Intelligent building aided design method, system, equipment and medium

Abstract

The invention relates to an intelligent building aided design method, system and device and a medium. The aided design method comprises the steps that a three-dimensional building information model is constructed based on building basic parameters; analyzing a preset technical contradiction matrix to generate a monitoring strategy, and outputting an enhanced building information model and a monitoring strategy file; dynamically adjusting sensor deployment parameters; performing data preprocessing in combination with external environment data to generate a space-time fusion database; performing risk trend prediction and dynamic calculation of an early warning threshold through a machine learning model, and outputting a risk early warning instruction set and a structure damage report; simulating mechanical distribution of different reinforcement schemes, calculating a cost-benefit ratio, and generating an interactive decision panel and a recommendation scheme report; and receiving a feedback score of the user on the recommendation scheme report and an early warning accuracy statistical result through the interactive decision panel, and updating a deployment rule in the monitoring strategy file and a parameter weight of the machine learning model. According to the invention, the safety monitoring precision and the early warning response speed of the building are improved.

Description

Technical Field

[0001] This application relates to the field of intelligent building design management technology, and in particular to an intelligent building auxiliary design method, system, device and medium. Background Technology

[0002] With the accelerating pace of urbanization and the continuous expansion of infrastructure construction, the structural safety risks faced by existing buildings and new projects under complex environments and long-term service conditions are becoming increasingly prominent. Especially under the combined effects of extreme climates, geological disasters, material aging, and improper human use, traditional management models relying on manual inspections and static monitoring are no longer sufficient to meet the modern building's demand for high reliability, real-time performance, and adaptive safety control. Against this backdrop, how to construct an intelligent auxiliary design system capable of autonomously sensing environmental changes, accurately identifying structural anomalies, scientifically predicting risk trends, and providing actionable decision-making suggestions has become a key technological challenge that urgently needs to be overcome in the field of intelligent buildings.

[0003] For a long time, building structural safety systems have primarily relied on passive protection. Monitoring system deployments have often been based on experience or minimum regulatory requirements, lacking differentiated configuration strategies for different building types and usage scenarios. This results in blind spots or redundancy in sensor deployment, making it difficult to guarantee the quality and representativeness of monitoring data. Furthermore, existing monitoring systems cannot adaptively adjust to dynamic factors such as environmental disturbances, load changes, or structural performance degradation, leading to frequent false alarms and missed alarms. This severely impacts the reliability and operability of early warning information, making it difficult to adapt to the ever-evolving safety needs throughout a building's lifespan. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides an intelligent building auxiliary design method, system, device, and medium.

[0005] Firstly, this application provides an intelligent building assisted design method, which adopts the following technical solution: A smart building assisted design method, the assisted design method comprising: Receive basic building parameters and scene labels, and construct a three-dimensional building information model based on the basic building parameters; Based on the scene label parsing preset technical contradiction matrix, a monitoring strategy is generated, and the sensor deployment coordinates and risk weights are marked in the three-dimensional building information model, outputting an enhanced building information model and a monitoring strategy file; Acquire real-time environmental data, dynamically adjust sensor deployment parameters based on the monitoring strategy file, perform self-calibration, and output sensor network topology diagram and real-time data stream after calibration. The system receives the real-time data stream after calibration and external environment data, performs spatiotemporal coordinate alignment and noise filtering, and generates a spatiotemporal fusion database. Based on the risk weights in the spatiotemporal fusion database and the enhanced building information model, a machine learning model is used to predict risk trends and dynamically calculate early warning thresholds, outputting a risk early warning instruction set and a structural damage report. Based on the risk warning instruction set, structural damage report, and pre-configured user decision constraint parameters, the mechanical distribution of different reinforcement schemes is simulated and the cost-benefit ratio is calculated, generating an interactive decision panel and a recommended scheme report; The interactive decision panel receives user feedback ratings and early warning accuracy statistics for the recommended solution report, and updates the deployment rules and parameter weights of the machine learning model in the monitoring strategy file.

[0006] By adopting the above technical solution, intelligent management of the entire building lifecycle has been achieved. This technical solution breaks through the limitations of traditional building monitoring systems, which are static, fixed, and lack self-adaptive capabilities. It innovatively integrates machine learning prediction algorithms, multi-source data fusion technology, and feedback-driven self-optimization mechanisms to form a highly intelligent building safety assurance system. This system not only effectively improves the accuracy of building safety monitoring and the speed of early warning response, but also continuously improves system performance based on actual operational results, providing important technical support and solutions for the development of modern intelligent buildings.

[0007] Secondly, this application provides an intelligent building auxiliary design system, which adopts the following technical solution: An intelligent building auxiliary design system, the auxiliary design system comprising: The data receiving module is used to receive building foundation parameters and scene labels; The building information model building module is used to build a three-dimensional building information model based on the building's basic parameters. The monitoring optimization module is used to generate a monitoring strategy based on the preset technical contradiction matrix parsed from the scene labels, and to mark the sensor deployment coordinates and risk weights in the three-dimensional building information model, and output an enhanced building information model and a monitoring strategy file; The sensor adjustment module is used to acquire real-time environmental data, dynamically adjust the sensor deployment parameters in conjunction with the monitoring strategy file, perform self-calibration operations, and output a sensor network topology diagram and real-time data stream after calibration. The data fusion processing module is used to receive the calibrated real-time data stream and external environmental data, perform spatiotemporal coordinate alignment and noise filtering, and generate a spatiotemporal fusion database. The risk warning module is used to predict risk trends and dynamically calculate warning thresholds based on the risk weights in the spatiotemporal fusion database and the enhanced building information model, and output risk warning instruction sets and structural damage reports. The decision module is used to simulate the mechanical distribution of different reinforcement schemes and calculate the cost-benefit ratio based on the risk warning instruction set, structural damage report and pre-configured user decision constraint parameters, and generate an interactive decision panel and recommended scheme report; The feedback optimization module is used to receive user feedback ratings and early warning accuracy statistics on the recommended solution report through the interactive decision panel, and update the deployment rules and parameter weights of the machine learning model in the monitoring strategy file.

[0008] Thirdly, this application provides a computer device, which adopts the following technical solution: A computer device includes a memory, a processor, and a computer program stored in the memory, the processor executing the computer program to perform the steps of the method as described in the first aspect.

[0009] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as in any of the methods in the first aspect. Attached Figure Description

[0010] Figure 1 This is a first flowchart of an intelligent building-aided design method according to one embodiment of this application.

[0011] Figure 2 This is a second flowchart of an intelligent building assisted design method according to one embodiment of this application.

[0012] Figure 3 This is a schematic diagram of the third process of an intelligent building assisted design method according to one embodiment of this application.

[0013] Figure 4 This is a schematic diagram of the fourth process of an intelligent building assisted design method according to one embodiment of this application.

[0014] Figure 5 This is a schematic diagram of the fifth process of an intelligent building assisted design method according to one embodiment of this application.

[0015] Figure 6 This is a schematic diagram of the sixth process of an intelligent building assisted design method according to one embodiment of this application.

[0016] Figure 7This is a schematic diagram of the seventh process of an intelligent building assisted design method according to one embodiment of this application. Detailed Implementation

[0017] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figures 1-7 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0018] This application discloses an intelligent building auxiliary design method.

[0019] Reference Figure 1 A smart building assisted design method, the assisted design method includes: Step S101: Receive basic building parameters and scene labels, and construct a three-dimensional building information model based on the basic building parameters; Among them, the basic building parameters cover key physical characteristics such as the building's geometric dimensions, structural material properties, and floor distribution, while the scene tags are used to identify the specific application scene characteristics of the building, such as different types of identification such as commercial complexes, residential communities, industrial plants, or historical and cultural heritage buildings.

[0020] Based on this, the system uses building information modeling technology to construct a three-dimensional building information model based on the building's basic parameters. This process involves converting two-dimensional architectural design drawings into a digital three-dimensional model with complete spatial topological relationships. This model not only contains the geometric shape information of the building, but also integrates multi-dimensional data such as the material properties of components, connection relationships, and functional zoning, providing an accurate spatial reference framework for subsequent intelligent analysis.

[0021] Step S102: Generate a monitoring strategy by parsing the preset technical contradiction matrix based on the scene labels, and mark the sensor deployment coordinates and risk weights in the 3D building information model, and output the enhanced building information model and monitoring strategy file; The system analyzes a pre-defined technical contradiction matrix based on received scene labels to generate targeted monitoring strategies. It then precisely marks the optimal deployment coordinates of sensors and the risk weight levels for each area within the constructed 3D building information model. In this process, the pre-defined technical contradiction matrix employs the contradiction resolution principle from the TRIZ theory framework, establishing a mapping relationship between monitoring needs and implementation constraints in different scenarios. For example, in the scenario of protecting old historical buildings, it is necessary to balance the contradiction between monitoring accuracy and construction damage. By querying the corresponding invention principle code, a non-contact sensor deployment scheme is determined.

[0022] Meanwhile, the system assigns corresponding risk weight values ​​to each monitoring point in the 3D building information model based on the structural importance, function, and potential risk factors of different areas. The final output is an enhanced building information model that integrates sensor deployment information and risk assessment data, as well as a monitoring strategy file containing specific monitoring parameter configurations. These two outputs provide a scientific basis and technical guidance for the subsequent deployment of sensor networks.

[0023] Step S103: Obtain real-time environmental data, dynamically adjust sensor deployment parameters based on monitoring strategy file and perform self-calibration operation, output sensor network topology diagram and real-time data stream after calibration; The system dynamically adjusts sensor deployment parameters and performs automatic calibration based on the generated monitoring strategy file and real-time environmental data collected on-site. This stage demonstrates the system's adaptive capability; when drastic changes in ambient temperature affect sensor accuracy, the system automatically adjusts sensor operating parameters, such as gain coefficient or sampling period, and compensates for measurement deviations through a built-in self-calibration algorithm. During calibration, the system also re-evaluates the sensor network topology, optimizes data transmission paths to ensure communication efficiency, and ultimately outputs a sensor network topology map reflecting the current optimal network state, as well as a high-quality real-time data stream after calibration, laying a reliable data foundation for subsequent data fusion and analysis.

[0024] Step S104: Receive the real-time data stream after calibration and external environment data, perform spatiotemporal coordinate alignment and noise filtering, and generate a spatiotemporal fusion database; The system receives calibrated real-time data streams and environmental data from external data sources such as meteorological departments and geological monitoring stations, and performs complex spatiotemporal coordinate alignment and noise filtering operations. Regarding spatiotemporal coordinate alignment, since sensor data is typically represented using a local coordinate system, while external environmental data is based on a global geographic coordinate system, a coordinate transformation matrix is ​​needed to unify data from different sources under the same reference frame, ensuring the consistency of spatial location information.

[0025] Meanwhile, considering that abnormal values ​​may be generated during the data acquisition process due to factors such as electromagnetic interference and equipment aging, the system can use digital filtering technology and statistical methods to denoise the raw data, eliminating the influence of random errors and system biases. After this series of data preprocessing operations, the system generates a high-quality spatiotemporal fusion database. This database integrates the monitoring data inside the building with the influencing factors of the external environment, providing comprehensive and accurate data support for subsequent risk analysis.

[0026] Step S105: Based on the risk weights in the spatiotemporal fusion database and the enhanced building information model, risk trend prediction and early warning threshold are dynamically calculated using a machine learning model, and a risk early warning instruction set and structural damage report are output. Based on the completed spatiotemporal fusion database and the pre-set risk weight information in the enhanced building information model, the system calls a pre-trained machine learning model to perform in-depth risk trend prediction analysis and dynamic calculation and adjustment of early warning thresholds. In this process, the machine learning model, through learning from historical monitoring data, can identify potential patterns and regularities in the development of structural damage and predict the potential risk evolution trends of buildings in the future.

[0027] At the same time, the system dynamically adjusts the threshold settings of various safety warning indicators based on changes in current environmental conditions and the risk weight distribution of different parts of the building, instead of using traditional fixed alarm limits. This adaptive threshold adjustment mechanism can significantly improve the accuracy and practicality of the warning system. Finally, the system outputs a risk warning instruction set containing specific countermeasures and a detailed structural damage assessment report, providing decision-makers with timely and effective safety protection recommendations.

[0028] Step S106: Based on the risk warning instruction set, structural damage report and pre-configured user decision constraint parameters, simulate the mechanical distribution of different reinforcement schemes and calculate the cost-benefit ratio, and generate an interactive decision panel and recommended scheme report. The system performs mechanical performance simulation analysis on various possible structural reinforcement schemes and calculates the corresponding cost-benefit ratio based on the generated risk warning instruction set, structural damage report, and user-preconfigured decision constraint parameters.

[0029] At this stage, the system integrates a professional finite element analysis engine, capable of accurately simulating the impact of different reinforcement measures on the structural mechanical behavior of buildings. For example, it calculates the deformation differences and stress distribution characteristics between carbon fiber reinforcement and steel structure reinforcement under load. Simultaneously, the system comprehensively considers multiple economic factors such as material costs, construction difficulty, schedule requirements, and maintenance costs, conducting a comprehensive cost-benefit assessment of various reinforcement schemes. Ultimately, it generates an intuitive and easy-to-understand interactive decision panel and a detailed recommendation report, enabling decision-makers to make the optimal choice from multiple alternatives, thus achieving an organic combination of engineering technology and economic benefits.

[0030] Step S107: Receive user feedback ratings and early warning accuracy statistics for the recommended solution report through the interactive decision panel, and update the deployment rules and parameter weights of the machine learning model in the monitoring strategy file.

[0031] The system collects user feedback and ratings on recommended hardening solutions through an interactive decision panel. It then combines this feedback with historical early warning accuracy statistics to intelligently update and optimize the deployment rules and core parameter weights of the machine learning model within the original monitoring strategy file. When user feedback indicates that a certain type of hardening solution is ineffective, the system reduces its priority weight in subsequent recommendations. Similarly, when the early warning system experiences a high false alarm rate, it automatically extracts relevant false alarm cases as new training samples to retrain the machine learning model and improve its predictive accuracy.

[0032] Understandably, through this feedback-based learning mechanism, the entire intelligent building auxiliary design system possesses the ability to self-improve and continuously evolve. It can continuously optimize its performance over time and ultimately output upgraded and improved monitoring strategy files and detailed algorithm update logs, forming a complete closed-loop optimization system.

[0033] The above implementation achieves intelligent management throughout the entire lifecycle of a building. This technical solution overcomes the limitations of traditional building monitoring systems, which are static, fixed, and lack adaptive capabilities. It innovatively integrates machine learning prediction algorithms, multi-source data fusion technology, and feedback-driven self-optimization mechanisms to form a highly intelligent building safety assurance system. This system not only effectively improves the accuracy of building safety monitoring and the speed of early warning response but also continuously improves system performance based on actual operational results, providing important technical support and solutions for the development of modern intelligent buildings.

[0034] Reference Figure 2 As one implementation of step S102, the steps of generating a monitoring strategy based on a preset technical contradiction matrix parsed from scene labels, marking sensor deployment coordinates and risk weights in the 3D building information model, and outputting an enhanced building information model and monitoring strategy file include: Step S201: Call the preset technical contradiction matrix library, match the corresponding technical contradiction pairs according to the scene label, and generate an optimization scheme code set; The scene tags may include at least one of the following: old houses, houses after disasters, and houses affected by construction. Specifically, building scene tags are used to characterize the specific contextual attributes of the building currently being monitored, such as whether it belongs to the category of old buildings, post-disaster reconstruction buildings, or buildings affected by nearby construction. These tags are essentially a high-level classification marker of building status, reflecting not only the physical aging or structural damage of the building itself, but also implying potential threats from its surrounding environment.

[0035] Subsequently, a pre-constructed technical contradiction matrix database is used to assist in formulating reasonable sensor configuration strategies. The technical contradiction matrix database is based on the dualistic theory of contradictions, meaning that any engineering scenario can be deconstructed into a pair of mutually constraining parameters (for example, in the scenario of old buildings, "improving monitoring accuracy" leads to "increased implementation costs," while "reducing implementation costs" may sacrifice "accuracy"). The matrix database solidifies these relationships through an index table, ensuring that scenario labels (such as "old buildings") can be directly mapped to predefined contradiction pairs. This mapping originates from statistical analysis of historical engineering data, ensuring the objectivity and repeatability of contradiction identification. The generation of the optimization scheme code set is an abstract expression of contradiction resolution; each code represents a type of optimization strategy (such as dynamic adjustment or component segmentation), aiming to bypass traditional trade-offs and achieve innovative reconciliation of contradictions.

[0036] In this embodiment, when the system receives building scene tags, it first executes a tag parsing algorithm to match the tags (such as "post-disaster housing") to the index table of the matrix library. The index table is stored in a key-value pair structure, where the key is the scene tag enumeration value and the value is the corresponding contradiction pair identifier (e.g., "post-disaster housing" maps to "contradiction pair ID: SP-DC", representing "Monitoring Speed ​​vs. Data Completeness"). Next, the system retrieves the contradiction matrix (a two-dimensional lookup table), where the rows and columns represent deterioration parameters and improvement parameters, respectively, and the cross cells output a set of optimized solution codes (e.g., the code "DYN" corresponds to a dynamic strategy, and "SEG" corresponds to a segmentation strategy). These codes are output in set form (e.g., {"DYN", "SEG"}), forming an optimized solution code set. The entire process is driven by a rule engine, ensuring efficiency and scalability—for example, when adding a new scene tag, only the index table needs to be expanded, not the core logic modified.

[0037] Step S202: Parse the optimized scheme code set to generate a sensor deployment rule set; wherein, the sensor deployment rule set includes sensor type, deployment density and accuracy level; Specifically, the optimized scheme coding set, as an abstract blueprint for resolving contradictions, does not have direct operability. Therefore, it needs to be transformed into a specific set of sensor deployment rules through a parsing mechanism. This set of rules defines the implementation details of the monitoring system, including sensor type selection, deployment density quantification, and accuracy level specifications.

[0038] The parsing process relies on a rule mapping system, which pre-establishes a database linking optimization scheme codes with sensor configuration templates. Each code (e.g., "DYN" representing a dynamic strategy) corresponds to one or more rule templates. These templates are designed based on domain knowledge (e.g., structural health monitoring standards) and emphasize parameterized output (e.g., density rules might be defined as functional expressions). The generation of sensor type, deployment density, and accuracy level is not an isolated decision but rather an optimization process using multi-objective optimization algorithms (e.g., weighted scoring models) to find the optimal solution under the constraints of the codes. For example, the dynamic code "DYN" implicitly contains the principle of "adaptive adjustment," which directly derives from the requirement that density rules include a real-time feedback mechanism; while the segmentation code "SEG" embodies the principle of "component decoupling," leading to sensor type combination rules. This parsing ensures the completeness and consistency of the rule set, avoiding rule conflicts.

[0039] In this embodiment, the system adopts a hierarchical parsing architecture: First, an optimization scheme encoding set (e.g., {"DYN", "SEG"}) is input, and the encoding query is used to match predefined templates in the rule base (templates are stored in JSON or XML format). Each template contains three core fields: SensorType (sensor type), DensityRule (deployment density rule), and PrecisionLevel (precision level requirement). The parser performs instantiation based on the encoding semantics—for example, if the encoding contains "DYN", the dynamic density adjustment rule generator is triggered, outputting a conditional statement such as "IF real-time vibration intensity > 5 mm / s THEN density increase 30%"; if the encoding contains "SEG", the combined unit rule is invoked to generate the instruction "deploy a combined unit of tilt sensors and strain gauges at the load-bearing wall node". The deployment density rule is usually calculated using a probability model (e.g., Poisson distribution), and the precision level is mapped according to international standard classifications. Finally, all rules are aggregated into a structured rule set, and the logical completeness is checked by a verification module (e.g., avoiding sensor type redundancy).

[0040] Step S203: Based on the 3D building information model, calculate the optimal deployment coordinate set according to the sensor deployment rule set; In this process, BIM modeling technology and finite element numerical simulation played a crucial role. On the one hand, by mining data from original drawings or on-site scanning results, high-fidelity spatial geometry and component connection relationships can be reconstructed. On the other hand, relying on this digital platform and supplemented by professional structural analysis software tools, refined stress field simulation calculations can be performed, thereby identifying several key areas with weak load-bearing capacity or prone to cumulative failure. Subsequently, around these hotspots, according to predetermined deployment density requirements, a series of uniformly distributed virtual anchor point sequences that satisfy coverage radius constraints are generated using a mesh generation algorithm.

[0041] It should be noted that this method differs from the traditional manual experience-based approach to site selection. It can make differentiated arrangements based on the actual stress characteristics of different individual buildings, ensuring that every sensitive area can be effectively monitored.

[0042] Step S204: Combine the historical risk data corresponding to the scene labels to annotate the risk weight distribution map in the 3D building information model; To further enhance the model's risk warning capabilities, it's necessary to combine historical risk data corresponding to building scene labels to create a risk weight distribution map within the 3D building information model. For this purpose, the system accessed a historical disaster database resource pool maintained by an authoritative institution, extracting a large number of similar case samples for training. Then, a convolutional neural network architecture from the field of deep learning was used to extract features and summarize patterns from these image samples, discovering potential correlations between recurring structural weaknesses and their surrounding micro-environmental factors. This learned knowledge was then transferred to the target object, assigning a corresponding danger level score to each pixel unit, thus creating an intuitive and visual heat map. Throughout this process, the emphasis is on the effective reuse of existing knowledge assets and the development of trend prediction capabilities under unknown conditions, rather than simply relying on subjective expert judgment for extensive labeling.

[0043] Step S205: Output the enhanced building information model and monitoring strategy file; wherein, the enhanced building information model integrates the optimal deployment coordinate set and risk weight distribution map, and the monitoring strategy file records the sensor deployment rule set.

[0044] Specifically, the enhanced building information model (BIM) is a product of overlaying two layers of new information elements onto the original BIM: first, the calculated optimal deployment coordinates of all sensors; and second, a rendered risk heat map layer. This integrated digital twin representation allows managers to simultaneously observe the spatial layout of physical entities, health status assessment conclusions, and recommended protective measures within a single viewport, greatly improving information exchange efficiency. The monitoring strategy document is a formally delivered document written in a structured format, detailing all the sensor selection principles, deployment density specifications, sampling frequency settings, and other special operating guidelines derived above. It serves as an important reference for downstream IoT hardware manufacturers to implement specific installation operations and also provides a standardized template for the company's internal operations and maintenance team to conduct regular inspections.

[0045] The above implementation methods demonstrate stronger adaptability and forward-looking advantages in proactively identifying potential hazards, accurately allocating sensor resources, and objectively quantifying risk exposure levels. They effectively compensate for the shortcomings of delayed response in the current system and lay a solid technical foundation for promoting the construction of a full life-cycle security system for smart city infrastructure.

[0046] Reference Figure 3 As one implementation of step S103, the steps of acquiring real-time environmental data, dynamically adjusting sensor deployment parameters based on the monitoring strategy file and performing self-calibration, and outputting a sensor network topology diagram and a real-time data stream after calibration include: Step S301: Receive monitoring strategy file and real-time environmental data; wherein, the monitoring strategy file contains a set of sensor deployment rules, and the real-time environmental data includes vibration intensity, temperature and humidity values, and geographic location coordinates; Specifically, the monitoring strategy document is essentially a set of pre-defined rules that define the appropriate sensor configuration and response behavior for different environments. Real-time environmental data comes from various field-deployed sensing devices, such as accelerometers for measuring vibration intensity and temperature and humidity sensors for recording the current environmental conditions. This data, combined with GPS or other positioning methods, determines the specific geographic coordinates of each sensor. These data collectively form the core input source for all subsequent decision-making. At this stage, it is crucial not only to ensure the accuracy and timeliness of data acquisition but also to guarantee that the strategy document possesses sufficient flexibility and adaptability to cover the widest possible range of application scenarios.

[0047] Step S302: parse the dynamic adjustment conditions in the deployment rule set, and generate a sensor parameter update instruction when the real-time environmental data meets the preset trigger threshold. The essence of this step is to transform abstract rules into concrete execution instructions. The key is to identify which external factors have reached preset critical points, thereby triggering corresponding adjustment measures. During this process, the system deeply analyzes the various judgment logics contained in the strategy document, especially those standards regarding the responses to specific working conditions. Once one or more environmental variables are detected to exceed predetermined thresholds—for example, when the vibration amplitude at the construction site exceeds safety limits or both temperature and humidity rise to dangerous levels—the corresponding response program is immediately activated, generating operation commands to modify a specific aspect (such as sensor deployment density or sampling period). This multi-dimensional, interconnected judgment approach makes the system's response more precise and efficient, far surpassing simple single-factor triggering mechanisms.

[0048] Step S303: Adjust the sensor deployment parameters according to the sensor parameter update command and generate a sensor network topology map; The key aspect of this step lies in how to modify the existing hardware layout or operating methods based on the received specific guidelines. A topology diagram is a graphical representation of the connections between sensor nodes and their distribution throughout the monitored area. To maintain optimal performance, it may be necessary to increase the number of detectors in certain sensitive areas to enhance local sensing capabilities, reduce the operating speed of some non-core components due to excessive energy consumption, or even replace them with new sensing devices more suitable for the current environmental characteristics. Each such change is reflected in the latest network architecture, resulting in a new topology that not only reflects the current optimized resource allocation scheme but also provides an accurate spatial reference framework for subsequent data processing.

[0049] Step S304: Based on the temperature and humidity values ​​in the real-time environmental data, call the self-calibration algorithm to perform error compensation on the raw data collected by the sensor and output the calibrated real-time data stream; External climate conditions can affect various electronic instruments to varying degrees, especially when exposed to harsh conditions for extended periods, leading to reading deviations. Therefore, a scientifically sound compensation mechanism is needed to eliminate the negative impact of these interference factors. To this end, an adaptive correction model specifically designed for such applications has been developed. This model comprehensively considers the changing trends of ambient air temperature and moisture content, derives a theoretically permissible range of deviation, and then applies this range back to the initial observation values ​​to accurately reflect the true situation. The results obtained after this refined processing are significantly more reliable than unprocessed raw data and are more conducive to subsequent big data analysis and anomaly early warning systems.

[0050] Step S305: Associate the sensor network topology map with the calibrated real-time data stream and store it in a distributed database.

[0051] Among them, associative storage emphasizes the need to establish a one-to-one correspondence between the newly generated map structure and the various indicators that have been purified within the corresponding time period. Only in this way can we have a basis for tracing the historical evolution trajectory or reviewing the cause of the accident in the future. Distributed databases are a choice made for performance considerations. They use multiple collaborative server clusters to distribute and store massive amounts of information, which can not only greatly improve access efficiency but also significantly enhance disaster recovery and backup capabilities.

[0052] The above implementation integrates multiple key technologies, including environmental perception, intelligent decision-making, flexible scheduling, data cleaning, and efficient archiving, forming a highly integrated yet hierarchical overall solution. Through effective control of the close coordination between its components, it achieves a transformation from passively receiving signals to proactively intervening and providing feedback, and then to continuous optimization and iteration—a full lifecycle management model.

[0053] Reference Figure 4 As one implementation of step S105, the steps of predicting risk trends and dynamically calculating early warning thresholds based on risk weights in the spatiotemporal fusion database and the enhanced building information model, and outputting a risk early warning instruction set and structural damage report, include: Step S401: Read the multi-source monitoring data from the spatiotemporal fusion database and the risk weight distribution from the enhanced building information model; The spatiotemporal fusion database refers to an integrated data platform that not only stores historical and real-time data collected from various sensors (such as inclinometers and settlement gauges), but also supports continuous tracking over time and precise matching of spatial locations. This type of database typically employs a distributed or edge computing architecture to ensure high concurrency access efficiency and combines Geographic Information System (GIS) and Building Information Modeling (BIM) for unified index management. Enhanced Building Information Modeling (BIM), on the other hand, introduces a structural vulnerability assessment module on top of traditional BIM. This module includes weighted values ​​for the failure probability of different components under stress conditions. These weights are obtained through learning from historical accident cases or statistical analysis of finite element simulation results. Therefore, the core of this step lies in establishing a comprehensive data view encompassing physical monitoring data and virtual structural characteristics, providing contextual support for subsequent data analysis.

[0054] Step S402: Extract the time-series feature vector and environmental feature vector from the multi-source monitoring data; Among them, the time-series feature vector includes structural settlement and the rate of change of tilt angle, and the environmental feature vector includes temperature, humidity, wind speed and rainfall. Specifically, the feature vector is not simply a set of raw numerical values, but rather a representative set of state parameters formed after signal processing and filtering / denoising. For example, regarding structural settlement, it is necessary not only to record the current cumulative settlement depth but also to pay attention to the rate of change and its fluctuation trend per unit time. Similarly, the rate of change of tilt angle must also consider the influence of periodic disturbances to avoid misjudgments caused by short-term abnormal interference. As for environmental characteristics, they cover indirect factors that may affect structural stability from external natural conditions, such as the thermal expansion and contraction effect caused by temperature gradients, the accumulation of fatigue loads under continuous strong winds, and the softening of the foundation caused by prolonged precipitation. Although these variables do not directly reflect the deterioration of the mechanical properties of the structure itself, they often exist as key factors inducing secondary disasters and must be included in the consideration. Through this step, the original messy data is abstracted into a standard mathematical expression that can be used for model training and inference, thereby improving the accuracy and robustness of subsequent analysis stages.

[0055] Step S403: Input the time-series feature vector, environmental feature vector and risk weight into the pre-trained machine learning model, and output the structural damage probability value and risk trend prediction curve. The pre-trained machine learning model employs a hybrid architecture combining a bidirectional long short-term memory network (Bi-LSTM) with fully connected neural layers. The advantage of Bi-LSTM lies in its ability to simultaneously capture the dependencies between past and future moments, which is particularly important for understanding complex nonlinear evolution processes.

[0056] Meanwhile, since environmental variables themselves do not possess obvious temporal characteristics, they are separately fed into several densely connected perceptron units for encoding and compression, generating a set of low-dimensional dense representations, namely environmental state codes. These two types of representations are then fed together into a higher-level fusion layer for integration, and finally, a normalized confidence score between 0 and 1 is output by a Softmax classifier to quantify the probability of structural damage occurring in a specific region. Furthermore, to help users grasp development trends over a future period, a set of smooth time function curves is additionally fitted to depict the evolution path of the expected risk level over time.

[0057] Step S404: Dynamically calculate the early warning threshold based on the risk trend prediction curve and environmental feature vector, and generate a risk early warning instruction set that includes the early warning level, risk coordinates and triggering basis; This scheme proposes a novel adaptive adjustment mechanism that allows the judgment criteria to be flexibly adjusted according to changes in external conditions. For example, when the meteorological department issues a typhoon warning, considering that storm surge may lead to an increase in groundwater levels and thus exacerbate the probability of soil liquefaction, the maximum allowable settlement limit should be appropriately reduced. Similarly, under continuous heavy rain conditions, the expansion of wall materials due to water absorption will weaken the original anchoring strength, and it is necessary to increase the upper limit of the tolerance for crack expansion width to prevent frequent false alarms.

[0058] It's important to note that all these correction coefficients are not arbitrarily set, but rather derived from empirical rules based on extensive analysis of real-world data. Furthermore, a coordinated assessment must be made based on whether the rate of risk growth exceeds a predetermined critical range. If signs of rapid deterioration in the short term are detected, the judgment boundaries should be tightened promptly to avoid missing the optimal intervention window. The resulting decision recommendations will be packaged into a series of executable command packages in a standardized format and sent to relevant operations and maintenance personnel to ensure they receive critical intelligence and take appropriate action as soon as possible.

[0059] Step S405: Based on the mapping relationship between the structural damage probability value and the spatial coordinates of the enhanced building information model, generate a structural damage report marking high-risk areas.

[0060] Although a closed-loop control system has been established, encompassing the entire process from perception to cognition and then to action guidance, its practical value is still difficult to realize without an intuitive and visual presentation method. Therefore, advanced graphics rendering technology and topological clustering algorithms are needed to further process the results.

[0061] Specifically, the probability estimates at each measurement point are projected back onto the corresponding 3D geometric nodes. Morphological operations are then used to expand and fill adjacent hotspots, constructing several interconnected danger zone outlines. This not only eliminates visual misleading information from isolated noise points but also more clearly distinguishes areas that are indeed at a higher threat level. The final electronic document, in addition to listing specific coordinates, includes stress cloud maps of varying color shades for interpretation and can even overlay arrows to indicate the most likely direction of further spread, helping managers develop more targeted emergency plans.

[0062] The above implementation method constructs a complete intelligent diagnosis-assessment-response chain, which effectively overcomes many drawbacks of the previous extensive management based on experience and intuition, and significantly improves the safety guarantee level of urban infrastructure operation.

[0063] Reference Figure 5 As one implementation of step S106, the steps of simulating the mechanical distribution of different reinforcement schemes and calculating the cost-benefit ratio based on the risk warning instruction set, structural damage report, and pre-configured user decision constraint parameters, and generating an interactive decision panel and recommended scheme report include: Step S501: Receive risk warning instruction set, structural damage report and user decision constraint parameters; wherein, the user decision constraint parameters include budget limit, schedule requirement and material preference; Specifically, risk warning instructions are usually issued by monitoring equipment or safety assessment platforms, reflecting whether there is a potential structural instability risk in a specific area within a certain time period; structural damage reports come from regular inspections or emergency investigations, recording specific information on structural defects such as crack locations, degree of steel reinforcement corrosion, and concrete spalling; while user decision-making constraints mainly include human-set factors such as budget limits, construction period requirements, and material preferences, which directly affect the selection direction of subsequent reinforcement strategies.

[0064] Step S502: Analyze the coordinates and stress distribution data of high-risk areas in the structural damage report to generate the finite element analysis input model; This process involves extracting key spatial geometric information from structural damage reports, such as the set of polygon vertex coordinates corresponding to the damaged areas, and mapping it onto the digital space constructed by Building Information Modeling (BIM). Subsequently, based on the measured stress distribution, each mesh node is assigned a corresponding physical attribute weight value, thereby establishing a finite element computational domain with realistic boundary conditions and material properties. This modeling method not only improves the accuracy of the simulation but also makes subsequent mechanical performance simulations more closely resemble actual conditions.

[0065] Step S503: Based on the type of high-risk area, call the predefined reinforcement scheme library and activate at least two candidate reinforcement schemes; wherein, the candidate reinforcement schemes include carbon fiber reinforcement, steel frame support or concrete replacement; Specifically, based on the type of high-risk area identified in the previous step (such as flexural failure, shear cracking, localized crushing, etc.), the system automatically retrieves pre-configured templates for various reinforcement methods. Typical examples include carbon fiber reinforcement, the addition of steel structure support systems, and localized concrete replacement reinforcement measures. Each solution corresponds to a standardized set of operating procedures and technical parameters. The system selects two or more alternative paths suitable for the target scenario, laying the foundation for further quantitative comparison.

[0066] Step S504: Perform mechanical simulation calculations based on the finite element analysis input model, and output the deformation control value and stress peak value of each candidate reinforcement scheme at the load-bearing node; For each candidate reinforcement scheme, corresponding structural modifications are applied (such as adding composite material layers, adding support members, or replacing concrete units). Nonlinear static or dynamic analyses are then performed under standard load combinations (including dead load, live load, wind load, and seismic action). The final output is the deformation control value and peak stress at key load-bearing nodes for each scheme. The deformation control value typically refers to the maximum displacement or rotation of the structure under the most unfavorable conditions, used to evaluate the overall stiffness improvement and performance recovery. The peak stress reflects the stress level at the most critical section within the material, directly related to the structure's safety margin and fatigue life.

[0067] Understandably, by comparing the changes in these two indicators before and after reinforcement, the actual improvement in structural performance of each scheme can be quantified. The core technology of this step lies in high-fidelity numerical simulation capabilities, requiring the model to accurately capture material constitutive relationships (such as the damage-plastic model of concrete and the elastoplastic hardening behavior of steel), interfacial bond-slip effects (such as the peel stress transfer between carbon fibers and the substrate), and geometric nonlinearity (internal force redistribution caused by large deformation).

[0068] Step S505: Based on the deformation control value and stress peak value of each candidate reinforcement scheme at the load-bearing node, and combined with the pre-stored material market price database and user decision constraint parameters, calculate the cost-benefit ratio of each candidate reinforcement scheme. The cost-benefit ratio is achieved by quantifying the ratio of the benefits of improved mechanical properties (including the reduction rate of deformation control values ​​and the reduction rate of peak stress) to cost parameters (materials, construction period). The formula can be defined as: Cost-benefit ratio = Overall score of mechanical property improvement / Total cost; In the above formula, the comprehensive score for improvement in mechanical properties = w1 × deformation control value reduction rate + w2 × stress peak reduction rate (w1 and w2 are weighting coefficients, which are dynamically adjusted based on the material preference in the user's decision constraint parameters).

[0069] Specifically, the cost-benefit ratio is not simply a cost-performance indicator. Instead, it is determined by establishing a multi-dimensional evaluation function. "Benefit" is defined as the quantitative result of improved mechanical performance, specifically including the reduction rate of deformation control values ​​(i.e., the percentage reduction in maximum displacement after reinforcement relative to the original structure) and the reduction rate of peak stress (reflecting the degree of improvement in safety margin). These two are weighted and synthesized into a comprehensive performance gain index. The "cost" component encompasses direct material costs (dynamically retrieved from a market price database and multiplied by the quantity used), labor and machinery inputs (converted to labor costs based on construction period requirements), construction cycle impacts (such as losses from traffic disruptions or production stoppages), and material preference penalties (such as weight attenuation factors introduced when using non-recommended materials). The final cost-benefit ratio is the ratio of the performance gain index to the total cost; a higher value indicates a more significant improvement in structural performance per unit of input. This calculation process is essentially an application of Multi-Criterion Decision Analysis (MCDA). Its scientific basis lies in transforming subjective preferences (such as users prioritizing rapid repair over minimum cost) into calculable weighted coefficients and eliminating incommensurability between indicators of different dimensions through normalization, thereby achieving fair comparison across different schemes.

[0070] Step S506: Output the interactive decision panel and recommended solution report.

[0071] The interactive decision panel integrates a 3D mechanical distribution visualization module and a cost-benefit comparison chart, while the recommended solution report marks the construction coordinates and material list of the optimal reinforcement solution according to the cost-benefit ratio from high to low.

[0072] Specifically, the interactive decision panel is a highly integrated graphical user interface with an embedded 3D visualization module to display the changing trends of the internal force field of the structure under different reinforcement strategies. It also features a dynamic radar chart to show the trade-offs between various economic and technical indicators. The recommended solution report is a detailed text document that not only indicates the specific construction coordinates for the optimal choice, but also includes a complete bill of materials to facilitate on-site implementation.

[0073] In the above implementation, based on a systematic analysis of the damage to existing building structures and incorporating user decision-making constraints, a scientific, reasonable, and economically feasible reinforcement proposal is automatically generated, thereby improving the scientific nature and transparency of the traditional reinforcement project planning work.

[0074] Reference Figure 6 As one implementation of step S107, the steps of receiving user feedback ratings and early warning accuracy statistics for the recommended solution report through the interactive decision panel, and updating the deployment rules and parameter weights of the machine learning model in the monitoring strategy file include: Step S601: Receive user feedback ratings and early warning accuracy statistics for the recommended solution report through the interactive decision panel; Among them, the user feedback rating includes the solution satisfaction score and the construction feasibility rating, and the early warning accuracy statistics include the false alarm rate, the missed alarm rate and the timeliness deviation value. Specifically, an interactive decision panel refers to a set of visual human-computer interface tools that allow engineers or managers to evaluate the performance of the current monitoring system based on actual operational experience. This subjective evaluation includes two aspects: first, the degree of satisfaction with the monitoring solution itself (i.e., the solution satisfaction score), and second, the construction feasibility rating of whether the implementation plan is easy to carry out.

[0075] Simultaneously, the system also collects objective performance data, such as false positive rate, false negative rate, and timeliness deviation. These indicators are derived from the results automatically aggregated from background logs during long-term operation and reflect the system's response accuracy and efficiency in real-world environments. The core function of this stage is to bridge the gap between human experience and machine behavior, enabling the previously isolated AI model to receive real feedback signals from frontline practice, thus providing a reliable input source for its subsequent behavior correction.

[0076] Step S602: Analyze the correlation between user feedback ratings and early warning accuracy statistics to generate a multi-dimensional rating matrix; Specifically, a five-dimensional spatial structure can be constructed to accommodate different types of quality attributes: user subjective ratings (solution satisfaction, construction feasibility) and system objective performance (false alarm rate, missed alarm rate, timeliness deviation). This high-dimensional abstract framework provides a unified measurement standard for subsequent comprehensive evaluation. By performing cluster analysis or correlation coefficient calculations on historical data accumulated over multiple project cycles, it's possible to reveal which combinations are prone to adverse consequences. For example, if certain building areas frequently experience false alarms while receiving low satisfaction ratings from most on-site engineers, it may indicate problems such as unreasonable deployment in that area or excessively sensitive algorithms.

[0077] Step S603: Analyze the rule defect items in the monitoring strategy file based on the multidimensional scoring matrix, and update the deployment rules in the sensor deployment rule set; Among these, rule defects refer to improper settings in the original configuration, which may be due to insufficient initial installation planning or changes in external conditions. These problems typically manifest in three ways: insufficient sensor density leading to coverage blind spots; the selection of sensors that do not match the specific scenario requirements; or excessively long sampling intervals causing event capture delays. To address these issues, the system will take corresponding measures based on the risk level marked in the aforementioned scoring matrix. These measures may include increasing the detector density in certain hotspot areas, replacing infrared imaging devices with those more suitable for high-temperature and humid environments, or shortening the sampling cycle of critical nodes to accelerate response speed. It is important to note that all adjustments adhere to preset safety boundaries to prevent excessively aggressive actions from causing new instabilities.

[0078] Step S604: Extract feature vectors as new training sample sets based on the false alarm period data in the early warning accuracy statistics. Specifically, the trends of various physical quantities within a time window before and after a false alarm are sliced ​​along a time axis. Then, advanced image processing techniques such as gradient-weighted class activation mapping are used to extract the most representative local feature patterns. These compressed and encoded vectors not only retain the main fluctuation characteristics of the original sequence but also significantly reduce storage overhead and computational complexity, making them ideal for expanding the size of existing databases. Furthermore, dedicated subsets can be established according to different types of misjudgment scenarios to ensure that each incremental learning focuses on the areas most in need of improvement.

[0079] Step S605: Incrementally train the machine learning model using the newly added training sample set, and adjust the parameter weight distribution of the convolutional neural network layer and the LSTM layer. Specifically, this application employs a more refined and controllable learning strategy. It keeps the first few layers of convolutional modules responsible for extracting basic texture and shape information unchanged, while only fine-tuning the recurrent neural units and their subsequent classifiers that handle time-dependent tasks such as predicting future vibration amplitude trends. This ensures that existing advantages are not lost while rapidly absorbing newly learned lessons, achieving a smooth evolutionary path. Furthermore, to avoid falling into local minima, it also incorporates momentum methods or other advanced optimizers to dynamically adjust the learning rate, improving convergence stability.

[0080] Step S606: Output the updated monitoring strategy file and algorithm update log.

[0081] The algorithm update log is used to record the changes in parameter weights and the fingerprints of training samples.

[0082] Specifically, the algorithm update log records changes in parameter weights and training sample fingerprints, completing the information archiving and version control functions for the entire process. The updated strategy document includes the latest hardware layout suggestions, software decision logic, and other content, which the operations team can refer to implement the new round of transformation plans. The accompanying log file records in detail every subtle change that occurred during this iteration, including the parameter drift magnitude caused by each round of gradient descent, the list of specific sample identifiers participating in this round of training, and other important details.

[0083] The above implementation combines the advantages of human subjective judgment and machine autonomous reasoning, solves the problem that traditional fixed configurations are difficult to cope with complex working conditions, and constructs a closed-loop control system with self-perception, judgment and optimization capabilities.

[0084] Reference Figure 7 As a further implementation of the intelligent building assisted design method, after generating the recommended solution report, it also includes: Step S701: Receive deformation data and material stress data of the reinforced structure collected by sensors during the construction phase; Specifically, sensors used during the construction phase include, but are not limited to, precision instruments such as vibrating wire strain gauges, fiber optic grating sensors, and laser displacement measuring instruments. These are embedded or attached to the surface or interior of critical load-bearing components to capture changes in mechanical response at a microscale, such as shrinkage and creep behavior of concrete structures during curing, and residual stress release in welded areas of steel structures due to cooling. These raw signals typically manifest as analog voltage fluctuations or frequency jumps. To facilitate subsequent data processing and transmission, edge computing nodes are needed to perform analog-to-digital conversion, filtering and noise reduction, and protocol encapsulation.

[0085] Step S702: Compare the deformation data of the reinforced structure with the design expectation values ​​of the corresponding coordinates in the enhanced building information model to output a heat map of structural performance offset. By quantitatively assessing the difference between actual monitored values ​​and their theoretical predictions, it is possible to reveal whether there are potential risks in local components that exceed the allowable error range. Specifically, this process relies on spatial registration algorithms to eliminate coordinate system inconsistencies caused by factors such as installation errors and modeling deviations. A commonly used method is the Iterative Closest Point (ICP) matching strategy, which approximates the optimal transformation matrix layer by layer to make the two sets of point clouds fit as closely as possible.

[0086] Subsequently, statistical methods can be used to measure the degree of deviation between the two. Mahalanobis distance is superior to the simple Euclidean distance evaluation method because it fully considers the correlation and variability between variables. The final visualization uses the HSL color model to map the magnitude and direction of the deviation, where hue represents the state of stretching or compression, saturation reflects the severity of exceeding the standard, and brightness reflects the potential danger level. The resulting heat map intuitively shows the current overall health status of the entire structural system and the location of its weak points.

[0087] Step S703: Compare the structural performance offset heatmap with a preset safety threshold. When the structural performance offset exceeds the preset safety threshold, generate a construction correction list. The construction correction list includes the coordinates of the components to be reinforced and the material replacement type. Specifically, once performance deviations in certain areas are found to exceed predetermined safety thresholds, a construction correction list is generated. This list includes the coordinates of components requiring reinforcement and the operational procedures for material replacement. At this point, the system activates a knowledge-driven auxiliary decision-making mechanism that combines a rule engine with case-based reasoning. On one hand, semantic matching can be performed based on the specific structural characteristics of the object to be addressed (such as component type, cross-sectional dimensions, and connection type) to derive preliminary feasible repair suggestions. On the other hand, successful experiences accumulated from similar past projects can be referenced, especially solutions that have experienced the same failure modes or faced similar geological and climatic backgrounds, from which best practice templates can be extracted for reference.

[0088] Step S704: Activate the preset emergency response protocol library based on the construction correction list, and match the corresponding engineering machinery dispatch instructions and personnel evacuation route planning.

[0089] The process involves determining the types and quantities of machinery and equipment needed based on the listed remedial measures, such as whether specialized tools like aerial work platforms, concrete pumping equipment, or welding robots are required. Then, an improved A* search algorithm is used to assign optimal evacuation routes. This optimization objective function comprehensively considers both transportation efficiency and safety, and specifically incorporates dynamic risk field modeling to mitigate the impact of high-risk areas such as areas with active crack propagation or areas filled with high-temperature smoke.

[0090] At the same time, it is also necessary to consider potential interference and conflicts with other operating construction units to ensure that the robotic arm's trajectory does not collide with existing structures and cause secondary damage. For the part involving the transfer of people, a complete evacuation network topology is built based on graph theory, and Dijkstra's shortest path algorithm is used to calculate the travel cost between each escape exit and the assembly point. The weight assignment fully reflects the effects of multiple influencing factors such as passage width restrictions, fire spread rate estimation, and real-time population density distribution.

[0091] The above implementation method achieves full-cycle, refined closed-loop monitoring and dynamic intervention of the structural reinforcement process during the construction phase. It can not only collect and process deformation and stress response data of the reinforced structure at the microscale in real time, but also accurately compare the measured values ​​with the design expectations through spatial registration and statistical analysis. It generates a structural performance offset heat map that integrates mechanical state and risk level, and then automatically triggers the construction correction mechanism based on the preset safety threshold to form a construction correction list that includes defect location and material optimization suggestions. It also links with the emergency response protocol library to realize intelligent scheduling of mechanical equipment and dynamic planning of personnel evacuation routes. While ensuring construction safety and structural reliability, it significantly improves on-site response efficiency and collaborative management level.

[0092] This application also discloses an intelligent building auxiliary design system.

[0093] An intelligent building auxiliary design system, specifically comprising: The data receiving module is used to receive building foundation parameters and scene labels; The Building Information Modeling (BIM) module is used to build three-dimensional BIM models based on basic building parameters. The monitoring optimization module is used to generate monitoring strategies by parsing the preset technical contradiction matrix based on scene labels, and to mark the sensor deployment coordinates and risk weights in the 3D building information model, and output the enhanced building information model and monitoring strategy file; The sensor adjustment module is used to acquire real-time environmental data, dynamically adjust sensor deployment parameters in conjunction with monitoring strategy files, perform self-calibration operations, and output sensor network topology diagram and real-time data stream after calibration. The data fusion processing module is used to receive real-time data streams after calibration and external environmental data, perform spatiotemporal coordinate alignment and noise filtering, and generate a spatiotemporal fusion database. The risk warning module is used to predict risk trends and dynamically calculate warning thresholds based on risk weights in a spatiotemporal fusion database and an enhanced building information model, and outputs a risk warning instruction set and structural damage report. The decision module is used to simulate the mechanical distribution of different reinforcement schemes and calculate the cost-benefit ratio based on the risk warning instruction set, structural damage report and pre-configured user decision constraint parameters, and generate an interactive decision panel and recommended scheme report; The feedback optimization module is used to receive user feedback ratings and early warning accuracy statistics on the recommended solution report through the interactive decision panel, and to update the deployment rules and parameter weights of the machine learning model in the monitoring strategy file.

[0094] The intelligent building auxiliary design system of this application embodiment can implement any of the above-mentioned intelligent building auxiliary design methods, and the specific working process of each module in the intelligent building auxiliary design system can refer to the corresponding process in the above-mentioned method embodiment.

[0095] In the several embodiments provided in this application, it should be understood that the provided methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for example, the division of a certain module is merely a logical functional division, and in actual implementation there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0096] This application also discloses a computer device.

[0097] A computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement an intelligent building-aided design method as described above.

[0098] This application also discloses a computer-readable storage medium.

[0099] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described above in any of the intelligent building aid design methods.

[0100] The computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device; the program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0101] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A method for intelligent building aided design, characterized by, The auxiliary design method comprises: receiving building foundation parameters and scene labels, and constructing a three-dimensional building information model based on the building foundation parameters; generating a monitoring strategy according to the preset technical contradiction matrix based on the scene label, and labeling sensor deployment coordinates and risk weights in the three-dimensional building information model, and outputting an enhanced building information model and a monitoring strategy file; obtain real-time environmental data, dynamically adjust sensor deployment parameters in combination with the monitoring strategy file, and execute self-calibration operations, output sensor network topology and calibrated real-time data flow; receive the calibrated real-time data flow and external environmental data, perform spatio-temporal coordinate alignment and noise filtering processing, and generate a spatio-temporal fusion database; based on the spatio-temporal fusion database and the risk weight in the enhanced building information model, the risk trend is predicted and the early warning threshold is dynamically calculated through the machine learning model, and the risk early warning instruction set and the structure damage report are output; according to the risk early warning instruction set, the structure damage report and the pre-configured user decision constraint parameter, the mechanical distribution of different reinforcement schemes is simulated and the cost benefit ratio is calculated, and the interactive decision panel and the recommended scheme report are generated; through the interactive decision panel, receive the feedback score of the user to the recommended scheme report and the early warning accuracy statistical result, update the deployment rule in the monitoring strategy file and the parameter weight of the machine learning model.

2. The intelligent building aided design method according to claim 1, wherein, The step of generating a monitoring strategy according to the preset technical contradiction matrix based on the scene label, and labeling sensor deployment coordinates and risk weights in the three-dimensional building information model, and outputting an enhanced building information model and a monitoring strategy file comprises: call a preset technical contradiction matrix library, match corresponding technical contradiction pairs according to the scene label, and generate an optimization scheme code set; analyze the optimization scheme code set to generate a sensor deployment rule set; wherein the sensor deployment rule set includes sensor type, deployment density and accuracy level; based on the three-dimensional building information model, calculate the optimal deployment coordinate set according to the sensor deployment rule set; in combination with the historical risk data corresponding to the scene label, label the risk weight distribution graph in the three-dimensional building information model; output the enhanced building information model and the monitoring strategy file; wherein the enhanced building information model integrates the optimal deployment coordinate set and the risk weight distribution graph, and the monitoring strategy file records the sensor deployment rule set.

3. The intelligent building aided design method according to claim 2, wherein, The step of obtaining real-time environmental data, dynamically adjusting sensor deployment parameters in combination with the monitoring strategy file, and executing self-calibration operations, outputting sensor network topology and calibrated real-time data flow comprises: receive the monitoring strategy file and real-time environmental data; wherein the monitoring strategy file includes a sensor deployment rule set, and the real-time environmental data includes vibration intensity, temperature and humidity value, and geographic position coordinates; analyze the dynamic adjustment conditions in the deployment rule set, and generate a sensor parameter update instruction when the real-time environmental data meets a preset trigger threshold; adjust the sensor deployment parameters according to the sensor parameter update instruction, and generate a sensor network topology; Call a self-calibration algorithm based on the temperature and humidity values in the real-time environment data to perform error compensation on the original data collected by the sensor, and output the calibrated real-time data stream; Store the sensor network topology graph and the calibrated real-time data stream in the distributed database.

4. The intelligent building aided design method according to claim 1, wherein, Based on the spatio-temporal fusion database and the risk weight in the enhanced building information model, the steps of risk trend prediction and dynamic calculation of warning threshold by machine learning model include: Read the multi-source monitoring data in the spatio-temporal fusion database and the risk weight distribution in the enhanced building information model; Extract the time series feature vector and environmental feature vector from the multi-source monitoring data; Input the time series feature vector, environmental feature vector and risk weight into the pre-trained machine learning model to output the structure damage probability value and risk trend prediction curve; Based on the risk trend prediction curve and environmental feature vector, dynamically calculate the warning threshold to generate a risk warning instruction set containing warning level, risk coordinate and trigger basis; According to the mapping relationship between the structure damage probability value and the spatial coordinate of the enhanced building information model, generate a structure damage report marking the high-risk area.

5. The intelligent building aided design method according to claim 4, wherein, According to the risk warning instruction set, structure damage report and pre-configured user decision constraint parameters, simulate the mechanical distribution of different reinforcement schemes and calculate the cost-benefit ratio to generate an interactive decision panel and a recommended scheme report. The steps include: Receive the risk warning instruction set, structure damage report and user decision constraint parameters; wherein the user decision constraint parameters include budget upper limit, construction period requirement and material preference; Parse the high-risk area coordinate and stress distribution data in the structure damage report to generate a finite element analysis input model; According to the type of the high-risk area, call the pre-defined reinforcement scheme library to activate at least two candidate reinforcement schemes; wherein the candidate reinforcement schemes include carbon fiber reinforcement, steel support or concrete replacement; Based on the finite element analysis input model, perform mechanical simulation calculation to output the deformation control value and stress peak value of each candidate reinforcement scheme at the bearing node; Based on the deformation control value and stress peak value of each candidate reinforcement scheme at the bearing node, combine the pre-stored material market price library and the user decision constraint parameters to calculate the cost-benefit ratio of each candidate reinforcement scheme; Output the interactive decision panel and the recommended scheme report.

6. The intelligent building aided design method according to claim 1, wherein, Through the interactive decision panel, receive the user's feedback score and warning accuracy statistical result of the recommended scheme report, and update the deployment rules in the monitoring strategy file and the parameter weight of the machine learning model. The steps include: Receive the user's feedback score and warning accuracy statistical result of the recommended scheme report through the interactive decision panel; Parse the association mapping relationship between the user's feedback score and the warning accuracy statistical result to generate a multi-dimensional score matrix; Based on the multi-dimensional score matrix, analyze the rule defects of the monitoring strategy file, and update the deployment rules in the sensor deployment rule set; According to the false alarm period data in the warning accuracy statistical result, extract the feature vector as a new training sample set; Incrementally train the machine learning model using the new training sample set, and adjust the parameter weight distribution of the convolutional neural network layer and the LSTM layer. Output the updated monitoring strategy file and algorithm update log.

7. The intelligent building aided design method according to any one of claims 1 to 6, characterized in that, The method further comprises the following steps after generating the recommended scheme report: Receiving reinforcement structure deformation data and material stress data collected by sensors in the construction phase; Comparing the reinforcement structure deformation data with the design expected value of the corresponding coordinates in the enhanced building information model, and outputting a structure performance offset heat map; Comparing the structure performance offset heat map with the preset safety threshold, and generating a construction correction list when the structure performance offset exceeds the preset safety threshold; wherein the construction correction list includes reinforced component coordinates and material replacement types; Activating the preset emergency response protocol library based on the construction correction list, and matching corresponding engineering machinery dispatching instructions and personnel evacuation path planning.

8. An intelligent building aided design system, characterized by comprising: The auxiliary design system comprises: a data receiving module configured to receive building basic parameters and scene labels; a building information model construction module configured to construct a three-dimensional building information model based on the building basic parameters; a monitoring optimization module configured to generate a monitoring strategy according to the scene labels by analyzing a preset technical contradiction matrix, and label sensor deployment coordinates and risk weights in the three-dimensional building information model, and output an enhanced building information model and a monitoring strategy file; a sensor adjustment module configured to obtain real-time environmental data, dynamically adjust sensor deployment parameters in combination with the monitoring strategy file, and perform a self-calibration operation, and output a sensor network topology graph and calibrated real-time data stream; a data fusion processing module configured to receive the calibrated real-time data stream and external environmental data, perform time-space coordinate alignment and noise filtering processing, and generate a time-space fusion database; a risk early warning module configured to perform risk trend prediction and dynamically calculate a warning threshold based on the time-space fusion database and the risk weights in the enhanced building information model by using a machine learning model, and output a risk early warning instruction set and a structure damage report; a decision module configured to simulate mechanical distribution of different reinforcement schemes and calculate a cost-benefit ratio according to the risk early warning instruction set, the structure damage report, and pre-configured user decision constraint parameters, and generate an interactive decision panel and a recommended scheme report; a feedback optimization module configured to receive feedback scores of the recommended scheme report and early warning accuracy statistical results through the interactive decision panel, and update deployment rules in the monitoring strategy file and parameter weights of the machine learning model.

9. A computer device, characterized by: A computer program product comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any one of claims 1 to 7 when executing the program.

10. A computer-readable storage medium, characterized in that: A computer program product comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any one of claims 1 to 7 when executing the program.

Images