Intelligent monitoring and early warning platform for whole high-altitude sliding process of large-span steel truss corridor

Abstract

The invention relates to the technical field of constructional engineering, and discloses a large-span steel truss corridor high-altitude slippage whole-process intelligent monitoring and early warning platform, which comprises a data acquisition module, a data fusion module, an intelligent early warning module and a visual monitoring module, when the whole process of high-altitude slippage of the large-span steel truss corridor is monitored, virtual mapping of the slippage state is constructed in real time through multi-source data of integrated structure health monitoring, environment monitoring and equipment state monitoring, key parameters in the slippage process can be detected in real time, and abnormal changes of stress, wind speed and thruster pressure are included. The monitoring comprehensiveness and timeliness are ensured, the problem of early warning lag caused by data islands is avoided, the accident risks of slippage offset and structural instability are reduced, the construction safety and continuity are improved, and the fault diagnosis and risk prediction of the slippage process are performed by matching a real-time state track with a historical fault library. Potential fault paths and risk levels are intelligently evaluated, and the accuracy of early fault intervention is ensured.

Description

Technical Field

[0001] This invention relates to the field of building engineering technology, specifically to an intelligent monitoring and early warning platform for the entire process of high-altitude sliding of large-span steel truss corridors. Background Technology

[0002] With the advancement of urban construction, more and more building projects are designed with the concept of integration and interconnection in mind, connecting the various buildings of the proposed project through sky bridges. These sky bridges connecting buildings typically use steel trusses. With the improvement of socio-economic levels, the demand for large-scale venues, sports centers, and large airports is increasing. The structural design of pier systems has become quite mature, and large-span trusses are being used more and more widely in architectural design. Large-span trusses have the advantages of light weight, good load-bearing capacity, and high strength, and can provide ultra-large spaces, save materials, and meet various shape requirements.

[0003] Currently, due to the dynamic coupling of multi-source heterogeneous data involved in the high-altitude sliding process of large-span steel truss corridors, the structural stress, environmental wind speed, and equipment pressure data collected by the sensor network lack a collaborative calibration mechanism when conducting real-time monitoring of the sliding state. This makes it impossible to verify the spatiotemporal consistency of the monitoring data in real time. When the timestamps and spatial coordinates of asynchronous data deviate, it will lead to distortion of the fusion model input, increase the risk of false alarms, and weaken the reliability of the monitoring system.

[0004] Therefore, a smart monitoring and early warning platform for the entire process of high-altitude sliding of large-span steel truss connecting corridors is proposed to solve the above problems. Summary of the Invention

[0005] (a) Technical problems to be solved

[0006] To address the shortcomings of existing technologies, this invention provides an intelligent monitoring and early warning platform for the entire process of high-altitude sliding of large-span steel truss corridors, solving the problems mentioned in the background section.

[0007] (II) Technical Solution

[0008] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor, the platform comprising:

[0009] Data acquisition module: Real-time acquisition of multi-source data during the sliding process, including structural health monitoring data (including stress, deformation and vibration parameters), environmental monitoring data (including wind speed and temperature parameters), equipment status monitoring data (including pusher pressure, displacement and oil temperature parameters), and video monitoring data. The data is deployed at key sliding points through a sensor network and preprocessed into a standardized dataset.

[0010] Data fusion module: It fuses the collected multi-source data to generate comprehensive state information, including time-stamp aligned feature vectors containing mean stress and peak wind speed. The input is a pre-trained data fusion model, which outputs a slip state matrix. The fusion weights are dynamically calculated based on the data standard deviation.

[0011] Intelligent early warning module: Utilizes artificial intelligence algorithms to analyze comprehensive status information to achieve early warning of key parameters exceeding limits, fault diagnosis, and risk prediction. Specifically, it triggers an early warning signal when key parameters, including stress, wind speed, and pusher pressure, exceed preset safety thresholds, and assesses the risk level by matching the historical fault database.

[0012] Visual monitoring module: Provides a real-time visual monitoring interface, dynamically displays multi-source data, early warning information and risk prediction results of the sliding process, and supports interactive operation by users;

[0013] The platform achieves closed-loop monitoring of the entire sliding process through multi-source data fusion and intelligent early warning collaboration, and generates intervention strategies when an early warning is triggered.

[0014] Preferably, the process of real-time acquisition of multi-source data in the data acquisition module is as follows:

[0015] A sensor network, including stress sensors, accelerometers, anemometers, temperature sensors, pressure transmitters, and cameras, is deployed at the starting, mid, and ending points of the slippage, and data is transmitted to a central server via an Internet of Things (IoT) protocol.

[0016] The sampling frequency was set to 10Hz to ensure data real-time performance. The raw data was preprocessed, including noise filtering, outlier removal, and data normalization, to generate a standardized dataset.

[0017] Preferably, the process of generating comprehensive status information in the data fusion module is as follows:

[0018] Extract feature vectors from multi-source data, including mean stress, peak wind speed, pressure variance, and deformation gradient;

[0019] The feature vectors are input into a data fusion model, which is trained based on a deep learning architecture and outputs a slip state matrix.

[0020] The fusion weights of the slip state matrix are calculated using the following formula:

[0021]

[0022] Where w i Let σ be the weight of the i-th class of data. i Let σ be the standard deviation of the i-th class of data. j Let be the standard deviation of the j-th data category, and n be the total number of data categories.

[0023] Preferably, the process of triggering the warning signal in the intelligent warning module is as follows:

[0024] Set safety thresholds for key parameters, including the stress threshold S. max Wind speed threshold W max and pressure threshold P max Calculate the current parameter P real The degree of deviation from the safety threshold:

[0025] ΔP=|P real -P thresh |;

[0026] Where P thresh Here, ΔP represents the safety threshold, and P is the deviation between the current parameter value and the safety threshold. real The actual value of the current parameter;

[0027] When ΔP>0, the evaluation checks whether the deviation meets the early warning activation condition. If it does, the early warning signal is activated and the system enters the fault diagnosis submodule.

[0028] Preferably, the fault diagnosis process in the intelligent early warning module is as follows:

[0029] The current slip state is matched with the historical fault database, which includes slip offset, structural resonance and pusher failure modes.

[0030] State features are extracted using a convolutional neural network, and the similarity with historical failure patterns is calculated. If the similarity exceeds 0.7, a potential failure path is identified.

[0031] Based on the fault type and risk level, generate a diagnostic report, including the fault location, cause, and urgency.

[0032] Preferably, the risk prediction process in the intelligent early warning module is as follows:

[0033] Based on the fault diagnosis results, predict the timing and probability of risk occurrence;

[0034] Use time series analysis models to fit the changing trends of key parameters;

[0035] The probability of risk is calculated using the following formula:

[0036] R = α·ΔP + β·sim;

[0037] Where R is the risk probability, ΔP is the parameter bias, sim is the similarity with historical failures, α is the weighting coefficient of parameter bias (0.6), and β is the weighting coefficient of similarity (0.4).

[0038] When R is greater than 0.5, it is considered high risk, and the predicted occurrence time is within the slippage period.

[0039] Preferably, the process of providing a real-time visual monitoring interface in the visual monitoring module is as follows:

[0040] A virtual mapping of the steel truss corridor is constructed based on 3D modeling technology;

[0041] Dynamically render multi-source data, including stress cloud maps, wind speed vectors, and equipment pressure heat maps;

[0042] It supports switching between multiple views, including floor plans, cross-sectional views, and real-time video, and can overlay warning information.

[0043] Users can adjust monitoring parameters and trigger data backtracking through the interface.

[0044] Preferably, the specific training process of the data fusion model in the data fusion module is as follows:

[0045] A convolutional autoencoder is trained using a historical slip dataset, with multi-source feature vectors as input and a fused state matrix as output.

[0046] The training objective is to minimize reconstruction error and feature correlation loss;

[0047] After the model is deployed, it is updated in real time based on the sliding progress feedback data.

[0048] Preferably, the process of classifying early warning levels in the intelligent early warning module is as follows:

[0049] The early warning system is divided into three levels:

[0050] Level 1 warning: Low risk, ΔP≤10%;

[0051] Level 2 warning: Medium risk, 10% < ΔP ≤ 30%;

[0052] Level 3 warning: High risk, ΔP>30%;

[0053] Different response strategies are generated for different levels: Level 1 warnings only log, Level 2 warnings trigger manual review, and Level 3 warnings trigger automatic intervention.

[0054] Preferably, the process by which the platform generates an intervention strategy when an early warning is triggered is as follows:

[0055] Based on the risk level and predicted occurrence time, combined with the current construction plan;

[0056] When the risk level is high and the event occurs within the slippage cycle, an intervention plan is generated, including adjusting the slippage speed, suspending construction, and activating backup equipment.

[0057] (III) Beneficial Effects

[0058] Compared with existing technologies, this invention provides an intelligent monitoring and early warning platform for the entire process of high-altitude sliding of large-span steel truss corridors, which has the following beneficial effects:

[0059] 1. In this invention, when monitoring the entire process of high-altitude sliding of a large-span steel truss corridor, a virtual mapping of the sliding state is constructed in real time by integrating multi-source data from structural health monitoring, environmental monitoring, and equipment status monitoring. This enables real-time detection of abnormal changes in key parameters during the sliding process, including stress, wind speed, and jacking pressure, ensuring comprehensive and timely monitoring, avoiding the problem of delayed early warning caused by data silos, reducing the risk of sliding deviation and structural instability accidents, and improving construction safety and continuity.

[0060] 2. In this invention, when performing fault diagnosis and risk prediction during the sliding process, the system intelligently assesses potential fault paths and risk levels by matching real-time status trajectories with historical fault databases. This enables the system to identify sliding offset, structural resonance, and jacking failure fault modes in advance. Furthermore, when the risk exceeds the equipment's health tolerance, the system automatically triggers a graded early warning mechanism to ensure the accuracy of early fault intervention, avoid false alarms and missed alarms, and improve the reliability and response efficiency of sliding construction.

[0061] 3. In this invention, when generating slip intervention strategies and making visual decisions, the optimal intervention plan is dynamically generated in conjunction with the current production plan, and multi-source data and early warning information are rendered in real time. This allows users to interactively adjust monitoring parameters and correct slip deviations, ensuring the scientific and economical nature of intervention measures such as adjusting slip speed and activating backup equipment, optimizing resource scheduling and schedule management, and improving overall construction efficiency and quality control. Attached Figure Description

[0062] Figure 1 This is a structural diagram of the intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss corridor according to the present invention. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] For specific implementation examples, please refer to: Figure 1 A smart monitoring and early warning platform for the entire process of high-altitude sliding of large-span steel truss connecting corridors. The platform includes:

[0065] Data acquisition module: Real-time acquisition of multi-source data during the sliding process, including structural health monitoring data (including stress, deformation and vibration parameters), environmental monitoring data (including wind speed and temperature parameters), equipment status monitoring data (including pusher pressure, displacement and oil temperature parameters), and video monitoring data. The data is deployed at key sliding points through a sensor network and preprocessed into a standardized dataset.

[0066] Data fusion module: It fuses the collected multi-source data to generate comprehensive state information, including time-stamp aligned feature vectors containing mean stress and peak wind speed. The input is a pre-trained data fusion model, which outputs a slip state matrix. The fusion weights are dynamically calculated based on the data standard deviation.

[0067] Intelligent early warning module: Utilizes artificial intelligence algorithms to analyze comprehensive status information to achieve early warning of key parameters exceeding limits, fault diagnosis, and risk prediction. Specifically, it triggers an early warning signal when key parameters, including stress, wind speed, and pusher pressure, exceed preset safety thresholds, and assesses the risk level by matching the historical fault database.

[0068] Visual monitoring module: Provides a real-time visual monitoring interface, dynamically displays multi-source data, early warning information and risk prediction results of the sliding process, and supports interactive operation by users;

[0069] The platform achieves closed-loop monitoring of the entire sliding process through multi-source data fusion and intelligent early warning collaboration, and generates intervention strategies when an early warning is triggered.

[0070] The process of real-time acquisition of multi-source data in the data acquisition module is as follows:

[0071] A sensor network, including stress sensors, accelerometers, anemometers, temperature sensors, pressure transmitters, and cameras, is deployed at the starting, mid, and ending points of the slippage, and data is transmitted to a central server via an Internet of Things (IoT) protocol.

[0072] The sampling frequency is set to 10Hz to ensure data real-time performance. The raw data is preprocessed, including noise filtering, outlier removal, and data normalization, to generate a standardized dataset.

[0073] To quantify data quality, the signal-to-noise ratio (SNR) of key parameters is calculated during preprocessing to ensure that high-noise data is filtered out. Specifically, this includes extracting the signal power and noise power from the acquired raw signal and calculating the SNR using the following formula:

[0074]

[0075] Where SNR is the signal-to-noise ratio, P signal For signal power, P noise Noise power;

[0076] If the SNR is lower than the preset threshold of 20dB, the data point is discarded and enhanced filtering is applied.

[0077] The process of generating comprehensive status information in the data fusion module is as follows:

[0078] Extract feature vectors from multi-source data, including mean stress, peak wind speed, pressure variance, and deformation gradient;

[0079] The feature vectors are input into a data fusion model, which is trained based on a deep learning architecture and outputs a slip state matrix.

[0080] The fusion weights of the slip state matrix are calculated using the following formula:

[0081]

[0082] Where w i Let σ be the weight of the i-th class of data. i Let σ be the standard deviation of the i-th class of data. j Let be the standard deviation of the j-th data category, and n be the total number of data categories.

[0083] The process of triggering an early warning signal in the intelligent early warning module is as follows:

[0084] Set safety thresholds for key parameters, including the stress threshold S. max Wind speed threshold W max and pressure threshold P max Calculate the current parameter P real The degree of deviation from the safety threshold:

[0085] ΔP=|P real -P thresh |;

[0086] Where P thresh Here, ΔP represents the safety threshold, and P is the deviation between the current parameter value and the safety threshold. real The actual value of the current parameter;

[0087] When ΔP>0, the evaluation checks whether the deviation meets the early warning activation condition. If it does, the early warning signal is activated and the system enters the fault diagnosis submodule.

[0088] The fault diagnosis process in the intelligent early warning module is as follows:

[0089] The current slip state is matched with the historical fault database, which includes slip offset, structural resonance and pusher failure modes.

[0090] State features are extracted using a convolutional neural network, and the similarity to historical failure patterns is calculated.

[0091] To improve fault matching accuracy, cosine similarity is used to quantify the similarity between the current state and historical fault patterns. Specifically, the similarity is calculated by taking the current state feature vector and the historical fault feature vector as inputs and using the following formula:

[0092]

[0093] Where sim is the similarity value, A is the current state feature vector, and B is the historical fault feature vector;

[0094] If the similarity exceeds 0.7, a potential fault path is identified.

[0095] Based on the fault type and risk level, generate a diagnostic report, including the fault location, cause, and urgency.

[0096] The risk prediction process in the intelligent early warning module is as follows:

[0097] Based on the fault diagnosis results, predict the timing and probability of risk occurrence;

[0098] Use time series analysis models to fit the changing trends of key parameters;

[0099] The probability of risk is calculated using the following formula:

[0100] R = α·ΔP + β·sim;

[0101] Where R is the risk probability, ΔP is the parameter bias, sim is the similarity with historical failures, α is the weighting coefficient of parameter bias (0.6), and β is the weighting coefficient of similarity (0.4).

[0102] When R is greater than 0.5, it is considered high risk, and the predicted occurrence time is within the slippage period.

[0103] The process of providing a real-time visual monitoring interface in the visual monitoring module is as follows:

[0104] A virtual mapping of the steel truss corridor is constructed based on 3D modeling technology;

[0105] Dynamically render multi-source data, including stress cloud maps, wind speed vectors, and equipment pressure heat maps;

[0106] To optimize visualization performance and calculate the real-time rendering frame rate (FPS) to ensure a smooth monitoring interface, the following steps are taken: Based on the data volume and processing time, the rendering frame rate is dynamically adjusted using the following formula:

[0107]

[0108] Where FPS is the frame rate, and N frame Δt is the number of frames rendered. render This is the rendering time interval;

[0109] When the FPS is below 30, the data resolution is automatically reduced and the model details are simplified.

[0110] It supports switching between multiple views, including floor plans, cross-sectional views, and real-time video, and can overlay warning information.

[0111] Users can adjust monitoring parameters and trigger data backtracking through the interface.

[0112] The specific training process of the data fusion model in the data fusion module is as follows:

[0113] A convolutional autoencoder is trained using a historical slip dataset, with multi-source feature vectors as input and a fused state matrix as output.

[0114] The training objective is to minimize reconstruction error and feature correlation loss;

[0115] To quantify the model training effect, a loss function is defined that combines reconstruction error and feature loss. Specifically, the total loss is calculated using the following formula to guide model optimization:

[0116]

[0117] Where L is the total loss value, and λ1 is the weighting coefficient of the reconstruction error term. For mean square error, y i Let i be the true value of the i-th sample. Let λ1 be the model prediction value of the i-th sample, λ2 be the weight coefficient of the feature correlation loss, CorrLoss be the feature correlation loss, calculate the mutual information loss between features, λ1 = 0.7, λ2 = 0.3, and the model converges when L is minimized.

[0118] After the model is deployed, it is updated in real time based on the sliding progress feedback data.

[0119] The process of classifying early warning levels in the intelligent early warning module is as follows:

[0120] The early warning system is divided into three levels:

[0121] Level 1 warning: Low risk, ΔP≤10%;

[0122] Level 2 warning: Medium risk, 10% < ΔP ≤ 30%;

[0123] Level 3 warning: High risk, ΔP>30%;

[0124] Different response strategies are generated for different levels: Level 1 warnings only log, Level 2 warnings trigger manual review, and Level 3 warnings trigger automatic intervention.

[0125] The process by which the platform generates an intervention strategy when an alert is triggered is as follows:

[0126] Based on the risk level and predicted occurrence time, combined with the current construction plan;

[0127] When the risk level is high and the event occurs within the slippage cycle, an intervention plan is generated, including adjusting the slippage speed, suspending construction, and activating backup equipment.

[0128] To minimize the impact of intervention on the project schedule, a cost optimization function is defined, which specifically includes: combining resource costs and downtime losses, and solving for the optimal intervention strategy using the following formula:

[0129] min(c resource ·R+c delay ·T);

[0130] Where R is the resource usage, T is the estimated delay time, and c resource c is the resource cost coefficient. delay The delay cost coefficient is denoted by T, where T is the estimated delay time. The solution must meet production plan constraints.

[0131] The solution optimizes the algorithm calculation to ensure that the impact on the construction period is minimized.

[0132] The operation steps of the intelligent monitoring and early warning platform for the entire process of high-altitude sliding of large-span steel truss connecting corridors are as follows:

[0133] Multi-source data collaborative sensing and dynamic calibration:

[0134] Before the sliding construction commenced, a heterogeneous sensor network was deployed at the sliding start point, key mid-span nodes, and end point of the steel truss corridor. This network included stress strain gauges, triaxial accelerometers, ultrasonic anemometers, temperature sensors, hydraulic jacking pressure transmitters, and high-definition industrial cameras. The system, via an IoT edge gateway, collected four types of core data in real time at a 10Hz sampling frequency: structural health monitoring data (including stress distribution, dynamic deformation, and vibration spectrum); environmental monitoring data (including real-time wind speed vectors, temperature and humidity gradients, and equipment operating status data); and jacking pressure peaks, displacement deviations, hydraulic oil temperature rise rates, and full-view video stream data. All raw data underwent spatiotemporal alignment through a preprocessing pipeline. A timestamp-based dynamic calibration algorithm eliminated clock drift between sensors, while adaptive filtering technology suppressed electromagnetic interference and mechanical noise, generating a spatiotemporally consistent multi-source standardized data stream to provide high-confidence input for virtual mapping construction.

[0135] Digital twin-driven state fusion and risk simulation

[0136] Based on preprocessed multi-source data streams, a pre-trained deep convolutional autoencoder model is invoked to construct a real-time virtual mapping of the steel truss corridor within the digital twin engine. This engine couples the structural deformation field, environmental load distribution, and equipment output status across domains to generate a state evolution trajectory matrix based on the time axis. The system synchronously loads a historical fault feature library and calculates the similarity probability between the current state and historical faults using a spatiotemporal trajectory pattern matching algorithm. When key parameters, including peak stress, wind speed pulses, and sudden oil temperature changes, exceed the dynamic safety boundary, a three-level early warning mechanism is immediately activated: Level 1 warning triggers log marking and data backtracking; Level 2 warning initiates the fault root cause analysis module to locate the anomaly source; and Level 3 warning, combined with the construction progress model and resource scheduling constraints, predicts the risk evolution path and time window, generating a risk quantification assessment report.

[0137] Intelligent decision generation and closed-loop control intervention

[0138] When the risk level reaches the intervention threshold, the system invokes a multi-objective optimization decision model to generate a tiered response strategy. For low-risk scenarios, the system automatically adjusts the jacking device speed curve to balance load distribution; for medium- to high-risk scenarios, it links the hydraulic system to reduce load and activates backup jacking units; for extreme risks, it triggers an emergency shutdown protocol and pushes a replanning scheme for obstacle avoidance paths. All decision commands are sent to the execution terminal via the Industrial Internet of Things (IIoT), while the visualization engine dynamically updates the 3D virtual scene: high-risk areas of the structure are rendered using stress cloud maps, equipment operating loads are projected using heat maps, and real-time warning locations and intervention trajectories are marked with AR. Users can correct control parameters through the human-machine interface, and the platform iterates the data fusion model weights in real time based on execution feedback to achieve closed-loop control of the entire process of "perception-diagnosis-decision-execution," ensuring efficient progress of sliding construction under dynamic risk management.

[0139] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0140] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart monitoring and early warning platform for the entire process of high-altitude sliding of large-span steel truss connecting corridors, characterized by: The platform includes: Data acquisition module: Real-time acquisition of multi-source data during the sliding process, including structural health monitoring data (including stress, deformation and vibration parameters), environmental monitoring data (including wind speed and temperature parameters), equipment status monitoring data (including pusher pressure, displacement and oil temperature parameters), and video monitoring data. The data is deployed at key sliding points through a sensor network and preprocessed into a standardized dataset. Data fusion module: It fuses the collected multi-source data to generate comprehensive state information, including time-stamp aligned feature vectors containing mean stress and peak wind speed. The input is a pre-trained data fusion model, which outputs a slip state matrix. The fusion weights are dynamically calculated based on the data standard deviation. Intelligent early warning module: Utilizes artificial intelligence algorithms to analyze comprehensive status information to achieve early warning of key parameters exceeding limits, fault diagnosis, and risk prediction. Specifically, it triggers an early warning signal when key parameters, including stress, wind speed, and pusher pressure, exceed preset safety thresholds, and assesses the risk level by matching the historical fault database. Visual monitoring module: Provides a real-time visual monitoring interface, dynamically displays multi-source data, early warning information and risk prediction results of the sliding process, and supports interactive operation by users; The platform achieves closed-loop monitoring of the entire sliding process through multi-source data fusion and intelligent early warning collaboration, and generates intervention strategies when an early warning is triggered.

2. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor as described in claim 1, characterized in that: The process of real-time acquisition of multi-source data in the data acquisition module is as follows: A sensor network, including stress sensors, accelerometers, anemometers, temperature sensors, pressure transmitters, and cameras, is deployed at the starting, mid, and ending points of the slippage, and data is transmitted to a central server via an Internet of Things (IoT) protocol. The sampling frequency was set to 10Hz to ensure real-time data processing. The raw data was preprocessed, including noise filtering, outlier removal, and data normalization, to generate a standardized dataset.

3. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor as described in claim 1, characterized in that: The process of generating comprehensive status information in the data fusion module is as follows: Extract feature vectors from multi-source data, including mean stress, peak wind speed, pressure variance, and deformation gradient; The feature vectors are input into a data fusion model, which is trained based on a deep learning architecture and outputs a slip state matrix. The fusion weights of the slip state matrix are calculated using the following formula: Where w i Let σ be the weight of the i-th class of data. i Let σ be the standard deviation of the i-th class of data. j Let be the standard deviation of the j-th data category, and n be the total number of data categories.

4. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor as described in claim 1, characterized in that: The process of triggering the early warning signal in the intelligent early warning module is as follows: Set safety thresholds for key parameters, including the stress threshold S. max Wind speed threshold W max and pressure threshold P max Calculate the current parameter P real The degree of deviation from the safety threshold: ΔP=|P real -P thresh |; Where P thresh Here, ΔP represents the safety threshold, and P is the deviation between the current parameter value and the safety threshold. real The actual value of the current parameter; When ΔP>0, the evaluation checks whether the deviation meets the early warning activation condition. If it does, the early warning signal is activated and the system enters the fault diagnosis submodule.

5. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor according to claim 4, characterized in that: The fault diagnosis process in the intelligent early warning module is as follows: The current slip state is matched with the historical fault database, which includes slip offset, structural resonance and pusher failure modes. State features are extracted using a convolutional neural network, and the similarity with historical failure patterns is calculated. If the similarity exceeds 0.7, a potential failure path is identified. Based on the fault type and risk level, generate a diagnostic report, including the fault location, cause, and urgency.

6. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor according to claim 5, characterized in that: The risk prediction process in the intelligent early warning module is as follows: Based on the fault diagnosis results, predict the timing and probability of risk occurrence; Use time series analysis models to fit the changing trends of key parameters; The probability of risk is calculated using the following formula: R = α·ΔP + β·sim; Where R is the risk probability, ΔP is the parameter bias, sim is the similarity with historical failures, α is the weighting coefficient of parameter bias (0.6), and β is the weighting coefficient of similarity (0.4). When R is greater than 0.5, it is considered high risk, and the predicted occurrence time is within the slippage period.

7. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor according to claim 1, characterized in that: The process of providing a real-time visual monitoring interface in the visual monitoring module is as follows: A virtual mapping of the steel truss corridor is constructed based on 3D modeling technology; Dynamically render multi-source data, including stress cloud maps, wind speed vectors, and equipment pressure heat maps; It supports switching between multiple views, including floor plans, cross-sectional views, and real-time video, and can overlay warning information. Users can adjust monitoring parameters and trigger data backtracking through the interface.

8. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor according to claim 3, characterized in that: The specific training process of the data fusion model in the data fusion module is as follows: A convolutional autoencoder is trained using a historical slip dataset, with multi-source feature vectors as input and a fused state matrix as output. The training objective is to minimize reconstruction error and feature correlation loss; After the model is deployed, it is updated in real time based on the sliding progress feedback data.

9. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor according to claim 4, characterized in that: The process of classifying early warning levels in the intelligent early warning module is as follows: The early warning system is divided into three levels: Level 1 warning: Low risk, ΔP≤10%; Level 2 warning: Medium risk, 10% < ΔP ≤ 30%; Level 3 warning: High risk, ΔP>30%; Different response strategies are generated for different levels: Level 1 warnings only log, Level 2 warnings trigger manual review, and Level 3 warnings trigger automatic intervention.

10. The intelligent monitoring and early warning platform for the entire process of high-altitude sliding of a large-span steel truss connecting corridor according to claim 1, characterized in that: The process by which the platform generates an intervention strategy when an early warning is triggered is as follows: Based on the risk level and predicted occurrence time, combined with the current construction plan; When the risk level is high and the event occurs within the slippage cycle, an intervention plan is generated, including adjusting the slippage speed, suspending construction, and activating backup equipment.

Images