A tower climbing anti-falling protection safety management and control system based on intelligent monitoring

Abstract

The application discloses a tower climbing anti-falling protection safety management and control system based on intelligent monitoring, and relates to the technical field of high-altitude operation safety management and control; the application deploys multiple types of sensors on the working personnel equipment and the upper part of the tower body through a physical perception layer to collect posture, protection equipment, structural deformation and environmental data in real time, and transmits the data to an edge node through a 5G industrial private network; a digital twin platform layer constructs a 1:1 real-time mapping digital twin model and adopts a multi-source data weighted synchronization algorithm to ensure accurate synchronization; a simulation deduction layer performs multi-dimensional falling risk simulation deduction before operation and generates an emergency plan; a real-time management and control layer identifies abnormalities based on an individual behavior fingerprint adaptive detection algorithm, triggers a four-level graded early warning and links a dynamic deviation correction of a proactive protection device; and an evaluation and analysis layer automatically generates a safety evaluation report after operation, so that the tower climbing operation whole-process intelligentization and closed-loop safety management and control are realized.

Description

Technical Field

[0001] This invention relates to the field of high-altitude operation safety management technology, specifically a tower climbing fall prevention and safety management system based on intelligent monitoring. Background Technology

[0002] With the rapid development of industries such as power and communications, the workload of installation, inspection, and maintenance of transmission line towers and communication base station towers continues to grow. As a crucial form of high-altitude work, tower climbing has always been a key focus of the industry in terms of safety management. Tower climbing operations involve complex environments and numerous risk factors. Workers face various safety threats during their ascents, including falls, tower structural failures, and adverse weather conditions. Therefore, an intelligent and systematic safety management system is urgently needed to ensure the safety of workers.

[0003] However, traditional tower climbing operation safety management mainly relies on manual monitoring and simple physical protective equipment, which has many shortcomings: First, it lacks the ability to perceive and intelligently identify the behavior of operators in real time, and cannot effectively distinguish between normal operation and violation operation. It often only responds after an accident occurs, which is a passive safety management. Second, the protective equipment has a single function and cannot dynamically adjust the protection strategy according to the risk level, making it difficult to achieve proactive correction. Third, it lacks a risk rehearsal before the operation and a system assessment mechanism after the operation. Safety management cannot form a closed loop, and the investigation of hidden dangers relies on manual experience, which is inefficient and prone to omissions. Fourth, the traditional management method cannot simultaneously coordinate the management of multiple towers, has poor scalability, and is difficult to meet the needs of large-scale operation scenarios.

[0004] In summary, existing tower climbing fall prevention safety management technologies have significant shortcomings in real-time perception, intelligent early warning, active protection, and closed-loop management throughout the entire process. There is an urgent need for a new safety management system that integrates technologies such as IoT perception, digital twins, intelligent simulation and deduction, and adaptive anomaly detection to realize the transformation of tower climbing operations from passive protection to active prevention, and comprehensively improve the intelligence and precision of high-altitude operation safety management. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a tower climbing fall protection safety management system based on intelligent monitoring. This system utilizes a physical sensing layer to deploy multiple sensors on personnel equipment and the tower body to collect real-time data on attitude, protective equipment, structural deformation, and the environment, transmitting this data to edge nodes via a 5G industrial private network. A digital twin platform layer constructs a 1:1 real-time mapped digital twin model and employs a multi-source data weighted synchronization algorithm to ensure accurate synchronization. A simulation and deduction layer performs multi-dimensional fall risk simulations before operations and generates emergency plans. A real-time control layer identifies anomalies based on an individual behavior fingerprint adaptive detection algorithm, triggering a four-level warning system and dynamically correcting deviations using active protection devices. An evaluation and analysis layer automatically generates a safety assessment report after operations, thereby achieving intelligent, closed-loop safety management throughout the entire tower climbing operation process.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a tower climbing fall protection safety management system based on intelligent monitoring, the system comprising the following components:

[0007] The physical sensing layer is deployed on the tower and the protective equipment of the workers to collect real-time data on the posture and position of the workers, the status of the protective equipment, the deformation data of the tower structure, and the environmental meteorological data.

[0008] The communication connection between the digital twin platform layer and the physical perception layer is used to construct a digital twin model that is a 1:1 real-time mapping of the physical tower climbing operation scenario, and to synchronize all data collected by the physical perception layer to the digital twin model in real time to achieve bidirectional linkage between the physical scene and the virtual scene;

[0009] The simulation and deduction layer is connected to the digital twin platform layer to load the digital twin model of the target tower before operation, input the operator information, operation content and predicted environmental parameters to conduct multi-dimensional safety risk simulation and deduction, and pre-simulate possible fall accidents under different working conditions and generate corresponding emergency response plans.

[0010] The real-time control layer is connected to the digital twin platform layer and the simulation and deduction layer to monitor and identify violations and dangerous conditions in real time based on the real-time updated digital twin model during the operation, and trigger graded early warnings. At the same time, it links the active protection device to perform dynamic correction.

[0011] The assessment and analysis layer is connected to the digital twin platform layer and the real-time control layer to automatically generate a safety assessment report based on the full-process data recorded by the digital twin model after the operation is completed. The report includes operation compliance analysis, risk point statistics, accident hazard investigation and improvement suggestions.

[0012] Furthermore, the physical sensing layer integrates a 1280×720 resolution 30fps binocular vision camera and a six-axis inertial measurement unit with a range of ±16g and a sampling frequency of 200Hz on the top of the worker's safety helmet. A tension sensor with a range of 0-5000N and an accuracy of 0.1N and a Hall effect locking sensor with a response time of 1ms are integrated into the main strap and hook of the safety belt, respectively. Distributed fiber optic strain sensors with a strain measurement range of 0-10000με are deployed every 2m along the tower ladder. Ultrasonic anemometers with a range of 0-60m / s and an accuracy of 0.1m / s are deployed at both ends of the tower crossarm and at the tower corners. A 77GHz millimeter-wave radar with a detection range of 0-100m and an angle of 120° is symmetrically deployed on both sides of the tower base. All sensors establish low-latency communication connections with the tower base edge computing nodes via a 5G industrial private network, and data transmission uses the AES-256 encryption algorithm to ensure data security.

[0013] Furthermore, the tower twin module of the digital twin platform layer uses a combination of ground-based LiDAR and UAV oblique photography to acquire tower point cloud data, with a point cloud density of no less than 50 points / m². Through point cloud denoising, registration, and segmentation algorithms, the structural features of the tower are extracted to construct a high-precision 3D model containing details of all angle steel, bolts, ladders, crossarms, and insulators. The model accuracy is controlled within ±5mm. The personnel twin module uses data from binocular vision cameras and inertial measurement units to construct a human skeleton model with 21 key points through human posture estimation algorithms. It maps the posture and position of the worker's head, torso, and limbs in real time, with an update frequency of no less than 50Hz. The environmental twin module synchronizes environmental parameters such as wind speed, temperature, humidity, and icing thickness in real time and visualizes the environmental risk level of different areas in the digital twin model with gradient colors.

[0014] Furthermore, the digital twin platform layer employs a multi-source data weighted synchronization error correction algorithm to achieve precise synchronization between the physical and virtual scenes. The formula is as follows: ,in The total synchronization error of the digital twin model. This represents the total number of sensor types. For the first Weighting coefficients for sensor-like devices For the first The actual data of the physical entity collected by the sensor-like device. This refers to the current data of the corresponding virtual entity in the digital twin model. For the first The transmission delay of sensor data For the data synchronization cycle, the synchronization error correction algorithm further includes: when the total synchronization error of the current cycle is calculated... Subsequently, the system uses a Kalman filter to perform timestamp alignment and interpolation compensation on the data streams from each sensor, dynamically adjusting the virtual model. The predicted value, and in the next period With minimization as the objective, the synchronization state of the virtual model is corrected in real time, and the weighting coefficients are adjusted accordingly. The result was obtained through the entropy weight method. The calculation process was based on historical synchronization error data from 1200 sets of different operating conditions over the past three years, including data transmission delay. The data synchronization period T is set to 20ms and is obtained by real-time measurement of the network latency between the sensor and the edge computing node.

[0015] Furthermore, the simulation and deduction layer divides fall accident scenarios into four major categories and 16 specific scenarios: personnel error, equipment failure, tower failure, and environmental change. Each scenario has three different severity parameter levels. The accident simulation process uses a discrete event simulation method, simulating the complete process of an accident from occurrence to end with a time step of 0.1 seconds. It records the position, posture, speed of the workers, and the state changes of protective equipment and tower structure within each time step. After the simulation, it generates quantitative assessment results of the accident's impact range, the degree of personnel injury, and economic losses. The contingency plan generation module automatically adjusts the trigger thresholds of different warning levels, the activation time of active protection devices, and the assembly location and rescue route of rescue personnel based on the simulation results. The generated contingency plan is automatically synchronized to the smart terminal devices of all participating workers.

[0016] Furthermore, the simulation layer employs a multi-dimensional coupled fall risk assessment model to calculate the comprehensive fall risk value during the operation process, using the following formula: ,in For dynamic anomaly thresholds, The baseline threshold for individual worker behavior is calculated using the average posture data, movement frequency data, and physiological data from the worker's past 10 normal tower climbing operations. This is the environmental impact factor. The influence coefficient of the tower structure state was determined by testing the changes in human behavior under different environmental and tower structure states using orthogonal experimental methods. The specific value was [value missing]. , , This represents the normalized environmental state risk value. The normalized tower status risk value ranges from 0 to 1. When the real-time detected personnel behavior parameters exceed the dynamic abnormal threshold, the system determines it as abnormal behavior.

[0017] Furthermore, the real-time control layer classifies abnormal situations into four levels: general violation, major danger, serious danger, and impending fall. The general violation level triggers an 85dB buzzer warning on the worker's safety helmet and a 100Hz vibration motor warning on the safety belt. The major danger level simultaneously triggers a red audible and visual warning on the ground monitoring terminal and a pop-up warning on the control center terminal. The serious danger level displays the danger location and corrective guidance to the worker through AR glasses and forcibly suspends the operation. The impending fall level automatically triggers the speed difference controller to lock and the safety net to pop up at the preset position on the tower. The linkage control module and the active protection device are connected by hard wiring, with a signal transmission delay of less than 10ms, ensuring that the protective action is completed within 100ms after the fall occurs.

[0018] Furthermore, the real-time control layer employs an adaptive anomaly detection algorithm based on individual behavioral fingerprints to identify abnormal behaviors of operators, as shown in the formula: ,in To adapt the anomaly detection threshold, This represents the average of the behavioral characteristics of the workers during their last 10 normal tower climbing operations. The standard deviation of the behavioral characteristics corresponding to the operator's last 10 normal tower climbing operations. This is the confidence coefficient, with a value of 3. The environmental impact factor is 0.02. The current ambient wind speed is expressed in m / s. Behavioral characteristics include tower climbing cadence, body tilt angle, and hand grip interval. The system calculates the individual values ​​for each of these behavioral characteristics. and , obtained multiple When any feature exceeds its corresponding If the behavior is deemed abnormal, the system will collect 30 minutes of normal work data when the operator first climbs the tower to establish an initial behavior fingerprint database. The behavior fingerprint database will be automatically updated after each subsequent operation.

[0019] Furthermore, the assessment and analysis layer automatically extracts the full-process data recorded in the digital twin model after the operation is completed. The data time resolution is 0.1s. The data statistics module compiles basic data such as total operation time, maximum climbing height, cumulative movement distance, number of violations, number of warnings, and number of interventions. The compliance analysis module compares the actual operation process with the standard operation process step by step, marks all non-compliant behaviors and records the occurrence time and duration. The hazard investigation module identifies tower structure hazards, protective equipment hazards, and management process hazards exposed during the operation through correlation analysis, generates a hazard list including hazard location, hazard level, and rectification suggestions, and finally automatically generates a PDF safety assessment report and stores it on the cloud server.

[0020] Furthermore, the system adopts an edge-cloud collaborative distributed architecture. Edge computing nodes are deployed near the tower base using industrial-grade embedded processors with a computing power of no less than 8 TOPS. They are responsible for local preprocessing of physical perception layer data, anomaly detection, and local early warning response, with a response latency of less than 50ms. The cloud platform is deployed on Alibaba Cloud ECS servers and is responsible for digital twin model construction, simulation, evaluation and analysis, and multi-tower collaborative management. It supports the simultaneous management of the safety of climbing operations on more than 100 towers in the same area. Model updates adopt an incremental update method, with each update data volume not exceeding 100MB. The update process does not affect the normal operation of the system. Data storage adopts a distributed storage method, and all data is retained for no less than 3 years.

[0021] Compared with existing technologies, this intelligent monitoring-based tower fall protection safety management system has the following advantages:

[0022] I. This system achieves intelligent safety management and control of the entire tower climbing operation process through a five-layer collaborative architecture: physical perception layer, digital twin platform layer, simulation and deduction layer, real-time control layer, and evaluation and analysis layer. The system employs a multi-source data weighted synchronization error correction algorithm to ensure a precise 1:1 mapping between the physical and virtual scenes. Combined with an adaptive anomaly detection algorithm based on individual behavioral fingerprints, it can dynamically adjust detection thresholds according to the behavioral habits of different operators, effectively identifying violations and dangerous conditions. Furthermore, it dynamically corrects deviations through a four-level graded early warning and linkage active protection device, significantly reducing the probability of tower climbing falls.

[0023] Second, this system adopts an edge-cloud collaborative distributed architecture. Edge computing nodes are responsible for local preprocessing and anomaly detection with a response latency of less than 50ms. The cloud platform supports the simultaneous management of tower climbing operations safety for more than 100 towers in the same area. After the operation is completed, it automatically generates a safety assessment report that includes operation compliance analysis, risk point statistics, accident hazard investigation and improvement suggestions. It realizes closed-loop management of the entire life cycle from risk simulation before operation, real-time control during operation to post-operation evaluation and improvement, which greatly improves the efficiency and precision of tower climbing operation safety management.

[0024] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0026] Figure 1 This is an overall flowchart of a tower fall protection safety management and control system based on intelligent monitoring;

[0027] Figure 2 A flowchart for digital twin multi-source data weighted synchronization error correction of a tower fall protection safety management and control system based on intelligent monitoring;

[0028] Figure 3 This is a flowchart illustrating the real-time management, four-level early warning, and active protection linkage of a tower fall protection safety management system based on intelligent monitoring. Detailed Implementation

[0029] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0030] Example

[0031] This operation was carried out on a straight-line tower of a 110kV transmission line in a certain area, numbered S107. The operation involved replacing three damaged insulators and tightening loose bolts on the crossarm. A total of three personnel participated in the operation, including two tower climbers and one ground monitor. The weather forecast before the operation indicated that the daytime was sunny with an average wind speed of 3 m / s and a maximum instantaneous wind speed of 8 m / s.

[0032] Thirty minutes before the operation, the system simulation layer loads the digital twin model of the tower, inputting the basic information of two workers, the specific content of the operation, and predicted environmental parameters. It addresses 16 specific fall accident scenarios across four categories: personnel error, equipment failure, tower malfunction, and sudden environmental changes. Each scenario has three severity levels: mild, moderate, and severe. A discrete event simulation method is used with a time step of 0.1 seconds to simulate the entire process from the occurrence to the end of the accident, recording the position, posture, and speed of the workers, as well as the changes in the state of protective equipment and the tower structure within each time step. During the simulation, a multi-dimensional coupled fall risk assessment model is used to calculate the comprehensive fall risk value during the operation; the formula is: ,in For dynamic anomaly thresholds, This serves as the baseline threshold for individual worker behavior. This is the environmental impact factor. The tower body condition influence coefficient. This represents the normalized environmental state risk value. To generate a normalized risk value for the tower body, the simulation results include quantitative assessments of the accident's impact range, the degree of personal injury, and economic losses. The contingency plan generation module automatically adjusts the trigger thresholds for different levels of warnings, the activation time of active protection devices, and the assembly locations and rescue routes of rescue personnel based on the simulation results. The generated contingency plan is automatically synchronized to the smart terminal devices of all personnel involved in the operation.

[0033] The digital twin platform layer had previously used a combination of ground-based lidar and UAV oblique photography to acquire point cloud data of the tower, with a point cloud density of no less than 50 points / m². The tower's structural features were extracted through point cloud denoising, registration, and segmentation algorithms, and a high-precision 3D model containing details of all angle steel, bolts, ladders, crossarms, and insulators was constructed. The model's accuracy was controlled within ±5mm. Incremental updates of the model were completed before the operation began, and the update process did not affect the normal operation of the system.

[0034] Workers wear protective gear integrated with smart sensors. The top of the safety helmet integrates a 1280×720 resolution 30fps binocular vision camera and a six-axis inertial measurement unit with a range of ±16g and a sampling frequency of 200Hz. The main strap and hook of the safety belt integrate a tension sensor with a range of 0-5000N and an accuracy of 0.1N, and a Hall lock sensor with a response time of 1ms, respectively. All sensors establish low-latency communication connections with the tower base edge computing nodes through a 5G industrial private network. Data transmission uses the AES-256 encryption algorithm to ensure data security. Every 2m along the tower ladder, a distributed fiber optic strain sensor with a strain measurement range of 0-10000με is deployed. Ultrasonic wind speed sensors with a range of 0-60m / s and an accuracy of 0.1m / s are deployed at both ends of the tower crossarm and at the tower corners. A 77GHz millimeter-wave radar with a detection range of 0-100m and an angle of 120° is symmetrically deployed on both sides of the tower base. The edge computing nodes are deployed near the tower base and use industrial-grade embedded processors to be responsible for local preprocessing of physical perception layer data, anomaly detection, and local early warning response.

[0035] After the operation begins, the physical sensing layer collects real-time data on the posture and position of the workers, the status of protective equipment, the deformation data of the tower structure, and environmental meteorological data, and transmits this data to the edge computing nodes. The digital twin platform layer synchronizes all the collected data to the digital twin model in real time, and uses a multi-source data weighted synchronization error correction algorithm to achieve accurate synchronization between the physical and virtual scenes; the formula is: ,in The total synchronization error of the digital twin model. This represents the total number of sensor types. For the first Weighting coefficients for sensor-like devices For the first The actual data of the physical entity collected by the sensor-like device. This refers to the current data of the corresponding virtual entity in the digital twin model. For the first The transmission delay of sensor data For the data synchronization period, the current period is calculated. Subsequently, the system uses a Kalman filter to perform timestamp alignment and interpolation compensation on the data streams from each sensor, dynamically adjusting the virtual model. The predicted value makes the next period's... Once the temperature drops below a preset threshold, the personnel twin module constructs a human skeleton model with 21 key points based on data from a binocular vision camera and an inertial measurement unit using a human posture estimation algorithm. This model maps the posture and position of the worker's head, torso, and limbs in real time, with an update frequency of no less than 50Hz. The environmental twin module synchronizes environmental parameters such as wind speed, temperature, humidity, and icing thickness in real time and visualizes the environmental risk level of different areas in the digital twin model using gradient colors.

[0036] The real-time control layer monitors the operational behavior of workers, the status of protective equipment, and environmental risks in real time based on a constantly updated digital twin model. It uses an adaptive anomaly detection algorithm based on individual behavioral fingerprints to identify abnormal worker behaviors; the formula is: ,in To adapt the anomaly detection threshold, This represents the average of the behavioral characteristics of the workers during their last 10 normal tower climbing operations. The standard deviation of the behavioral characteristics corresponding to the operator's last 10 normal tower climbing operations. Here is the confidence coefficient. This is the environmental impact factor. Given the current ambient wind speed, the system calculates the tower climbing cadence, body tilt angle, and hand gripping interval time separately. When any feature exceeds the corresponding threshold, it is judged as abnormal. Abnormal situations are classified into four levels: general violation, major danger, serious danger, and impending fall. During the operation, a worker failed to hook his safety belt hook to the fall arrest rail of the ladder while climbing the tower. After the system identifies this, it triggers a general violation warning. The 85dB buzzer on the worker's safety helmet and the 100Hz vibration motor on the safety belt start simultaneously. The worker immediately corrects the violation after receiving the warning. When the worker climbs to the middle of the tower, the ultrasonic wind speed sensor detects that the instantaneous wind speed reaches 12m / s. The system triggers a major danger warning. At the same time, the red sound and light warning on the ground monitoring terminal and the pop-up warning on the control center terminal are activated. The ground monitoring personnel remind the worker to stop the operation and take shelter in a safe position. The operation can continue after the wind speed drops to a safe value. When the worker is replacing an insulator at the crossarm, he slips and loses his balance briefly. The system identifies this as a serious danger level. The system displays the dangerous location and correction instructions to the worker through AR glasses and forcibly stops the operation. The stop command is lifted after the worker regains a stable posture. The entire operation linkage control module and the active protection device are connected by hard wiring to ensure the reliability of command transmission. In the simulation test, the time interval from the moment the system determines that a fall is about to occur to the completion of the speed difference automatic controller locking is always controlled within 100ms.

[0037] After the operation is completed, the assessment and analysis layer automatically extracts the full-process data recorded in the digital twin model. The data statistics module compiles basic data such as the total operation time of 2 hours and 15 minutes, the highest climbing height of 28m, the cumulative movement distance of 126m, the number of violations of 1, the number of warnings of 3, and the number of interventions of 1. The compliance analysis module compares the actual operation process with the standard operation process step by step, marks all non-compliant behaviors and records the occurrence time and duration. The hazard investigation module identifies two loose bolt hazards on the tower crossarm, one safety belt tension sensor sensitivity hazard, and the hazard of insufficient equipment inspection before operation in the management process through correlation analysis. It generates a hazard list including the location of the hazard, the hazard level, and rectification suggestions. Finally, it automatically generates a safety assessment report in PDF format and stores it on the cloud server. The cloud platform is deployed on Alibaba Cloud ECS server and supports the simultaneous management of tower climbing operation safety of more than 100 towers in the same area. Data storage adopts a distributed storage method.

[0038] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A tower fall protection safety management system based on intelligent monitoring, characterized in that, The system includes the following components: The physical sensing layer is deployed on the tower body and the protective equipment of the workers to collect real-time data on the posture and position of the workers, the status data of the protective equipment, the deformation data of the tower structure, and the environmental meteorological data. The communication connection between the digital twin platform layer and the physical perception layer is used to construct a digital twin model that is a 1:1 real-time mapping of the physical tower climbing operation scenario, and to synchronize all data collected by the physical perception layer to the digital twin model in real time to achieve bidirectional linkage between the physical scene and the virtual scene; The simulation and deduction layer is connected to the digital twin platform layer to load the digital twin model of the target tower before operation, input the operator information, operation content and predicted environmental parameters to conduct multi-dimensional safety risk simulation and deduction, and pre-simulate possible fall accidents under different working conditions and generate corresponding emergency response plans. The real-time control layer is connected to the digital twin platform layer and the simulation and deduction layer to monitor and identify violations and dangerous conditions in real time based on the real-time updated digital twin model during the operation, and trigger graded early warnings. At the same time, it links the active protection device to perform dynamic correction. The assessment and analysis layer is connected to the digital twin platform layer and the real-time control layer to automatically generate a safety assessment report based on the full-process data recorded by the digital twin model after the operation is completed. The report includes operation compliance analysis, risk point statistics, accident hazard investigation and improvement suggestions.

2. The intelligent monitoring-based tower fall protection safety management system according to claim 1, characterized in that, The physical sensing layer integrates a 1280×720 resolution, 30fps binocular vision camera and a six-axis inertial measurement unit with a range of ±16g and a sampling frequency of 200Hz on the top of the worker's safety helmet. A tension sensor with a range of 0-5000N and an accuracy of 0.1N, and a Hall effect locking sensor with a response time of 1ms are integrated into the main strap and hook of the safety belt, respectively. Distributed fiber optic strain sensors with a strain measurement range of 0-10000με are deployed every 2m along the tower ladder. Ultrasonic anemometers with a range of 0-60m / s and an accuracy of 0.1m / s are deployed at both ends of the tower crossarm and at the tower corners. A 77GHz millimeter-wave radar with a detection range of 0-100m and an angle of 120° is symmetrically deployed on both sides of the tower base. All sensors establish low-latency communication connections with the tower base edge computing nodes via a 5G industrial private network, and data transmission uses the AES-256 encryption algorithm to ensure data security.

3. The intelligent monitoring-based tower fall protection safety management system according to claim 1, characterized in that, The tower twin module of the digital twin platform layer uses a combination of ground-based LiDAR and UAV oblique photography to acquire tower point cloud data with a point cloud density of no less than 50 points / m². Through point cloud denoising, registration, and segmentation algorithms, the structural features of the tower are extracted to construct a high-precision 3D model containing details of all angle steel, bolts, ladders, crossarms, and insulators. The model accuracy is controlled within ±5mm. The personnel twin module uses data from binocular vision cameras and inertial measurement units to construct a human skeleton model with 21 key points through human posture estimation algorithms. It maps the posture and position of the worker's head, torso, and limbs in real time with an update frequency of no less than 50Hz. The environmental twin module synchronizes environmental parameters such as wind speed, temperature, humidity, and icing thickness in real time and visualizes the environmental risk level of different areas in the digital twin model with gradient colors.

4. The intelligent monitoring-based tower fall protection safety management system according to claim 1, characterized in that, The digital twin platform layer employs a multi-source data weighted synchronization error correction algorithm to achieve accurate synchronization between the physical and virtual scenes. The formula is as follows: ,in The total synchronization error of the digital twin model. This represents the total number of sensor types. For the first Weighting coefficients for sensor-like devices For the first The actual data of the physical entity collected by the sensor-like device. This refers to the current data of the corresponding virtual entity in the digital twin model. For the first The transmission delay of sensor data This refers to the data synchronization cycle.

5. A tower fall protection safety management system based on intelligent monitoring according to claim 1, characterized in that, The simulation and deduction layer divides fall accident scenarios into four major categories and 16 specific scenarios: personnel error, equipment failure, tower failure, and environmental change. Each scenario has three different severity parameter levels. The accident simulation process adopts the discrete event simulation method, simulating the complete process of the accident from occurrence to end with a time step of 0.1s. It records the position, posture, speed of the workers, and the state changes of protective equipment and tower structure within each time step. After the simulation, it generates quantitative assessment results of the accident's impact range, personnel injury degree, and economic loss. The contingency plan generation module automatically adjusts the trigger thresholds of different warning levels, the activation time of active protection devices, and the assembly position and rescue route of rescue personnel based on the simulation results. The generated contingency plan is automatically synchronized to the smart terminal devices of all participating workers.

6. A tower fall protection safety management system based on intelligent monitoring according to claim 1, characterized in that, The simulation layer uses a multi-dimensional coupled fall risk assessment model to calculate the comprehensive fall risk value during the operation process. The formula is as follows: ,in For dynamic anomaly thresholds, This serves as the baseline threshold for individual worker behavior. This is the environmental impact factor. The tower body condition influence coefficient. This represents the normalized environmental state risk value. This is the normalized risk value of the tower body condition.

7. A tower fall protection safety management system based on intelligent monitoring according to claim 1, characterized in that, The real-time control layer classifies abnormal situations into four levels: general violation, major danger, serious danger, and impending fall. The general violation level triggers an 85dB buzzer warning on the worker's safety helmet and a 100Hz vibration motor warning on the safety belt. The major danger level triggers a red audible and visual warning on the ground monitoring terminal and a pop-up warning on the control center terminal. The serious danger level displays the danger location and correction guidance to the worker through AR glasses and forcibly suspends the operation. The impending fall level automatically triggers the speed difference controller to lock and the safety net to pop up at the preset position on the tower. The linkage control module and the active protection device are connected by hard wiring.

8. A tower fall protection safety management system based on intelligent monitoring according to claim 1, characterized in that, The real-time control layer employs an adaptive anomaly detection algorithm based on individual behavioral fingerprints to identify abnormal behaviors of operators. The formula is as follows: ,in To adapt the anomaly detection threshold, This represents the average of the behavioral characteristics of the workers during their last 10 normal tower climbing operations. The standard deviation of the behavioral characteristics corresponding to the operator's last 10 normal tower climbing operations. Here is the confidence coefficient. This is the environmental impact factor. This refers to the current ambient wind speed.

9. A tower fall protection safety management system based on intelligent monitoring according to claim 1, characterized in that, After the operation is completed, the assessment and analysis layer automatically extracts the full-process data recorded in the digital twin model. The data statistics module compiles basic data such as total operation time, maximum climbing height, cumulative movement distance, number of violations, number of warnings, and number of interventions. The compliance analysis module compares the actual operation process with the standard operation process step by step, marks all non-compliant behaviors and records the time of occurrence and duration. The hazard investigation module identifies tower structure hazards, protective equipment hazards, and management process hazards exposed during the operation through correlation analysis, generates a hazard list including hazard location, hazard level, and rectification suggestions, and finally automatically generates a safety assessment report in PDF format and stores it on the cloud server.

10. A tower fall protection safety management system based on intelligent monitoring according to claim 1, characterized in that, The system adopts an edge-cloud collaborative distributed architecture. Edge computing nodes are deployed near the tower base using industrial-grade embedded processors, which are responsible for local preprocessing of physical perception layer data, anomaly detection, and local early warning response. The cloud platform is deployed on Alibaba Cloud ECS servers, which are responsible for digital twin model construction, simulation, evaluation and analysis, and multi-tower collaborative management and control. It supports the simultaneous management of the safety of climbing operations on more than 100 towers in the same area. The model is updated incrementally, and the update process does not affect the normal operation of the system. Data storage adopts a distributed storage method.

Images