, the causal analysis model identifies the causal chain existing in the construction process by performing causal reasoning on the anomaly detection data, and generates construction strategy adjustment suggestions based on the causal analysis results. The construction strategy adjustment suggestions include: resource reallocation and process optimization.

[0137] In the present invention, the causal analysis model is a tool for identifying and understanding the complex causal relationships in the construction process. This model reveals the causal relationship chain in the construction process by conducting in-depth analysis of anomaly detection data, and provides construction strategy adjustment suggestions to achieve efficient resource allocation and optimization of construction processes.

[0138] Causal analysis models are based on the principle of causal reasoning and are designed to identify causal relationships between events from data, rather than just correlations. For construction management, when a monitoring system detects an anomaly, a causal analysis model can help identify the root cause of the anomaly and analyze its potential impact on the construction process. This analysis usually involves building a causal graph or causal model to connect different construction events and variables in the form of a causal chain.

[0139] By performing causal reasoning on anomaly detection data, the model can identify which factors lead to the occurrence of a certain anomaly. For example, abnormal vibration during construction may be caused by the failure or improper operation of construction equipment. Causal analysis can further identify whether this anomaly is caused by improper regular maintenance of equipment or changes in the on-site environment.

[0140] Based on the results of the causal analysis, the model can generate construction strategy adjustment suggestions. These suggestions usually include resource reallocation and process optimization. Resource reallocation refers to adjusting the use of manpower, equipment and materials according to actual needs to maximize the efficiency of resource utilization. Process optimization refers to making adjustments in the construction process to improve construction efficiency and quality.

[0141] Through the application of causal analysis models, construction management can more accurately identify the root causes of problems and take corresponding measures, which significantly improves the efficiency and accuracy of problem solving on the construction site. In addition, the data support provided by causal analysis makes the adjustment of construction strategies more scientific and reasonable, reducing unnecessary waste of resources and construction delays.

[0142] In practical applications, the causal analysis model implements its functions through the following steps:

[0143] (1) Abnormal data collection and collation: Obtain abnormal detection data at the construction site from a multi-scale sensor network. This data may include information such as equipment failures and changes in environmental conditions.

[0144] (2) Causal reasoning: Use causal reasoning algorithms to analyze abnormal data, construct causal graphs, and identify causal relationships between abnormal events. For example, analyzing abnormal vibration data may reveal the causal chain between it and equipment wear.

[0145] (3) Generate adjustment suggestions: Generate construction strategy adjustment suggestions based on the causal analysis results. For vibration abnormalities caused by equipment wear, the suggestions may include adjusting the equipment maintenance plan or replacing equipment parts.

[0146] (4) Implementation strategy adjustment: Apply the adjustment suggestions to the actual construction process to reallocate resources and optimize processes to improve construction efficiency and safety.

[0147] Through these steps, the causal analysis model in this invention provides a powerful analysis and decision support tool for the construction team. This method not only improves the responsiveness of the construction site, but also provides technical support for dynamic management in complex construction projects.

[0148] like Figure 4 As shown, quality standards are defined according to project requirements, quality control plans are developed and implemented, sensor monitoring systems are used for quality assessment and automated acceptance, acceptance reports are generated, and long-term structural monitoring and maintenance plan development are performed through health monitoring systems.

[0149] In this invention, quality control is a key step to ensure that the construction project meets the expected standards and specifications. The process involves all-round management from the definition of quality standards to implementation monitoring, and finally to automated acceptance and long-term structural health monitoring.

[0150] Quality control first defines clear quality standards based on project requirements. These standards cover quality requirements in all aspects of the construction process, including material quality, construction technology, structural strength, etc. Clearly defined quality standards are the basis for the development of quality control plans. They ensure that construction activities can be carried out in accordance with established quality requirements and provide a basis for evaluation and acceptance.

[0151] The next step of quality control is to formulate a quality control plan, which describes in detail how to implement the various steps of quality monitoring, evaluation and acceptance. It usually includes specific contents such as inspection frequency, inspection points, evaluation indicators and acceptance criteria. The implementation of the quality control plan relies on advanced sensor monitoring systems to ensure that the construction quality meets the standards by monitoring various parameters in the construction process in real time.

[0152] The sensor monitoring system can install sensors at key locations on the construction site to collect real-time environmental and structural parameters such as temperature, humidity, and vibration. These data are used to conduct quality assessments and detect construction activities that do not meet quality standards. The sensor monitoring system not only provides accurate data support, but also enables acceptance in an automated manner, that is, automatically judging whether the construction meets the acceptance standards based on the data fed back by the sensors.

[0153] After construction is completed, the structure is monitored and maintained over the long term through a health monitoring system. The health monitoring system continues to collect dynamic data of the structure and monitor its performance changes during long-term use. These data are not only used to regularly assess the health of the structure, but also provide a basis for the formulation of maintenance plans to ensure the safety and stability of the structure throughout its life cycle.

[0154] Through clear quality standards and effective quality control plans, construction projects can meet the expected quality requirements and quickly confirm quality compliance through automated acceptance after construction is completed. The application of sensor monitoring systems and health monitoring systems improves the accuracy and efficiency of construction quality assessment and reduces human intervention and errors. In addition, long-term structural monitoring provides important data support for subsequent maintenance and operations, helping to extend the service life of the structure and reduce maintenance costs.

[0155] In practical applications, the implementation of quality control can be carried out through the following steps:

[0156] (1) Define quality standards: Based on the project requirements, define the quality standards that must be achieved during the construction process. This includes selecting appropriate material specifications, construction process requirements, and the final structural performance standards.

[0157] (2) Develop a quality control plan: Prepare a detailed quality control plan, including inspection frequency, quality assessment indicators, and acceptance criteria.

[0158] (3) Sensor monitoring and quality assessment: A sensor monitoring system is installed at the construction site to collect real-time environmental and structural parameters during the construction process. By analyzing this data, the construction quality can be automatically assessed.

[0159] (4) Automated acceptance and report generation: Based on the data from the sensor system, it automatically determines whether the construction quality meets the acceptance standards and generates an acceptance report.

[0160] (5) Health monitoring and maintenance plan: Use a health monitoring system to conduct long-term structural monitoring, collect data to assess the health of the structure, and develop appropriate maintenance plans.

[0161] Through these steps, the quality control of the curtain wall construction phase can be achieved accurately, automatically and with long-term quality assurance. This method not only improves the transparency and controllability of the construction process, but also ensures the high-quality delivery and long-term safety of the project.

[0162] like Figure 5 As shown, a system for implementing the curtain wall construction phase management method comprises:

[0163] The data acquisition module is configured to acquire multi-source data of the construction site through drones, robots and IoT sensor networks. The module is used to fuse and correct the multi-source data to generate a comprehensive data set; the data acquisition module is the basic part of the system and is responsible for acquiring multi-source data of the construction site through drones, robots and IoT sensor networks. These data include image data, point cloud data, environmental parameters, etc. Drones provide data from an aerial perspective, robots are used to capture ground details, and IoT sensors monitor environmental parameters in real time. After data acquisition, the module fuses and corrects these multi-source data to eliminate noise and redundant information to generate a comprehensive and accurate comprehensive data set.

[0164] The modeling module is configured to use the comprehensive data set to build and optimize a dynamic three-dimensional model to reflect the actual conditions of the construction site; the modeling module uses the comprehensive data set to build a dynamic three-dimensional model that truly reflects the actual conditions of the construction site. By using advanced modeling technology, the module is able to capture the complexity of the building structure and eliminate redundancy through optimization algorithms to improve the accuracy and computational efficiency of the model. The dynamic three-dimensional model provides reliable basic data support for subsequent design and optimization.

[0165] The design and optimization module is configured to generate a preliminary design scheme based on the dynamic three-dimensional model using a generative adversarial network, perform multi-objective optimization through a reinforcement learning algorithm and a genetic algorithm, and use fuzzy logic reasoning to comprehensively evaluate the optimized design scheme, generate and update the final design scheme in real time; in the design and optimization module, based on the dynamic three-dimensional model, the system uses a generative adversarial network (GAN) to generate a preliminary design scheme. GAN quickly generates innovative and practical design schemes through adversarial training of generators and discriminators. Then, the design scheme is optimized for multiple objectives through reinforcement learning algorithms and genetic algorithms, which can balance multiple objectives such as structural stability, material cost and construction efficiency. Fuzzy logic reasoning further comprehensively evaluates the optimized design scheme to ensure the rationality and superiority of the design scheme in multiple indicators, and finally generates and updates the design scheme in real time.

[0166] The digital twin module is configured to integrate the construction site data in real time according to the final design plan, generate a construction twin model, and dynamically adjust the construction strategy and resource allocation through adaptive control algorithms and sensor networks; the digital twin module integrates the construction site data in real time according to the final design plan to generate a construction twin model. This model is a virtual mapping of the physical construction site and can reflect the dynamic changes of the site through real-time updates. Through the adaptive control algorithm, the module dynamically adjusts the construction strategy and resource allocation according to the data provided by the sensor network to ensure the efficiency and flexibility of the construction process.

[0167] The monitoring and feedback module is configured to deploy a multi-scale sensor network to monitor the construction environment in real time, integrate monitoring data and perform anomaly detection through deep learning, and provide a basis for adjusting the construction strategy in combination with the causal analysis model to achieve dynamic feedback control and adaptive optimization of construction strategies; The monitoring and feedback module deploys a multi-scale sensor network to monitor the construction environment in real time. By integrating monitoring data, the module uses a deep learning model to detect anomalies and identify potential construction risks. Combined with the causal analysis model, the module performs causal relationship analysis on the anomaly detection data, provides a basis for adjusting the construction strategy, and helps construction management achieve dynamic feedback control and adaptive optimization of construction strategies.

[0168] The quality control and maintenance module is configured to define quality standards according to project requirements, develop and execute quality control plans, use sensor monitoring systems for quality assessment and automated acceptance, generate acceptance reports, and perform long-term structural monitoring and maintenance plan formulation through the health monitoring system. The quality control and maintenance module defines quality standards according to project requirements, develops and executes detailed quality control plans. Through the sensor monitoring system, the module performs quality assessment and automated acceptance to ensure that the construction quality meets the set standards. After the acceptance is completed, the module generates a detailed acceptance report. In addition, long-term structural monitoring is performed through the health monitoring system, real-time collection of structural health data, and maintenance plans are formulated to ensure the safety and stability of the structure throughout its life cycle.

[0169] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0170] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A curtain wall construction phase management method, It is characterized in that The following steps are involved: Acquire multi-source data of the construction site through drones, robots and IoT sensor networks, generate a comprehensive data set after data fusion and correction, and use the comprehensive data set to build and optimize a dynamic three-dimensional model to reflect the actual conditions of the construction site; Based on the dynamic three-dimensional model, a preliminary design scheme is generated by using a generative adversarial network, and multi-objective optimization is performed by using a reinforcement learning algorithm and a genetic algorithm. Fuzzy logic reasoning is used to comprehensively evaluate the optimized design scheme, and a final design scheme is generated and updated in real time. According to the final design plan, the digital twin technology is used to integrate the construction site data in real time to generate a construction twin model, and the construction strategy and resource allocation are dynamically adjusted through adaptive control algorithms and sensor networks; Deploy a multi-scale sensor network to monitor the construction environment in real time, integrate monitoring data and detect anomalies through deep learning, combine causal analysis models to provide a basis for adjusting construction strategies, and achieve dynamic feedback control and adaptive optimization of construction strategies; Define quality standards according to project requirements, develop and implement quality control plans, use sensor monitoring systems for quality assessment and automated acceptance, generate acceptance reports, and conduct long-term structural monitoring and maintenance plan development through health monitoring systems.

2. The method according to claim 1, It is characterized in that The comprehensive data set generated after data fusion and correction of the multi-source data includes point cloud data, image data and environmental parameter data fused by a weighted average algorithm, and the weight coefficient in the weighted average algorithm is dynamically adjusted according to the reliability and signal-to-noise ratio of each data type.

3. The method according to claim 1, It is characterized in that The generative adversarial network consists of a generator and a discriminator. The generator maximizes the misjudgment probability of the discriminator by generating potential design schemes, while the discriminator compares the generated design schemes with actual construction data to minimize the misjudgment probability and generates a preliminary design scheme through adversarial training optimization.

4. The method according to claim 1, It is characterized in that The process of multi-objective optimization by using a reinforcement learning algorithm and a genetic algorithm comprises the following steps: defining structural stability, material cost and construction efficiency as optimization objectives, searching for optimal design parameters in a design space using a reinforcement learning algorithm, performing iterative optimization in a population using selection, crossover and mutation operations of a genetic algorithm, and generating an optimized design parameter combination to achieve a multi-objective optimization effect.

5. The method according to claim 4, It is characterized in that The specific calculation process of the multi-objective function is to calculate the contribution of each optimization target by linear combination, wherein the calculation formula of the multi-objective function is: F (x) =α·S(x)+β·C(x)+γ·E(x) Among them, F (x) is the multi-objective function value; S(x) is the structural stability function value; C(x) is the material cost function value; E(x) is the construction efficiency function value; α, β and γ are the weight coefficients of structural stability, material cost and construction efficiency respectively; and the sum of these weight coefficients is one.

6. The method according to claim 1, It is characterized in that The fuzzy logic reasoning comprehensively evaluates the optimization design scheme by establishing a fuzzy rule base, the fuzzy rule base performs reasoning based on the fuzzy indicator matrix of at least two optimization design schemes, calculates the membership of each design indicator through a fuzzy membership function, and uses a fuzzy reasoning mechanism to comprehensively evaluate all candidate design schemes, and selects the scheme with the highest membership as the final design scheme.

7. The method according to claim 1, It is characterized in that The digital twin technology collects and integrates construction site data in real time, builds a construction twin model, simulates and predicts the construction process, and dynamically adjusts construction strategies and resource allocation through adaptive control algorithms to respond to site changes and optimize the construction process.

8. The method according to claim 1, It is characterized in that The multi-scale sensor network includes temperature, humidity and vibration sensors deployed at the construction site. It integrates data from various sensors to monitor the construction environment in real time, and uses deep learning models to extract features and detect anomalies from the monitoring data to identify potential risks in the construction process.

9. The method according to claim 8, It is characterized in that The causal analysis model identifies the causal chain in the construction process by performing causal reasoning on the anomaly detection data, and generates construction strategy adjustment suggestions based on the causal analysis results. The construction strategy adjustment suggestions include: resource reallocation and process optimization.

10. A system for implementing the curtain wall construction phase management method according to any one of claims 1 to 9, It is characterized in that include: A data acquisition module, configured to acquire multi-source data of the construction site through drones, robots and IoT sensor networks, and the module is used to perform data fusion and correction on the multi-source data to generate a comprehensive data set; a modeling module configured to construct and optimize a dynamic three-dimensional model using the comprehensive data set to reflect the actual conditions of the construction site; The design and optimization module is configured to generate a preliminary design scheme based on the dynamic three-dimensional model using a generative adversarial network, perform multi-objective optimization using a reinforcement learning algorithm and a genetic algorithm, perform a comprehensive evaluation of the optimized design scheme using fuzzy logic reasoning, and generate and update the final design scheme in real time; A digital twin module is configured to integrate construction site data in real time according to the final design plan, generate a construction twin model, and dynamically adjust construction strategies and resource allocation through adaptive control algorithms and sensor networks; The monitoring and feedback module is configured to deploy a multi-scale sensor network to monitor the construction environment in real time, integrate monitoring data and perform anomaly detection through deep learning, and provide a basis for adjusting the construction strategy in combination with a causal analysis model to achieve dynamic feedback control and adaptive optimization of the construction strategy; The quality control and maintenance module is configured to define quality standards according to project requirements, develop and execute quality control plans, use sensor monitoring systems to perform quality assessment and automated acceptance, generate acceptance reports, and perform long-term structural monitoring and maintenance plan development through health monitoring systems.

Images