Elevator monitoring system based on Beidou satellite positioning

Abstract

The invention discloses an elevator monitoring system based on Beidou satellite positioning, and the system comprises an elevator person trapping event recognition module which is used for judging whether a person trapping event occurs in an elevator or not; the elevator operation characteristic data acquisition module is used for acquiring elevator operation characteristic data; the elevator fault recognition module is used for constructing an elevator fault recognition model and recognizing the elevator fault type; the active alarm analysis module is used for judging whether emergency alarm is needed or not according to the elevator operation characteristic data; the automatic alarm module is used for sending rescue data of people trapped in the elevator through short message services of the rescue unit in the elevator car and the Beidou satellite navigation system; the elevator posture monitoring module is used for continuously monitoring elevator posture data and recognizing the abnormal state of the elevator posture; and the elevator fault maintenance module is used for generating a maintenance work order of the equipment and synchronously updating the maintenance execution state of the fault elevator. According to the method, the comprehensive management level of all parties on elevator equipment is improved, the elevator operation management cost is reduced, and the elevator operation economic benefits are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of elevator monitoring and control, and in particular to an elevator monitoring system based on Beidou satellite positioning. Background Art

[0002] With the continuous increase of high-rise buildings, elevators, as an important vertical transportation tool in modern buildings, have become an indispensable part of daily life. However, elevator failures and entrapment incidents still occur from time to time, posing serious hidden dangers to the life safety and property of passengers. At present, the elevator monitoring system faces many challenges in ensuring the safety of elevators, especially when the elevator has abnormal posture, failure or entrapment incidents, the existing technology often fails to detect and make effective emergency responses in time, resulting in the difficulty of timely warning of elevator failures, affecting the reliability and safety of elevators. In addition, when an elevator entraps people, the traditional emergency call button is usually the only means of asking for help, and this single alarm method is less efficient in complex environments, especially when the communication signal in the elevator car is interrupted, it is impossible to effectively transmit the trapped information, resulting in rescue delays. Especially in complex fault conditions, the elevator's operating status, posture and dynamic changes of passengers are difficult to capture, which brings difficulties to rescue work. In addition, the maintenance of elevators still relies on traditional manual inspection and regular testing modes, which makes it difficult to achieve real-time fault warning and intelligent fault diagnosis. Although some elevator monitoring systems have integrated basic sensor data, the existing systems lack sufficient monitoring accuracy and response capabilities in the face of complex elevator operation conditions, such as sudden tilt, vibration or abnormal load. This maintenance mode that relies on manual inspection often lacks sufficient data analysis and fault prevention mechanisms, resulting in the inability to predict elevator failures in a timely manner, affecting the normal operation of the elevator. There is a lack of effective technical means to solve problems such as inadequate maintenance and difficulty in tracing elevator safety responsibilities, and it is difficult to effectively monitor, warn and prevent them, making it difficult to fundamentally improve the safety performance of elevators. With the continuous expansion of the application scope of Internet of Things technology in elevator safety monitoring and management, its refined management functions such as foreseeable fault intelligent alarm, automatic disposal of trapped people and paperless maintenance have gradually become key technologies for improving elevator safety supervision and promoting the improvement of elevator maintenance levels.

[0003] How to integrate more sensor data and optimize the elevator fault diagnosis algorithm is still a major challenge in elevator safety management. Especially when someone is trapped in an elevator, how to transmit fault information in a timely and accurate manner through other means when the signal is interrupted and ensure the efficiency of rescue is still a problem that needs to be solved urgently. Therefore, how to solve the early warning and diagnosis problems of elevator faults, especially to deal with the interruption of the signal of the rescue unit in the elevator, improve the intelligent monitoring level of the elevator, shorten the fault response time and improve the rescue efficiency, has become a key technical problem that the elevator industry urgently needs to solve. Summary of the invention

[0004] The present invention solves the problems existing in the above-mentioned prior art and provides an elevator monitoring system based on Beidou satellite positioning, which mainly includes:

[0005] The elevator entrapment event recognition module is used to obtain real-time car monitoring images, identify the number of passengers in the elevator car, and determine whether an elevator entrapment event has occurred based on the door opening time after the car takes off and lands vertically;

[0006] The elevator operation characteristic data acquisition module is used to obtain the elevator operation status, elevator entrapment status, elevator parking position and elevator posture data through the sensors in the elevator to form the elevator operation characteristic data;

[0007] The elevator fault identification module is used to build an elevator fault identification model based on the elevator operation characteristic data, identify the elevator fault type, match the predetermined maintenance plan as a reference maintenance plan, and build elevator trapped person rescue data;

[0008] Active alarm analysis module, used to determine whether emergency alarm is needed based on the facial expressions and movements of elevator passengers combined with elevator operation characteristic data;

[0009] Automatic alarm module, used to send rescue data of trapped people in the elevator through the rescue unit in the car and the short message service of Beidou satellite navigation system;

[0010] The elevator posture monitoring module is used to continuously monitor the elevator posture data, identify abnormal elevator posture status, and issue real-time alarm reminders;

[0011] The elevator fault maintenance module is used to generate equipment maintenance work orders based on elevator trapped person rescue data, and synchronously update the maintenance execution status and detection information of the faulty elevator.

[0012] Furthermore, the elevator entrapment event recognition module is used to obtain real-time car monitoring images, identify the number of passengers in the elevator car, and determine whether an elevator entrapment event occurs based on the door opening time after the car vertically takes off and lands, including:

[0013] Through the camera in the elevator car, real-time car monitoring images are obtained, and the real-time car monitoring images are transmitted to the car monitoring image database; through the car monitoring image database, historical car monitoring images are obtained, and the number of passengers is marked. The convolutional neural network is used for model training, and a passenger number recognition model is constructed to identify the number of passengers in the elevator car; if there are passengers in the elevator car, the timer starts timing when the car takes off and lands vertically, and the door opening and closing status of the elevator door is obtained in real time using the door opening and closing sensor. If the elevator door is not opened within the preset time, it is determined that a person is trapped in the elevator.

[0014] Furthermore, the elevator operation characteristic data acquisition module is used to acquire the elevator operation status, elevator entrapment status, elevator parking position and elevator posture data through sensors in the elevator to form elevator operation characteristic data, including:

[0015] The current data of the elevator circuit is obtained through the power sensor, the elevator operation status corresponding to the current signal is marked, the decision tree algorithm is used for model training, and an elevator operation status recognition model is constructed to identify the current operation status of the elevator; the vertical take-off and landing speed of the elevator is obtained by the Hall sensor to determine the state of the elevator being trapped; the leveling photoelectric sensor is used to determine whether the elevator that stops trapping people is parked at the same level, and the Beidou positioning system is used to obtain the elevator's parking position; the elevator attitude data is obtained through the attitude sensor, and the elevator attitude data includes X-axis acceleration, Y-axis acceleration, Z-axis acceleration, X-axis angle, Y-axis angle and Z-axis angle; the elevator operation status, elevator trapped status, elevator parking position and elevator attitude data are standardized and combined into an elevator operation feature vector to constitute elevator operation feature data, which are transmitted to the elevator operation monitoring database.

[0016] Furthermore, the elevator fault identification module is used to construct an elevator fault identification model according to the elevator operation characteristic data, identify the elevator fault type, match the predetermined maintenance plan as a reference maintenance plan, and construct elevator trapped person rescue data, including:

[0017] Through the elevator operation monitoring database, historical elevator operation characteristic data is obtained, and the elevator fault types are marked. A recurrent neural network is used for model training to build an elevator fault recognition model. According to the elevator operation characteristic data obtained in real time, the elevator fault recognition model is used to identify the elevator fault type. Based on the elevator fault type, the scheduled maintenance plan is matched as the reference maintenance plan. The elevator operation characteristic data, the elevator fault type recognition results and the reference maintenance plan are integrated into elevator trapped people rescue data.

[0018] Furthermore, the active alarm analysis module is used to determine whether an emergency alarm is needed based on the facial expressions and movements of the elevator passengers in combination with the elevator operation characteristic data, including:

[0019] If the elevator passenger uses the one-touch alarm button to actively alarm, the elevator passenger's face recognition is performed through the elevator car camera and Faceboxs model, and the face recognition result is matched with the car face monitoring database to determine whether the elevator passenger is a person who has made a historical malicious rescue call; if the elevator passenger is not a person who has made a historical malicious rescue call, an emergency alarm is made through the elevator car rescue unit; if the elevator passenger is a person who has made a historical malicious rescue call, the historical elevator operation feature data and the historical car monitoring images are obtained, and the model is trained using a convolutional neural network based on the car monitoring images to identify the facial expressions and movements of the historical malicious rescue callers; based on the facial expressions and movements of the historical malicious rescue callers and the elevator operation feature data, it is determined whether an emergency alarm is required through the elevator car rescue unit; if the judgment result is that an emergency alarm is not required, the elevator passenger's face image is sent to the car face monitoring database, and the number of historical malicious calls by the elevator passenger is recorded through the car face monitoring database.

[0020] Furthermore, the automatic alarm module is used to send elevator trapped person rescue data through the rescue unit in the car and the short message service of the Beidou satellite navigation system, including:

[0021] The elevator trapped person rescue data is sent to the rescue end through the rescue unit in the elevator car; if the signal of the rescue unit in the elevator car is interrupted, the elevator trapped person rescue data is sent to the rescue end using the short message service of the Beidou satellite navigation system.

[0022] Furthermore, the elevator posture monitoring module is used to continuously monitor the elevator posture data, identify abnormal elevator posture states, and provide real-time alarm reminders, including:

[0023] According to the attitude sensor installed in the elevator, the real-time elevator attitude data of the elevator during operation is obtained; the normal attitude range is set by using the threshold detection algorithm. If the real-time elevator attitude data exceeds the preset normal attitude range, the current elevator attitude state is judged to be abnormal; according to the abnormal elevator attitude state detection result, the alarm information and the corresponding elevator attitude data are sent to the elevator fault maintenance module through the automatic alarm module.

[0024] Furthermore, the elevator fault maintenance module is used to generate a maintenance work order for the equipment according to the elevator trapped person rescue data, and synchronously update the maintenance execution status and detection information of the faulty elevator, including:

[0025] Based on the elevator trapped person rescue data and elevator posture abnormality data, generate the equipment maintenance work order, including equipment model, fault type, maintenance steps and elevator stop position; use the dispatching system at the rescue end to transmit the maintenance work order and maintenance report to the maintenance unit, assign maintenance personnel and maintenance tools to the task, and record the dispatching time and maintenance personnel information; based on the progress and result information fed back by the maintenance personnel during the execution process, record the maintenance execution status information in real time and record it in the paperless maintenance database, and simultaneously update the maintenance execution status and detection information of the faulty elevator. The maintenance execution status information includes the fault repair completion mark, maintenance record and the confirmation result of the executor; formulate a paperless maintenance plan for the elevator based on the historical elevator operation feature vector and historical maintenance records, including regular maintenance and emergency maintenance.

[0026] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0027] The present invention provides an elevator monitoring system based on Beidou satellite positioning. The present invention can accurately identify the number of passengers in the elevator and the operating status of the elevator by acquiring car monitoring images and sensor data in real time, and timely determine whether a trapped person event occurs, effectively solving the problem that the existing system cannot accurately capture the abnormal state of the elevator. The present invention combines elevator operation characteristic data and intelligent fault identification models, can automatically identify the type of elevator fault, and recommend corresponding maintenance measures according to a predetermined maintenance plan, thereby improving the accuracy of fault prediction and maintenance decision-making. In terms of emergency response, the present invention can timely detect abnormalities and trigger alarms by real-time monitoring of the elevator posture and passenger dynamics, especially when a trapped person event occurs in the elevator, even if there is a communication signal interruption, with the help of the short message service of Beidou satellite navigation, the accurate transmission of rescue data is ensured, and the timeliness and effectiveness of the rescue are improved. The present invention integrates a paperless intelligent monitoring system for elevator maintenance, and scientifically generates additional maintenance work order items according to the elevator safety status and operating conditions to improve the quality of maintenance and improve the safety of the elevator, and simultaneously updates the maintenance progress, thereby improving the maintenance efficiency and reducing the probability of repeated failures. The present invention also establishes a complete elevator monitoring and maintenance management database, realizes the digitization of various management and services, improves the information level of elevator safety management, and greatly improves elevator safety management services and emergency rescue capabilities. The present invention integrates the relevant resources of elevator users, maintenance units, and rescue units, strengthens corporate self-regulation and self-discipline, timely discovers elevator equipment safety hazards, reduces elevator equipment accidents, uses the Internet of Things, big data, and mobile Internet technologies to understand the elevator operation status and the situation of passengers in the elevator in real time, reduces the number of false alarms and spoof rescues, and realizes the automatic discovery of elevator abnormalities by the elevator intelligent monitoring system. The maintenance management system generates project work orders in a targeted manner, organizes targeted maintenance and rescue in a timely manner, and at the same time improves the quality of elevator maintenance and rescue, improves the comprehensive management level of elevator equipment by all parties, reduces the cost of elevator operation and management, and improves the economic benefits of elevator operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a flow chart of an elevator monitoring system based on Beidou satellite positioning of the present invention;

[0029] Figure 2 A schematic diagram of an elevator monitoring system based on Beidou satellite positioning according to the present invention;

[0030] Figure 3 This is another schematic diagram of an elevator monitoring system based on Beidou satellite positioning according to the present invention. DETAILED DESCRIPTION

[0031] In order to make the objectives, technical strategies and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] like Figure 1-3 In this embodiment, an elevator monitoring system based on Beidou satellite positioning may specifically include:

[0033] Step S101, an elevator entrapment event recognition module is used to obtain real-time car monitoring images, identify the number of passengers in the elevator car, and determine whether an elevator entrapment event occurs based on the door opening time after the car vertically takes off and lands.

[0034] Through the camera in the elevator car, the real-time car monitoring image is obtained and transmitted to the car monitoring image database. Through the car monitoring image database, the historical car monitoring image is obtained and the number of passengers is marked. The convolutional neural network is used for model training to build a passenger number recognition model to identify the number of passengers in the elevator car. If there are passengers in the elevator car, the timer starts when the car takes off and lands vertically, and the door opening and closing status of the elevator door is obtained in real time using the door opening and closing sensor. If the elevator door does not open within the preset time, it is determined that a person is trapped in the elevator.

[0035] Exemplarily, the camera in the elevator car is used to obtain the car monitoring image in real time. For example, during a certain elevator operation, the camera captures an image showing that there are three passengers in the car. The image is then stored in the car monitoring image database as historical monitoring data. The images in the car monitoring image database are labeled and trained using a convolutional neural network. The set training set includes 5,000 images. Each image is labeled and used to train the model to accurately determine the number of passengers in the car. After the training is completed, a passenger number recognition model is established. If a passenger is detected in the car, the timer starts from zero when the elevator starts to take off and land vertically. At the same time, the door opening and closing sensor monitors the status of the elevator door in real time. The preset time is set to 45 seconds. If the elevator door does not open within 45 seconds, it is determined that a person is trapped in the elevator.

[0036] Step S102, the elevator operation characteristic data acquisition module is used to obtain the elevator operation status, elevator entrapment status, elevator parking position and elevator posture data through sensors in the elevator to form the elevator operation characteristic data.

[0037] If a person is trapped in an elevator, the current data of the elevator circuit is obtained through the power sensor. The historical elevator current data is obtained through the elevator current monitoring database, and the elevator operation status corresponding to the current signal is marked. The decision tree algorithm is used for model training to build an elevator operation status recognition model to identify the current operation status of the elevator. The elevator operation status includes normal operation, power outage, maintenance operation and abnormal operation. The vertical take-off and landing speed of the elevator is obtained by the Hall sensor to determine the state of the elevator being trapped. If the vertical take-off and landing speed of the elevator is 0, the elevator state of being trapped is judged as being stopped and trapped. If the vertical take-off and landing speed of the elevator is not 0, the elevator state of being trapped is judged as being running and trapped. The leveling photoelectric sensor is used to determine whether the elevator that stops trapping people is stopped at the same level. If the elevator that stops trapping people is stopped at the same level, the elevator stop floor is obtained. If the elevator does not stop at the same level, the Beidou positioning system is used to obtain the elevator stop position. The elevator attitude data is obtained through the attitude sensor. The elevator attitude data includes X-axis acceleration, Y-axis acceleration, Z-axis acceleration, X-axis angle, Y-axis angle and Z-axis angle. The air pressure sensor installed in the elevator shaft can obtain the shaft air pressure data in real time. According to the shaft air pressure data, the elevator Z-axis acceleration correction unit is used to predict the elevator Z-axis acceleration drift. The elevator operation status, elevator trapped state, elevator parking position and elevator posture data are standardized and combined into an elevator operation feature vector to form the elevator operation feature data, which is then transmitted to the elevator operation monitoring database.

[0038] For example, if an elevator has a trapped person event, the current data of the elevator circuit is obtained through the power sensor, and the current current is 20 amperes. At the same time, the historical current data is extracted from the elevator current monitoring database, and it is found that the average current during normal operation is 2 to 5 amperes. This signal is marked as an abnormal operation state. The elevator operation state includes normal operation, power outage, maintenance operation and abnormal operation. The decision tree algorithm is used for model training. The training data set contains the current state and corresponding operation status of the elevator operation for the past one thousand times. The elevator operation state recognition model is constructed to identify the current operation state of the elevator. If the Hall sensor detects that the vertical take-off and landing speed of the elevator is 0 meters / second, it is judged that the elevator is in a stopped trapped state. If the vertical take-off and landing speed of the elevator is not 0 meters / second, it is judged that the elevator is in a running trapped state. Use the leveling photoelectric sensor to determine whether the elevator is leveled and stopped. If the elevator is indeed leveled and stopped on the 5th floor, the stop floor is recorded and the corresponding elevator stop position is obtained. If the elevator is not leveled, the Beidou positioning system is used to obtain the precise position of the elevator, such as the floor position obtained, which is 2 meters away from the bottom of the elevator shaft. The attitude data of the elevator is obtained through the WT901C six-axis attitude sensor, where the acceleration of the X axis is 0.2 m / s2, the acceleration of the Y axis is 0.1 m / s2, the acceleration of the Z axis is 9.81 m / s2, the angle of the X axis is 0°, the angle of the Y axis is 1°, and the angle of the Z axis is 0°. At the same time, the air pressure sensor monitors the air pressure in the shaft in real time to be 1013hPa. The WT901C six-axis attitude sensor integrates an attitude solver and a dynamic Kalman filter algorithm. The attitude solution algorithm includes quaternion or Euler angle method, which can accurately output the current attitude of the module in a dynamic environment. It has sensor modules such as a three-axis gyroscope and a three-axis accelerometer, supports the 485-level Modbus standard communication protocol, and is installed in the elevator car to move with it. The obtained elevator attitude data is directly connected to the 485 interface of the DTU. The elevator's operating status is abnormal operation, the trapped person status is stopped and trapped persons are trapped, the stop position is the 5th floor, and the elevator posture data are standardized to generate an elevator operation feature vector, such as [abnormal operation, stop and trapped persons, 5th floor, 0.2m / s2, 0.1m / s2, 9.81m / s2], and the elevator operation feature data is transmitted to the elevator operation monitoring database.

[0039] Step S103, the elevator fault identification module is used to build an elevator fault identification model according to the elevator operation characteristic data, identify the elevator fault type, match the predetermined maintenance plan as a reference maintenance plan, and build elevator trapped person rescue data.

[0040] Through the elevator operation monitoring database, historical elevator operation feature data is obtained, and the elevator fault type is marked. The recurrent neural network is used for model training to build an elevator fault recognition model. According to the elevator operation feature data obtained in real time, the elevator fault recognition model is used to identify the elevator fault type. Based on the elevator fault type, the scheduled maintenance plan is matched as the reference maintenance plan. The elevator operation feature data, the elevator fault type recognition results and the reference maintenance plan are integrated into the elevator trapped rescue data.

[0041] Exemplarily, through the elevator operation monitoring database, historical elevator operation characteristic data is obtained, including X-axis acceleration, Y-axis acceleration, Z-axis acceleration, X-axis angle, Y-axis angle and Z-axis angle, and the elevator fault types corresponding to the historical data are marked, including brake fault, current abnormality, speed sensor fault and door control system fault, etc. The historical elevator operation characteristic data with the elevator fault types marked corresponding to the historical data are used to train the recurrent neural network model to learn the operation characteristics of the elevator under different fault conditions and build an elevator fault recognition model. According to the real-time acquisition of elevator operation characteristic data, such as [abnormal operation, stop trapped, 5th floor, 0.2m / s 2 ,0.1m / s 2 ,9.71m / s 2 ]. Through the pre-trained elevator fault identification model, the elevator fault type is identified as brake fault based on real-time data. Based on the identified fault type, it is matched with the predetermined maintenance plan, and it is found that the reference maintenance plan for brake fault includes checking the wear degree of the brake components and their working status, ensuring that the brake response time and braking force meet the standards, checking the electrical connection of the brake to eliminate faults caused by abnormal current signals, replacing the brake fluid, and performing necessary lubrication. The elevator's operating characteristic data, the identified fault type, and the corresponding reference maintenance plan are integrated into elevator trapped rescue data.

[0042] Step S104, an active alarm analysis module is used to determine whether an emergency alarm is needed based on the facial expressions and movements of the elevator passengers in combination with the elevator operation characteristic data.

[0043] If the elevator passenger uses the one-touch alarm button to actively alarm, the elevator passenger's face recognition is performed through the camera in the elevator car and the Faceboxs model, and the face recognition result is matched with the car face monitoring database to determine whether the elevator passenger is a person who has made a malicious rescue call in the past. If the elevator passenger is not a person who has made a malicious rescue call in the past, an emergency alarm is made through the rescue unit in the elevator car. If the elevator passenger is a person who has made a malicious rescue call in the past, the historical elevator operation feature data and the historical car monitoring image are obtained, and the model is trained using a convolutional neural network based on the car monitoring image to identify the facial expressions and movements of the person who has made a malicious rescue call in the past. Based on the facial expressions and movements of the person who has made a malicious rescue call in the past, combined with the elevator operation feature data, it is determined whether an emergency alarm needs to be made through the rescue unit in the elevator car. If the judgment result is that an emergency alarm is not required, the elevator passenger's face image is sent to the car face monitoring database, and the number of historical malicious calls by the elevator passenger is recorded through the car face monitoring database.

[0044] Exemplarily, when a passenger presses the emergency alarm button in the elevator, the camera in the elevator car immediately captures the passenger's facial image and performs face recognition. The captured passenger facial image is preprocessed, including face detection and normalization, and the face area is extracted using Haar features or the face detector in the Dlib library. The face feature vector is extracted using the Faceboxs model, and the facial image is converted into a 128-dimensional or higher-dimensional feature vector to identify the passenger's identity. The extracted feature vector is compared with the historical data in the car face monitoring database, and the similarity is calculated using cosine similarity. If the similarity exceeds the set similarity threshold, it is judged to be the same identity, thereby identifying whether the elevator passenger is a historical malicious rescue caller. If the passenger's image features are processed by the Facebox model, the extracted face feature vector is [0.85, 0.78, 0.90, ...]. This feature vector is matched with the historical features in the elevator face monitoring database. It is found that the passenger's features do not match the malicious caller record in the database. It is confirmed that the passenger is not a historical malicious rescue caller. The rescue unit in the elevator car sends an alarm signal to notify relevant personnel to rescue. If the elevator passenger is a historical malicious rescue caller, the elevator operation feature data is obtained, [abnormal operation, stop trapped, 5th floor, 0.2m / s 2 ,0.1m / s 2 ,9.71m / s 2], the rescue unit in the elevator car automatically sends out an alarm signal to notify relevant personnel for rescue. Obtain historical elevator operation feature data and historical car monitoring images, use historical car monitoring images with facial expressions and movements of malicious callers marked, use convolutional neural networks to train models, and after the historical malicious rescue caller presses the alarm button, obtain his facial image in real time and identify it with the trained convolutional neural network model to determine the current passenger's expression and movement. If an expression of anxiety or fear is recognized, accompanied by specific body movements, it is judged as a possible emergency, and an alarm signal is sent out through the rescue unit in the elevator car. If it is determined that the passenger's emotions and behaviors do not constitute an emergency, and the elevator operation feature data is normal, no alarm signal is sent, but the passenger's facial image is uploaded to the car face monitoring database, and the number of historical malicious calls by the passenger is recorded, such as 3 times.

[0045] Step S105, the automatic alarm module is used to send elevator trapped person rescue data through the rescue unit in the car and the short message service of the Beidou satellite navigation system.

[0046] The rescue data of people trapped in the elevator is sent to the rescue end through the rescue unit in the elevator car. If the signal of the rescue unit in the elevator car is interrupted, the short message service of the Beidou satellite navigation system is used to send the rescue data of people trapped in the elevator to the rescue end.

[0047] For example, the elevator trapped person rescue data is sent to the rescue end through the rescue unit in the elevator car. If the signal of the rescue unit in the elevator car is interrupted, causing the rescue unit to fail to work normally, it switches to the Beidou short message alarm module for emergency alarm and sends emergency rescue data to the rescue center.

[0048] Step S106, the elevator posture monitoring module is used to continuously monitor the elevator posture data, identify abnormal elevator posture states, and issue real-time alarm reminders.

[0049] According to the attitude sensor installed in the elevator, the real-time elevator attitude data of the elevator during operation is obtained; the normal attitude range is set by using the threshold detection algorithm. If the real-time elevator attitude data exceeds the preset normal attitude range, the current elevator attitude state is judged to be abnormal; according to the abnormal elevator attitude state detection result, the alarm information and the corresponding elevator attitude data are sent to the elevator fault maintenance module through the automatic alarm module.

[0050] Exemplarily, the elevator's operating status is monitored in real time according to the attitude sensor installed in the elevator. When the elevator is operating normally, the attitude data recorded by the sensor is that the X-axis angle is 1.2°, the Y-axis angle is 0.8°, and the Z-axis angle is 0°. The normal attitude range is set by using the threshold detection algorithm. The preset X-axis angle should be between -2° and 2°, the preset Y-axis angle should be between -1.5° and 1.5°, and the preset Z-axis angle should be between -1° and 1°. During a certain operation, the real-time attitude data detected by the elevator's attitude sensor is abnormal. The data shows that the X-axis angle is 2.5°, the Y-axis angle is 1.7°, and the Z-axis angle remains at 0°. Since the X-axis angle of 2.5° and the Y-axis angle of 1.7° are both beyond the normal range, the current elevator attitude state is judged to be abnormal. After the abnormal attitude state is detected, an alarm message is immediately sent to the elevator fault maintenance module through the automatic alarm module, including a detailed record of the current elevator attitude data.

[0051] Among them, according to the weight distribution in the elevator, the elevator's X-axis angle offset and Y-axis angle offset are predicted to identify abnormal elevator posture states.

[0052] According to the attitude sensor installed in the elevator, the real-time elevator attitude data of the elevator during operation is obtained. Through the center of gravity sensor, the weight distribution diagram in the elevator is obtained, and the model is trained using a convolutional neural network to predict the X-axis angle offset and Y-axis angle offset under different weight distribution conditions. According to the real-time weight distribution in the elevator, the X-axis angle offset and Y-axis angle offset under the current elevator weight distribution are predicted. If the difference between the predicted X-axis angle offset and Y-axis angle offset and the real-time detected X-axis angle offset and Y-axis angle offset is greater than the preset difference threshold, the elevator attitude is judged to be abnormal. According to the abnormal elevator attitude detection result, the alarm information and the corresponding abnormal elevator attitude data are sent to the elevator fault maintenance module through the automatic alarm module.

[0053] Exemplarily, the operating status and internal weight distribution of the elevator are monitored in real time by the attitude sensor and center of gravity sensor installed in the elevator. When the elevator is running normally, the real-time elevator attitude data recorded by the sensor is that the X-axis angle offset is 0.5° and the Y-axis angle offset is 0.3°, and the weight distribution map in the elevator is obtained through the center of gravity sensor. According to historical data, a convolutional neural network is used to train different weight distribution conditions to obtain a prediction model, which can predict the X-axis and Y-axis angle offsets of the elevator under different elevator weight distribution conditions. If the total load in the elevator is 400kg in the current weight distribution of the elevator, and the left front weighing of the elevator is high, the model predicts that the X-axis angle offset of the elevator should be 0.6° and the Y-axis angle offset should be 0.4°. The X-axis angle offset detected by the elevator real-time monitoring system is 0.8° and the Y-axis angle offset is 0.5°. Compare the real-time measured angle data with the predicted angle data to calculate the difference between them. For the X-axis angle, the difference is |0.8°-0.6°|=0.2°, and for the Y-axis angle, the difference is |0.5°-0.4°|=0.1°. If the preset difference threshold is 0.15°. Since the X-axis difference of 0.2° is greater than the threshold of 0.15°, the elevator posture state is judged to be abnormal. Through the automatic alarm module, the alarm information is immediately sent, and the current elevator posture data, including the X-axis angle offset of 0.8° and the Y-axis angle offset of 0.5°, is sent to the elevator fault maintenance module, providing real-time feedback to the maintenance team, so that they can quickly locate the problem and deal with it.

[0054] Step S107, the elevator fault maintenance module is used to generate a maintenance work order for the equipment according to the elevator trapped person rescue data, and synchronously update the maintenance execution status and detection information of the faulty elevator.

[0055] Generate equipment maintenance work orders based on elevator rescue data and elevator posture abnormality data, including equipment model, fault type, maintenance steps and elevator parking position. Use the dispatching system at the rescue end to transmit maintenance work orders and maintenance reports to the maintenance unit, assign maintenance personnel and maintenance tools to the task, and record the dispatch time and maintenance personnel information. According to the progress and result information fed back by the maintenance personnel during the execution process, record the maintenance execution status information in real time and record it in the paperless maintenance database, and simultaneously update the maintenance execution status and detection information of the faulty elevator. The maintenance execution status information includes the fault repair completion mark, maintenance record and the confirmation result of the execution personnel. According to the historical elevator operation feature vector and historical maintenance records of the elevator, formulate a paperless maintenance plan for the elevator, including regular maintenance and emergency maintenance.

[0056] For example, in an elevator failure incident, the elevator was trapped on the 5th floor, and the passengers notified the rescue end through the emergency alarm system. Based on the obtained elevator trapped rescue data, a maintenance work order was generated, including the equipment model XYZ-123, the fault type was brake failure, and the maintenance steps included checking the brake status, replacing damaged parts, and testing the elevator function. The stop position was the 5th floor. After receiving this maintenance work order, the dispatch system at the rescue end quickly sent the maintenance report and work order information to the maintenance team. The dispatch system assigned two maintenance personnel to the task, namely maintenance personnel A and maintenance personnel B, and recorded the dispatch time as 15:45 on May 1, 2020. The maintenance tools included a multimeter, a wrench, and a spare brake assembly. After maintenance personnel A and maintenance personnel B arrived at the elevator location, they provided real-time feedback on the maintenance progress, including feedback from maintenance personnel A that the brake had been checked and there was no obvious wear, and maintenance personnel B that the damaged brake components had been replaced. This information was recorded in real time in the paperless maintenance database, and the maintenance execution status of the faulty elevator was updated. The records included the fault repair completion mark, the maintenance record that the brake had been replaced, and the execution personnel confirmed that both maintenance personnel A and maintenance personnel B had confirmed it. According to the historical operation feature vector of the elevator, the Y-axis acceleration value of the elevator under high load conditions exceeded 0.5m / s2, and it had approached or exceeded the set threshold many times, and signs of wear of the brake system were detected in many operations. The historical maintenance records of the elevator showed that the brake system of the elevator had been repaired twice in the past six months, but there was still a problem of rapid wear. Therefore, a paperless maintenance plan for the elevator was developed, including increasing the frequency of regular inspections of the brake system, adjusting the original inspection every six months to once a quarter, and including an emergency maintenance plan, adding real-time monitoring for the detection of abnormal conditions of the brake system, ensuring that when the Y-axis acceleration exceeds 0.5m / s2 again, the emergency alarm is immediately triggered and maintenance personnel are arranged for on-site inspection and repair. In addition, the plan also arranges for the calibration of the car posture sensor to ensure the accuracy of posture detection in the next three months.

[0057] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept of the present application. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) to form a technical solution.

Claims

1. An elevator monitoring system based on Beidou satellite positioning, characterized in that: The system comprises: The elevator entrapment event recognition module is used to obtain real-time car monitoring images, identify the number of passengers in the elevator car, and determine whether an elevator entrapment event has occurred based on the door opening time after the car takes off and lands vertically; The elevator operation characteristic data acquisition module is used to obtain the elevator operation status, elevator entrapment status, elevator parking position and elevator posture data through the sensors in the elevator to form the elevator operation characteristic data; The elevator fault identification module is used to build an elevator fault identification model based on the elevator operation characteristic data, identify the elevator fault type, match the predetermined maintenance plan as a reference maintenance plan, and build elevator trapped person rescue data; Active alarm analysis module, used to determine whether emergency alarm is needed based on the facial expressions and movements of elevator passengers combined with elevator operation characteristic data; Automatic alarm module, used to send rescue data of trapped people in the elevator through the rescue unit in the car and the short message service of Beidou satellite navigation system; The elevator posture monitoring module is used to continuously monitor the elevator posture data, identify abnormal elevator posture status, and issue real-time alarm reminders; The elevator fault maintenance module is used to generate equipment maintenance work orders based on elevator trapped person rescue data, and synchronously update the maintenance execution status and detection information of the faulty elevator.

2. The system according to claim 1, wherein: The elevator entrapment event recognition module is used to obtain real-time car monitoring images, identify the number of passengers in the elevator car, and determine whether an elevator entrapment event occurs based on the door opening time after the car vertically takes off and lands, including: Through the camera in the elevator car, real-time car monitoring images are obtained, and the real-time car monitoring images are transmitted to the car monitoring image database; through the car monitoring image database, historical car monitoring images are obtained, and the number of passengers is marked. The convolutional neural network is used for model training, and a passenger number recognition model is constructed to identify the number of passengers in the elevator car; if there are passengers in the elevator car, the timer starts timing when the car takes off and lands vertically, and the door opening and closing status of the elevator door is obtained in real time using the door opening and closing sensor. If the elevator door is not opened within the preset time, it is determined that a person is trapped in the elevator.

3. The system according to claim 1, wherein: The elevator operation characteristic data acquisition module is used to acquire the elevator operation status, elevator entrapment status, elevator parking position and elevator posture data through sensors in the elevator to form elevator operation characteristic data, including: The current data of the elevator circuit is obtained through the power sensor, the elevator operation status corresponding to the current signal is marked, the decision tree algorithm is used for model training, and an elevator operation status recognition model is constructed to identify the current operation status of the elevator; the vertical take-off and landing speed of the elevator is obtained by the Hall sensor to determine the state of the elevator being trapped; the leveling photoelectric sensor is used to determine whether the elevator that stops trapping people is parked at the same level, and the Beidou positioning system is used to obtain the elevator's parking position; the elevator attitude data is obtained through the attitude sensor, and the elevator attitude data includes X-axis acceleration, Y-axis acceleration, Z-axis acceleration, X-axis angle, Y-axis angle and Z-axis angle; the elevator operation status, elevator trapped status, elevator parking position and elevator attitude data are standardized and combined into an elevator operation feature vector to constitute elevator operation feature data, which are transmitted to the elevator operation monitoring database.

4. The system according to claim 1, wherein: The elevator fault identification module is used to construct an elevator fault identification model according to the elevator operation characteristic data, identify the elevator fault type, match the predetermined maintenance plan as a reference maintenance plan, and construct elevator trapped person rescue data, including: Through the elevator operation monitoring database, historical elevator operation characteristic data is obtained, and the elevator fault types are marked. A recurrent neural network is used for model training to build an elevator fault recognition model. According to the elevator operation characteristic data obtained in real time, the elevator fault recognition model is used to identify the elevator fault type. Based on the elevator fault type, the scheduled maintenance plan is matched as the reference maintenance plan. The elevator operation characteristic data, the elevator fault type recognition results and the reference maintenance plan are integrated into elevator trapped people rescue data.

5. The system according to claim 1, wherein: The active alarm analysis module is used to determine whether an emergency alarm is needed based on the facial expressions and movements of elevator passengers in combination with elevator operation characteristic data, including: If the elevator passenger uses the one-touch alarm button to actively alarm, the elevator passenger's face recognition is performed through the elevator car camera and Faceboxs model, and the face recognition result is matched with the car face monitoring database to determine whether the elevator passenger is a person who has made a historical malicious rescue call; if the elevator passenger is not a person who has made a historical malicious rescue call, an emergency alarm is made through the elevator car rescue unit; if the elevator passenger is a person who has made a historical malicious rescue call, the historical elevator operation feature data and the historical car monitoring images are obtained, and the model is trained using a convolutional neural network based on the car monitoring images to identify the facial expressions and movements of the historical malicious rescue callers; based on the facial expressions and movements of the historical malicious rescue callers and the elevator operation feature data, it is determined whether an emergency alarm is required through the elevator car rescue unit; if the judgment result is that an emergency alarm is not required, the elevator passenger's face image is sent to the car face monitoring database, and the number of historical malicious calls by the elevator passenger is recorded through the car face monitoring database.

6. The system according to claim 1, wherein: The automatic alarm module is used to send elevator trapped person rescue data through the rescue unit in the car and the short message service of the Beidou satellite navigation system, including: The rescue data of people trapped in the elevator is sent to the rescue terminal through the rescue unit in the elevator car; If the signal from the rescue unit in the elevator car is interrupted, the short message service of the Beidou satellite navigation system will be used to send the elevator trapped person rescue data to the rescue end.

7. The system according to claim 1, wherein: The elevator attitude monitoring module is used to continuously monitor the elevator attitude data, identify abnormal elevator attitude states, and provide real-time alarm reminders, including: According to the attitude sensor installed in the elevator, the real-time elevator attitude data during the operation of the elevator is obtained; the normal attitude range is set by using the threshold detection algorithm. If the real-time elevator attitude data exceeds the preset normal attitude range, the current elevator attitude state is judged to be abnormal; According to the abnormal state detection result of the elevator posture, the alarm information and the corresponding elevator posture data are sent to the elevator fault maintenance module through the automatic alarm module.

8. The system according to claim 1, wherein: The elevator fault maintenance module is used to generate a maintenance work order for the equipment according to the elevator trapped person rescue data, and synchronously update the maintenance execution status and detection information of the faulty elevator, including: Based on the elevator trapped person rescue data and elevator posture abnormality data, generate the equipment maintenance work order, including equipment model, fault type, maintenance steps and elevator stop position; use the dispatching system at the rescue end to transmit the maintenance work order and maintenance report to the maintenance unit, assign maintenance personnel and maintenance tools to the task, and record the dispatching time and maintenance personnel information; based on the progress and result information fed back by the maintenance personnel during the execution process, record the maintenance execution status information in real time and record it in the paperless maintenance database, and simultaneously update the maintenance execution status and detection information of the faulty elevator. The maintenance execution status information includes the fault repair completion mark, maintenance record and the confirmation result of the executor; formulate a paperless maintenance plan for the elevator based on the historical elevator operation feature vector and historical maintenance records, including regular maintenance and emergency maintenance.

Images