Method and system for monitoring operation state of elevator in large venue and disposing emergency failure

Abstract

This invention discloses a method and system for monitoring the operational status and handling emergency faults of elevators in large venues, belonging to the field of intelligent monitoring and emergency operation and maintenance technology for building electromechanical systems. It aims to solve the problems of lagging fault identification and passive emergency response in existing large venue elevator systems. The method includes: multi-source data acquisition and interoperability with the venue's existing systems via an open protocol; diagnosing faults and outputting handling solutions using a three-layer fusion algorithm of LSTM, CNN, and random forest; hierarchical emergency response; recording fault data chains and optimizing strategies through DQN and PPO algorithms; and continuous optimization through iterative verification using digital twins and actual testing. The system includes modules for multi-source data acquisition, intelligent diagnosis, and emergency response. This invention achieves real-time elevator monitoring, accurate fault early warning, and rapid handling, adapting to dynamic operating conditions, reducing operation and maintenance costs, ensuring safety and traffic efficiency, and is applicable to various large venues.

Description

Technical Field

[0001] This invention relates to the field of building electromechanical intelligent monitoring and emergency operation and maintenance technology, and in particular to a method and system for monitoring the operation status and handling emergency faults of elevators in large venues. Background Technology

[0002] In large venues such as stadiums, convention centers, airport terminals, and large commercial complexes—public buildings with high population density—elevators (including escalators and vertical elevators) are core facilities ensuring efficient personnel passage. Their operational status directly affects public safety, venue operational efficiency, and service quality. During peak periods such as sporting events and exhibitions, passenger flow in these venues surges dramatically, placing elevators under constant high load. This imposes extremely high demands on elevator operational safety, emergency response speed, passenger flow adaptability, and remote control capabilities. Current elevator maintenance models and existing monitoring systems have significant shortcomings in large venue scenarios, failing to meet the complex maintenance needs under these conditions. At least the following technical issues exist:

[0003] First, fault identification is delayed and inaccurate, failing to achieve "early identification" and making it difficult to deal with hidden and sudden faults. Current elevator operation and maintenance largely rely on manual periodic inspections and fixed threshold alarms, which can only monitor a few basic operating parameters such as speed and load. They do not integrate safety parameters, environmental data, passenger flow information, and video visual data. The data collection dimension is single, making it impossible to capture early hidden faults caused by equipment aging, component fatigue, and drastic fluctuations in passenger flow. At the same time, there is a lack of efficient intelligent diagnostic algorithms, which cannot accurately identify visual faults such as door system jamming and passenger abnormalities. Coupled faults are prone to misjudgment and missed judgment. The identification of sudden faults is delayed for a long time, making it impossible to achieve early warning and accurate location of faults, which can easily lead to safety risks such as entrapment and congestion.

[0004] Second, the emergency response is passive and inefficient, failing to achieve rapid handling and easily escalating the impact of accidents. Current technology heavily relies on manual reporting by on-site personnel and on-site investigation and handling by maintenance staff after a malfunction occurs. It lacks tiered automatic handling capabilities and multi-system linkage mechanisms. For emergency malfunctions such as people trapped in the elevator car, electrical short circuits, and escalator reversal, it cannot achieve rapid automatic braking, personnel evacuation, and coordinated rescue. Furthermore, problems such as information asymmetry and unclear rescue routes exist during entrapment rescue, leading to excessively long rescue times and potentially causing panic among personnel.

[0005] Third, the operation and maintenance strategies are rigid and lack self-learning and dynamic adaptation capabilities, making it impossible to achieve "strong adaptability". The fault diagnosis rules and troubleshooting strategies of traditional systems are preset fixed patterns, which cannot be iteratively optimized according to dynamic operating conditions such as elevator service life, component aging, and periodic changes in passenger flow. As the equipment operating time increases, the operation and maintenance accuracy continues to decline, making it difficult to adapt to complex dynamic scenarios such as peak passenger flow in large venues and equipment aging.

[0006] Fourth, the system has poor synergy and lacks unified remote collaborative management and control capabilities, making it impossible to achieve "manageable and controllable". Existing elevator operation and maintenance systems are independent of the venue building management system (BMS), video surveillance platform, fire emergency system, etc. Data interoperability is poor, and multi-system linkage disposal and global collaborative scheduling cannot be achieved. At the same time, there is a lack of multi-terminal interaction interfaces and multi-channel alarm mechanisms. Managers and maintenance personnel cannot real-time grasp the elevator operation status, quickly receive fault alarms, and remotely control, making it difficult to meet the needs of centralized operation and maintenance of multiple elevators in large venues.

[0007] Fifth, existing systems have poor compatibility and high transformation costs. Most elevator monitoring systems cannot be compatible with the existing information systems in the venue. If data interoperability is to be achieved, large-scale transformation of existing equipment and lines is required, increasing the operation and maintenance costs and implementation difficulties of the venue.

[0008] In summary, considering the characteristics of large venues, such as dense personnel, a large number of elevators, complex working conditions, high safety requirements, and fast response speed requirements, existing elevator operation and maintenance technologies cannot solve key technical problems such as lagging fault identification, passive emergency disposal, rigid strategies, and insufficient synergy. Therefore, there is an urgent need to develop a method and system for monitoring the operation status of elevators in large venues and emergency fault disposal to solve the above technical problems. Summary of the Invention

[0009] The technical problem to be solved by this invention is to provide a method and system for monitoring the operation status of elevators in large venues and emergency fault disposal, mainly used for elevator operation and maintenance in large venues, to achieve real-time monitoring of elevator operation status, accurate identification of emergency faults, automatic disposal, and remote management and control, and to ensure personnel safety and traffic efficiency.

[0010] A method for monitoring the operation status of elevators in large venues and emergency fault disposal disclosed by this invention includes the following steps:

[0011] S1. Multi-source data collection: Obtain elevator operation data, safety data, environmental data, passenger flow data, and video data of emergencies, generate multi-source data, and use an open protocol to achieve data interoperability between multi-source data and the existing information systems in the venue.

[0012] S2. Intelligent fault diagnosis: Use a three-layer fusion algorithm of LSTM time series prediction, CNN visual feature extraction, and random forest classification to analyze the multi-source data collected in step S1, and determine the elevator fault type, fault level, and fault location based on the historical fault database and the preset fault and emergency disposal plan database, and output the corresponding emergency disposal plan.

[0013] S3. Tiered emergency response: Based on the fault level and fault type determined in step S2, the corresponding response operation is automatically executed based on the preset initial fault and solution database.

[0014] S4. Strategy self-learning optimization: Record the complete data chain of each failure in step S3 and store it in the failure case database. Based on the failure case database, use DQN deep Q network and PPO near-end strategy optimization algorithm to train the emergency response strategy model and generate an optimized strategy that adapts to dynamic working conditions.

[0015] S5. Strategy verification iteration: The optimization strategy generated in step S4 is verified for feasibility through digital twin simulation and actual testing. After the test is passed, it is updated to the fault and emergency response plan database and the fault case database.

[0016] S6. Continuous iterative optimization: Repeat steps S1-S5 to achieve continuous iterative optimization of elevator operation status monitoring and emergency response strategies.

[0017] Furthermore, in step S1, the elevator operation data is collected through an operation parameter sensor; the safety data is collected through a safety parameter sensor; the environmental data is collected through an environmental sensor; the passenger flow data is collected through a passenger flow counter at the elevator entrance; the emergency video data is collected through a camera; and the multi-source data is interconnected with the venue's existing information system through open protocols such as Modbus, OPC UA, and BACnet.

[0018] Further, step S2 includes the following steps:

[0019] S201.LSTM time series prediction predicts the parameter range for normal elevator operation based on historical operation database, and identifies abnormal parameters in elevator operation data, safety data, environmental data, and passenger flow data.

[0020] S202.CNN visual feature extraction performs image processing on video frames of emergency event video data to identify visual malfunctions such as elevator door jamming and passenger falls;

[0021] S203. Random forest classification, which integrates abnormal parameters and visual faults, determines the fault type, fault level and fault location based on a preset initial fault and emergency response plan database, and outputs the corresponding emergency response plan.

[0022] The preset initial fault and emergency response plan database includes elevator fault types, characteristic parameters and corresponding response plans; the fault levels are divided into general faults, serious faults and emergency faults.

[0023] As a preferred approach, step S3, the tiered emergency response includes,

[0024] In the event of a general malfunction, the adjustment strategy is automatically triggered to handle the elevator malfunction.

[0025] In the event of a serious malfunction, the elevator will automatically stop at the floor where passengers will be evacuated and a maintenance work order will be sent to the maintenance personnel.

[0026] In case of an emergency, the elevator will automatically brake and activate, simultaneously triggering an emergency fault warning system for the venue and sending a rescue route to maintenance personnel, along with the elevator's location, fault type, and emergency response plan.

[0027] As a preferred method, when an emergency occurs and passengers are trapped in the car, the system automatically activates a voice reassurance broadcast inside the car and establishes a real-time video call between the passenger and the monitoring room via a camera and microphone.

[0028] Furthermore, in step S4, the complete data chain includes passenger flow status at the time of the fault, equipment operating parameters, environmental conditions, video clips, handling process, rescue time, and passenger feedback.

[0029] This invention discloses a system for monitoring the operating status and handling emergency malfunctions of elevators in large venues, used to implement the method for monitoring the operating status and handling emergency malfunctions of elevators in large venues as described above, including:

[0030] The multi-source data acquisition module is configured to execute step S1;

[0031] The intelligent monitoring and diagnostic module is configured to execute step S2;

[0032] An automatic emergency fault handling module is configured to execute step S3;

[0033] The self-learning strategy optimization module is configured in execution step S4;

[0034] The optimization strategy verification iteration module is configured to run steps S5 and S6.

[0035] Furthermore, it also includes a remote collaborative management and control module, which is configured to handle elevator malfunctions, push maintenance work orders, issue early warnings for venue emergency malfunctions, and push rescue routes in the graded emergency response process in step S3.

[0036] As a preferred approach, the remote collaborative management module includes an elevator control cabinet, a cloud platform, a terminal interface, and an alarm system. The elevator control cabinet is responsible for real-time data processing and local emergency response for corresponding fault types. The cloud platform is responsible for multi-source data storage, emergency response strategy model training, global status monitoring, and iterative updates of optimization strategies, supporting centralized management of elevators in multiple venues. The terminal interface includes a PC terminal in the monitoring room, a mobile APP for maintenance personnel, and a mobile APP for venue management personnel. The alarm system includes audible and visual alarms in the monitoring room, mobile phone push notifications, and venue broadcast alerts.

[0037] As a preferred approach, the monitoring room PC is a large-screen visual interface used to display the real-time operating status of all elevators, passenger flow heat map, and fault handling progress, and supports remote control operations such as remote door opening, emergency braking, and backup elevator dispatching.

[0038] The beneficial effects of this invention are as follows: By employing multimodal data fusion acquisition and a three-layer intelligent diagnostic algorithm combining LSTM, CNN, and random forest, it can identify hidden faults caused by equipment aging and passenger flow fluctuations in advance, solving the pain points of traditional systems such as delayed fault identification, misjudgment, and missed judgment, and significantly reducing safety risks; achieving accurate and efficient fault identification, realizing early warning and zero missed judgment. A three-level automatic handling mechanism of general, severe, and emergency is established. Emergency faults can automatically brake, cut off power, and link fire protection and broadcasting; in the case of people trapped, voice reassurance, video calls, and one-click push of the optimal rescue route are realized, shortening the average time for rescue and improving the efficiency of handling complex faults, shifting from passive handling to proactive and rapid rescue, realizing proactive and intelligent emergency handling, and significantly improving rescue speed. Based on DQN, PPO reinforcement learning, and digital twin verification, fault cases are automatically accumulated and handling strategies are continuously optimized; thresholds and solutions can be dynamically adjusted according to equipment aging, peak or off-peak passenger flow, and event or exhibition scenarios, maintaining high operational accuracy over the long term, solving the problems of rigid strategies and poor adaptability in traditional systems, realizing self-learning and iteration of strategies, and adapting to dynamic operating conditions over the long term. Supporting Modbus, OPC UA, and BACnet open protocols, it seamlessly integrates with existing venue systems such as BMS, fire protection, and video surveillance without requiring large-scale modifications to wiring and equipment. With a short deployment cycle and strong applicability, it is suitable for various large venues including stadiums, airports, and commercial complexes, achieving high system compatibility, low modification costs, and rapid deployment. Automated monitoring and handling replace a large number of manual inspections, reducing maintenance labor costs and equipment downtime. Combined with passenger flow thermal optimization strategies, elevator efficiency is improved, maintenance costs are significantly reduced, achieving a balance between safety, efficiency, and low-carbon smart operation goals. The elevator control cabinet is responsible for local rapid response, while the cloud platform is responsible for global optimization and model training, achieving three-level collaborative management and control. Even in the event of a network outage, it can still complete core operations such as fault diagnosis, emergency braking, and passenger reassurance, meeting the high reliability and uninterrupted operation requirements of large venues. It also provides multi-terminal interaction through a monitoring room large screen, maintenance APP, and management terminal, supporting remote control, work order dispatch, and data backtracking. The entire fault data chain is traceable, verifiable, and analyzable, enabling centralized and unified operation and maintenance of multiple elevators in large venues, multi-terminal collaborative management and control, and global visualization and traceability. Attached Figure Description

[0039] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0040] The present invention will be further described below.

[0041] This invention provides a method for monitoring the operating status and handling emergency faults of elevators in large venues, mainly used for elevator operation and maintenance in large venues, and includes the following steps:

[0042] S1. Multi-source data acquisition: Acquire elevator operation data, safety data, environmental data, passenger flow data, and emergency video data to generate multi-source data. Utilize open protocols to achieve data interoperability between multi-source data and the venue's existing information systems.

[0043] The elevator operation data is collected through operation parameter sensors; the safety data is collected through safety parameter sensors; the environmental data is collected through environmental sensors; the passenger flow data is collected through passenger flow counters at the elevator entrance; the emergency video data is collected through cameras; and the multi-source data is interconnected with the venue's existing information system through Modbus, OPC UA, and BACnet open protocols.

[0044] S2. Intelligent fault diagnosis: LSTM time series prediction, CNN visual feature extraction and random forest classification three-layer fusion algorithm are used to analyze the multi-source data collected in step S1. Based on the historical fault database and the preset fault and emergency response plan database, the elevator fault type, fault level and fault location are determined and the corresponding emergency response plan is output.

[0045] S201.LSTM time series prediction predicts the parameter range for normal elevator operation based on historical operation database, and identifies abnormal parameters in elevator operation data, safety data, environmental data, and passenger flow data.

[0046] S202.CNN visual feature extraction performs image processing on video frames of emergency event video data to identify visual malfunctions such as elevator door jamming and passenger falls;

[0047] S203. Random forest classification, which integrates abnormal parameters and visual faults, determines the fault type, fault level and fault location based on a preset initial fault and emergency response plan database, and outputs the corresponding emergency response plan.

[0048] The preset initial fault and emergency response plan database includes elevator fault types, characteristic parameters and corresponding response plans; the fault levels are divided into general faults, serious faults and emergency faults.

[0049] S3. Tiered emergency response: Based on the fault level and fault type determined in step S2, the corresponding response operation is automatically executed based on the preset initial fault and solution database.

[0050] The tiered emergency response includes,

[0051] In the event of a general malfunction, the adjustment strategy is automatically triggered to handle the elevator malfunction.

[0052] In the event of a serious malfunction, the elevator will automatically stop at the floor where passengers will be evacuated and a maintenance work order will be sent to the maintenance personnel.

[0053] In case of an emergency, the elevator will automatically brake and activate, simultaneously triggering an emergency fault warning system for the venue and sending a rescue route to maintenance personnel, along with the elevator's location, fault type, and emergency response plan.

[0054] In case of an emergency malfunction resulting in passengers being trapped in the elevator car, the system will automatically activate the voice reassurance broadcast inside the car and establish a real-time video call between the passenger and the monitoring room via a camera and microphone.

[0055] S4. Strategy self-learning optimization: Record the complete data chain of each failure in step S3 and store it in the failure case database. Based on the failure case database, use DQN deep Q network and PPO near-end strategy optimization algorithm to train the emergency response strategy model and generate an optimized strategy that adapts to dynamic working conditions.

[0056] The complete data chain includes passenger flow status at the time of the failure, equipment operating parameters, environmental conditions, video clips, handling process, rescue time, and passenger feedback.

[0057] S5. Strategy verification iteration: The optimization strategy generated in step S4 is verified for feasibility through digital twin simulation and actual testing. After the test is passed, it is updated to the fault and emergency response plan database and the fault case database.

[0058] S6. Continuous iterative optimization: Repeat steps S1-S5 to achieve continuous iterative optimization of elevator operation status monitoring and emergency response strategies.

[0059] like Figure 1 As shown, the method for monitoring the elevator operation status and handling emergency malfunctions in this large venue includes the following steps:

[0060] S1. Multi-source data acquisition, executed by the multi-source data acquisition module, acquires multi-source data for the entire elevator scenario and achieves data interoperability with the venue's existing information system, providing comprehensive and accurate data support for subsequent fault diagnosis.

[0061] In practice, the multi-source data acquisition module acquires various types of data through different acquisition devices:

[0062] Elevator operation data: Collected through operation parameter sensors, including car speed / acceleration, traction machine speed / temperature, guide rail vibration frequency, door opening and closing time, load, floor positioning, etc., to ensure comprehensive capture of core parameters during elevator operation;

[0063] Safety data: Collected through safety parameter sensors, including safety clamp action detection signals, emergency stop button status, door lock signals, electrical circuit current / voltage, braking system pressure, etc., to monitor the elevator safety status in real time;

[0064] Environmental data: Collected through environmental sensors, including data such as temperature and humidity inside the elevator car and smoke concentration, to promptly detect abnormal situations in the elevator operating environment;

[0065] Passenger flow data: Collected through elevator entrance passenger flow counters, the number of passengers getting on and off the elevator entrance is counted in real time, generating passenger flow data to support subsequent fault handling priority determination and strategy optimization;

[0066] Emergency video data: High-definition cameras installed inside the elevator car and at the elevator door area capture real-time emergencies during elevator operation, such as elevator door jamming, passenger falls, and people trapped in the car, providing video material for visual fault identification.

[0067] After data collection, multi-source data in a unified format is generated. The multi-source data acquisition module uses three open protocols: Modbus, OPC UA, and BACnet to achieve data exchange between the multi-source data and the venue's existing information systems, including the BMS building management system, video surveillance platform, fire alarm system, and energy consumption monitoring platform. This eliminates the need for large-scale modifications to existing equipment and wiring, reducing implementation costs. The data update interval is ≤30 seconds to ensure real-time performance.

[0068] S2. Intelligent fault diagnosis, executed by the intelligent monitoring and diagnosis module, utilizes a three-layer fusion algorithm of LSTM time series prediction, CNN visual feature extraction, and random forest classification to analyze the multi-source data collected in S1. Combined with a historical fault database and a pre-set fault and emergency response plan database, it accurately determines the elevator fault type, fault level, and fault location, and outputs the corresponding emergency response plan. Specifically, it consists of the following sub-steps:

[0069] S201.LSTM Time Series Prediction: The intelligent monitoring and diagnosis module calls the elevator's historical operation data stored in the historical fault database, and establishes a prediction model for the elevator's normal operation parameters based on the LSTM time series prediction algorithm. It predicts the normal range of various elevator operation parameters, safety parameters, environmental parameters, and passenger flow parameters. The real-time multi-source data collected by S1 is compared with the predicted normal parameter range to identify abnormal parameters that exceed the normal range, thus completing the preliminary identification of parameter-related anomalies.

[0070] S202. CNN Visual Feature Extraction: The intelligent monitoring and diagnosis module extracts frames from the video data of sudden events collected by S1, and uses the CNN visual feature extraction algorithm to process the video frames, including image noise reduction, feature extraction, target recognition and other operations, to accurately identify visual faults such as elevator door jamming and passenger falls.

[0071] S203. Random Forest Classification: The intelligent monitoring and diagnosis module fuses the abnormal parameters identified in S201 with the visual faults identified in S202, and calls the preset initial fault and emergency response plan database. This database pre-integrates high-frequency fault types, corresponding feature parameters, and corresponding emergency response plans for elevators in large venues. Through the random forest classification algorithm, the fused fault features are classified and identified to accurately determine the fault type, fault level, and fault location, and outputs the emergency response plan that matches the fault.

[0072] The fault levels are clearly divided into three categories: general faults, serious faults, and emergency faults. General faults refer to faults such as slight elevator vibration, load approaching the threshold, and abnormal temperature and humidity in the car, which can be eliminated or alleviated by automatically adjusting parameters, adjusting the operating mode, and providing voice prompts to passengers. These faults pose no threat to passenger safety and can be handled without manual intervention. Serious faults refer to faults such as elevator doors failing to close, abnormal noise from the traction machine, and slight guide rail jamming, which cannot be completely eliminated by automatic adjustment, pose a risk of equipment damage and personnel being trapped, threaten passenger safety, and require maintenance personnel to handle. Emergency faults refer to faults such as people trapped in the car, electrical short circuits, car overspeeding, and escalator reversal, which seriously threaten passenger life safety and require immediate braking, emergency rescue, and multi-system coordinated handling.

[0073] Specifically, the aforementioned pre-set initial fault and emergency response plan database contains at least 168 high-frequency fault characteristics of elevators in large venues, covering various faults such as elevator door systems, traction systems, electrical systems, and safety systems, ensuring the comprehensiveness and accuracy of fault diagnosis.

[0074] S3. Tiered emergency response, executed by the automatic emergency fault handling module in conjunction with the remote collaborative management module, automatically executes corresponding handling operations based on the fault level and type determined in S2 and a preset initial fault and solution database, achieving rapid fault response and handling. The specific tiered handling method is as follows:

[0075] General Faults: When S2 determines a general fault, such as slight elevator vibration, load approaching the threshold, or abnormal temperature and humidity inside the car, the emergency fault automatic handling module automatically triggers adjustment strategies without manual intervention. These operations include reducing elevator speed, adjusting ventilation or temperature and humidity inside the car, and issuing voice reminders inside the car to complete the fault handling.

[0076] Serious Fault: When S2 determines a serious fault, such as the elevator door failing to close, abnormal noise from the traction machine, or slight jamming of the guide rail, the emergency fault automatic handling module automatically controls the elevator to stop at the nearest floor, opens the elevator door to evacuate passengers, and at the same time pushes a priority maintenance work order to the maintenance personnel's mobile APP through the remote collaborative management module. The work order includes the location of the faulty elevator, the fault type, and basic handling instructions, so that maintenance personnel can quickly arrive at the scene to handle the problem.

[0077] Emergency Faults: When S2 determines an emergency fault, such as passenger entrapment, electrical short circuit, car overspeed, or escalator reversal, the emergency fault automatic handling module immediately controls the elevator to perform operations such as brake braking, cutting off non-emergency power, and activating emergency lighting and ventilation. The escalator will immediately stop running and activate the anti-reversal device. At the same time, the remote collaborative control module synchronizes with the venue's emergency fault early warning system, issuing alarms through three methods: audible and visual alarms in the monitoring room, SMS or mobile APP notifications to management and maintenance personnel, and venue broadcast prompts, reminding relevant personnel to respond immediately. In addition, the remote collaborative control module automatically calculates the optimal rescue route and pushes it to the maintenance personnel's mobile APP, along with the elevator location, fault type, and emergency handling plan.

[0078] When an emergency malfunction results in passengers being trapped in the elevator car, the emergency malfunction automatic handling module will automatically activate the voice reassurance broadcast inside the car, such as "Please do not panic, the system has been alarmed, and rescue personnel will arrive within 3 minutes. Please wait patiently." At the same time, a real-time video call will be established between the passenger and the monitoring room through the camera and microphone inside the car, so that the monitoring room personnel can understand the situation inside the car and calm the passengers.

[0079] S4. Strategy self-learning optimization, executed by the self-learning strategy optimization module, generates optimized handling strategies adapted to the dynamic working conditions of elevators through the accumulation of fault cases and reinforcement learning.

[0080] In practice, the self-learning strategy optimization module synchronously records the complete data chain of each fault in S3 and stores it in the fault case database. The complete data chain specifically includes the passenger flow status (peak / off-peak) when the fault occurs, equipment operating parameters, environmental conditions, video clips, handling process, rescue time and passenger feedback, so as to realize the full traceability of the fault.

[0081] Based on a large amount of fault data accumulated in the fault case database, the self-learning strategy optimization module adopts a fusion algorithm of DQN deep Q network and PPO near-end strategy optimization to train the emergency response strategy model, and explore the mapping relationship of "fault type-passenger flow status-equipment status-optimal handling solution". For dynamic operating conditions such as venue passenger flow fluctuations and equipment aging, it generates appropriate optimized handling strategies, such as prioritizing rescue of trapped people during peak exhibition periods, adjusting fault judgment thresholds after equipment aging, and optimizing elevator door opening and closing delays during periods of high passenger flow.

[0082] S5. Strategy Verification Iteration, executed by the optimization strategy verification iteration module, verifies the feasibility and security of the optimization strategy generated in S4 to ensure that the optimization strategy can adapt to actual operating conditions. The specific verification process is as follows:

[0083] First, the optimization strategy verification and iteration module utilizes digital twin technology to construct a simulation environment consistent with the actual operation of elevators in Chengdu Phoenix Mountain Sports Park. This environment recreates scenarios with varying passenger flow densities, elevator operating loads, and equipment aging levels. The optimization strategy generated by S4 is then imported into the simulation environment to verify its feasibility and safety. After successful simulation verification, 3-5 elevators within the venue are selected for small-scale practical application. Feedback data, such as rescue time, equipment wear and tear, and passenger satisfaction, are collected during the execution of the optimization strategy. If the feedback data meets the standards, such as reduced rescue time, decreased equipment wear and tear, and improved passenger satisfaction, the optimization strategy is updated in the fault and emergency response plan database and the fault case database, completing the iterative update of the strategy.

[0084] S6. Continuous Iterative Optimization: The optimization strategy verification and iteration module controls the entire system to repeatedly execute steps S1-S5, achieving continuous monitoring of the elevator's operating status. Simultaneously, by continuously accumulating fault cases, training optimization strategies, and verifying updates, the emergency response strategy undergoes continuous iterative optimization, ensuring the system can adapt to the dynamic operating conditions of elevators in large venues for extended periods, maintaining high operational accuracy and response efficiency. Specifically, in this embodiment, a complete iteration cycle is completed every 3 months to ensure the timeliness of strategy optimization.

[0085] The method for monitoring the operational status and handling emergency faults of elevators in large venues employs multimodal data fusion acquisition and a three-layer intelligent diagnostic algorithm combining LSTM, CNN, and random forest. This approach can identify hidden faults caused by equipment aging and passenger flow fluctuations in advance, addressing the pain points of traditional systems such as delayed fault identification, misjudgments, and missed diagnoses, significantly reducing safety risks. It achieves accurate and efficient fault identification, enabling early warning and zero missed diagnoses. A three-level automatic handling mechanism (general, severe, and emergency) is established. Emergency faults can automatically brake, cut off power, and link fire and broadcast systems. In cases of passenger entrapment, voice reassurance, video calls, and one-click push of the optimal rescue route are implemented, shortening the average time for passenger entrapment rescue and improving the efficiency of handling complex faults. This transforms passive handling into proactive and rapid rescue, achieving proactive and intelligent emergency response and significantly improving rescue speed. Based on DQN, PPO reinforcement learning, and digital twin verification, fault cases are automatically accumulated and handling strategies are continuously optimized. Thresholds and solutions can be dynamically adjusted according to equipment aging, peak or off-peak passenger flow, and event or exhibition scenarios, maintaining high operational accuracy over the long term. This solves the problems of rigid strategies and poor adaptability in traditional systems, enabling self-learning and iteration of strategies to adapt to dynamic operating conditions over the long term. Supporting Modbus, OPC UA, and BACnet open protocols, it seamlessly integrates with existing venue systems such as BMS, fire protection, and video surveillance without requiring large-scale modifications to wiring and equipment. With a short deployment cycle and strong applicability, it is suitable for various large venues including stadiums, airports, and commercial complexes, achieving high system compatibility, low modification costs, and rapid deployment. Automated monitoring and handling replace a large number of manual inspections, reducing maintenance labor costs and equipment downtime. Combined with passenger flow thermal optimization strategies, elevator efficiency is improved, maintenance costs are significantly reduced, achieving a balance between safety, efficiency, and low-carbon smart operation goals.

[0086] This invention also provides a system for monitoring the operating status and handling emergency faults of elevators in large venues, used to implement a method for monitoring the operating status and handling emergency faults of elevators in large venues as described above, including: a multi-source data acquisition module configured in execution step S1; an intelligent monitoring and diagnosis module configured in execution step S2; an automatic emergency fault handling module configured in execution step S3; a self-learning strategy optimization module configured in execution step S4; and an optimization strategy verification and iteration module configured in execution steps S5 and S6.

[0087] The aforementioned multi-source data acquisition module is configured in execution step S1, specifically including operating parameter sensors, safety parameter sensors, environmental sensors, elevator entrance passenger flow counters, high-definition cameras, and a data communication unit. Among these, the load sensor accuracy is ±5kg, the infrared entrapment detection (integrated into the environmental sensor) recognition accuracy is ≥99%, and the data acquisition update interval is ≤30 seconds. The data communication unit incorporates three open protocols: Modbus, OPC UA, and BACnet, enabling seamless data exchange with the venue's existing information systems. The intelligent monitoring and diagnosis module is configured in execution step S2, incorporating LSTM time-series prediction algorithms, CNN visual feature extraction algorithms, random forest classification algorithms, as well as a historical fault database and a pre-set fault and emergency response plan database. It can achieve real-time analysis of multi-source data, anomaly identification, fault judgment, and output of response plans, with a fault diagnosis accuracy rate ≥98%. The emergency fault automatic handling module is configured in execution step S3, linked with the elevator control cabinet, the venue's fire protection system, and the escalator control system. It can automatically execute corresponding handling operations based on the fault level, including elevator speed adjustment and stop at different floors. The system controls layer control, brake braking, and emergency lighting activation, while simultaneously linking with a self-learning strategy optimization module to record the complete fault data chain. The self-learning strategy optimization module, configured in execution step S4, incorporates a fault case database, a DQN deep Q-network algorithm, and a PPO near-end strategy optimization algorithm. It enables the storage of the complete fault data chain, training of emergency response strategy models, and generation of optimization strategies adapted to dynamic operating conditions. The optimization strategy verification and iteration module, configured in execution steps S5 and S6, incorporates a digital twin simulation unit and an actual test data acquisition unit. It can perform simulation verification, actual testing, and database updates of the optimization strategy, and control the system to cyclically execute steps S1-S5 to achieve continuous iterative optimization.

[0088] Furthermore, it also includes a remote collaborative management and control module, which is configured in step S3 of the graded emergency response process for elevator fault handling, maintenance work order push, venue emergency fault early warning, and rescue route push. The remote collaborative management and control module includes an elevator control cabinet, a cloud platform, a terminal interface, and an alarm system. The elevator control cabinet is responsible for real-time data processing and local emergency response for corresponding fault types. The cloud platform is responsible for multi-source data storage, emergency response strategy model training, global status monitoring, and iterative updates of optimization strategies, supporting centralized management of elevators in multiple venues. The terminal interface includes a monitoring room PC, a maintenance personnel mobile APP, and a venue management personnel mobile APP. The alarm system includes a monitoring room audible and visual alarm, mobile phone push notifications, and venue broadcast alerts. The monitoring room PC is a large-screen visual interface used to display the real-time operating status of all elevators, passenger flow heatmaps, and fault handling progress, supporting remote door opening, emergency braking, and remote control operations for backup elevator dispatch.

[0089] The remote collaborative management module, serving as a supplementary module to the system for monitoring the operational status and handling emergency faults of elevators in large venues, adopts a cloud-edge collaborative architecture. It is configured in step S3 of the hierarchical emergency response process, including elevator fault handling, maintenance work order push notifications, venue emergency fault early warnings, and rescue route push notifications. This achieves three-level collaboration between the cloud, edge, and on-site, with the edge ensuring real-time response, the cloud supporting global optimization, and core functions remaining uninterrupted even in network outages, thus meeting the high reliability requirements of large venues. Specifically, it includes an elevator control cabinet, a cloud platform, a terminal interface, and an alarm system. The specific implementation details of each part are as follows:

[0090] The elevator control cabinet, acting as an edge computing unit, is responsible for receiving real-time data from the multi-source data acquisition module. It works in conjunction with the emergency fault automatic handling module for local emergency response, while simultaneously processing real-time data to ensure the normal operation of core emergency response functions during network outages. The cloud platform is responsible for multi-source data storage, emergency response strategy model training, global status monitoring, and optimization strategy iteration updates. It supports centralized management of elevators in multiple venues and can simultaneously manage elevator operation and maintenance data and strategies in multiple large venues. The terminal interaction interfaces include a monitoring room PC, a mobile app for maintenance personnel, and a mobile app for venue management personnel. The monitoring room PC features a large-screen visualization. The interface displays the real-time operating status of all elevators, passenger flow heatmaps, and fault handling progress, and supports remote control operations such as remote door opening, emergency braking, and backup elevator dispatching. Maintenance personnel's mobile app receives fault work orders, views rescue routes and handling guidelines, and provides feedback on processing results. Venue management personnel's mobile app receives important fault alarms, views elevator operation and maintenance reports, and approves optimization strategies. The alarm system includes audible and visual alarms in the monitoring room, mobile push notifications (SMS + app notifications), and venue broadcast alerts. In the event of an emergency, all three alarm methods are triggered simultaneously to ensure that monitoring room personnel, maintenance personnel, and venue management personnel respond immediately.

[0091] The elevator control cabinet is responsible for local rapid response, while the cloud platform is responsible for global optimization and model training, achieving three-level collaborative management and control. Even in the event of a network outage, it can still complete core operations such as fault diagnosis, emergency braking, and passenger reassurance, meeting the high reliability and uninterrupted operation requirements of large venues. It also provides multi-terminal interaction through a monitoring room large screen, maintenance APP, and management terminal, supporting remote control, work order dispatch, and data backtracking. The entire fault data chain is traceable, verifiable, and analyzable, enabling centralized and unified operation and maintenance of multiple elevators in large venues, multi-terminal collaborative management and control, and global visualization and traceability.

[0092] Example

[0093] When the method and system for monitoring the operation status and handling emergency faults of elevators in a large venue were applied to a large sports park, the venue had a total building area of ​​350,000 square meters, including 86 vertical elevators and 42 escalators. It hosted more than 100 exhibitions annually, and the peak hourly passenger flow exceeded 50,000 people. The requirements for elevator operation safety, traffic efficiency and emergency response capabilities were extremely high.

[0094] The implementation steps are as follows:

[0095] Data acquisition deployment; install operating parameter sensors (speed, load, vibration, etc.), safety sensors (door locks, braking systems, etc.), and environmental and passenger flow sensors (temperature, humidity, smoke, infrared entrapment detection, passenger flow counters) in all elevators; install high-definition cameras and two-way voice equipment in the elevator cars; and connect to the venue's existing BMS system and video monitoring platform via the OPC UA protocol to achieve data interoperability.

[0096] A database of pre-defined fault and emergency response plans was constructed. Elevator fault data of the venue over the past three years was compiled, such as door jamming, traction machine overheating, and people being trapped during peak exhibition periods, to form a historical fault database. Combined with the high-frequency fault types of elevators in large venues in the industry, a database of pre-defined fault and emergency response plans containing at least 168 fault characteristics and basic handling strategies was established.

[0097] Model training and system debugging were conducted based on one year of historical operational data from the venue, including extreme passenger flow scenarios such as peak exhibition periods and event dispersal periods. LSTM, CNN, and random forest fusion diagnostic models were trained, and reinforcement learning algorithm parameters were debugged. 25 types of emergency faults were simulated, such as people trapped in the elevator car, electrical short circuits, and escalator reversal, to verify the system's diagnostic accuracy and emergency response speed.

[0098] After the system goes live, it automatically records actual fault data chains and generates optimization strategies through reinforcement learning, such as prioritizing rescue of people trapped during peak hours and adjusting door opening and closing delays from 3 seconds to 2 seconds. After verification by digital twins, these strategies are applied to all elevators in the building. The strategy is updated every 3 months.

[0099] After 12 months of operation, actual verification has shown significant technical results: the elevator fault diagnosis accuracy rate reached 98.7%, the emergency fault response delay was ≤1.8 seconds, the average rescue time for trapped people was reduced from 15 minutes to 4 minutes, the elevator maintenance labor cost was reduced by 82%, no safety accidents occurred due to untimely fault handling, elevator traffic efficiency was increased by 30%, the efficiency of handling complex faults was increased by 70%, equipment downtime was reduced by 65%, and maintenance costs were significantly reduced; at the same time, it adapts to dynamic operating conditions such as venue passenger flow fluctuations and equipment aging, ensuring the long-term safe and efficient operation of elevators, meeting the safety operation and service quality requirements of densely populated places, and conforming to the trend of green, low-carbon, and intelligent operation and maintenance of smart venues; it fully meets the core requirements of large venue elevators: "safety first, rapid response, and passenger flow adaptation".

Claims

1. A method for monitoring the running state of an elevator in a large venue and disposing emergency faults, characterized in that, Including steps, S1. Multi-source data acquisition: Acquire elevator operation data, safety data, environmental data, passenger flow data, and emergency video data to generate multi-source data. Utilize open protocols to achieve data interoperability between multi-source data and the venue's existing information systems. S2. Intelligent fault diagnosis: LSTM time series prediction, CNN visual feature extraction and random forest classification three-layer fusion algorithm are used to analyze the collected multi-source data. Based on the historical fault database and the preset fault and emergency response plan database, the elevator fault type, fault level and fault location are determined and the corresponding emergency response plan is output. S3. Tiered emergency response: Based on the fault level and fault type determined in step S2, the corresponding response operation is automatically executed based on the preset initial fault and solution database. S4. Strategy self-learning optimization: Record the complete data chain of each failure in step S3 and store it in the failure case database. Based on the failure case database, use DQN deep Q network and PPO near-end strategy optimization algorithm to train the emergency response strategy model and generate an optimized strategy that adapts to dynamic working conditions. S5. Strategy verification iteration: The optimization strategy generated in step S4 is verified for feasibility through digital twin simulation and actual testing. After the test is passed, it is updated to the fault and emergency response plan database and the fault case database. S6. Continuous iterative optimization: Repeat steps S1-S5 to achieve continuous iterative optimization of elevator operation status monitoring and emergency response strategies.

2. A method for monitoring the operation state of an elevator in a large venue and disposing emergency faults, according to claim 1, characterized in that, In step S1, the elevator operation data is collected by an operation parameter sensor; the safety data is collected by a safety parameter sensor; the environmental data is collected by an environmental sensor; the passenger flow data is collected by a passenger flow counter at the elevator entrance; the emergency video data is collected by a camera; and the multi-source data is interconnected with the venue's existing information system through open protocols such as Modbus, OPC UA, and BACnet.

3. The method for monitoring the operating status and handling emergency faults of elevators in large venues as described in claim 1, characterized in that, Step S2 includes the following steps. S201.LSTM time series prediction predicts the parameter range for normal elevator operation based on historical operation database, and identifies abnormal parameters in elevator operation data, safety data, environmental data, and passenger flow data. S202.CNN visual feature extraction performs image processing on video frames of emergency event video data to identify visual malfunctions such as elevator door jamming and passenger falls; S203. Random forest classification, which integrates abnormal parameters and visual faults, determines the fault type, fault level and fault location based on a preset initial fault and emergency response plan database, and outputs the corresponding emergency response plan. The preset initial fault and emergency response plan database includes elevator fault types, characteristic parameters and corresponding response plans; the fault levels are divided into general faults, serious faults and emergency faults.

4. The method for monitoring the operating status and handling emergency faults of elevators in large venues as described in claim 3, characterized in that, In step S3, the tiered emergency response includes, In the event of a general malfunction, the adjustment strategy is automatically triggered to handle the elevator malfunction. In the event of a serious malfunction, the elevator will automatically stop at the floor where passengers will be evacuated and a maintenance work order will be sent to maintenance personnel. In case of an emergency, the elevator will automatically stop and activate, simultaneously triggering an emergency fault warning system for the venue and sending a rescue route to maintenance personnel, along with the elevator's location, fault type, and emergency response plan.

5. The method for monitoring the operating status and handling emergency faults of elevators in large venues as described in claim 4, characterized in that, In case of an emergency malfunction resulting in passengers being trapped in the elevator car, the system will automatically activate the voice reassurance broadcast inside the car and establish a real-time video call between the passenger and the monitoring room via a camera and microphone.

6. The method for monitoring the operating status and handling emergency faults of elevators in large venues as described in claim 1, characterized in that, In step S4, the complete data chain includes passenger flow status at the time of the fault, equipment operating parameters, environmental conditions, video clips, handling process, rescue time, and passenger feedback.

7. A system for monitoring the operating status and handling emergency faults of elevators in large venues, characterized in that, A method for monitoring the operating status and handling emergency malfunctions of elevators in large venues as described in any one of claims 1-6, comprising: The multi-source data acquisition module is configured to execute step S1; The intelligent monitoring and diagnostic module is configured to execute step S2; An automatic emergency fault handling module is configured to execute step S3; The self-learning strategy optimization module is configured in execution step S4; The optimization strategy verification iteration module is configured to run steps S5 and S6.

8. The system for monitoring the operating status and handling emergency faults of elevators in large venues as described in claim 7, characterized in that, It also includes a remote collaborative management and control module, which is configured in the graded emergency response process of step S3 to handle elevator malfunctions, push maintenance work orders, issue early warnings for venue emergency malfunctions, and push rescue routes.

9. The system for monitoring the operating status and handling emergency faults of elevators in large venues as described in claim 8, characterized in that, The remote collaborative management and control module includes an elevator control cabinet, a cloud platform, a terminal interface and an alarm system. The elevator control cabinet is responsible for real-time data processing and local emergency response for corresponding fault types. The cloud platform is responsible for multi-source data storage, emergency response strategy model training, global status monitoring, optimization strategy iteration and updates, and supports centralized management of elevators in multiple venues; the terminal interaction interface includes a monitoring room PC, a maintenance personnel mobile APP, and a venue management personnel mobile APP; the alarm system includes monitoring room audible and visual alarms, mobile phone push notifications, and venue broadcast prompts.

10. The system for monitoring the operating status and handling emergency faults of elevators in large venues as described in claim 9, characterized in that, The monitoring room PC is a large-screen visual interface used to display the real-time operating status of all elevators, passenger flow heat map, and fault handling progress. It supports remote control operations such as remote door opening, emergency braking, and backup elevator dispatching.

Images