Construction safety risk early warning system and early warning method thereof

Abstract

The application relates to the technical field of risk early warning, in particular to a construction safety risk early warning system, which comprises a data acquisition module, which is used for acquiring historical monitoring data of a plurality of monitoring points in a continuous time period in a construction site; a trend analysis module, which is connected with the data acquisition module, is used for calculating change trend information of at least one monitoring parameter of each monitoring point in the continuous time period based on the historical monitoring data; a rule base, which pre-stores a judgment rule corresponding to at least one risk early warning scene, and the judgment rule defines a logical combination relationship between a plurality of monitoring parameters and the change trend information related to the risk early warning scene; and an early warning judgment module, which is connected with the trend analysis module and the rule base respectively, is used for matching the currently calculated change trend information with the judgment rule in the rule base, and outputting early warning information of a corresponding grade according to a matching result. The application improves the timeliness and foresight of risk early warning.

Description

Technical Field

[0001] This application relates to the technical field of risk warning, and in particular to a construction safety risk warning system and its warning method. Background Technology

[0002] With the development of information technology, construction safety management increasingly relies on various intelligent monitoring systems. These systems deploy sensors and cameras at the construction site to collect various parameters in real time, such as displacement, settlement, and images, and compare them with preset safety thresholds. This triggers alarms when parameters are abnormal, prompting managers to take timely action and ensure construction safety.

[0003] Currently, common construction site monitoring systems typically employ threshold alarm mechanisms. For example, in a foundation pit monitoring system, displacement sensors are installed on the support structure. The system continuously acquires displacement data and directly compares the real-time displacement value with a preset fixed threshold (e.g., 50 mm). Once the real-time displacement value exceeds this fixed threshold, the system immediately issues an alarm signal.

[0004] However, this alarm mechanism based on a single, fixed threshold has significant limitations. The formation of construction safety risks is often a gradual process involving the interaction of multiple parameters. For example, the risk of a foundation pit landslide may manifest itself several days before significant displacement occurs, through subtle, continuously increasing displacement rates, combined with other factors such as rainfall and surrounding loads. Existing threshold-based alarm methods only focus on whether a parameter exceeds a critical point at a certain moment, essentially a "post-event" or "late-stage" judgment. They cannot comprehensively analyze the early, subtle trend changes of multiple related parameters, leading to a lag in risk warnings and hindering the acquisition of valuable lead time for risk intervention. Summary of the Invention

[0005] To address at least one of the aforementioned technical problems, this application provides a construction safety risk early warning system and its early warning method.

[0006] Firstly, this application provides a construction safety risk early warning system, which adopts the following technical solution: The data acquisition module is used to acquire historical monitoring data from multiple monitoring points at the construction site over a continuous time period; the historical monitoring data includes multi-source monitoring data collected from the same monitoring object by sensors with at least two different measurement principles; A data fusion preprocessing module, connected to the data acquisition module, is used to perform confidence-weighted fusion processing on multi-source monitoring data of the same monitoring object to obtain fused monitoring data. The trend analysis module, connected to the data acquisition module, is used to calculate the change trend information of at least one monitoring parameter of each monitoring point within the continuous time period based on the historical monitoring data. The rule base pre-stores at least one judgment rule corresponding to a risk warning scenario. The judgment rule defines the logical combination relationship between multiple monitoring parameters and their changing trend information related to the risk warning scenario. The early warning judgment module is connected to the trend analysis module and the rule base respectively. It is used to match the currently calculated trend information with the judgment rules in the rule base and output the corresponding level of early warning information according to the matching result. The BIM data collaboration module is connected to the early warning judgment module. It is used to obtain the corresponding construction status or design parameter information from the Building Information Modeling (BIM) platform based on the monitoring point location information associated with the early warning information, and to output the early warning information in association with the construction status or design parameter information.

[0007] By adopting the above technical solution, and by adding a trend analysis module and a rule base with pre-stored multi-parameter logical combination judgment rules, the traditional alarm mode based on single threshold comparison has been changed. It can comprehensively analyze the historical trends of multiple parameters and match them with preset risk scenario rules, achieving a qualitative leap from single-point threshold triggering to multi-dimensional trend assessment, significantly improving the timeliness and foresight of risk warnings.

[0008] This system, based on the traditional early warning architecture, introduces a data fusion preprocessing module and a BIM data collaboration module. The former, through confidence-weighted fusion processing, ensures the quality and anti-interference capability of input data from the source; the latter, by associating with engineering context information from the BIM platform, upgrades isolated alarms to collaborative diagnosis. These two modules fundamentally improve the system from the "input end" and "output end" respectively, together forming a complete technical solution that further enhances the accuracy, timeliness, and decision support capabilities of early warnings, achieving a qualitative leap from passive alarms to proactive diagnosis.

[0009] In one possible implementation, the trend information includes the rate of change and / or acceleration of change of the monitored parameters.

[0010] By adopting the above technical solution, the dimension of early warning judgment is expanded from the parameter value itself to its rate of change and acceleration. The rate of change is the first derivative, and the acceleration is the second derivative. This enables the system to more sensitively capture early signals of accelerating risk development, thereby enhancing the sensitivity and accuracy of early warning.

[0011] Secondly, this application provides a construction safety risk early warning method, comprising the following steps: Acquire historical monitoring data from multiple monitoring points at the construction site over a continuous time period; The confidence-weighted fusion process is performed on the multi-source monitoring data of the same monitoring object to obtain monitoring data for trend analysis. Specifically, this includes dynamically evaluating the confidence weight of each data source based on at least one dimension, including the device's own status, the degree of environmental interference, and the consistency of historical data; and weighting and fusing the multi-source monitoring data according to the confidence weight to obtain monitoring data for trend analysis. Based on the historical monitoring data, analyze and calculate the change trend information of at least one monitoring parameter of each monitoring point within the continuous time period; The calculated trend information is matched with pre-stored judgment rules; wherein, the judgment rules define the logical combination relationship between multiple monitoring parameters and their trend information under different risk warning scenarios. Based on the matching results, output the warning information corresponding to the risk warning scenario and level; In addition to outputting the warning information, it also includes, Based on the monitoring point location information associated with the early warning information, the corresponding construction status or design parameter information is obtained from the Building Information Modeling (BIM) platform, and the early warning information is associated with the construction status or design parameter information and output; specifically, Based on the pre-stored mapping relationship between monitoring points and BIM components, determine the BIM component identifier corresponding to the monitoring point that triggers the early warning; The system initiates a query request carrying the BIM component identifier to the BIM platform through the application programming interface (API) to obtain at least one of the design parameters and construction status information associated with the component. The warning information is associated and encapsulated with the acquired information to generate a collaborative diagnostic report and output it.

[0012] By adopting the above technical solution, and through the steps of acquisition, analysis, matching, and output, a core method for achieving progressive early warning is defined, utilizing historical data for trend analysis and matching combination rules. This method has clear steps and solves the problem of delayed early warning.

[0013] This technical solution effectively resists data noise or failure caused by environmental interference from a single sensor by weighting the data with confidence levels from sensors based on different principles. This ensures that the data quality of the input trend analysis module is higher and more reliable, thereby improving the accuracy and anti-interference ability of the final early warning conclusion.

[0014] Meanwhile, by integrating the early warning method with the BIM platform, it not only outputs early warnings but also proactively links and pushes background information from the BIM, such as design loads and construction phases. This provides managers with comprehensive decision support that includes spatial location and project context, achieving a leap from isolated alarms to collaborative information diagnosis and greatly improving the efficiency and accuracy of risk management.

[0015] In one possible implementation, prior to the step of outputting the warning information, the following is also included: The historical monitoring data and the calculated trend information are stored for subsequent trend analysis.

[0016] By adopting the above technical solutions, the continuity of historical data is ensured, providing a data foundation for long-term trend analysis and backtracking, making trend analysis more reliable, and early warning judgments more based on long-term data patterns.

[0017] In one possible implementation, the output warning information includes: When the matching result meets the low-level risk conditions, a risk attention prompt message is output; When the matching result meets the high-level risk conditions, a risk alarm message is output.

[0018] By adopting the above technical solution, a tiered output mechanism for early warning information is clarified, optimizing the simple binary judgment of "alarm / no alarm" into a progressive risk alert system of attention, warning, and alarm. This aligns with the tiered response principle in risk management, avoiding fatigue caused by frequent strong alarms while ensuring sufficient attention is paid when risks escalate.

[0019] In one possible implementation, the step of acquiring historical monitoring data may further include data from mobile monitoring devices or monitoring nodes temporarily deployed at new work sites.

[0020] By adopting the above technical solution, the monitoring data source is extended from fixed points to mobile and temporary nodes, enabling the method to acquire data from traditional monitoring blind spots or dynamic operation areas. This greatly enriches the original data dimensions used for trend analysis, making the coverage of risk assessment more comprehensive, and is especially suitable for tracking dynamic risk processes such as mobile device operations.

[0021] In one possible implementation, the method further includes a reconfiguration step: In response to changes in instructions during the construction phase, the monitoring parameter types and corresponding decision rules for the specified monitoring nodes are reconfigured.

[0022] By adopting the above technical solutions, the same set of hardware can be reconfigured through software to undertake different monitoring tasks at different construction stages, such as switching from foundation pit displacement monitoring to scaffolding tilt monitoring. This can flexibly adapt to the entire life cycle of engineering projects with long spans and large stage differences, significantly improving the maintainability and long-term economic efficiency of the system.

[0023] Thirdly, this application provides an electronic device including a memory and a processor, wherein the memory is used to store computer program code, and the processor is used to execute the computer program code stored in the memory to implement the methods in the first aspect and any one of the first aspects, or in the second aspect and any possible implementation of the second aspect.

[0024] Fourthly, this application provides a computer-readable storage medium storing a computer program or instructions that, when executed, implement the methods described in the first aspect and any one thereof, or the second aspect and any possible implementation thereof. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating a construction safety risk early warning method provided in an embodiment of this application.

[0026] Figure 2 This is a structural schematic diagram of a construction safety risk early warning system provided in an embodiment of this application.

[0027] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0028] The technical solutions in this application will now be described with reference to all the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0029] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. "And / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Furthermore, in the description of the embodiments of this application, "plural" or "multiple" refers to two or more than two.

[0030] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this embodiment, unless otherwise stated, "a plurality of" means two or more.

[0031] The terminology used in the following embodiments is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to also include expressions such as “one or more,” unless the context clearly indicates otherwise. It should also be understood that in the following embodiments of this application, “at least one” and “one or more” refer to one, two, or more than two.

[0032] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "one embodiment," "some embodiments," "another embodiment," "other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0033] This application provides a construction safety risk early warning method, executed by an electronic device. This electronic device can be a standalone physical electronic device, a cluster of multiple physical electronic devices, a distributed system, or a cloud electronic device providing cloud computing services. This application does not impose limitations on this method. Figure 1 As shown, the method includes the following steps: S1. Obtain historical monitoring data from multiple monitoring points at the construction site over a continuous time period.

[0034] Specifically, to achieve this step, various types of monitoring equipment need to be deployed at key locations on the construction site.

[0035] In this embodiment, the hardware facilities include tilt sensors and displacement sensors installed on key parts such as the foundation pit support structure, the tall formwork support system, and the slope, for monitoring structural deformation; environmental sensors deployed at the construction site, such as rain gauges, anemometers, and soil moisture sensors; and fixed or mobile video acquisition equipment for acquiring visual information.

[0036] All sensors and data acquisition devices are connected to a data server at the construction site or in the cloud via wired or wireless communication modules. The data acquisition and storage module in the server continuously receives and stores raw data from all monitoring points at preset time intervals, such as once per minute, forming a structured historical monitoring database. This database records at least the timestamp, monitoring point ID, parameter type, and value for each data point.

[0037] Furthermore, data sources are not limited to fixed equipment.

[0038] In this embodiment, for example, images and vibration data of the high-altitude working surface can be obtained by means of a mobile monitoring node attached to the tower crane hook, or the soil pressure of the newly excavated area can be monitored by a wireless sensing unit that can be temporarily deployed.

[0039] Data from these mobile or temporary nodes is fed into the backbone network via ad-hoc networking technology and ultimately converges on the data server. This demonstrates the system's ability to monitor the network's dynamism and spatial separation, ensuring comprehensive spatiotemporal coverage of data collection.

[0040] When receiving data, the server performs preliminary cleaning and verification, such as filtering out obvious communication errors and aligning the timestamps of multi-source data for the same physical quantity, such as data from the same point from two different brands of displacement gauges, to provide a high-quality, multi-dimensional unified data source for subsequent analysis.

[0041] Based on this, through the coordination of the aforementioned hardware network and software logic, this step systematically and continuously captures panoramic historical information on the physical state of the construction site. This solves the problem of traditional systems having a single data source and potential data gaps, laying a solid and reliable data foundation for subsequent in-depth trend analysis. The acquisition of a large amount of continuous historical data is the fundamental prerequisite for achieving the leap from instantaneous point judgment to time series analysis, constituting the first significant advancement of this invention's method compared to simple threshold comparison methods.

[0042] In some embodiments, in order to make the risk assessment more comprehensive, the data acquired in S1 may also include data from mobile monitoring devices or monitoring nodes temporarily deployed at new work sites.

[0043] Specifically, to achieve this step, the system hardware configuration expands upon the traditional fixed monitoring network by adding mobile monitoring equipment and temporary monitoring nodes. The mobile monitoring equipment is typically an integrated module with a dustproof and waterproof casing. Internally, it integrates specific sensors such as high-definition cameras, vibration sensors, and tilt sensors; wireless communication modules; positioning modules such as GPS / BeiDou; and energy harvesting units such as rechargeable batteries or small solar panels. It can be easily installed on key parts of mobile construction machinery such as tower crane hooks, excavator arms, and concrete pump trucks, moving with the equipment.

[0044] The temporary monitoring nodes have a simpler structure, likely pre-packaged with miniature sensors such as crack gauges and earth pressure cells, as well as low-power wireless transmitters, for rapid deployment at newly emerging risk points such as newly excavated surfaces of foundation pits and temporary loading areas. These nodes are typically temporarily fixed using piles, magnetic attachment, or brackets.

[0045] In terms of software and communication logic, the system needs to support the access and management of dynamic, heterogeneous networks. The data acquisition and access module running on the central server needs to be able to identify and register newly joined mobile / temporary nodes. After powering on, the node sends a registration request to the central system via the wireless network, reporting its device ID, sensor type, current location, and other information.

[0046] Subsequently, it actively uploads monitoring data according to a preset cycle or triggering conditions such as vibration exceeding limits. Its data packet contains device ID, timestamp, sensor readings, and possible location information.

[0047] For areas with weak network edge signals, nodes can use a multi-hop relay mode to relay data through adjacent nodes to a fixed gateway. Furthermore, the system's monitoring network topology and sensor resources are no longer static, but can be dynamically adjusted and reconfigured according to changes in construction phases and work areas, achieving adaptive matching of monitoring capabilities with risk locations in both space and time.

[0048] Based on this, by integrating data from mobile and temporary nodes, the dimensions and spatial coverage of monitoring data have been expanded, resulting in significant progress.

[0049] Specifically, firstly, it directly overcomes the visual and sensor blind spots inherent in fixed monitoring systems. For example, cameras attached to tower crane hooks can acquire high-definition images of personnel activities and obstacles below the hoisting area in real time, a perspective that ground-based fixed cameras cannot achieve. Another example is the tilt sensor installed on the boom of an excavator, which can continuously monitor the stability of the boom during excavation operations and prevent tipping.

[0050] Secondly, it enables continuous monitoring of dynamic and ongoing operations. Traditional methods struggle to effectively track the entire operation of mobile machinery, but this embodiment perfectly resolves this challenge by attaching the monitoring device to the moving object itself. It captures changes in risk parameters under various working conditions, such as movement, rotation, and loading. This significantly enriches the input information for the trend analysis module, enabling the system to model and provide early warnings for risks that were previously difficult to quantify, such as collision risks with mobile equipment and temporary structural instability under dynamic loads. This fundamentally improves the comprehensiveness, timeliness, and accuracy of safety risk identification at construction sites.

[0051] In this embodiment, the method further includes: S2. Based on historical monitoring data, analyze and calculate the changing trend information of at least one monitoring parameter of each monitoring point over a continuous time period.

[0052] Specifically, this step is performed by the trend analysis module deployed on the server. This module retrieves all historical data sequences from the historical monitoring database, specifying the monitoring points and parameters, within a selected time window such as the most recent 24 hours or 7 days. Its core computing unit processes each data sequence according to a predetermined algorithm model.

[0053] In this embodiment, for example, for a data sequence of a displacement monitoring point, the module calculates its rate of change within a time window, i.e., the amount of displacement change per unit time, such as millimeters per hour, and the acceleration, i.e., how fast the rate of change itself changes. The calculation can be performed using the difference method or by fitting a simple curve to obtain the slope and curvature.

[0054] For environmental parameters such as rainfall, it is possible to calculate its cumulative amount and duration.

[0055] Furthermore, instead of performing calculations on a single latest data point in isolation, this trend analysis module places the current moment on a continuous time axis to examine the behavior patterns of parameter values ​​over time.

[0056] In this embodiment, for example, it is known not only that the current displacement is 30 mm, but more importantly, how this 30 mm was achieved, namely, whether it was a sudden increase of 10 mm in the past hour, i.e., a high rate of change, or whether it was a slow accumulation of 10 mm in the past three days, i.e., a low rate of change but continuous.

[0057] This analysis of process rather than state is one of the core essential characteristics of this method. The calculated trend information, such as the rate of change, acceleration, and continuous growth indicators, is stored or transmitted to the next module in real time as new, higher-dimensional feature data.

[0058] Based on this, by transforming raw monitoring values ​​into trend information that characterizes their dynamic evolution, the system achieves a quantitative assessment of the kinetic and potential energy of risk development. Even if the absolute value of a parameter has not yet reached the danger threshold, its rapid growth or accelerating change trend itself may indicate that potential risks are accumulating. This is equivalent to giving the system predictive capabilities, enabling it to identify characteristic signals of risks in their early, nascent stages. This step is a key technical link in transforming early warning actions from reactive response to proactive prevention, bringing significant progress.

[0059] In some embodiments, to ensure higher and more reliable data quality for the input trend analysis module, thereby improving the accuracy and anti-interference capability of the final early warning conclusion, the following is also included before S2: S201. Perform confidence-weighted fusion processing on the multi-source monitoring data of the same monitoring object to obtain monitoring data for trend analysis.

[0060] Specifically, to achieve this step, the system needs to be equipped with at least two sensors based on different measurement principles for the key monitoring objects at the hardware level, forming a multi-source sensing array. For example, to monitor the position of tower crane hooks, a high-definition visible light camera and millimeter-wave radar are deployed simultaneously; to monitor the horizontal displacement of deep foundation pits, a hydrostatic level and an inclinometer-based tilt sensor array are installed simultaneously; to monitor the safety status of personnel in critical areas, video surveillance and a positioning system based on UWB or Bluetooth beacons are set up simultaneously. The above sensors are connected to the field data acquisition gateway through their respective interfaces or protocol converters.

[0061] At the software level, a data fusion preprocessing module was added to the data server. This module first performs spatiotemporal alignment, synchronizing all incoming data streams with a unified high-precision time source, such as a GPS clock or Network Time Protocol (NTP). For spatially related data, it uses pre-defined coordinate transformation relationships to map the data to a unified two-dimensional or three-dimensional spatial coordinate system, such as images of the same area from different cameras.

[0062] Furthermore, after spatiotemporal alignment, the core task of the data fusion preprocessing module is to perform confidence assessment and weighted fusion.

[0063] The confidence level assessment is based on multiple dynamically updated dimensions. First, the device's own status, such as the health status of the sensor's self-test report, battery level, and communication signal strength. Second, environmental interference, for example, in nighttime, fog, or heavy rain, the confidence level of visible light cameras should be dynamically lowered, while the confidence level of millimeter-wave radar remains relatively stable. In areas with strong electromagnetic interference, the confidence level of specific types of electronic sensors is attenuated. Third, historical data consistency, which compares the current sensor reading with its own historical short-term predictions and with the degree of conformity with the current readings of other sensors at the same location. The greater the deviation of the data source, the lower its current confidence level is temporarily.

[0064] Based on these dimensions, the module calculates a dynamic confidence weight between 0 and 1 for each data source at each time step. Its weighted fusion algorithm employs a simple weighted average or a more robust algorithm, such as weighting after removing the lowest confidence data, ultimately outputting a fused, more representative monitoring data value for subsequent trend analysis.

[0065] By merging similar or auxiliary operations and pre-assessing data reliability to offset the impact of potential interference, a data awareness layer with intrinsic fault tolerance and adaptive capabilities is constructed.

[0066] Based on this, significant progress has been made in controlling the quality of data sources by implementing this multi-source data confidence weighted fusion process.

[0067] First, it can effectively resist the specific interference of the complex environment at the construction site on a single sensing technology. For example, when the camera image becomes blurry and the recognition position error increases due to dust, its confidence weight will automatically decrease, while the weight of radar data, which is not affected by dust, will increase accordingly. This ensures that the fused position data still maintains high accuracy and avoids misjudgment or data loss caused by the temporary failure of a single sensor.

[0068] Secondly, it enhances the robustness and continuity of the underlying data for trend analysis. Even if a sensor completely fails, as long as other data sources are available, the system can still adjust the weights and output usable fused data, ensuring the continuity of the monitoring data stream. This is crucial for trend analysis that relies on continuous historical data.

[0069] Ultimately, it purifies raw data from sensors based on different principles, which may contain noise or errors, into more reliable and consistent high-quality data. This directly and significantly improves the accuracy of subsequent trend analysis results, thereby making the early warning conclusions based on this more authoritative and credible. It greatly enhances the anti-interference capability and stable performance of the entire early warning system under harsh conditions.

[0070] In this embodiment, the method further includes: S3. Match the calculated trend information with the pre-stored judgment rules.

[0071] The judgment rules define the logical combination relationship between multiple monitoring parameters and their changing trend information under different risk warning scenarios.

[0072] Specifically, this step relies on a pre-generated risk warning rule base stored in a database. This rule base is constructed based on industry safety knowledge, accident case analysis, and expert experience. Each rule corresponds to a specific risk warning scenario, such as "foundation pit landslide risk" or "formwork support instability risk" in this embodiment. These rules are defined using a computer-executable logic language.

[0073] In this embodiment, for example, a rule may be defined as, "IF the displacement change rate of monitoring point A > R1 and the duration > T1, AND the cumulative rainfall during the same period > W1, THEN trigger 'foundation pit landslide - intermediate warning'".

[0074] The conditions in the rules not only include the instantaneous values ​​of the monitored parameters, but more importantly, they include information on their changing trends, such as the rate of change and duration, as well as logical combinations of relationships between different parameters, such as "AND" and "OR".

[0075] Furthermore, the early warning judgment module in the server acts as a rule matching engine, running periodically. It acquires the latest trend information generated by the trend analysis module and uses it as input to perform match-and-condition judgment against all rules in the rule base. This matching process is parallel and efficient.

[0076] Its essential feature is that the logical unit of judgment has been upgraded from the traditional single parameter and threshold comparison to a complex multi-parameter trend feature and logical rule matching.

[0077] Furthermore, the system no longer simply asks "Has the displacement exceeded the limit?", but comprehensively asks "Does the rate of increase and duration of displacement, combined with rainfall conditions, match the profile of a known risk pattern?" This enables the system to identify more complex, hidden, and multi-factor coupled risk precursors.

[0078] Based on this, the system achieves intelligent risk identification through this rule-based matching mechanism. It encodes and solidifies the experience and knowledge of human experts, enabling computers to automatically and continuously perform comprehensive assessments that previously required senior security engineers. This not only significantly improves the consistency and efficiency of risk identification, but more importantly, it makes early detection of multi-factor-related and progressive risks possible.

[0079] The rule base itself can be continuously expanded and optimized, enabling the system to continuously learn and evolve. This is the core logic behind the progressive early warning system of this invention, demonstrating significant intelligent advancements.

[0080] S4. Based on the matching results, output the warning information corresponding to the risk warning scenario and level.

[0081] Specifically, after the early warning judgment module completes the rule matching, it will execute the output action according to the THEN part of the triggering rule.

[0082] Its output interface includes both software and hardware components. On the software side, the system generates structured early warning messages, including the risk scenario name (e.g., "foundation pit landslide"), the warning level (e.g., "attention," "warning," "alarm"), the relevant monitoring points, the key triggering conditions, and time information. On the hardware side, these early warning messages are distributed to multiple terminals via a communication network. They are visually displayed on the central monitoring screen at the construction site as highlighted icons and pop-up windows; pushed to the mobile smart terminals of project safety managers, supervising engineers, and other relevant personnel via SMS or a dedicated application; and can automatically trigger on-site audible and visual alarms.

[0083] Furthermore, its early warning information is tiered and categorized.

[0084] In this embodiment, for example, a slight abnormality in the trend of only one parameter may trigger a low-level "risk concern" alert for management personnel to take note; if the combination of multiple parameters matches a clear risk rule, a medium-level "risk warning" will be triggered, requiring on-site verification; if the trend continues to deteriorate and matches a more serious rule, the highest-level "alarm" will be triggered, requiring immediate emergency measures to be taken.

[0085] This output method differentiates single alarm signals according to severity and urgency, changing the crude interaction of traditional systems that are "either silent or shrill," and providing a smoother, more instructive risk communication channel.

[0086] Based on this, by providing accurate, timely, and tiered early warning information, this invention ultimately transforms the technical value of the aforementioned steps—data collection, trend analysis, and rule matching—into practical safety management benefits. It enables managers to anticipate the evolution of risk situations and obtain different levels of intervention windows and decision-making basis before risks truly reach a critical point. This directly solves the fundamental problem of delayed early warning in the background technology, shifting the safety management model from post-event remediation to pre-event prevention.

[0087] The tiered output avoids alarm fatigue, ensures the seriousness and response priority of the highest level alarm, significantly improves the timeliness, scientific nature and operability of safety risk management at construction sites, and achieves remarkable progress in safety and economic benefits.

[0088] In some embodiments, to make trend analysis more reliable and early warning judgments more based on long-term data patterns, the following are included before S4: S401. Store historical monitoring data and calculated trend information for subsequent trend analysis.

[0089] Specifically, to achieve this step, the system needs to be equipped with high-capacity and highly reliable data storage facilities. Its core hardware consists of data servers deployed in a local server room or cloud service provider, along with their associated storage arrays, such as disk arrays or solid-state storage arrays. At the software level, data storage and management modules run on the data servers.

[0090] This module establishes and maintains two core databases: the raw monitoring database and the trend feature database.

[0091] Its raw monitoring database continuously receives and stores raw data streams uploaded in real time from each monitoring point. Each record includes a timestamp, a unique identifier for the monitoring point, the sensor type, and the raw measurement value.

[0092] Its trend feature database is specifically used to store the results periodically calculated and output by the trend analysis module. Each record includes a timestamp, the corresponding monitoring point and parameters, and the calculated trend information, such as the average rate of change, acceleration of change, and coefficients of the fitted curve over the past 24 hours.

[0093] The two databases are linked together by timestamps and monitoring point identifiers, forming a complete, timestamped historical data chain.

[0094] Furthermore, the key to implementing this step lies in building a long-term data warehouse system that is time-series oriented and supports efficient querying. Its data storage and management module not only performs simple write operations, but also implements data lifecycle management strategies.

[0095] In this embodiment, for example, raw data collected at a high frequency once per second is automatically aggregated into low-frequency statistical data such as hourly average, maximum, and minimum values ​​after being stored for a period of time such as 30 days, and then stored for a long time to balance storage costs and data availability.

[0096] Simultaneously, the module establishes multi-dimensional indexes, such as indexes by monitoring point ID, parameter type, and time range, ensuring that historical data across any time span can be quickly located and retrieved during subsequent trend analysis or manual backtracking queries. This reflects the continuous and effective use of the system's information resources, ensuring the sustained accumulation and usability of data value.

[0097] Therefore, by implementing this systematic data storage procedure, firstly, it ensures the robustness of the analytical foundation. The reliability of trend analysis highly depends on the continuity and completeness of historical data.

[0098] This implementation method, through dedicated hardware and software facilities, avoids data loss or interruptions caused by temporary equipment failures, network outages, or system restarts, providing a continuous and consistent data source for trend calculation.

[0099] Secondly, it enables deeper pattern discovery. Long-term, comprehensive data accumulation allows the system to transcend short-term fluctuations and identify more macroscopic and fundamental patterns. For example, by analyzing several months of data, the system may discover that the displacement of a certain support structure always exhibits a specific gradual change pattern after a particular rainfall pattern. This helps to optimize and refine the corresponding risk assessment rules and distinguish between normal seasonal deformation and abnormal precursors of danger.

[0100] Therefore, this step is a key support for upgrading the early warning system from relying on short-term fragment judgments to being based on long-term pattern cognition, significantly improving the scientific nature, accuracy, and foresight of early warning judgments.

[0101] In some embodiments, in order to avoid fatigue caused by frequent strong alarms while ensuring sufficient attention is paid when risks escalate, the warning information output in S4 includes: S402. When the matching result meets the low-level risk condition, output a risk concern prompt message.

[0102] S403. When the matching result meets the high-level risk conditions, output risk alarm information.

[0103] In order to achieve tiered early warning output, the system's early warning judgment module needs to have built-in tiered judgment criteria in its logic.

[0104] Specifically, at the hardware level, the system needs to connect to multiple output terminals with different warning intensities. For low-level "risk concern alerts," the output terminals mainly include the user interface of the project safety management software, a large display screen in the central monitoring room, and a dedicated application on the smartphones or tablets of relevant personnel.

[0105] For high-level "risk alarm information", in addition to the above, the output terminal must also include audible and visual alarms deployed in key areas of the construction site, such as the edge of the foundation pit, the tower crane control room, and the project management office, and can be connected to the broadcast system.

[0106] At the software level, after completing rule matching, the early warning judgment module not only determines whether an early warning is triggered, but also generates early warning message packages with different levels and content structures based on predefined level identifiers such as "attention level", "early warning level" and "alarm level" in the rule base.

[0107] Furthermore, the generation and distribution of early warning information follow differentiated communication logic and presentation formats. Specifically, when a low-level risk condition is matched, the risk attention alert information generated by the system typically includes a brief risk description, the relevant monitoring points, and a summary of parameter trends. This information is pushed through the internal network in a non-blocking and non-forced manner. On the software interface, it may highlight relevant monitoring point icons with a specific color such as yellow, or display text prompts scrolling in the information sidebar, accompanied by a gentle alert sound effect; while messages pushed to mobile terminals use a standard notification style, without forced pop-ups or continuous ringing.

[0108] When a high-level risk condition is matched, the system generates a risk alarm message that includes a clear risk type, specific location, severity, and recommended handling measures. Its distribution is mandatory and high-priority, immediately triggering the audible and visual alarms in the corresponding area. On the software interface, a prominent red pop-up occupies the visual center and continuously emits a rapid alarm sound; when pushed to mobile terminals, it uses the highest-priority message channel, resulting in repeated ringing, strong vibration, and a screen lock pop-up until confirmation is received.

[0109] By dividing the single alarm function according to its degree and dynamically changing the system's response intensity with the risk level, a hierarchical and matched information output system is constructed.

[0110] Based on this, by implementing this strictly differentiated and logically clear hierarchical output mechanism, significant progress has been made in human-computer interaction and risk management effectiveness.

[0111] First, it effectively solves the alarm fatigue problem caused by the traditional binary alarm mode. For a large number of early, low-level risk signs that have not yet posed a direct threat but require attention, the system notifies them in a non-intrusive manner, fulfilling its obligation to inform while avoiding frequent interruptions and interference to the daily work of managers, thus maintaining their sensitivity and trust in the system's alarms.

[0112] Secondly, it ensures efficient and authoritative response in truly critical and high-level situations. When the system triggers a high-level alarm, it conveys information through multiple channels, in a high-intensity and mandatory manner, which can alert all relevant personnel on site in the most immediate and unmistakable way, thus buying valuable time for activating emergency plans, organizing evacuations, or carrying out rescue operations.

[0113] This progressive system of monitoring, early warning, and alarm elevates risk communication from a rough assessment of presence or absence to a refined level of communication, making safety management decisions more hierarchical and scientific, and significantly improving the practicality, acceptability, and ultimate safety assurance effectiveness of the entire early warning system.

[0114] In some embodiments, in order to achieve the leap from isolated alarms to collaborative information-based diagnosis and greatly improve the efficiency and accuracy of risk handling, in addition to S4, the following are also included: S404. Based on the location information of the monitoring points associated with the early warning information, obtain the corresponding construction status or design parameter information from the Building Information Modeling (BIM) platform, and associate the early warning information with the construction status or design parameter information for output.

[0115] Specifically, to achieve this step, the system's hardware architecture must ensure that the early warning system server can establish high-speed and stable network communication with an independent BIM platform server or cloud-based BIM services. At the software level, a BIM data collaboration module needs to be added to the early warning system. One of the core tasks of this module is to establish and maintain a mapping table between monitoring points and BIM components.

[0116] This table is generated during system initialization or monitoring point deployment. It explicitly records the unique identifier of one or more logical components corresponding to each physical monitoring point, such as the sensor device ID, in the BIM 3D spatial model, such as the component's GUID.

[0117] For example, the displacement gauge deployed on the capping beam of the south side support pile of the foundation pit of Building 1 will be associated with the component named "Building 1 - Foundation Pit Support System - South Side - Capping Beam - Section 3" in the BIM model.

[0118] When the early warning judgment module generates an early warning message, the message already contains the monitoring point ID that triggered the early warning, and the BIM data collaboration module then finds the corresponding BIM component identifier according to the mapping table.

[0119] Furthermore, based on the acquired BIM component identifiers, the BIM data collaboration module initiates a data query request to the BIM platform through a predefined application programming interface. This request carries the target component identifier and the required information type.

[0120] The BIM platform then responds to the request, retrieving and returning multidimensional background information related to the component that is currently available from its database. This information may include: 1) design parameters, such as the design bending strength, design allowable displacement, and concrete strength grade of the support pile; 2) construction status, obtained from the progress management module associated with BIM, such as whether the area is currently planned to be in the "earthwork excavation stage" or the "foundation slab pouring stage"; 3) related component information, such as the status of components adjacent to the warning point and the current planned construction load value; and 4) historical change information, such as whether the part has undergone design changes.

[0121] After receiving the above information, the BIM data collaboration module does not simply list it, but intelligently associates and encapsulates it with the original early warning information such as risk type, level, and trend data to generate a structured data package called "Collaborative Diagnosis Report", which is then pushed and displayed through the unified output interface of the early warning system.

[0122] Based on this, by implementing a deeply integrated BIM information association output process, its risk decision support capabilities are enhanced.

[0123] First, it fundamentally changes the value chain of safety early warning. Traditional alarms only indicate that there is a problem at a certain location, while the collaborative diagnostic report output by this solution further explains "what this problem means in the engineering context".

[0124] For example, the system not only alerts that "the foundation pit displacement exceeds the limit," but also displays that "the design allowable displacement at this location is 50mm, currently it is 52mm, and this area is in the most unfavorable stage of earthwork excavation, with a temporary shovel area designed around it." This provides managers with extremely critical decision-making context.

[0125] Secondly, it significantly improves the efficiency and accuracy of risk assessment. Managers no longer need to manually query and compare information across multiple discrete systems, such as monitoring systems, BIM models, drawings, and schedules; instead, all key decision-making elements are automatically aggregated and presented in a correlated manner. This avoids misjudgments and delays caused by incomplete information or time-consuming queries, making the formulation of response measures such as reinforcement, unloading, and evacuation more precise and rapid.

[0126] This transforms the early warning system from a passive signaler into a proactive diagnostic assistant, achieving deep integration and collaboration between safety management and engineering management information.

[0127] In some embodiments, to improve the maintainability and long-term economic efficiency of the system, the method further includes a reconfiguration step: S5. In response to changes in the construction phase, reconfigure the monitoring parameter types and corresponding judgment rules for the specified monitoring nodes.

[0128] Specifically, to achieve this step, the system's hardware architecture needs to support functional reconfigurability. The core lies in deploying intelligent monitoring nodes that can be reconfigured in the field. These nodes employ a modular design, typically consisting of two parts: a fixed core base containing a core processor, wireless communication module, power management unit, and standard electrical interfaces; and a pluggable functional cartridge integrating specific types of sensors, such as tilt sensor modules, visual acquisition modules, vibration sensor modules, and gas detection modules. Users can replace the functional cartridge as needed, much like changing a camera lens.

[0129] At the software level, a node configuration and management module runs on the central server. This module provides a graphical human-computer interaction interface, allowing authorized users to view the status of all online monitoring nodes and perform remote software configuration on them.

[0130] Furthermore, the specific software communication logic for reconfiguration is as follows: when the construction phase changes, such as from "foundation pit construction" to "main structure construction", the management personnel can select one or more monitoring nodes whose tasks need to be changed through the interface of the configuration management module. For example, the node located in the backfilled foundation pit area can be reused for the newly erected external scaffolding.

[0131] Subsequently, the operations include: 1) Parameter redefinition: Selecting a new function cartridge type for the chosen node, such as switching from "earth pressure cell" to "tilt sensor," and the system updates the node's data parsing protocol, sampling frequency, range, and other parameters in the software accordingly. 2) Rule reassociation: In the risk warning rule base, for the new monitoring object of the node, such as "scaffolding on the south side of Building 2," binding a new set of judgment rules applicable to the object and the current construction stage, such as multi-level warning rules for scaffolding tilt.

[0132] Its configuration instructions are sent to the core base of the target node via the network. The firmware in the base will update its working mode to adapt to the new function cartridge and start collecting data according to the new parameters and uploading according to the new identifier.

[0133] This enables the system to be functionally adjustable at each stage of its lifecycle and to integrate replaceable functional modules into a general entity, thereby achieving flexible reconfiguration of hardware functions through software definition and monitoring strategies.

[0134] Based on this, significant progress has been made in terms of system lifecycle adaptability and economy by implementing reconfiguration steps.

[0135] First, it greatly improves the maintainability and reusability of the system. In traditional systems, dedicated monitoring equipment for different construction phases is often idle or scrapped after the phase ends. However, this solution, by replacing function cards and configuring remote software, enables the same set of core hardware infrastructure, such as base, network, and server, to repeatedly perform monitoring tasks at different stages of the project and for different risk objects. This achieves the recycling of hardware assets and significantly reduces the total equipment investment cost throughout the entire project cycle.

[0136] Secondly, it endows the system with unprecedented flexibility and scalability. Faced with changes in construction plans or sudden monitoring needs, managers can quickly redeploy and adjust the functions of existing monitoring nodes without waiting for the procurement and installation of new equipment, greatly shortening response time.

[0137] This enables a single system to efficiently adapt to complex engineering projects that span several years and have vastly different risk characteristics at each stage. It has evolved from a dedicated tool customized for a specific stage to an adaptive intelligent platform that runs through the entire project lifecycle, significantly improving the return on investment and the forward-looking planning capabilities for safety management.

[0138] The following describes a construction safety risk early warning system provided in an embodiment of this application. The construction safety risk early warning system described below can be referred to in correspondence with the construction safety risk early warning method described above.

[0139] refer to Figure 2 The construction safety risk early warning system should be physically deployed at the construction site or in the cloud, and it includes: Data acquisition module 1 is used to acquire historical monitoring data from multiple monitoring points at the construction site over a continuous period of time.

[0140] The hardware foundation of data acquisition module 1 is a heterogeneous sensor network, which includes tilt, displacement, and pressure sensors fixedly installed on key structures, as well as environmental sensors and video acquisition equipment deployed at the construction site.

[0141] All sensors are connected to the field industrial gateway via wired or wireless communication modules or uploaded directly to the cloud server.

[0142] The data acquisition and storage submodule software running on the server is responsible for polling or receiving data from each node and storing it in the historical monitoring database according to a unified timestamp.

[0143] The data fusion preprocessing module runs as an independent software service on the server. It acquires multi-source data streams from the same monitoring object and the same time period from the data acquisition module 1, performs spatiotemporal alignment, that is, unifies the timestamp based on the NTP protocol, unifies the spatial coordinate system based on the calibration parameters, and dynamically calculates the real-time confidence weight of each data source based on multiple dimensions such as the device's own status, the degree of environmental interference, and the consistency of historical data. Finally, it uses a weighted average or a robust weighted algorithm after removing outliers to output a fused, highly reliable monitoring data value.

[0144] Trend analysis module 2 is connected to data acquisition module 1. Trend analysis module 2 is used to calculate the change trend information of at least one monitoring parameter of each monitoring point in a continuous time period based on historical monitoring data.

[0145] Trend Analysis Module 2, as a computing service running on the server, periodically retrieves historical data sequences from the database for specified monitoring points and time windows.

[0146] Its core algorithm performs mathematical processing on each data sequence, such as using the least squares method to perform linear fitting to calculate the rate of change, i.e., the slope, or calculating the second difference to evaluate the acceleration of change.

[0147] Rule base 3 contains at least one judgment rule corresponding to a risk warning scenario. The judgment rule defines the logical combination relationship between multiple monitoring parameters and their changing trend information related to the risk warning scenario.

[0148] Rule Base 3 is a structured collection of knowledge stored in a database, with entries predefined by domain experts based on incident cases, standards, and safety experience.

[0149] Each rule clearly points to a risk warning scenario, such as "foundation pit landslide" or "scaffolding overturning," and is written in the logical form of "IF-THEN." The "IF" part is composed of multiple conditions combined through "AND" and "OR" relationships. The condition items can be the absolute values ​​of the monitoring parameters, but more importantly, their trend information, such as "displacement change rate continuously greater than X for Y hours," as well as combinations with other parameters, such as "cumulative rainfall in the same period is greater than Z."

[0150] The early warning judgment module 4 is connected to the trend analysis module 2 and the rule base 3 respectively. The early warning judgment module 4 is used to match the currently calculated trend information with the judgment rules in the rule base 3, and output the corresponding level of early warning information according to the matching result.

[0151] The early warning judgment module 4 is an independent logical processing service that continuously monitors the output of the trend analysis module 2 and takes the latest set of changing trend information as input to perform matching and condition judgment on each rule in the rule base 3.

[0152] The BIM data collaboration module is a key output enhancement component of the system. Internally, it maintains a mapping table between monitoring points and BIM components, recording the correspondence between each physical monitoring point ID and the GUID of a logical component in the BIM model. When the early warning judgment module 4 generates an early warning message, this module finds the corresponding BIM component identifier based on the mapping table, queries the BIM platform server through a standard API interface to obtain design parameters such as allowable displacement and design strength, and construction status such as the current construction stage and schedule. It then intelligently encapsulates this engineering context information with the original early warning message to generate a structured collaborative diagnostic report, which is ultimately output hierarchically through various channels such as large display screens, mobile terminal push notifications, and on-site audible and visual alarms.

[0153] Furthermore, the modules communicate and collaborate through standard software interfaces and message queues.

[0154] The data acquisition module 1 ensures the continuity and integrity of the raw data, provides sensing input for the system, and the multi-source heterogeneous raw data it provides is the basis for the system to achieve anti-interference capability.

[0155] The data fusion preprocessing module employs a dynamic confidence weighting algorithm to construct an adaptive protection layer at the data source. This algorithm does not statically assign weights but rather senses the health status of sensors in real time, such as self-test errors and low battery levels. For example, if a camera is obstructed by dust, its visual data weight is reduced; if radar is exposed to rain or fog, it maintains a high weight for environmental interference intensity; and the consistency of the data itself is also considered. If a sensor reading deviates from its historical pattern or conflicts with readings from other sensors at the same location, its weight is temporarily reduced.

[0156] Through this multi-dimensional dynamic evaluation, the fused data possesses intrinsic fault tolerance. Even if a single sensor temporarily fails or is disturbed, the system can still extract reliable information from other data sources, ensuring the continuity and accuracy of the data stream.

[0157] Trend Analysis Module 2 elevates raw, state-reflecting data into higher-order information characterizing process dynamics. This is the essential characteristic that enables the system to perform trend assessment rather than threshold comparison. It no longer views a data point in isolation, but rather analyzes the behavior pattern of that data point over time.

[0158] Rule Base 3, as the knowledge brain of the system, encapsulates human experts' understanding of the evolution of complex risks and solidifies the multi-parameter, cross-time dimension correlation logic.

[0159] The early warning judgment module 4 serves as the inference engine. It utilizes the feature vectors provided by the trend analysis module 2 to quickly search and match in the knowledge graph of the rule base 3, thereby completing the intelligent mapping from data features to risk conclusions.

[0160] The refined, high-quality fusion data is fed into trend analysis module 2 to calculate its rate of change and acceleration of change, which elevates the cognitive dimension of the system from static state values ​​to dynamic process dynamics characteristics.

[0161] Based on these high-order features, the early warning judgment module 4 performs multi-parameter logical combination matching in the rule base 3, realizing the identification of complex risk patterns, such as the combination pattern of "accelerated displacement growth accompanied by rainfall".

[0162] The introduction of the BIM data collaboration module adds value to the information at the final stage of early warning output. It automatically associates an isolated numerical over-limit signal with the BIM platform and presents it as collaborative diagnostic information with complete engineering semantics, such as "an anomaly occurring in a certain component during a specific construction stage and under a specific design background".

[0163] Based on this, through the organic cooperation of the above modules, this system has achieved a comprehensive leap from early warning accuracy to decision support capability, and has made significant progress.

[0164] First, at the input end, the multi-source confidence weighting mechanism of the data fusion preprocessing module fundamentally solves the problem of interference from the complex environment of the construction site on a single sensing technology. For example, when dust causes video displacement monitoring to fail, the radar data weight is automatically increased, and the fused displacement data remains accurate, avoiding false alarms or missed alarms. When a sensor completely fails, as long as there are other available data sources, the system can maintain the continuity of the data stream, ensuring the reliability of trend analysis.

[0165] This ensures that subsequent early warning conclusions are based on high-quality, robust data, significantly improving the accuracy and credibility of the early warnings.

[0166] Secondly, on the output side, the introduction of the BIM data collaboration module upgrades early warning information from simple signals to diagnostic reports. When managers receive an alarm for "exceeding the limit of foundation pit displacement," the system automatically pushes key contextual information such as "the design allowable displacement at this location is 50mm, currently 52mm, and we are in the most unfavorable stage of earthwork excavation."

[0167] This solution frees managers from the tedious task of manually checking drawings, progress, and designs across multiple systems, allowing them to immediately focus on the essence and urgency of the problem and quickly make scientific decisions such as whether reinforcement, unloading, or evacuation is necessary. This qualitative leap from passive alarm to proactive diagnosis greatly improves the efficiency and accuracy of risk management, truly realizing intelligent construction safety risk management.

[0168] In some embodiments, rule base 3 adopts a layered architecture design, with the bottom layer being a basic general rule layer, such as displacement exceeding limits and change rate thresholds, and the upper layer being a scene-specific rule layer. During system deployment, the corresponding dedicated rule set is automatically loaded first through the project type configuration, such as selecting "super high-rise building" or "mountain tunnel".

[0169] Furthermore, the rule set for high-altitude operations can be defined as follows: 1) Fall from Height Risk Rule. Combining personnel positioning tags and safety belt smart hook sensors. The rule is defined as: IF personnel enter the edge area (e.g., distance from the edge < 1.5m) AND safety belt unfastened for a duration > T (e.g., 30 seconds) AND personnel moving at an abnormal speed (e.g., sudden acceleration, possibly indicating a fall), THEN triggers "Fall from Height - Warning Level". 2) Tower Crane Collision Risk Rule. Monitoring the slewing angle, amplitude, and height data of multiple tower cranes, and calculating the spatial distance and approach speed between them. The rule is defined as: IF the spatial distance between any two tower crane booms or hooks < safe distance (e.g., 10m) AND relative approach speed > V (e.g., 0.5m / s), THEN triggers "Tower Crane Collision - Alarm Level" and outputs anti-collision operation suggestions. 3) Scaffolding Instability Risk Rule. Monitoring the tilt angle, strain, and vibration frequency of key nodes of the scaffolding. The rule is defined as follows: If the rate of change of tilt angle continuously accelerates (e.g., acceleration > 0.01° / h²) AND abnormal low-frequency components appear in the vibration frequency (which may be a precursor to overall instability) AND local strain exceeds the limit, then "scaffolding instability - warning level" is triggered.

[0170] Furthermore, the tunnel construction scenario rule set example can be defined as follows: 1) Face collapse risk rule. Integrate crown settlement, peripheral convergence, surrounding rock stress, and advanced geological prediction (such as TSP, ground-penetrating radar) data. The rule is defined as: IF crown settlement rate > R1 AND peripheral convergence rate > R2 AND sudden change in surrounding rock stress (such as stress value sudden increase > 20%) AND advanced prediction shows that there is a fracture zone ahead, THEN triggers "face collapse - attention level"; if the above indicators continue to deteriorate and the rate of change is positive, it is upgraded to warning level. 2) Gas over-limit risk rule: Monitor gas concentration, wind pressure, and wind speed at multiple points. The rule is defined as: IF gas concentration > 0.5% AND concentration change rate > 0.1% / min AND local wind pressure sudden drop (possibly due to fan failure), THEN triggers "gas over-limit - alarm level" and automatically associates with ventilation system control commands.

[0171] Furthermore, Rule Base 3 supports dynamic expansion. During project implementation, security managers can use a graphical rule editor to customize rules based on newly discovered risk cases and categorize them into the corresponding scenario rule sets, enabling the continuous evolution of Rule Base 3.

[0172] Based on this, a general risk early warning architecture is deeply integrated with the specific risks of different construction scenarios. The types of risks, key monitoring parameters, and their evolution patterns vary greatly across different stages of construction sites, such as foundation pits, main structures, and decoration, and among different engineering types, such as building construction, tunnels, and bridges. Traditional early warning systems often use fixed, general rule bases, making it difficult to accurately adapt to the specificities of each scenario. This embodiment, by pre-setting dedicated risk rule sets for typical scenarios such as high-altitude operations and tunnel construction, transforms industry experts' profound understanding of specific risk mechanisms into computable and executable rule models.

[0173] For example, the rule set for high-altitude operations incorporates unique monitoring indicators such as safety belt status and tower crane collision parameters, and combines these with their changing trends, such as the duration of unfastened safety belts and the approach speed of tower cranes, to make comprehensive judgments. The rule set for tunnel scenarios integrates trend information from surrounding rock stress and advanced geological forecast data to achieve early warnings of sudden risks such as tunnel face collapses. This ensures that the warning rules are highly matched with the actual risk characteristics on site, significantly improving the accuracy and adaptability of warnings, and avoiding false alarms or missed alarms caused by general rules.

[0174] In some embodiments, the AR interaction module is tightly integrated with the BIM data collaboration module, and its workflow is as follows: S10. Spatial Registration and Positioning. On-site workers wear AR glasses or use AR mobile terminals to enter the construction site. The AR device uses built-in real-time positioning and mapping technology to create a 3D map of the environment in real time, and uses GPS, BeiDou, or UWB positioning tags to obtain its approximate position in the global coordinate system.

[0175] Meanwhile, AR devices scan pre-set QR codes or natural feature points on-site to perform precise registration with the spatial coordinate system in the BIM model, ensuring accurate overlay of virtual content onto the real world.

[0176] S20. Warning Triggering and Data Push. When the system's warning judgment module generates new warning information, the BIM data collaboration module generates a collaborative diagnostic report in the aforementioned manner. This report, along with the BIM component identifiers and spatial coordinate information associated with the warning, is pushed to the AR interaction module in real time.

[0177] S30. Virtual Content Generation and Rendering. The AR interaction module retrieves the 3D geometric model of the BIM component associated with the warning from the local cache or the cloud. Combining the AR device's current position and orientation, it calculates the component's projection position on the screen, generates a semi-transparent, highlighted warning outline (e.g., a flashing red border), and generates an information label containing the risk level, current value, and recommended measures. The label is anchored near the component and always faces the user to avoid obstruction.

[0178] S40. Multimodal Interaction and Feedback. Operators can interact with virtual content through gestures, voice, or eye contact. For example, tapping a label with a gesture expands a more detailed collaborative diagnostic report; asking "risk cause" via voice prompts the system to announce the main factors triggering the rule; and staring at a risk point for more than 2 seconds automatically displays a detailed monitoring curve. AR devices also support taking photos and recording videos, which can be uploaded to the system with a single click as evidence for on-site verification.

[0179] S50, multi-user collaboration and remote guidance. When multiple AR devices are online simultaneously, the virtual markers in each wearer's field of vision remain synchronized. Project managers in the office can see the field of vision of personnel through a third-person perspective and can draw auxiliary markers such as arrows and circles in their field of vision to achieve precise remote guidance.

[0180] Based on this, abstract digital early warning information can be seamlessly integrated with real physical scenarios. Traditional early warning information has a gap with actual objects on site, requiring managers to mentally map information to location—a time-consuming and error-prone process. This embodiment utilizes AR technology to directly overlay early warning signs, risk areas, and handling guidelines onto the actual field of vision of workers, achieving intuitive guidance.

[0181] For example, when a scaffolding structure shows signs of instability, the AR glasses can immediately highlight a flashing red outline on the corresponding member and display floating labels such as "Risk Level: Warning," "Change Rate: 0.5mm / h," and "Recommendation: Reinforcement." This interactive method greatly improves the speed and depth of risk perception for on-site personnel, shortens response time, and is particularly suitable for scenarios such as emergency evacuation and locating hidden risk points. It achieves precise coupling of safety information and physical space, extending the value of digital twins to the on-site operational level.

[0182] Furthermore, for example, in a subway tunnel construction project in a certain city, the system of this application was fully applied to verify the effect of AR collaborative early warning.

[0183] If the tunnel excavation reaches a fault zone, the system will trigger a "face collapse - early warning level" warning by integrating data on crown settlement, convergence meter readings, and advanced geological predictions. The BIM data collaboration module will generate a diagnostic report that includes the current settlement rate (2.5 mm / d, allowable 1.5 mm / d), design support parameters, and emergency measures such as "immediately stop face operations and construct a temporary invert arch".

[0184] The safety officer, wearing AR glasses while patrolling the tunnel, immediately received the warning. Within his field of vision, the monitoring section within 10 meters behind the tunnel face was highlighted in red, with prominent labels such as "Collapse Warning," "Settlement Rate Exceeds Standard," and "Stop Work" displayed. Without needing to consult drawings or his phone, the safety officer could immediately pinpoint the risk area and, following the "recommended evacuation range" indicated by the labels, directed the three workers in that area to evacuate along the safety passage indicated by the green arrows.

[0185] Meanwhile, the safety officer reported the situation on-site via voice command, and the AR glasses automatically captured video of the scene and uploaded it to the project command center along with location information. Center experts then used the remote collaboration function to overlay virtual markers of "temporary support points" into the safety officer's field of vision, guiding on-site emergency response.

[0186] The entire process, from triggering the warning to completing the evacuation, can be significantly shortened. In contrast, under the traditional model, safety officers must first receive a mobile phone alarm, then confirm the location using a map, then walk to the location, and finally coordinate via walkie-talkie, which takes at least 20 minutes. This embodiment fully demonstrates the significant progress that AR collaborative early warning makes in improving the speed and accuracy of on-site risk response.

[0187] In some embodiments, the trend information includes the rate of change and / or acceleration of change of the monitored parameters.

[0188] In summary, by specifically implementing the calculation and utilization of the rate of change and acceleration of change, firstly, the sensitivity of the early warning system is qualitatively improved. The system is able to capture weak signals of the rate of change or acceleration of change in risk during the accumulation phase of quantitative change.

[0189] For example, before a foundation pit landslide, the displacement may first undergo a subtle but continuously accelerating creep process that is difficult to detect. Traditional threshold alarms are completely unaware of this, while this system can trigger early attention when the displacement acceleration changes from zero to a tiny positive value, even before the displacement exceeds the limit.

[0190] Secondly, the accuracy of early warnings is thus enhanced. By combining multi-dimensional trend information on state, speed, and acceleration, richer and more profound feature-based evidence is provided for the rule base's judgment.

[0191] This system can define combined rules such as "displacement change rate is greater than A and change acceleration is greater than B". Such rules can more accurately identify truly dangerous development patterns and effectively filter out relatively safe situations that have reached a stable level, i.e., the change rate and acceleration are close to zero. This reduces false alarms and shifts risk warning from post-event tracing and in-event emergency response to pre-event prevention.

[0192] This application provides an electronic device, such as... Figure 3 As shown, Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 3 The illustrated electronic device 300 includes a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302. Optionally, the electronic device 300 may also include a transceiver 304. It should be noted that in practical applications, the transceiver 304 is not limited to one type, and the structure of this electronic device 300 does not constitute a limitation on the embodiments of this application.

[0193] Processor 301 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in connection with the embodiments of this application. Processor 301 may also be a combination that implements computing functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0194] Bus 302 may include a pathway for transmitting information between the aforementioned components. Bus 302 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 302 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0195] The memory 303 may be a ROM (Read-Only Memory) or other type of static storage device capable of storing static information and instructions, RAM (Random Access Memory) or other type of dynamic storage device capable of storing information and instructions, or it may be an EEPROM (Electrically Erasable Programmable Read-Only Memory), a CD-ROM (Compact Disc Read-Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0196] The memory 303 is used to store application code that executes the scheme of the embodiments of this application, and its execution is controlled by the processor 301. The processor 301 is used to execute the application code stored in the memory 303 to implement the content shown in the foregoing method embodiments.

[0197] Among them, electronic devices include, but are not limited to: mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), and in-vehicle terminals (such as in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality or scope of the embodiments described in this application.

[0198] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps of the construction safety risk warning method described above.

[0199] Since the embodiments of the computer-readable storage medium portion correspond to the embodiments of the method portion, please refer to the description of the embodiments of the method portion for the embodiments of the computer-readable storage medium portion.

[0200] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0201] The above are only some embodiments of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A construction safety risk early warning system, characterized in that, include: The data acquisition module is used to acquire historical monitoring data from multiple monitoring points at the construction site over a continuous period of time. The historical monitoring data includes multi-source monitoring data collected from the same monitoring object by sensors with at least two different measurement principles; A data fusion preprocessing module, connected to the data acquisition module, is used to perform confidence-weighted fusion processing on multi-source monitoring data of the same monitoring object to obtain fused monitoring data. The trend analysis module, connected to the data acquisition module, is used to calculate the change trend information of at least one monitoring parameter of each monitoring point within the continuous time period based on the historical monitoring data. The rule base pre-stores at least one judgment rule corresponding to a risk warning scenario. The judgment rule defines the logical combination relationship between multiple monitoring parameters and their changing trend information related to the risk warning scenario. The early warning judgment module is connected to the trend analysis module and the rule base respectively. It is used to match the currently calculated trend information with the judgment rules in the rule base and output the corresponding level of early warning information according to the matching result. The BIM data collaboration module is connected to the early warning judgment module. It is used to obtain the corresponding construction status or design parameter information from the Building Information Modeling (BIM) platform based on the monitoring point location information associated with the early warning information, and to output the early warning information in association with the construction status or design parameter information.

2. The construction safety risk early warning system according to claim 1, characterized in that, The rule base includes pre-set sets of specialized risk rules for different construction scenarios; each specialized risk rule set includes at least one of the rule sets for high-altitude operations and tunnel construction scenarios; wherein... The high-altitude operation scenario rule set includes judgment rules related to the risks of falls from height, tower crane collisions, and scaffold instability. The monitoring parameters involved in the judgment rules include at least one of the following: personnel positioning data, safety belt fastening status, tower crane rotation angle and speed, scaffold tilt angle and vibration frequency, as well as the corresponding trend information. The tunnel construction scenario rule set includes judgment rules related to the risks of face collapse, gas exceedance, and support deformation. The monitoring parameters involved in the judgment rules include at least one of the following: crown settlement, perimeter convergence, gas concentration, surrounding rock stress, and advanced geological prediction data, as well as the corresponding trend information.

3. The construction safety risk early warning system according to claim 1, characterized in that, It also includes an augmented reality (AR) interaction module, which is connected to the BIM data collaboration module and is used for: Acquire the location and posture information of AR glasses or mobile terminals of on-site workers; Based on the location information of the monitoring points associated with the early warning information, a virtual identifier is generated and rendered onto the display screen of AR glasses or mobile terminals, so that the virtual identifier is superimposed on the location of the corresponding physical component in the real scene. The information in the collaborative diagnostic report is displayed in the form of visual labels associated with the virtual identifier.

4. The system according to claim 1, characterized in that: The trend information includes the rate of change and / or acceleration of change of the monitored parameters.

5. A construction safety risk early warning method, based on the system described in any one of claims 1-4, characterized in that, Includes the following steps: Acquire historical monitoring data from multiple monitoring points at the construction site over a continuous time period; The confidence-weighted fusion processing of multi-source monitoring data of the same monitoring object is performed to obtain monitoring data for trend analysis; specifically, the confidence weight of each data source is dynamically evaluated based on at least one dimension, including the device's own status, the degree of environmental interference, and the consistency of historical data. Based on the aforementioned confidence weights, the multi-source monitoring data are weighted and fused to obtain monitoring data for trend analysis. Based on the historical monitoring data, analyze and calculate the change trend information of at least one monitoring parameter of each monitoring point within the continuous time period; The calculated trend information is matched with pre-stored judgment rules; wherein, the judgment rules define the logical combination relationship between multiple monitoring parameters and their trend information under different risk warning scenarios. Based on the matching results, output the warning information corresponding to the risk warning scenario and level; In addition to outputting the warning information, it also includes, Based on the monitoring point location information associated with the early warning information, the corresponding construction status or design parameter information is obtained from the Building Information Modeling (BIM) platform, and the early warning information is associated with the construction status or design parameter information and output; specifically, Based on the pre-stored mapping relationship between monitoring points and BIM components, determine the BIM component identifier corresponding to the monitoring point that triggers the early warning; The system initiates a query request carrying the BIM component identifier to the BIM platform through the application programming interface (API) to obtain at least one of the design parameters and construction status information associated with the component. The warning information is associated and encapsulated with the acquired information to generate a collaborative diagnostic report and output it.

6. The construction safety risk early warning method according to claim 5, characterized in that, The confidence-weighted fusion processing of multi-source monitoring data for the same monitoring object specifically includes: The confidence weight of each data source is dynamically evaluated based on at least one dimension, including the device's own status, the degree of environmental interference, and the consistency of historical data. Based on the confidence weights, the multi-source monitoring data are weighted and fused to obtain monitoring data for trend analysis.

7. The construction safety risk early warning method according to claim 5, characterized in that, The step of obtaining corresponding construction status or design parameter information from the Building Information Modeling (BIM) platform based on the monitoring point location information associated with the output early warning information specifically includes: Based on the pre-stored mapping relationship between monitoring points and BIM components, determine the BIM component identifier corresponding to the monitoring point that triggers the early warning; The system initiates a query request carrying the BIM component identifier to the BIM platform through the application programming interface to obtain at least one of the design parameters and construction status information associated with the component. The warning information is associated and encapsulated with the acquired information to generate a collaborative diagnostic report and output it.

8. The method according to claim 5, characterized in that, The output warning information includes: When the matching result meets the low-level risk conditions, a risk attention prompt message is output; When the matching result meets the high-level risk conditions, a risk alarm message is output.

9. The method according to claim 5, characterized in that, The steps for acquiring historical monitoring data also include acquiring data from mobile monitoring devices or monitoring nodes temporarily deployed on new work sites.

10. The method according to claim 5, characterized in that, In addition to outputting the warning information, it also includes: Acquire the position and orientation information of augmented reality (AR) devices used by on-site workers; Based on the location information of the monitoring points associated with the early warning information, a virtual identifier is generated and rendered onto the display screen of the AR device, so that the virtual identifier is superimposed on the location of the corresponding physical component in the real scene. The information in the collaborative diagnostic report is displayed in the form of visual labels associated with the virtual identifier.

Images