An ultra-high-rise building construction period data-driven integrated management and control method and system

Abstract

This invention discloses a data-driven integrated management and control method and system for the construction period of super high-rise buildings. By constructing a closed-loop management and control process—from spatiotemporal alignment and fusion of multi-source data, comprehensive state quantitative assessment, dynamic deduction of risk transmission paths, to intelligent resource scheduling and multi-task collaborative execution—it effectively solves the problems of lack of situational awareness and delayed risk response caused by data fragmentation in existing technologies. By generating multi-level instructions with priority ranking and linking multiple trades and intelligent equipment clusters, it breaks down the silo effect between various operational links, achieving proactive reshaping and rapid recovery of the on-site safety situation. The archiving of end-to-end data and the construction of a standardized knowledge base provide reusable strategy templates for risk management in similar subsequent projects, significantly improving the overall safety level and emergency response efficiency during the construction period of super high-rise buildings.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to a data-driven integrated management and control method and system for the construction period of super high-rise buildings. Background Technology

[0002] The construction cycle of super high-rise buildings is long, the working environment is complex and involves multiple trades working together. The traditional construction management mode mainly relies on manual on-site inspection and experience judgment, or adopts a decentralized management architecture in which each subsystem (such as environmental monitoring, equipment monitoring, personnel positioning, etc.) operates independently.

[0003] In existing technologies, although some projects have introduced sensor data acquisition, they often lack a unified data fusion mechanism, resulting in problems such as time asynchrony, format incompatibility and noise interference in multi-source heterogeneous data, making it difficult to form a real-time perception of the overall situation of the construction site. At the same time, risk identification is mostly based on static threshold alarms, lacking the ability to dynamically extrapolate the transmission path of risks in spatial and temporal dimensions, which makes resource scheduling and isolation measures often lag behind the occurrence of risks, and unable to achieve proactive, collaborative and precise control. Summary of the Invention

[0004] This invention aims to provide a data-driven integrated management and control method and system for the construction period of super high-rise buildings. It effectively solves the problems of lack of situational awareness and delayed risk response caused by data fragmentation in the prior art. It realizes the proactive prediction, accurate positioning and efficient handling of risks during the construction period of super high-rise buildings, and significantly improves the safety management and emergency response efficiency of the construction site.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a data-driven integrated management and control method for the construction period of super high-rise buildings, comprising: Real-time acquisition, timestamp alignment, anomaly cleaning, and normalization of multi-source heterogeneous data from the construction site are performed, and a spatiotemporal correlation matrix is ​​constructed. Based on the aforementioned spatiotemporal correlation matrix and normalized multidimensional features, a comprehensive state index is quantitatively calculated to assess the construction status and locate high-risk core areas. Based on the high-risk core area and the spatiotemporal correlation matrix, the risk transmission path is deduced, and multi-level control instructions with priority ranking are generated; Based on the multi-level control instructions and the risk transmission path, dynamic resource scheduling and physical isolation zone deployment are carried out; By executing the aforementioned multi-level control commands and through multi-task collaborative operations and the linkage of intelligent equipment clusters, the on-site safety situation can be reshaped; Archive all data across the entire chain and build a standardized knowledge base that includes typical cases and management strategy templates.

[0006] Preferably, the construction of the spatiotemporal correlation matrix includes: The raw data streams from devices with different sampling frequencies are timestamped using linear interpolation to generate a continuous dataset with a unified time index. After completing the timestamp alignment, outliers are identified using the mean and standard deviation within the sliding window, and the identified outliers are corrected using cubic spline interpolation. After correcting the abnormal data, the maximum-minimum normalization method is used to map all monitoring indicators to a dimensionless range of 0 to 1. After data normalization, the construction site is divided into grid cells, and the Pearson correlation coefficient between adjacent grid cells is calculated to construct the spatiotemporal correlation matrix.

[0007] Preferably, the location of the high-risk core area includes: Multiple independent risk factors are decoupled from the normalized multidimensional features, and the dynamic weight of each risk factor is calculated using the entropy weight method. The comprehensive state index is obtained by weighting and summing the dynamic weights of each risk factor with their corresponding normalized eigenvalues. After calculating the comprehensive state index, the deviation of each grid cell from the global average state is calculated. After calculating the deviation, the grid cell is combined with its centrality score in the spatiotemporal correlation matrix, and grid cells whose deviation exceeds a preset threshold are marked as high-risk core areas.

[0008] Preferably, the multi-level control instructions for generating priority sorting include: Based on the high-risk core area, a risk propagation dynamics model is constructed, in which the risk propagation rate is determined by the elements in the spatiotemporal correlation matrix. The risk propagation dynamics model is solved by numerical integration to obtain the risk propagation intensity distribution of each grid cell over a future period of time. The priority score for each control instruction is calculated based on the speed of risk transmission, the potential scale of loss, and the number of people affected. All control instructions to be executed are sorted according to the priority scores, and a set of non-conflicting cooperative instructions is generated.

[0009] Preferably, the dynamic resource scheduling and physical isolation zone deployment includes: Based on the risk intensity and operational area of ​​the high-risk core area, the required resource types and quantities are quantitatively calculated. By combining a vertical logistics network model, the optimal resource path is planned under multiple constraints, including travel time, energy consumption, and path risks. The boundaries of the physical isolation zone are determined based on the tangential direction of risk propagation, the width of the isolation zone is dynamically adjusted according to the speed of risk propagation, and mobile equipment is directed to be deployed.

[0010] Preferably, the execution of the multi-level control instructions, through multi-task collaborative operations and linkage with intelligent equipment clusters, reshapes the on-site safety situation, including: The tasks of each job type are divided into time slices, non-overlapping collaborative operation sequences are arranged, and synchronization control is achieved through a unified clock signal; Activate surrounding smart surveillance cameras, dust suppression sprinkler systems, and emergency lighting equipment, enabling them to respond autonomously based on preset logic rules; Evacuation route guidance based on risk simulation results is disseminated to workers through a broadcasting system and personnel positioning wristbands; Real-time monitoring of key indicator changes; when the overall situation score remains within the safe range and the intensity of the risk source drops below the threshold, it is determined that the on-site situation has been reshaped.

[0011] Preferably, the construction of a standardized knowledge base containing typical cases and management strategy templates includes: The entire data chain is cleaned, and ontology-based methods are used to transform unstructured logs into structured records. Based on structured records, unsupervised learning algorithms are used to extract feature labels for event type, environmental conditions, involved devices, and response strategies. Events with similar characteristic labels are clustered into typical cases. Based on successful cases, management strategy templates containing adjustable parameters are extracted and stored as configurable rule sets.

[0012] Preferably, when constructing the spatiotemporal correlation matrix, for continuous missing data caused by network interruption, a probability density function is generated based on the data distribution pattern under the same historical period and working conditions, and random sampling is performed to fill the missing interval.

[0013] Preferably, when quantifying the comprehensive state index, a hyperbolic tangent function is introduced to perform a nonlinear transformation on the normalized index of the asymmetric distribution, so as to enhance the sensitivity of the critical region.

[0014] On the other hand, this invention proposes a data-driven integrated management and control system for the construction period of super high-rise buildings, comprising: The data preprocessing module is used to collect multi-source heterogeneous data from the construction site in real time, perform timestamp alignment, anomaly cleaning, normalization processing, and construct a spatiotemporal correlation matrix. The status assessment module is used to quantitatively calculate the comprehensive status index based on the spatiotemporal correlation matrix and normalized multidimensional features to assess the construction status and locate high-risk core areas. The risk simulation module is used to simulate the risk transmission path based on the high-risk core area and the spatiotemporal correlation matrix, and generate multi-level control instructions with priority ranking. The resource scheduling module is used to perform dynamic resource scheduling and physical isolation zone deployment based on the multi-level control instructions and the risk transmission path. The collaborative execution module executes the multi-level control commands and reshapes the on-site safety situation through multi-task collaborative operations and linkage with intelligent equipment clusters. The knowledge accumulation module is used to archive data across the entire chain and build a standardized knowledge base containing typical cases and management strategy templates.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention effectively solves the problems of lack of situational awareness and delayed risk response caused by data fragmentation in existing technologies by constructing a closed-loop management process that integrates multi-source data spatiotemporal alignment and fusion, comprehensive state quantitative assessment, dynamic deduction of risk transmission paths, intelligent resource scheduling, and multi-task collaborative execution. Specifically, by real-time cleaning of multi-source heterogeneous data and construction of spatiotemporal correlation matrices, time deviations and format barriers between different sensors are eliminated, achieving accurate insight into the overall state of the construction site. Based on this dynamically deduced risk transmission path, the system can predict the spread trend of risks in advance, thereby transforming traditional remediation into pre-emptive prevention and guiding resources to be deployed at the optimal time and along the optimal path. At the same time, by generating multi-level instructions with priority ranking and linking multiple trades and intelligent equipment clusters, the silo effect between various work links is broken, achieving proactive reshaping and rapid recovery of the on-site safety situation. Finally, the archiving of the entire data chain and the construction of a standardized knowledge base not only accumulate valuable engineering experience but also provide reusable strategy templates for risk management in similar projects in the future, significantly improving the overall safety level and emergency response efficiency during the construction of super high-rise buildings. Attached Figure Description

[0016] Figure 1 This is a flowchart of the data-driven integrated management and control method for the construction period of super high-rise buildings according to the present invention; Figure 2 This is a block diagram of the data-driven integrated management and control system for the construction period of super high-rise buildings according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a specific posture. If the specific posture changes, the directional indicators will also change accordingly.

[0019] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the use of "and / or" or "and / or" throughout the text includes three parallel solutions. For example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0020] Please see Figure 1 This invention provides a technical solution: a data-driven integrated management and control method for the construction period of super high-rise buildings, comprising: The process involves real-time acquisition, timestamp alignment, anomaly cleaning, and normalization of multi-source heterogeneous data from the construction site, followed by the construction of a spatiotemporal correlation matrix. Specifically, this includes: real-time timestamp alignment of raw data streams from devices with different sampling frequencies using linear interpolation to generate a continuous dataset with a unified time index; identifying outliers using the mean and standard deviation within a sliding window after timestamp alignment, and correcting these outliers using cubic spline interpolation; mapping all monitoring indicators to a dimensionless interval of 0 to 1 using the min-max normalization method after anomaly correction; and dividing the construction site into grid cells after data normalization, calculating the Pearson correlation coefficient between adjacent grid cells to construct the spatiotemporal correlation matrix.

[0021] When constructing the spatiotemporal correlation matrix, for continuous missing data caused by network interruption, a probability density function is generated based on the data distribution pattern under the same historical period and working conditions, and random sampling is performed to fill the missing interval.

[0022] By combining linear interpolation and cubic spline interpolation, the time deviation caused by sampling frequency differences among multi-source devices is effectively eliminated and outlier noise is smoothed, ensuring the continuity and consistency of the data stream. By combining max-min normalization and a probability density function filling mechanism based on historical patterns, not only is the problem of long-term data loss caused by network interruptions solved, but the integrity and statistical representativeness of the data are also significantly improved, laying a solid and reliable data foundation for the subsequent construction of a high-precision spatiotemporal correlation matrix and accurate risk situation perception.

[0023] Based on the spatiotemporal correlation matrix and normalized multidimensional features, a comprehensive state index is quantitatively calculated to assess the construction status and locate high-risk core areas. Specifically, this includes: decoupling multiple independent risk factors from the normalized multidimensional features and calculating the dynamic weight of each risk factor using the entropy weight method; weighting and summing the dynamic weights of each risk factor with their corresponding normalized feature values ​​to obtain the comprehensive state index; calculating the deviation of each grid cell from the global average state after calculating the comprehensive state index; and combining the deviation with the centrality score of the grid cell in the spatiotemporal correlation matrix to mark grid cells whose centrality scores exceed a preset threshold as high-risk core areas.

[0024] When quantifying the comprehensive state index, a hyperbolic tangent function is introduced to perform a nonlinear transformation on the normalized index with an asymmetric distribution, thereby enhancing the sensitivity of the critical region.

[0025] By combining the dynamic weighting method of entropy weighting with the synergistic application of hyperbolic tangent nonlinear transformation, the shortcomings of traditional static weights in adapting to changes in operating conditions and the insensitivity of linear assessment to critical risks are effectively overcome. This achieves adaptive quantification of multi-source risk factors and keen capture of minute abnormal fluctuations. At the same time, by combining the dual screening mechanism of state deviation and network centrality, it can accurately locate core risk areas with strong transmission and high destructive power, thereby transforming fuzzy overall assessment into precise risk positioning, providing a scientific and reliable decision-making basis for subsequent targeted resource scheduling and control command generation.

[0026] Based on the high-risk core area and the spatiotemporal correlation matrix, the risk transmission path is deduced, and multi-level control instructions with priority ranking are generated. Specifically, this includes: constructing a risk propagation dynamics model based on the high-risk core area, where the risk propagation rate is determined by the elements in the spatiotemporal correlation matrix; solving the risk propagation dynamics model through numerical integration to obtain the risk propagation intensity distribution of each grid unit over a future period; calculating the priority score of each control instruction based on the risk propagation speed, potential loss scale, and number of affected persons; ranking all control instructions to be executed according to the priority scores and generating a set of non-conflicting collaborative instructions.

[0027] By constructing a risk propagation dynamics model based on a spatiotemporal correlation matrix, the diffusion trend and intensity distribution of risks at the construction site can be accurately quantified, thereby transforming passive response into proactive prediction. Combining comprehensive consideration of propagation speed, loss scale, and personnel impact, the system can intelligently generate multi-level collaborative control instructions with priority ranking and no conflicts, effectively solving the problems of chaotic resource scheduling and delayed response under complex working conditions, ensuring that key risk points are dealt with with the highest priority, and significantly improving the scientific nature of emergency response and the overall efficiency of on-site management.

[0028] Based on multi-level control instructions and risk transmission paths, dynamic resource scheduling and physical isolation zone deployment are carried out. Specifically, this includes: quantitatively calculating the required resource types and quantities based on the risk intensity and operational area of ​​high-risk core areas; planning the optimal resource path under multiple constraints, including travel time, energy consumption, and path risks, in conjunction with a vertical logistics network model; determining the boundary of the physical isolation zone based on the tangential direction of risk propagation; dynamically adjusting the width of the isolation zone according to the risk propagation speed; and directing mobile equipment for deployment.

[0029] By quantifying risk intensity and accurately matching resource needs with the work area, and combining the optimal path planning with a multi-constraint vertical logistics model, the timeliness and economy of material allocation are unified. At the same time, the boundaries and width of the physical isolation zone are dynamically adjusted according to the risk propagation dynamics, so that the protective measures can adaptively expand or contract in real time with the evolution of the risk situation, effectively blocking the risk diffusion path. Under the premise of ensuring construction safety, the cost loss caused by ineffective isolation is greatly reduced, and a flexible, efficient and rapid-response on-site defense system is constructed.

[0030] Execute multi-level control commands and reshape the on-site safety situation through multi-job collaborative operations and intelligent equipment cluster linkage. Specifically, this includes: dividing the tasks of each job into time slices, arranging non-overlapping collaborative operation sequences, and achieving synchronous control through a unified clock signal; activating surrounding intelligent monitoring cameras, sprinkler dust suppression systems, and emergency lighting equipment, enabling them to respond autonomously based on preset logic rules; issuing evacuation route guidance based on risk simulation results to workers through a broadcast system and personnel positioning wristbands; and monitoring changes in key indicators in real time. When the comprehensive situation score is continuously maintained within the safe range and the intensity of the risk source drops below the threshold, it is determined that the on-site situation has been reshaped.

[0031] By using time-slicing and a unified clock synchronization mechanism, timing conflicts and waiting losses in multi-task collaborative operations are eliminated, significantly improving on-site execution efficiency. The autonomous linkage response of the intelligent equipment cluster has achieved a leap from passive monitoring to active defense, greatly shortening emergency response delays. Combined with precise evacuation guidance based on risk simulation and a dynamic situation assessment closed loop, it not only effectively avoids secondary disasters and accidental injuries, but also quickly verifies and confirms the substantial restoration of the safety status after the source of risk is under control, thereby building an efficient, orderly, and self-verifying on-site safety reshaping system.

[0032] The entire data chain is archived to build a standardized knowledge base containing typical cases and management strategy templates. Specifically, this includes: cleaning the entire data chain and using ontology methods to transform unstructured logs into structured records; based on the structured records, using unsupervised learning algorithms to extract feature tags for event types, environmental conditions, involved devices, and response strategies; clustering events with similar feature tags into typical cases; extracting management strategy templates containing adjustable parameters based on successful cases and storing them as configurable rule sets.

[0033] Through ontology transformation and unsupervised learning mining, we have achieved deep cleaning and feature reconstruction of massive unstructured historical data, effectively breaking down data silos and establishing a standardized knowledge expression system. The resulting typical case library and configurable strategy templates not only accumulate valuable practical experience, but also support rapid rule matching and adaptive adjustment in new scenarios, significantly reducing the cost of repetitive decision-making and providing continuously evolving data-driven support for intelligent prediction and precise control of similar risks in the future.

[0034] On the other hand, this invention proposes a data-driven integrated management and control system for the construction period of super high-rise buildings, such as... Figure 2 As shown, it includes: The data preprocessing module is used to collect multi-source heterogeneous data from the construction site in real time, perform timestamp alignment, anomaly cleaning, normalization processing, and construct a spatiotemporal correlation matrix. The status assessment module is used to quantitatively calculate the comprehensive status index based on the spatiotemporal correlation matrix and normalized multidimensional features to assess the construction status and locate high-risk core areas. The risk simulation module is used to simulate the risk transmission path based on the high-risk core area and the spatiotemporal correlation matrix, and generate multi-level control instructions with priority ranking. The resource scheduling module is used to perform dynamic resource scheduling and physical isolation zone deployment based on the multi-level control instructions and the risk transmission path. The collaborative execution module executes the multi-level control commands and reshapes the on-site safety situation through multi-task collaborative operations and linkage with intelligent equipment clusters. The knowledge accumulation module is used to archive data across the entire chain and build a standardized knowledge base containing typical cases and management strategy templates.

[0035] Specifically, in this embodiment, the modules described above are also used to implement other steps of the aforementioned data-driven integrated management and control method for the construction period of super high-rise buildings, as follows: Step 1: Real-time perception and standardized preprocessing of multi-source heterogeneous construction site data; During the construction of super high-rise buildings, construction sites are characterized by large spatial spans, complex working environments, numerous participating entities, and extremely high data generation frequency. To establish an effective management and control foundation, the primary task is to collect and clean raw data from various sensor networks, manual data entry terminals, and external meteorological monitoring systems. This step aims to establish a unified data access framework, converging multi-source heterogeneous data scattered across tower crane monitoring, elevator operation status, environmental monitoring instruments, personnel positioning tags, and video analysis equipment to a central processing node, and eliminating data fragmentation caused by differences in sampling frequencies and inconsistent communication protocols. Specifically, this includes: 1.1 Concurrent acquisition and timestamp alignment of multi-source heterogeneous data streams; First, IoT gateways deployed at key nodes on the construction site establish connection channels to various sensing devices. These devices include torque limiters for monitoring tower crane lifting weight, amplitude, and slewing angle; sensors for recording the operating speed and load status of construction hoists; and environmental monitoring stations for detecting PM2.5, noise, wind speed, and wind direction. Simultaneously, indoor positioning base stations based on UWB technology acquire the location coordinates of workers, and receive structural deformation visual feature vectors extracted by high-definition cameras via an edge computing gateway. Because the sampling frequencies of different devices vary significantly—for example, environmental monitoring data might be collected once per minute, while tower crane attitude data might be collected ten times per second—direct splicing would lead to time series misalignment. Therefore, the system uses linear interpolation to complete the low-frequency data, ensuring its time granularity is consistent with the high-frequency data. Let the data stream of the i-th type of device collected at time t be... Its original sampling interval is The target uniform sampling interval is (Usually, the least common multiple or the interval of the highest frequency equipment is taken). For any point in time... If there are no original observations for a given point, then the observations from two adjacent points will be used. and (satisfy The weighted average of the values ​​is estimated. The specific calculation formula is as follows: ; in, This indicates that the i-th type of data after interpolation is in Standardized values ​​of time, and These represent the actual measurements at adjacent time points. This formula ensures that all data streams are strictly synchronized in the time dimension, eliminating time-series deviations caused by different acquisition frequencies and providing a unified time benchmark for subsequent multidimensional data fusion. After this step, the originally chaotic raw data stream is transformed into a continuous dataset with a unified time index, enabling subsequent analysis to be carried out based on the same time slice.

[0036] 1.2. Identification of constant data and dynamic imputation of missing values; After data time alignment, outliers caused by equipment malfunction, signal interference, or human error must be removed, and data segments lost during transmission must be repaired. Given the strong winds characteristic of high-rise buildings, wind speed sensors are prone to momentary fluctuations; directly using such data in calculations would severely distort structural stress assessment results. The system employs a statistically based dynamic threshold mechanism, utilizing the mean and standard deviation within a sliding window to define the normal fluctuation range. Let the data set within the sliding window at the current time t be... Its sample mean is The sample standard deviation is If the data point at a certain moment Meet the conditions ,in The confidence coefficient, typically set to 3, indicates that the data point is an outlier and is marked. Marked outliers are not discarded directly but corrected based on contextual information. If valid data exists before and after the data point, cubic spline interpolation is used for smooth replacement; if the point is at the beginning or end of the data sequence, the value from the previous valid time step is used.

[0037] In addition, for continuous missing data caused by network interruption, the system generates a probability density function f(x) based on the data distribution pattern under the same historical period and working conditions, and randomly samples and fills the missing intervals to maintain the integrity of the data distribution.

[0038] 1.3. Normalization and dimensional unification of multimodal data; Even after cleaning, the data retains its individual physical dimensions, such as displacement in millimeters, stress in megapascals, and temperature in degrees Celsius. This makes it impossible to directly perform mathematical calculations or comparisons between data from different dimensions. To achieve deep fusion of multi-source data, all indicators must be mapped to a unified dimensionless space.

[0039] This step uses the max-min normalization method to compress the values ​​of each indicator to a range of 0 to 1. For the j-th monitoring indicator, let its maximum value within the current time window be... The minimum value is The original observation value is Normalized values The calculation is as follows: ; This formula converts physical quantities of different orders of magnitude into comparable relative proportions, eliminating the influence of unit systems. However, for some asymmetrically distributed indicators, such as concrete strength growth curves or settlement rates, simple linear normalization may lead to excessive information compression. In such cases, a nonlinear transformation using the hyperbolic tangent function is introduced as a supplementary strategy. This enhances the sensitivity of data in critical regions. Through the above normalization process, all monitoring indicators are transformed into scalar values ​​under the same dimension, forming the basic elements of the high-dimensional feature vector. This step not only solves the problem of data heterogeneity but also lays the mathematical foundation for subsequent feature weight allocation and comprehensive state evaluation, enabling physical phenomena of different properties to be quantitatively described within the same logical framework.

[0040] 1.4 Data feature extraction and spatiotemporal correlation matrix construction; After data cleaning and normalization, it is necessary to extract key features reflecting the construction status of super high-rise buildings from massive time-series data. These features not only include numerical changes at single points but also involve spatial correlations between multiple points. The system constructs a spatiotemporal correlation matrix, dividing the construction site into several grid cells, each corresponding to a specific construction area and equipment cluster. Let the feature vector of the m-th grid cell at the n-th time step be... This vector consists of normalized readings from all sensors within the region. To capture spatial coupling effects, the Pearson correlation coefficient between adjacent grid cells is calculated. The correlation coefficient is used to characterize the similarity of states in different regions at the same time. The formula for calculating the correlation coefficient is: ; in, and , representing the characteristic mean values ​​of grids m and k over N time steps, respectively. This coefficient reflects the degree of synchronization between the two regions in terms of construction load, environmental disturbance, etc. When A value close to 1 indicates that both regions are significantly affected by the same external factors (such as strong winds and seismic waves); a value close to -1 indicates that they exhibit opposite response trends. Based on this, the system constructs an M×M spatiotemporal correlation matrix R, where M is the total number of grid cells. The elements in the matrix not only include static spatial distance weights but also incorporate dynamic correlation coefficients, thus forming a topological structure that can automatically evolve with the construction progress. This matrix transforms discrete single-point data into a holistic field with spatial structure, revealing the intrinsic connections between different height levels and different work surfaces of the super high-rise building, providing structured input for the next step of comprehensive state quantification.

[0041] Step 2: Quantitative assessment of construction status based on multi-dimensional feature fusion; After completing data perception, cleaning, normalization, and spatiotemporal correlation construction, the system enters the in-depth analysis stage of the overall status of the super high-rise building construction site. The core of this step lies in transforming the massive feature data and spatiotemporal correlation matrix generated in the previous step into comprehensive evaluation indicators that can intuitively reflect the construction safety, quality, and progress. By introducing a weighted fusion algorithm, the single-dimensional monitoring data is upgraded to a multi-dimensional status profile, thereby achieving a precise quantitative description of the complex construction environment, specifically including: 2.1 Decoupling of multi-dimensional risk factors and adaptive weight allocation; To accurately assess the current construction status, it is first necessary to decouple the independent risk factors affecting safety from the complex feature vectors. These factors include, but are not limited to, structural tilt, tower crane load rate, falling object risk index, environmental severity, and frequency of personnel violations. Since the impact of each risk factor on overall safety dynamically changes with the construction progress—for example, structural tilt is the primary concern during the main structure construction phase, while falling object and fire risks are more prominent during the finishing phase—the system employs the entropy weight method to dynamically calculate the weights of each factor, abandoning the traditional practice of manually assigning fixed weights. Let the normalized eigenvalue of the i-th risk factor at the j-th time step be... Its entropy value during that time period The calculation is as follows: ; in, Let be the feature proportion of the i-th factor at time j, i.e. K represents the total number of factors. A smaller entropy value indicates greater fluctuation in the factor, providing more information, and therefore a higher weight should be assigned. The weight of the i-th factor is then calculated. : ; This formula allows the system to automatically adjust the contribution of each risk factor based on real-time data fluctuations. When a risk factor experiences drastic fluctuations, its entropy decreases, and its weight automatically increases, thus gaining greater attention in the subsequent comprehensive scoring. This adaptive mechanism ensures that the assessment results always focus on the most pressing safety hazards, avoiding misjudgments caused by fixed weights. After weight allocation, each risk factor is assigned a dynamic importance that aligns with the current construction scenario, providing a scientific basis for subsequent comprehensive scoring.

[0042] 2.2 Calculation and classification of comprehensive state index; After obtaining the dynamic weights of each risk factor, the system performs a weighted summation with the corresponding feature values ​​to calculate the comprehensive state index at the current moment. This index, a value between 0 and 100, represents the overall health and safety level at the construction site of a super high-rise building. The calculation formula is as follows: ; in, Let i be the dynamic weight of the i-th factor at time t. This is the normalized eigenvalue of the factor. The safety sensitivity coefficient for this factor is pre-set by an expert database; for example, the structural stability coefficient is set to 1.5, and the environmental noise coefficient is set to 0.8. This formula integrates multi-dimensional risk factors into a single scalar value, greatly simplifying the logic of state judgment. Subsequently, the system applies the preset threshold range... Perform a classification determination. If... If it is rated as excellent, it means that the construction is in an ideal state; if The condition was assessed as good, with minor but manageable potential risks; if If it is judged to be moderate, monitoring needs to be strengthened; if If the condition is poor, it is judged as "poor" and an emergency warning is triggered. This hierarchical system transforms abstract numerical values ​​into intuitive level labels, enabling managers to quickly grasp the on-site situation. Through the output of the comprehensive status index, the system achieves a quantitative summary of the complex construction environment, providing clear guidance for subsequent decision-making.

[0043] 2.3 Evolution analysis of spatiotemporal correlation matrix and local anomaly localization; While the overall state index reflects the overall situation, it cannot pinpoint specific areas of anomaly. Therefore, it is necessary to use the spatiotemporal correlation matrix R constructed in step one for backtracking analysis to locate the specific sources causing the overall state decline. The system calculates the deviation of each grid cell from the global average state. And combined with its centrality score in the association matrix. Deviation is defined as the difference between the weighted average of all risk factors within the grid and the overall average. Centrality score, on the other hand, is derived from the Laplace spectral decomposition of matrix R, reflecting the grid's pivotal role in the entire construction network. If a certain grid's... Larger and A higher value indicates that the area is not only the source of the anomaly, but its anomaly is also highly likely to spread to the surrounding areas. The specific judgment logic is as follows: If and ,in and If an empirical threshold is used, the grid is marked as a high-risk core area. The system further tracks the spatiotemporal propagation path of this core area and analyzes the trajectory of its impact range over time. In this way, the system not only points out where the problem occurred but also reveals how the problem propagates spatially. This positioning capability enables control measures to be precisely targeted to specific areas, avoiding the waste of resources caused by blind investigations and achieving a leap from macro-level overview to micro-level positioning.

[0044] 2.4 Historical Status Comparison and Trend Extrapolation Prediction; To determine whether the current state is an occasional fluctuation or the beginning of a long-term deterioration trend, the system introduces a historical state comparison mechanism. This mechanism compares the current comprehensive state index... The numerical sequence of the data and its constituent factors is compared with historical data over the past T periods. A baseline trend line is constructed by calculating the moving average MA(t) and the exponentially smoothed value ES(t). If the current value remains below the baseline by a certain margin and the slope diverges negatively, the trend is considered to be deteriorating. The prediction model uses linear regression to fit the state trend over the next H time steps, and the prediction formula is: ; Where h is the prediction step size, a is the trend slope, and b is the intercept, which are obtained by the least squares method based on the most recent N historical data points. If the predicted value... If the price falls below the warning line in the near future, a trend warning will be issued in advance. This step extends the static assessment to the dynamic future, making management more proactive. By comparing historical data, the system can distinguish between normal construction fluctuations and genuine risk accumulation, avoiding false alarms caused by short-term disturbances. The trend extrapolation results provide a time window for subsequent resource allocation and contingency plan activation, ensuring that intervention measures are taken before the risk actually erupts, thereby guaranteeing a smooth and orderly construction process.

[0045] Step 3: Risk transmission path deduction and control strategy generation based on state quantification; After completing a comprehensive quantitative assessment of the construction site status and identifying the sources and trends of anomalies, the system enters the crucial stage of risk transmission simulation and control strategy generation. This step aims to simulate the propagation mechanism of different risk factors in a spatiotemporal network, deduce the evolution path of potential accidents, and generate targeted control instructions accordingly. By constructing a risk propagation dynamics model, the system can predict the scope and speed of risk diffusion, thereby formulating the optimal containment plan and realizing the transformation from passive response to proactive defense. Specifically, this includes: 3.1 Risk propagation dynamics modeling and path deduction; Based on the spatiotemporal correlation matrix and high-risk core area determined in step two, the system constructs a risk propagation dynamic equation to simulate the diffusion process of risk factors in the construction site network. It is assumed that risk propagation follows an SI model similar to an infectious disease outbreak, where the risk source is the infection source and the surrounding unaffected areas are susceptible zones. Let the risk propagation intensity of the m-th grid be... Its rate of change with time is determined by both the propagation rate of neighboring grids and its own attenuation rate. The propagation equation is expressed as: ; in, Denotes the neighborhood set of grid m. The risk coefficient for propagation from grid k to grid m is determined by the elements in the spatiotemporal correlation matrix R from step one. Decide, Let be the risk attenuation coefficient of grid m itself, representing the inherent risk resistance of the area. By solving this system of differential equations through numerical integration, a risk propagation intensity distribution map of each grid over a future period can be obtained. The simulation results show that risks do not spread uniformly, but rather form risk beams along highly correlated paths, rapidly impacting specific areas. For example, the risk of tower crane overturning may rapidly propagate vertically to the work surface below, while the risk of fire may spread along horizontal ventilation ducts. Through visualization of the simulation results, the system clearly displays the spatiotemporal trajectory of risk evolution and identifies secondary disaster areas that are about to be affected. This simulation capability enables managers to predict the consequences of accidents, plan evacuation routes and isolation zones in advance, and minimize casualties and property losses.

[0046] 3.2 Multi-level control threshold triggering and priority ranking; Based on the results of risk propagation simulations, the system dynamically adjusts the triggering conditions for control thresholds at each level and generates corresponding control instruction sets. Unlike fixed threshold settings, this system adopts a dynamic threshold mechanism, automatically adjusting according to the current construction stage and environmental conditions. Let the control instruction corresponding to the j-th level risk be... Its triggering condition is the risk intensity obtained through deduction. Exceeding the dynamic threshold The formula for calculating the dynamic threshold is: ; in, Based on the basic security threshold, These are environmental disturbance factors (such as wind speed and rainfall). and This is an adjustment coefficient. When the simulation shows that the risk in a certain area will be... When the threshold is exceeded at time t, the system generates an early warning instruction. Simultaneously, the system prioritizes all pending control instructions. The prioritization criteria include the speed of risk propagation, the potential scale of loss, and the number of people affected. Priority score. The calculation is as follows: ; in, To speed up the spread of risk, To estimate economic losses, For the number of people affected, This is a weighting coefficient. Higher-priority instructions will be issued and executed first, ensuring that limited human and material resources are invested in the most critical aspects. Through this dynamic adjustment and prioritization, the system optimizes the allocation of control resources, avoids instruction conflicts and resource idleness, and ensures the efficient implementation of control measures in complex and ever-changing environments.

[0047] 3.3 Collaborative generation and conflict resolution of multi-source control commands; Once the priorities of control instructions are determined, the system needs to translate these instructions into specific action guidelines, ensuring that there are no logical conflicts between them. For example, evacuation instructions may overlap with hoisting operation instructions in terms of time and space; lack of coordination will lead to chaos on site. The system employs a constraint satisfaction problem (CSP) solver, treating various control instructions as constraint variables, and the physical space of the construction site and operation time as constraint conditions. By searching the feasible solution space, a set of mutually non-conflicting coordinated instructions is generated. For unavoidable conflicts, the system resolves them according to priority rules, prioritizing personnel safety, then protecting critical equipment, and finally considering project progress.

[0048] For example, when a fire risk is detected, the system automatically suspends all hoisting operations and forcibly activates the fire sprinkler system, while simultaneously generating an evacuation route map. The instruction generation process considers actual site conditions, such as physical limitations like passageway width and equipment operating radius, ensuring the instructions' executability. Furthermore, the system refines the instructions into specific operational steps, including responsible personnel, execution time, and required materials, forming standardized work instructions. Through the collaborative generation of multi-source instructions, the system achieves seamless integration from macro-level strategy to micro-level operation, ensuring the accurate communication and execution of control intentions.

[0049] 3.4 Simulation verification of control strategies and preset feedback parameters; Before officially issuing control instructions, the system uses digital twin technology to simulate and verify the generated strategy, predicting its execution effect. By reproducing the 3D model and physical characteristics of the construction site in a virtual environment, the system simulates the changes in the site state after the implementation of control instructions. During the simulation, the system records the changing trends of various key indicators, such as the reduction in evacuation time, equipment downtime, and the speed of risk dissipation. If the simulation results show that the strategy has failed to effectively contain the risk or has created new safety hazards, the system will automatically roll back and readjust the parameters. At the same time, the system presets feedback parameters based on the simulation results, providing an initial state for the subsequent step four. The preset parameters include the expected state distribution after the instruction is executed, resource consumption prediction, and potential secondary risk points. The generation of preset parameters also lays the foundation for subsequent real-time monitoring and dynamic adjustment, forming a complete closed-loop logical chain from deduction to verification to execution.

[0050] Step 4: Dynamic resource scheduling and physical isolation deployment based on simulation results; After completing the deduction of risk transmission paths and the generation of control strategies, the system enters the crucial stage of translating abstract instructions into concrete physical actions. The core task of this step is to dynamically schedule manpower, machinery, and materials at the construction site in real time, based on the high-risk areas, transmission paths, and priority rankings determined in previous steps, and to implement precise physical isolation measures. By constructing a resource spatiotemporal matching model, it ensures that critical resources can accurately reach designated locations before the risk spreads, forming an effective defensive barrier and thus preventing further spread of the risk. Specifically, this includes: 4.1 Precise quantification of resource demand and spatiotemporal distribution mapping; Based on the control instruction set generated in step three and the projected risk propagation range, the system first precisely quantifies the required resources. For each grid cell marked as a "high-risk core area" or "secondary disaster area," the system calculates the required resource type, quantity, and arrival time. Let the resource demand vector of the k-th grid cell at time t be... This vector includes elements such as the number of personnel, types of protective equipment, number of firefighting equipment, and number of heavy machinery units. The demand calculation is based on risk intensity. The product of the risk consumption coefficient and the unit risk consumption coefficient is calculated using the following formula: ; in, The resource consumption coefficient per unit of risk intensity is determined by historical data and engineering experience; Let k be the working area of ​​grid k. Simultaneously, the system integrates the construction progress timeline, mapping resource requirements to specific spatiotemporal coordinates. Considering the complexity of vertical transportation in super high-rise buildings, the system establishes a vertical logistics network model, connecting ground storage yards, tower crane lifting paths, floor working surfaces, and refuge floors. For each type of resource, the system calculates the estimated time from its current location to the target location. The time taken is affected by the elevator's operating speed, the hoisting radius, and the degree of congestion along the route.

[0051] 4.2 Resource-optimal path planning under multiple constraints; After clarifying the resource requirements and spatiotemporal distribution, the system needs to solve the path planning problem for resource scheduling. Due to the narrow space, overlapping work areas, and numerous moving obstacles (such as other construction vehicles and temporary scaffolding) at the construction site, simple straight-line distance planning is often infeasible. The system uses an improved ant colony algorithm to construct a path search graph, dividing the construction site into several nodes, with edge weights representing the difficulty and time cost of passage. Let P be the set of paths from the starting point S to the ending point D, and let the total cost of each path be... It consists of travel time, energy consumption, and potential interference risks. The cost function is defined as: ; in, Let e ​​be the time it takes to travel along edge e. The energy consumption of this section of road. The risk factor for this road section (e.g., areas near hazardous sources have a higher risk factor). These are weight parameters. The system, while satisfying resource timeliness constraints, seeks to... The minimum optimal path. In particular, to address the vertical transportation bottleneck in super high-rise buildings, the system introduces a staggered scheduling mechanism. By utilizing the time window information in step three, the resource scheduling times for different areas are staggered to avoid congestion in narrow vertical passages.

[0052] 4.3 Dynamic Physical Isolation Zone Construction and Buffer Zone Settings; To physically block the spread of risk, the system automatically plans the deployment of physical safety barriers based on the predicted direction of the risk beams. Unlike traditional fixed barriers, this step emphasizes the dynamic and targeted nature of the safety barriers. The system utilizes drones or mobile robots to rapidly deploy flexible safety nets, water curtain systems, and audible and visual warning devices around high-risk areas. The shape and extent of the safety barriers are determined by the tangential direction of risk propagation, aiming to cut off the propagation path of risk factors. Let the area covered by the safety barriers be... Its boundary is determined by the direction vector of the fastest risk gradient descent. Confirmed. Width of the isolation strip. The minimum safe distance requirement must be met, and the calculation formula is as follows: ; in, To standardize the minimum safe distance, For safety redundancy coefficient, This formula represents the speed of risk propagation. Using this formula, the system can dynamically adjust the width of the safety barrier according to the speed of risk spread, ensuring an effective physical barrier is formed before the risk arrives. Simultaneously, the system includes a buffer zone inside the safety barrier to accommodate potentially leaked hazardous materials or serve as an emergency escape space. The safety barrier is deployed using a modular splicing method, ensuring rapid expansion and contraction. The system also monitors the integrity of the safety barrier in real time, immediately triggering a reinforcement command upon detecting damage or gaps.

[0053] 4.4 Real-time verification and anomaly correction of resource deployment status; When resources begin to be scheduled according to plan and isolation zones begin to be deployed, the system enters a real-time verification phase to ensure that the actual execution is consistent with the preset plan. Through IoT sensors and visual recognition technology, the system collects real-time data on the displacement and status of on-site resources, as well as the construction progress of isolation facilities. Let the actual deployment state vector be... The planned state vector is The system calculates the deviation vector between the two. If the deviation exceeds the allowable threshold The system immediately activates the anomaly correction mechanism.

[0054] For example, if a crane is delayed due to a malfunction, the system quickly reallocates nearby available backup equipment and updates the route plan; if the construction of the safety barrier is behind schedule, the system automatically allocates more manpower or adjusts the priority of subsequent tasks. The correction process uses feedback control logic, but is limited to single-step, immediate adjustments and does not involve long-term system iterations. The system records the cause of each deviation and the corrective measures, forming a log of on-site execution data.

[0055] Step 5: Three-dimensional coordinated execution of multi-dimensional control commands and reshaping of the on-site situation; After resources are deployed and physical isolation is established, the system enters the full implementation phase of control commands. This step focuses on integrating the scattered commands generated in the previous steps into a unified on-site action plan. Through human-machine collaboration, it aims to rapidly restore construction order and fundamentally reverse the safety situation. The system no longer focuses solely on the actions of individual devices but rather on the systemic reconstruction of the entire construction site. Through the coordinated efforts of multiple trades and equipment, it eliminates the sources of potential hazards and reshapes a safe and controllable construction environment. Specifically, this includes: 5.1. Multi-task collaborative operation sequence arrangement and synchronization control; The construction site involves multiple trades such as civil engineering, installation, and decoration, and during risk management, the actions of each trade must be highly coordinated to avoid mutual interference or even new conflicts. Based on the resource deployment plan and priority ranking determined in step four, the system finely arranges the tasks for each trade. Let the task sequence for the i-th trade be... The system uses time-slicing to arrange the operation periods of different jobs in a non-overlapping manner, ensuring that no more than two work groups will be in the same spatial area at the same time. The synchronization control mechanism uses high-precision clock signals to direct the start and stop actions of each job.

[0056] For example, when dismantling a damaged structure, the hoisting team is responsible for removing components, the welding team for reinforcing the surrounding area, and the cleaning team for removing debris. All three teams must coordinate with millisecond-level timing accuracy. The system issues synchronization commands via a wireless communication network, requiring each team's terminal to receive a unified timestamp signal to ensure consistency of actions. This collaborative orchestration avoids secondary damage caused by disordered procedures and improves overall work efficiency. Through refined sequence orchestration, the system transforms complex on-site operations into orderly assembly line operations, ensuring the smooth flow of control commands at the execution level.

[0057] 5.2 Intelligent device cluster linkage and automated processing; In addition to manual operation, the system deeply integrates various intelligent devices, making them the main force in executing control commands. After confirming a risk area, the system automatically activates surrounding intelligent monitoring cameras, dust suppression sprinkler systems, emergency lighting equipment, and automatic alarm devices. These devices form a distributed execution cluster that responds autonomously according to preset logical rules.

[0058] For example, when a temperature sensor detects an abnormally high temperature, the nearby sprinkler system automatically activates, the fire alarm broadcasts evacuation instructions, and the video surveillance automatically switches to the fire source's perspective. The linkage logic between the devices is based on Boolean algebra operations; that is, when condition A (high temperature) and condition B (smoke) are simultaneously met, action C (activating sprinklers and triggering an alarm) is executed. The system uses a central controller to manage all devices in a unified manner, ensuring they work collaboratively as a whole, rather than operating independently. Furthermore, for large mechanical equipment, the system implements a remote takeover mode, restricting its dangerous actions and forcing it into a safe standby state. This cluster linkage mechanism greatly improves response speed, reduces delays and errors caused by human operation, and enables a powerful automated defense network to be formed instantly on-site, effectively curbing further escalation of risks.

[0059] 5.3. Mechanism for guiding personnel behavior and psychological intervention; During risk management, panic among personnel can lead to disorder and exacerbate the consequences of an accident. Therefore, the system incorporates a dedicated process for personnel behavior guidance and psychological intervention. Through large screens, a broadcast system, and personnel positioning wristbands, the system provides clear evacuation routes and safety instructions to workers. The instructions are based on the simulation results from step three, avoiding risk transmission paths and selecting the safest routes. The system utilizes speech synthesis technology to deliver instructions in a calm and firm tone, avoiding words that could induce panic.

[0060] Meanwhile, the system monitors the behavioral characteristics of people, such as running speed and gathering density. If abnormal gatherings or disorderly running are detected, it immediately uses broadcasts to direct people out of the area. For the few people stranded in high-risk areas, the system automatically dispatches the nearest rescue team or guiding robot to assist them. This humane guidance mechanism not only ensures the physical safety of people but also stabilizes the psychological state on site, preventing secondary stampedes caused by panic.

[0061] 5.4 On-site situation reshaping and phased acceptance evaluation; With the full implementation of various control measures, the situation at the construction site has begun to undergo a fundamental change. The system continuously monitors changes in key indicators through real-time data streams to evaluate the effectiveness of the control measures. Let the current overall situation score at the site be... This score is based on the comprehensive state index calculation method in step two, but with a greater emphasis on risk elimination rate and order restoration rate. If the value remains in the excellent or good range for M consecutive time steps, and the intensity of the risk source is... When the alert level drops to zero or near zero, the system determines that the on-site situation has been successfully rebuilt. At this point, the system initiates a phased acceptance evaluation procedure to comprehensively verify the completed isolation facilities, resource deployment, and personnel evacuation. Acceptance criteria include the integrity of physical isolation, the stability of equipment operation, and personnel occupancy rates. Only when all indicators meet the acceptance criteria can the system allow the partial lifting of the alert level and gradually resume normal production activities. This process marks the transition of risk management from the emergency blocking phase to the order restoration phase, ensuring that the construction site can quickly return to normal after experiencing a risk impact, providing a safety guarantee for subsequent continuous construction.

[0062] Step Six: Manage the entire process of data archiving and build a standardized knowledge base; After completing the on-site situation reshaping and acceptance, the system's core task shifts to in-depth data mining and knowledge accumulation throughout the entire process. This step aims to transform the massive amounts of data, decision-making logic, execution results, and lessons learned from this management practice into structured knowledge assets and store them in a standardized knowledge base. Specifically, this includes: 6.1 End-to-end data cleaning and structured reassembly; First, the system comprehensively collects and organizes all raw data, intermediate calculation results, and final decision logs generated in steps one through five. This data covers the entire data chain, from sensor readings at the perception layer, model outputs at the calculation layer, device commands at the execution layer, to situational assessments at the result layer. Due to the diverse sources and formats of the data, the system first performs data cleaning to remove invalid noise and duplicate records. Then, it uses ontology-based methods to construct unified data, transforming unstructured text descriptions (such as field logs and voice recordings) into structured fields.

[0063] For example, a natural language description like "severe tower crane tilt" is mapped to a structured record such as "Equipment ID: T01, Risk Type: Tilt, Value: 0.05, Time: 2023-10-01 10:00". The reconstructed data is indexed according to timeline, spatial location, and event type, forming a vast time-series database. This structured processing method enables efficient retrieval and correlation analysis of historical data, providing a solid data foundation for subsequent pattern recognition and case matching.

[0064] 6.2 Construction of a typical case library and extraction of feature labels; Based on the reorganized end-to-end data, the system begins constructing a typical case library. Each case corresponds to a complete risk management event, encompassing four dimensions: cause, development, handling, and outcome. The system utilizes unsupervised learning algorithms to extract key feature tags from massive amounts of cases, such as tower crane overturning due to strong winds, illegal electricity use during nighttime construction, and crack propagation during concrete curing. The tag extraction process is based on text mining and pattern matching techniques, automatically identifying the core elements of the event. Each case is assigned a set of multi-dimensional tag vectors to characterize its uniqueness. For example, case... The label vector is These represent the event type, environmental conditions, involved equipment, and response strategy, respectively. In this way, the system can cluster events of similar nature together, forming representative case clusters. The establishment of this typical case library means that future risk management will no longer rely on the experience of individual experts, but can instead utilize a standardized case library for rapid matching and reference, greatly improving the ability to handle complex operating conditions.

[0065] 6.3. Management strategy templates and parameter generalization; Based on the case library, the system further refines general management strategy templates. By analyzing the execution process of multiple successful cases, the system summarizes the most effective resource allocation models, buffer zone deployment schemes, and collaborative operation processes under different risk scenarios. These strategies are solidified into standardized templates, including a series of adjustable parameters.

[0066] For example, to address the risk of falling objects from heights, the system generates a standard operating procedure template that includes setting a warning radius, planning evacuation routes, and reinforcing the top safety net. The parameters in the template (such as the warning radius and the proportion of evacuees) are not fixed values, but rather generalized ranges derived from statistical patterns, adaptable to building projects of different sizes and heights. The system stores these templates as configurable rule sets for future use. When encountering new events, the system can automatically match the most suitable strategy template based on event characteristics and fine-tune parameters according to the actual situation, thereby achieving rapid response. This strategy template approach transforms tacit experience into explicit knowledge, reducing reliance on external experts and improving the system's autonomous decision-making capabilities.

[0067] 6.4 Knowledge base version management and standardized interface encapsulation; Finally, the system implements version management and interface encapsulation for the constructed knowledge base, ensuring its long-term availability and interoperability. Each time a case is added to the knowledge base and a strategy is updated, the system generates a new version number and records a change log for easy traceability and rollback. The knowledge base uses standardized data exchange formats (such as JSON or XML) and provides services externally through well-defined API interfaces. Other subsystems or external applications can query specific types of cases, obtain strategy templates, or download historical statistics through these interfaces. This open architecture breaks down information silos and promotes knowledge sharing and circulation.

[0068] Meanwhile, the system regularly conducts integrity checks and consistency verifications of the knowledge base to ensure the accuracy of stored data and that policy templates comply with the latest specifications. Through standardized version management and interface encapsulation, the system has built a sustainably evolving knowledge ecosystem, providing continuous intellectual support for data-driven integrated management and control during the construction phase of super high-rise buildings, ensuring continuous improvement in management and control capabilities and long-term stable operation.

[0069] The above description is merely an exemplary embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention specification and drawings under the technical concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A data-driven integrated management and control method for the construction period of super high-rise buildings, characterized in that, include: Real-time acquisition, timestamp alignment, anomaly cleaning, and normalization of multi-source heterogeneous data from the construction site are performed, and a spatiotemporal correlation matrix is ​​constructed. Based on the aforementioned spatiotemporal correlation matrix and normalized multidimensional features, a comprehensive state index is quantitatively calculated to assess the construction status and locate high-risk core areas. Based on the high-risk core area and the spatiotemporal correlation matrix, the risk transmission path is deduced, and multi-level control instructions with priority ranking are generated; Based on the multi-level control instructions and the risk transmission path, dynamic resource scheduling and physical isolation zone deployment are carried out; By executing the aforementioned multi-level control commands and through multi-task collaborative operations and the linkage of intelligent equipment clusters, the on-site safety situation can be reshaped; Archive all data across the entire chain and build a standardized knowledge base that includes typical cases and management strategy templates.

2. The data-driven integrated management and control method for the construction period of super high-rise buildings according to claim 1, characterized in that, The construction of the spatiotemporal correlation matrix includes: The raw data streams from devices with different sampling frequencies are timestamped using linear interpolation to generate a continuous dataset with a unified time index. After completing the timestamp alignment, outliers are identified using the mean and standard deviation within the sliding window, and the identified outliers are corrected using cubic spline interpolation. After correcting the abnormal data, the maximum-minimum normalization method is used to map all monitoring indicators to a dimensionless range of 0 to 1. After data normalization, the construction site is divided into grid cells, and the Pearson correlation coefficient between adjacent grid cells is calculated to construct the spatiotemporal correlation matrix.

3. The data-driven integrated management and control method for the construction period of super high-rise buildings according to claim 1, characterized in that, The high-risk core area to be located includes: Multiple independent risk factors are decoupled from the normalized multidimensional features, and the dynamic weight of each risk factor is calculated using the entropy weight method. The comprehensive state index is obtained by weighting and summing the dynamic weights of each risk factor with their corresponding normalized eigenvalues. After calculating the comprehensive state index, the deviation of each grid cell from the global average state is calculated. After calculating the deviation, the grid cell is combined with its centrality score in the spatiotemporal correlation matrix, and grid cells whose deviation exceeds a preset threshold are marked as high-risk core areas.

4. The data-driven integrated management and control method for the construction period of super high-rise buildings according to claim 1, characterized in that, The multi-level control instructions for generating priority sorting include: Based on the high-risk core area, a risk propagation dynamics model is constructed, in which the risk propagation rate is determined by the elements in the spatiotemporal correlation matrix. The risk propagation dynamics model is solved by numerical integration to obtain the risk propagation intensity distribution of each grid cell over a future period of time. The priority score for each control instruction is calculated based on the speed of risk transmission, the potential scale of loss, and the number of people affected. All control instructions to be executed are sorted according to the priority scores, and a set of non-conflicting cooperative instructions is generated.

5. The data-driven integrated management and control method for the construction period of super high-rise buildings according to claim 1, characterized in that, The dynamic resource scheduling and physical isolation zone deployment includes: Based on the risk intensity and operational area of ​​the high-risk core area, the required resource types and quantities are quantitatively calculated. By combining a vertical logistics network model, the optimal resource path is planned under multiple constraints, including travel time, energy consumption, and path risks. The boundaries of the physical isolation zone are determined based on the tangential direction of risk propagation, the width of the isolation zone is dynamically adjusted according to the speed of risk propagation, and mobile equipment is directed to be deployed.

6. The data-driven integrated management and control method for the construction period of super high-rise buildings according to claim 1, characterized in that, The execution of the multi-level control commands, through multi-task collaborative operations and the linkage of intelligent equipment clusters, reshapes the on-site safety situation, including: The tasks of each job type are divided into time slices, non-overlapping collaborative operation sequences are arranged, and synchronization control is achieved through a unified clock signal; Activate surrounding smart surveillance cameras, dust suppression sprinkler systems, and emergency lighting equipment, enabling them to respond autonomously based on preset logic rules; Evacuation route guidance based on risk simulation results is disseminated to workers through a broadcasting system and personnel positioning wristbands; Real-time monitoring of key indicator changes; when the overall situation score remains within the safe range and the intensity of the risk source drops below the threshold, it is determined that the on-site situation has been reshaped.

7. The data-driven integrated management and control method for the construction period of super high-rise buildings according to claim 1, characterized in that, The construction of a standardized knowledge base containing typical cases and management strategy templates includes: The entire data chain is cleaned, and ontology-based methods are used to transform unstructured logs into structured records. Based on structured records, unsupervised learning algorithms are used to extract feature labels for event type, environmental conditions, involved devices, and response strategies. Events with similar characteristic labels are clustered into typical cases. Based on successful cases, management strategy templates containing adjustable parameters are extracted and stored as configurable rule sets.

8. The data-driven integrated management and control method for the construction period of a super high-rise building according to claim 2, characterized in that, When constructing the spatiotemporal correlation matrix, for continuous missing data caused by network interruption, a probability density function is generated based on the data distribution pattern under the same historical period and working conditions, and random sampling is performed to fill the missing interval.

9. A data-driven integrated management and control method for the construction period of a super high-rise building according to claim 3, characterized in that, When quantifying the comprehensive state index, a hyperbolic tangent function is introduced to perform a nonlinear transformation on the normalized index with an asymmetric distribution, thereby enhancing the sensitivity of the critical region.

10. A data-driven integrated management and control system for the construction period of super high-rise buildings, used to implement the method as described in any one of claims 1-9, characterized in that, include: The data preprocessing module is used to collect multi-source heterogeneous data from the construction site in real time, perform timestamp alignment, anomaly cleaning, normalization processing, and construct a spatiotemporal correlation matrix. The status assessment module is used to quantitatively calculate the comprehensive status index based on the spatiotemporal correlation matrix and normalized multidimensional features to assess the construction status and locate high-risk core areas. The risk simulation module is used to simulate the risk transmission path based on the high-risk core area and the spatiotemporal correlation matrix, and generate multi-level control instructions with priority ranking. The resource scheduling module is used to perform dynamic resource scheduling and physical isolation zone deployment based on the multi-level control instructions and the risk transmission path. The collaborative execution module executes the multi-level control commands and reshapes the on-site safety situation through multi-task collaborative operations and linkage with intelligent equipment clusters. The knowledge accumulation module is used to archive data across the entire chain and build a standardized knowledge base containing typical cases and management strategy templates.

Images