Intelligent safety monitoring system and method for construction hoist

Abstract

This invention discloses an intelligent safety monitoring system and method for construction hoists, belonging to the field of construction safety monitoring technology. It constructs a multi-dimensional risk feature set encompassing "meteorology-equipment-space" through the collaborative work of a hoist status determination module, an operation channel obstacle perception module, an operation channel obstacle risk assessment module, a meteorological environment monitoring module, and a fusion analysis and hierarchical control module. This enables comprehensive perception and fusion analysis of equipment operating status, personnel behavior / cargo status, physical working space of the operation channel, and meteorological environment, filling the gap in physical space obstacle monitoring. Based on the multi-dimensional risk feature set, a hierarchical priority control architecture is adopted to achieve an organic combination of independent triggering of emergency risks and multi-risk coupled analysis, representing a leap from "single threshold comparison" to "hierarchical progressive control" and from "isolated risk determination" to "multi-source coupled analysis," thereby improving the safety level of construction hoists under complex working conditions.

Description

Technical Field

[0001] This invention relates to the field of construction safety monitoring technology, and in particular to an intelligent safety monitoring system and method for construction hoists. Background Technology

[0002] Construction hoists, also known as building construction elevators, outdoor elevators, or construction site hoisting cages, are commonly used construction machinery for carrying people and goods. They are mainly used for the construction of interior and exterior decoration of high-rise buildings, bridges, chimneys, and other structures. Traditional safety monitoring of construction hoists mainly relies on manual experience and regular maintenance, which has problems such as lagging supervision, limited risk identification, and low level of intelligence.

[0003] In recent years, with the development of the Internet of Things and sensor technology, some intelligent safety monitoring solutions have been proposed. For example, Chinese patent document CN118561120A discloses "An Intelligent Construction Hoist and Its Safety Monitoring and Management System". This system collects the equipment's own operating data, image data inside the hoist cage, and environmental data outside the hoist cage to assess the risk values ​​of tipping over, failure, and worker behavior, and calculates the hoist's overall operating risk value. Based on the analysis results, it generates operational suggestions and triggers alarms when the operating risk exceeds the standard. This solution achieves a comprehensive assessment of the hoist's operating risk from multiple dimensions, which is a significant improvement over traditional manual supervision methods.

[0004] However, the solution presented in the above technical content still has the following technical defects: First, there is a lack of active detection capability for physical obstacles: the system does not cover the detection of physical obstacles in the elevator's operating channel. In actual construction scenarios, collision accidents caused by scaffolding not being removed in time, steel pipes extending unexpectedly, or protective materials intruding into the elevator's operating trajectory occur frequently.

[0005] Second, there is a lack of monitoring capabilities for interference from surrounding operations: Modern construction sites often involve multiple tower cranes and elevators working together. During the rotation of the tower crane boom, it may encroach on the operating space of the elevator, causing serious safety accidents.

[0006] Third, the risk control logic is simplistic: although warning thresholds are set, the control logic is mainly "to trigger an alarm when the overall risk value exceeds the preset threshold" and "to provide corresponding operational suggestions when a single risk exceeds the limit". This control method is a passive fixed threshold response type, lacking a gradual and proactive intervention mechanism strategy based on early warning of risk trends.

[0007] To address the aforementioned technical issues, it is necessary to develop an intelligent construction hoist safety monitoring system and method that can comprehensively perceive physical space obstacles, monitor interference from surrounding operations, and perform integrated analysis and hierarchical progressive control of multi-source risk factors. Summary of the Invention

[0008] The purpose of this invention is to address the aforementioned problems by providing an intelligent safety monitoring system and method for construction hoists. Based on the multi-dimensional risk feature set of "meteorology-equipment-space" constructed according to this invention, it achieves comprehensive perception and fusion analysis of equipment operating status, personnel behavior / cargo status, physical working space of the operating channel, and meteorological environment, filling the gap in physical space obstacle monitoring. Furthermore, based on the multi-dimensional risk feature set, a hierarchical priority control architecture is adopted to achieve an organic combination of independent triggering of emergency risks and multi-risk coupling analysis, representing a leapfrog improvement from "single threshold comparison" to "hierarchical progressive control" and from "isolated risk judgment" to "multi-source coupling analysis".

[0009] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: According to one aspect of the present invention, an intelligent safety monitoring system and method for construction hoists are provided, comprising: The hoist status determination module is used to collect equipment operation data and image data inside the cage of the construction hoist. Based on the equipment operation data and image data, it performs equipment operation safety determination, personnel behavior and load status safety determination, and constructs equipment operation risk feature set and status risk feature set based on the determination results. The obstacle perception module for the operating channel is used to collect three-dimensional spatial data within the operating channel of the hoist cage, identify intruding obstacles within the operating channel, and distinguish obstacle types based on the motion characteristics of the intruding obstacles, including static obstacles and dynamic mechanical obstacles; The passage obstacle risk assessment module is used to receive three-dimensional spatial data and equipment operation data, perform static intrusion safety judgment for static obstacles, and dynamic interference safety judgment for dynamic mechanical obstacles, and construct static collision risk feature sets and dynamic interference risk feature sets based on the judgment results; The work environment monitoring module is used to acquire environmental data outside the hoist cage and the current height of the hoist cage, make a safety judgment on the work environment based on the environmental data and the current height of the hoist cage, and construct a meteorological risk feature set based on the judgment results; The fusion analysis and hierarchical control module is used to receive all risk feature sets and perform multi-priority hierarchical early warning and control based on various risk feature sets.

[0010] Furthermore, the specific procedures for determining equipment operation safety and personnel behavior and cargo status safety based on equipment operation data and image data include: Collect equipment operation data, including cage load, tilt angle, and actual operating speed. Compare the cage load with the rated load. If it exceeds the preset threshold, it is determined to be an overload state. Compare the tilt angle with the preset safety angle. If it exceeds the preset threshold, it is determined to be an abnormal tilt state. Compare the actual operating speed with the theoretical operating speed. If the deviation exceeds the preset threshold, it is determined to be an abnormal speed state. Image recognition algorithms are used to identify whether personnel inside the cage are wearing safety helmets; if not, it is considered a helmet-missing behavior. The algorithms also identify whether personnel inside the cage are running, jumping, or engaging in any unauthorized operations; if so, it is considered a dangerous operation. The algorithms further identify the location of materials stacked inside the cage; if the offset between the center of the materials and the center of the cage exceeds a safety threshold, it is considered an off-center load.

[0011] Furthermore, the specific process for extracting the equipment operation risk feature set and behavioral risk feature set includes: When an overload condition is determined, the ratio of the current load to the rated load is extracted as the load deviation rate feature. When an abnormal tilt condition is determined, the rate of change of tilt angle is extracted as the tilt change rate feature. When an abnormal speed condition is determined, the speed deviation is extracted as the speed fluctuation rate feature. All of these are uniformly summarized into the equipment operation risk feature set. When a safety helmet is deemed missing, the proportion of personnel not wearing safety helmets within a preset time window is counted as the safety helmet missing rate feature. When a dangerous operation is deemed to be dangerous, the number of times the dangerous operation occurs within a preset time window is counted as the dangerous operation frequency feature. When an off-center load is deemed to be, the offset of the material center relative to the cage center is calculated as the off-center load feature. All of these are uniformly summarized into the state risk feature set.

[0012] Preferably, the process of performing static intrusion safety assessment and extracting the static collision risk feature set includes: Based on three-dimensional spatial data, static obstacle point clouds that intrude into the operating space are extracted. The minimum static relative distance between the static obstacle and the outermost edge of the operating channel is calculated. If it is less than the preset static safety distance threshold, it is determined that there is a risk of static obstacle intrusion into the operating channel. The coordinates of the static obstacle point cloud corresponding to the minimum static relative distance are taken as the static collision risk location. The time required for the elevator to run to this location is calculated as the expected collision time feature and summarized into the static collision risk feature set.

[0013] Preferably, the process of performing dynamic interference safety assessment and extracting dynamic interference risk feature set includes: The real-time position and speed of the dynamic mechanical obstacle are obtained. Combined with the real-time speed of the elevator body and the preset operating time window, a motion trajectory prediction model of the dynamic mechanical obstacle and the elevator is established. The minimum dynamic approach distance between the two is calculated. If it is less than the preset dynamic safety distance threshold, it is determined that there is a risk of dynamic mechanical cross-interference in the operating channel. The predicted meeting position corresponding to the minimum dynamic approach distance is taken as the trajectory interference intersection point. The time required to reach the trajectory interference intersection point from the current time is taken as the expected interference time feature and summarized into the dynamic interference risk feature set.

[0014] Preferably, the work environment monitoring module is used to collect real-time wind speed and wind direction data as environmental data, and to make a work environment safety judgment based on the environmental data and the current height of the cage. Specifically, the real-time wind speed, wind direction angle and the current height of the cage are input into the built-in wind impact risk level model, and the output wind impact risk level and real-time wind speed are used as meteorological impact features and summarized into the meteorological risk feature set. The wind impact risk level includes four levels: low risk, medium risk, high risk and emergency risk.

[0015] Preferably, the specific process for multi-priority hierarchical early warning and control based on various risk characteristic sets includes: The first priority control unit is used to determine the wind impact risk level. If the wind impact risk level is an emergency risk level, it generates an emergency warning signal and executes an emergency braking command. The second priority control unit is activated when the wind impact risk level is non-emergency risk level and the static estimated collision time and / or dynamic estimated intervention time are received. It is used to calculate the longest adjustable time window based on the estimated collision time and estimated intervention time, and execute the corresponding graded control commands based on the interval in which the longest adjustable time window is located, including active deceleration and emergency braking. The third priority control unit is activated when the wind impact risk level is non-emergency risk level and the longest adjustable time window is greater than the preset warning time threshold. It is used to take the equipment operation risk feature set as input, introduce real-time wind speed as coupling factor, perform comprehensive risk judgment through a preset multi-output machine learning model, generate probability warning signals for various specific risk types, and execute corresponding differentiated equipment control instructions. The status risk alert unit is used to monitor in real time based on the status risk feature set. When each status risk feature exceeds the corresponding preset warning threshold, it generates the corresponding emergency warning signal.

[0016] The present invention also proposes an intelligent safety monitoring method for construction hoists, comprising the following steps: Step 1: Acquisition of multi-source data: Collect equipment operation data, image data, 3D spatial data, environmental data, and the current height of the hoist cage; Step 2: Multi-source data security assessment and feature set acquisition: Perform multi-channel security assessments based on various types of data, and construct corresponding risk feature sets based on the assessment results; Step 3: Receive all risk feature sets and conduct multi-priority hierarchical early warning and control based on various risk feature sets.

[0017] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention introduces an obstacle perception module and an obstacle risk assessment module for the operating channel to achieve active detection and accurate identification of static and dynamic mechanical obstacles within the operating channel. For static obstacles, it calculates the minimum static relative distance and predicts the collision time; for dynamic mechanical obstacles, it calculates the minimum dynamic approach distance and predicts the intervention time. This enables the system to identify spatial obstacle risks such as scaffolding not being removed, protruding steel pipes, and tower crane boom intrusion in advance, and to take proactive intervention measures before a collision occurs, fundamentally solving the blind spot problem of traditional technologies being unable to perceive physical spatial obstacles.

[0018] 2. This invention also constructs a multi-dimensional risk feature set of "meteorology-equipment-space", adopts a hierarchical priority control architecture, sets meteorological emergency risks as the highest priority to ensure immediate response, sets physical space collision risks as the second priority to achieve hierarchical control based on time urgency, sets equipment-meteorology coupling risks as the third priority, and uses machine learning models to achieve hidden risk mining and targeted regulation. At the same time, it provides parallel operation status risk reminders to achieve personnel behavior monitoring, realizing the organic combination of independent triggering of emergency risks and multi-risk coupling analysis, overcoming the shortcomings of the single control logic of traditional technology. Attached Figure Description

[0019] Figure 1 This is a block diagram illustrating the system module principle of the present invention; Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and preferred embodiments. However, it should be noted that many details listed in the specification are merely to provide the reader with a thorough understanding of one or more aspects of the invention, and these aspects of the invention can be achieved even without these specific details.

[0021] Please see Figure 1 The present invention provides an intelligent safety monitoring system for construction hoists, the technical solution of which is as follows: the system includes a hoist status determination module, an operation channel obstacle perception module, an operation channel obstacle risk assessment module, an operation environment monitoring module, and a fusion analysis and hierarchical control module.

[0022] The hoist status determination module is used to collect equipment operation data and image data inside the cage of the construction hoist. Based on the equipment operation data and image data, it performs equipment operation safety determination, personnel behavior and load status safety determination, and constructs equipment operation risk feature set and status risk feature set based on the determination results. The judgment process of this module includes: collecting equipment operation data using load sensors, encoders, and tilt sensors, including cage load, tilt angle, and actual operating speed; comparing the cage load with the rated load, and determining an overload state if it exceeds a preset threshold; comparing the tilt angle with a preset safety angle, and determining an abnormal tilt state if it exceeds a preset threshold; and comparing the actual operating speed with the theoretical operating speed, and determining an abnormal speed state if the deviation exceeds a preset threshold. High-definition cameras are used to collect image data inside the cage. Image recognition algorithms are used to identify whether personnel inside the cage are wearing safety helmets. If they are not wearing them, it is determined to be a helmet-missing behavior. The system also identifies whether personnel inside the cage are running, jumping, or engaging in any unauthorized operations. If such actions are found, they are determined to be dangerous operations. The system also identifies the location of materials stacked inside the cage. If the offset between the center of the materials and the center of the cage exceeds a safety threshold, it is determined to be an off-center load. The specific process for extracting equipment operation risk feature sets and behavioral risk feature sets includes: When an overload condition is determined, the ratio of the current load to the rated load is extracted as the load deviation rate feature. When an abnormal tilt condition is determined, the rate of change of tilt angle is extracted as the tilt change rate feature. When an abnormal speed condition is determined, the speed deviation is extracted as the speed fluctuation rate feature. All of these are uniformly summarized into the equipment operation risk feature set. When a safety helmet is deemed missing, the proportion of personnel not wearing safety helmets within a preset time window is counted as the safety helmet missing rate feature. When a dangerous operation is deemed to be dangerous, the number of times the dangerous operation occurs within a preset time window is counted as the dangerous operation frequency feature. When an off-center load is deemed to be, the offset of the material center relative to the cage center is calculated as the off-center load feature. All of these are uniformly summarized into the state risk feature set.

[0023] The obstacle perception module for the running channel is used to collect three-dimensional spatial data within the running channel of the hoist cage, identify intruding obstacles within the running channel, and distinguish obstacle types based on the motion characteristics of the intruding obstacles, including static obstacles and dynamic mechanical obstacles.

[0024] The passage obstacle risk assessment module is connected to the elevator status determination module and the multi-source obstacle fusion perception module of the operating passage. It is used to receive three-dimensional spatial data and equipment operation data, perform static intrusion safety determination for static obstacles, and dynamic interference safety determination for dynamic mechanical obstacles. Based on the determination results, it constructs a static collision risk feature set and a dynamic interference risk feature set. The process of determining static intrusion safety and extracting static collision risk feature sets for static obstacles includes: Based on three-dimensional spatial data, the inherent structural features of the pre-set guide rail frame and wall-mounted frame in the operating channel are removed by the point cloud comparison algorithm. The static obstacle point cloud that intrudes into the operating space is extracted. The minimum static relative distance between the static obstacle point cloud and the outermost edge of the pre-set operating channel is calculated. If the minimum static relative distance is less than the pre-set static safety distance threshold, it is determined that there is a risk of static obstacle intrusion in the operating channel. The coordinates of the static obstacle point cloud corresponding to the minimum static relative distance are taken as the static collision risk position. The time required for the elevator to run to the static collision risk position is calculated as the expected collision time feature and summarized into the static collision risk feature set. The process of dynamically determining the safety of mechanical obstacles and extracting a dynamic interference risk feature set includes: The real-time position and speed of the dynamic mechanical obstacle are obtained. Combined with the real-time speed of the elevator body and the preset operating time window, a motion trajectory prediction model of the dynamic machinery and the elevator is established. The minimum dynamic approach distance between the two within the preset time window is calculated. If it is less than the preset dynamic safety distance threshold, it is determined that there is a risk of dynamic mechanical cross-interference in the operating channel. The predicted meeting position of the elevator and the dynamic machinery when they reach the minimum dynamic approach distance is obtained. The predicted meeting position corresponding to the minimum dynamic approach distance is taken as the trajectory interference intersection point. The time required to reach the trajectory interference intersection point from the current time is taken as the expected interference time feature and summarized into the dynamic interference risk feature set.

[0025] The obstacle perception module in the operating channel collects three-dimensional spatial data in the operating channel of the hoist cage through multi-line lidar and ultra-wideband positioning unit, identifies intruding obstacles and distinguishes between static obstacles and dynamic mechanical obstacles. On the one hand, it fills the technical gap in physical space obstacle monitoring, and on the other hand, it realizes intelligent differentiation of obstacle types, providing a basis for differentiated risk assessment. The obstacle risk assessment module makes safety judgments for both static and dynamic mechanical obstacles, and extracts the expected collision time and expected interference time as core risk features. For static obstacles, spatial distance is converted into a time dimension to achieve risk quantification based on urgency. For dynamic mechanical obstacles, dynamic trajectory prediction technology is used to accurately capture the risk of "tower-ladder interference".

[0026] The work environment monitoring module is used to acquire environmental data outside the hoist cage and the current height of the hoist cage, make a safety judgment on the work environment based on the environmental data and the current height of the hoist cage, and construct a meteorological risk feature set based on the judgment results; The work environment monitoring module collects real-time wind speed and wind direction data as environmental data. Based on the environmental data and the current height of the hoist cage, it makes a safety judgment on the work environment. Specifically, it inputs the real-time wind speed, wind direction angle and the current height of the hoist cage into the built-in wind impact risk level model. The output wind impact risk level and real-time wind speed are used as meteorological impact features and summarized into the meteorological risk feature set. The wind impact risk level includes four levels: low risk, medium risk, high risk and emergency risk.

[0027] The integrated analysis and hierarchical control module is connected to the elevator status determination module, the passage obstacle risk assessment module, and the work environment monitoring module, respectively. It is used to receive all risk feature sets and perform multi-priority hierarchical early warning and control based on various risk feature sets. The fusion analysis and hierarchical control module includes a first priority control unit, a second priority control unit, a third priority control unit, and a status risk alert unit; The first priority control unit is used to determine the wind impact risk level based on the meteorological risk feature set. If the wind impact risk level is an emergency risk level, it generates an emergency warning signal and executes an emergency braking command. The second priority control unit is activated when the wind impact risk level is non-emergency risk level and the static estimated collision time and / or dynamic estimated intervention time are received. It is used to calculate the longest adjustable time window based on the estimated collision time and estimated intervention time, and to execute the corresponding graded control strategy based on the interval of the longest adjustable time window and the calculated threshold, including active deceleration and emergency braking. The third priority control unit is activated when the wind impact risk level is non-emergency and the longest adjustable time window is greater than the preset warning time threshold. It is used to take the equipment operation risk feature set as input and introduce real-time wind speed as a coupling factor. It performs comprehensive risk judgment through a preset machine learning model and generates probability warning signals for at least a variety of specific risk types, including rollover risk, structural fatigue risk, motor overheating risk, and transmission failure risk. It then executes corresponding differentiated equipment control instructions based on the probability warning signals of the specific risk types. The status risk alert unit operates in parallel with the first, second, and third priority control units. It is used to monitor the status risk feature set in real time. When each status risk feature exceeds the corresponding preset warning threshold, it generates the corresponding emergency warning signal. The emergency warning signal does not interfere with the equipment operation control. For the second priority control unit, the process of obtaining the longest adjustable time window includes: When only the predicted collision time feature or only the predicted interference time feature is received, the received single feature value is used as the longest adjustable time window T_control; The calculation thresholds of the longest adjustable time window include the warning time threshold T_warning and the deceleration time threshold T_decelerate, and T_warning > T_decelerate; Executing corresponding hierarchical control instructions according to the interval where the longest adjustable time window is located includes: When T_control > T_warning, continuously monitor and do not execute intervention; When T_decelerate < T_control ≤ T_warning, execute the active deceleration instruction; When T_control ≤ T_decelerate, execute the emergency braking instruction; When both the predicted collision time and the predicted interference time are received, compare their magnitudes, and use the smaller value as the longest adjustable time window T_control. This is because the smaller time represents a more urgent risk and needs to be processed first; Calculate the time interval Δt between the two. This interval is used to determine whether the two risks occur "continuously". If the time interval Δt is less than the preset continuous risk interval threshold, it is determined that there is a continuous collision risk, and an upgraded control strategy is executed according to the interval where the longest adjustable time window is located. Specifically: If the longest adjustable time window T_control is in the deceleration interval (the longest adjustable time window T_control is between the deceleration time threshold T_decelerate and the warning time threshold T_warning), then upgrade and execute the emergency braking instruction, indicating that deceleration may not be sufficient to handle the second risk immediately following, and direct braking is safer; If the longest adjustable time window T_control is in the warning interval (the longest adjustable time window T_control is greater than the warning time threshold T_warning), then upgrade and execute the active deceleration instruction. T_control being greater than T_warning means there is充裕 time, and only monitoring is needed. However, considering there are subsequent risks, decelerate in advance for prevention; The core problem solved by the second-priority control unit is: when there is one or more collision risks, how to determine the most urgent risk and take appropriate control measures according to the urgency of the risk. It considers both the urgency of a single risk (taking the minimum value) and the temporal correlation of multiple risks (interval judgment), and ensures sufficient safety redundancy in the worst case through upgraded control; The machine learning model for the third priority control unit is used to evaluate the impact of wind load on the stability of equipment operation. A multi-output gradient boosting tree is used as the core classification model. The model learns multiple risk types at the same time, captures the correlation between risks, and sets independent early warning thresholds and alarm thresholds for each risk type. The thresholds can be dynamically adjusted according to the environment. All features are normalized to eliminate the influence of dimensions. Real-time wind speed is used as a key coupling factor and is input into the model along with other features. Through multi-feature interaction, the model can capture coupling effects that cannot be reflected by a single feature, output the corresponding specific risk type, and execute differentiated control instructions.

[0028] This module adopts a hierarchical priority control architecture based on a multi-dimensional risk feature set, realizing an organic combination of independent triggering of emergency risks and multi-risk coupled analysis, a leapfrog improvement from "single threshold comparison" to "hierarchical progressive control", and from "isolated risk judgment" to "multi-source coupled analysis".

[0029] Example 2: This invention also proposes an intelligent safety monitoring method for construction hoists. Please refer to [link / reference]. Figure 2 It includes the following steps: Step 1: Acquisition of multi-source data: Collect equipment operation data, image data, 3D spatial data, environmental data, and the current height of the hoist cage; Step 2: Multi-source data security assessment and feature set acquisition: Perform multi-channel security assessments based on various types of data, and construct corresponding risk feature sets based on the assessment results; Based on equipment operation data, determine equipment operation safety and extract equipment operation risk feature set; based on image data, determine personnel behavior and cargo status safety and extract status risk feature set; based on 3D spatial data, determine static intrusion safety for static obstacles and extract static collision risk feature set; based on 3D spatial data and equipment operation data, determine dynamic interference safety for dynamic mechanical obstacles and extract dynamic interference risk feature set; based on environmental data and the current height of the hoist cage, determine the working environment safety and extract meteorological risk feature set. Step 3: Receive all risk feature sets and perform multi-priority hierarchical early warning and control based on various risk feature sets; The first priority is determined based on the meteorological risk characteristic set; if it is an emergency risk, emergency braking is implemented. Otherwise, a second priority determination is made based on the static collision risk characteristics and dynamic interference risk characteristics to determine the longest adjustable time window and execute the corresponding active deceleration or emergency braking. If there is no collision risk or the collision risk is acceptable, the equipment operation risk characteristics are comprehensively judged in the third priority manner based on the meteorological risk characteristics, and the corresponding equipment control instructions are executed. Simultaneously, in parallel with the above steps, state risk alerts are issued based on the state risk feature set.

[0030] It should be added that the article involves comparisons of various thresholds. Thresholds, preset values, preset ranges, etc., are set for result comparison and analysis to determine good or bad. The magnitude of these thresholds is determined by a combination of large-scale model analysis of sample data and human experience. They can also be appropriately adjusted based on seasonal or common-sense influences.

[0031] In summary, this system includes: a hoist status determination module, used to collect equipment operation data and cage image data, and extract equipment operation risk feature sets and status risk feature sets; an operation channel obstacle perception module, used to collect three-dimensional spatial data and distinguish between static obstacles and dynamic mechanical obstacles; a channel obstacle risk assessment module, used to extract the estimated collision time for static obstacles and the estimated intervention time for dynamic mechanical obstacles; a work environment monitoring module, used to extract the risk level of wind impact based on environmental data and cage height; and a fusion analysis and hierarchical control module, used to perform multi-priority hierarchical early warning and control based on various feature sets.

[0032] This invention also constructs a multi-dimensional risk feature set of "meteorology-equipment-space" to achieve comprehensive perception and fusion analysis of equipment operating status, personnel behavior / cargo status, physical working space of operating channels, and meteorological environment, filling the gap in physical space obstacle monitoring. Based on the multi-dimensional risk feature set, a hierarchical priority control architecture is adopted, setting meteorological emergency risks as the highest priority, physical space collision risks as the second priority to achieve hierarchical control based on time urgency, and setting equipment-meteorological coupling risks as the third priority. Hidden risks are mined through machine learning models, and parallel operation status risk alerts are provided. This enables comprehensive perception and intelligent hierarchical control of multi-dimensional risks of construction hoists, significantly improving the safety of equipment operation.

[0033] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A construction hoist intelligent safety monitoring system, characterized in that: include: The hoist status determination module is used to collect equipment operation data and image data inside the cage of the construction hoist. Based on the equipment operation data and image data, it performs equipment operation safety determination, personnel behavior and load status safety determination, and constructs equipment operation risk feature set and status risk feature set based on the determination results. The obstacle perception module for the operating channel is used to collect three-dimensional spatial data within the operating channel of the hoist cage, identify intruding obstacles within the operating channel, and distinguish obstacle types based on the motion characteristics of the intruding obstacles, including static obstacles and dynamic mechanical obstacles; The passage obstacle risk assessment module is used to receive three-dimensional spatial data and equipment operation data, perform static intrusion safety judgment for static obstacles, and dynamic interference safety judgment for dynamic mechanical obstacles, and construct static collision risk feature sets and dynamic interference risk feature sets based on the judgment results; The work environment monitoring module is used to acquire environmental data outside the hoist cage and the current height of the hoist cage, make a safety judgment on the work environment based on the environmental data and the current height of the hoist cage, and construct a meteorological risk feature set based on the judgment results; The fusion analysis and hierarchical control module is used to receive all risk feature sets and perform multi-priority hierarchical early warning and control based on various risk feature sets.

2. The intelligent safety monitoring system for construction hoist according to claim 1, characterized in that: The specific procedures for determining equipment operational safety and personnel behavior and cargo status safety based on equipment operation data and image data include: Collect equipment operation data, including cage load, tilt angle, and actual operating speed. Compare the cage load with the rated load. If it exceeds the preset threshold, it is determined to be an overload state. Compare the tilt angle with the preset safety angle. If it exceeds the preset threshold, it is determined to be an abnormal tilt state. Compare the actual operating speed with the theoretical operating speed. If the deviation exceeds the preset threshold, it is determined to be an abnormal speed state. Image recognition algorithms are used to identify whether personnel inside the cage are wearing safety helmets; if not, it is considered a helmet-missing behavior. The algorithms also identify whether personnel inside the cage are running, jumping, or engaging in any unauthorized operations; if so, it is considered a dangerous operation. The algorithms further identify the location of materials stacked inside the cage; if the offset between the center of the materials and the center of the cage exceeds a safety threshold, it is considered an off-center load.

3. The intelligent safety monitoring system for construction hoists according to claim 2, characterized in that: The specific process for extracting equipment operation risk feature sets and behavioral risk feature sets includes: When an overload condition is determined, the ratio of the current load to the rated load is extracted as the load deviation rate feature. When an abnormal tilt condition is determined, the rate of change of tilt angle is extracted as the tilt change rate feature. When an abnormal speed condition is determined, the speed deviation is extracted as the speed fluctuation rate feature. All of these are uniformly summarized into the equipment operation risk feature set. When a safety helmet is deemed missing, the proportion of personnel not wearing safety helmets within a preset time window is counted as the safety helmet missing rate feature. When a dangerous operation is deemed to be dangerous, the number of times the dangerous operation occurs within a preset time window is counted as the dangerous operation frequency feature. When an off-center load is deemed to be, the offset of the material center relative to the cage center is calculated as the off-center load feature. All of these are uniformly summarized into the state risk feature set.

4. The intelligent safety monitoring system for construction hoists according to claim 3, characterized in that: The process of determining static intrusion safety and extracting static collision risk feature sets includes: Based on three-dimensional spatial data, static obstacle point clouds that intrude into the operating space are extracted. The minimum static relative distance between the static obstacle and the outermost edge of the operating channel is calculated. If it is less than the preset static safety distance threshold, it is determined that there is a risk of static obstacle intrusion into the operating channel. The coordinates of the static obstacle point cloud corresponding to the minimum static relative distance are taken as the static collision risk location. The time required for the elevator to run to this location is calculated as the expected collision time feature and summarized into the static collision risk feature set.

5. The intelligent safety monitoring system for construction hoists according to claim 4, characterized in that: The process of determining the safety of dynamic interference and extracting the dynamic interference risk feature set includes: The real-time position and speed of the dynamic mechanical obstacle are obtained. Combined with the real-time speed of the elevator body and the preset operating time window, a motion trajectory prediction model of the dynamic mechanical obstacle and the elevator is established. The minimum dynamic approach distance between the two is calculated. If it is less than the preset dynamic safety distance threshold, it is determined that there is a risk of dynamic mechanical cross-interference in the operating channel. The predicted meeting position corresponding to the minimum dynamic approach distance is taken as the trajectory interference intersection point. The time required to reach the trajectory interference intersection point from the current time is taken as the expected interference time feature and summarized into the dynamic interference risk feature set.

6. The intelligent safety monitoring system for construction hoists according to claim 5, characterized in that: The work environment monitoring module is used to collect real-time wind speed and wind direction data as environmental data. Based on the environmental data and the current height of the hoist cage, the module makes a safety judgment on the work environment. Specifically, the real-time wind speed, wind direction angle and the current height of the hoist cage are input into the built-in wind impact risk level model. The output wind impact risk level and real-time wind speed are used as meteorological impact features and summarized into a meteorological risk feature set. The wind impact risk level includes four levels: low risk, medium risk, high risk and emergency risk.

7. The intelligent safety monitoring system for construction hoists according to claim 6, characterized in that: The specific process for multi-priority hierarchical early warning and control based on various risk characteristic sets includes: The first priority control unit is used to determine the wind impact risk level. If the wind impact risk level is an emergency risk level, it generates an emergency warning signal and executes an emergency braking command. The second priority control unit is activated when the wind impact risk level is non-emergency risk level and the static estimated collision time and / or dynamic estimated intervention time are received. It is used to calculate the longest adjustable time window based on the estimated collision time and estimated intervention time, and execute the corresponding graded control commands based on the interval in which the longest adjustable time window is located, including active deceleration and emergency braking. The third priority control unit is activated when the wind impact risk level is non-emergency risk level and the longest adjustable time window is greater than the preset warning time threshold. It is used to take the equipment operation risk feature set as input, introduce real-time wind speed as coupling factor, perform comprehensive risk judgment through a preset multi-output machine learning model, generate probability warning signals for various specific risk types, and execute corresponding differentiated equipment control instructions. The status risk alert unit is used to monitor in real time based on the status risk feature set. When each status risk feature exceeds the corresponding preset warning threshold, it generates the corresponding emergency warning signal.

8. A method for intelligent safety monitoring of construction hoists, employing an intelligent safety monitoring system for construction hoists as described in any one of claims 1-7, characterized in that, Includes the following steps: Step 1: Acquisition of multi-source data: Collect equipment operation data, image data, 3D spatial data, environmental data, and the current height of the hoist cage; Step 2: Multi-source data security assessment and feature set acquisition: Perform multi-channel security assessments based on various types of data, and construct corresponding risk feature sets based on the assessment results; Step 3: Receive all risk feature sets and conduct multi-priority hierarchical early warning and control based on various risk feature sets.

Images