Elevator self-diagnosis fault early warning system based on information fusion

Abstract

The invention discloses an elevator self-diagnosis fault early warning system based on information fusion, which belongs to the technical field of elevator fault diagnosis and comprises an information acquisition subsystem and a data analysis processing subsystem. According to the method, the acceleration and deceleration and vibration in all directions in the elevator running process, the action condition of a brake, the temperature rise condition of the brake, the insufficient braking force condition, the counter-weight over-travel distance and the like are continuously measured, self-learning training is carried out, a model is established, and when abnormal deviation occurs in the comparison result of monitored data and the established model, the abnormal deviation is detected. And the potential safety hazards of the elevator are found and eliminated in time, and the problems that the liability consciousness of safety management personnel is not high, the level of maintenance personnel is insufficient, and the potential safety hazards of the elevator are difficult to find in advance are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of elevator fault diagnosis, and in particular to an elevator self-diagnosis fault early warning system based on information fusion. Background Art

[0002] With the development of modern society, elevators have experienced rapid growth as a means of transporting people up and down high-rise buildings. While enjoying the convenience of elevator service, people are also placing greater demands on ride comfort and safety. Analysis of numerous elevator accidents reveals that before an accident occurs due to component failure, operating parameters such as starting and braking acceleration, normal operating data curves, and motor temperature rise can differ from those during normal operation. If these operating parameter changes can be monitored in advance and early warnings issued to the user or manufacturer, along with professional investigations, elevator accidents can be effectively prevented in the bud.

[0003] The safety of existing elevators depends not only on correct passenger operation but also on routine inspections by user safety managers and preventive maintenance by maintenance personnel. This requires a high level of responsibility from safety managers and the expertise and experience of maintenance personnel. However, in reality, routine inspections by user units are difficult to implement, and maintenance personnel vary widely in their skills, often failing to detect safety hazards. This can lead to elevators operating with defects, posing a serious threat to passenger safety. Therefore, there is an urgent need for a system that can monitor elevator operating parameters in real time, around the clock, and promptly notify relevant personnel for targeted inspections when an elevator experiences an operational anomaly. This system would provide early warning of elevator failures and improve the inherent safety of elevators. Summary of the Invention

[0004] The technical problem to be solved by the present invention is: how to improve the current situation where automatic early warning of elevator faults cannot be achieved, and then solve the problems of weak sense of responsibility of safety management personnel, insufficient level of maintenance personnel, and difficulty in discovering elevator safety hazards in advance, and provide an elevator self-diagnosis fault early warning system based on information fusion.

[0005] The present invention solves the above technical problems through the following technical solutions. The present invention includes an information acquisition subsystem and a data analysis and processing subsystem;

[0006] The information collection subsystem includes an acceleration and vibration monitoring subsystem, a brake action monitoring subsystem, a brake operation monitoring subsystem, a braking force deficiency monitoring subsystem, and a counterweight overtravel distance monitoring subsystem;

[0007] Among them, the acceleration and vibration monitoring subsystem is used to obtain acceleration and vibration monitoring data in all directions of the elevator car, the brake action monitoring subsystem is used to obtain judgment data on whether the logic of the brake contactor action and the brake action is correct, that is, the brake logic monitoring data, and the brake operation monitoring subsystem is used to obtain brake temperature monitoring data; the insufficient braking force monitoring subsystem is used to obtain the acceleration and vibration monitoring data of the elevator car in the vertical direction within a preset time after the elevator car is leveled, as well as the door opening and closing time data after leveling, and judge whether the braking force is insufficient based on the acceleration and vibration monitoring data and the door opening and closing time data to obtain the braking force monitoring data; the counterweight overtravel distance monitoring subsystem is used to monitor the counterweight overtravel distance when the elevator car is leveled at the top floor;

[0008] The data analysis and processing subsystem includes an analysis subsystem, which is used to input a five-dimensional array consisting of acceleration and vibration monitoring data, braking logic monitoring data, temperature monitoring data, braking force monitoring data and counterweight overtravel distance monitoring data into a deep network. Through continuous learning through the deep network, it ultimately realizes elevator fault warning and determines whether the elevator has a fault; wherein, the deep network adopts an improved ResNet-12 network.

[0009] Furthermore, in the acceleration and vibration monitoring subsystem, acceleration and vibration sensors are installed at set positions on the elevator car bottom to measure the acceleration and vibration of the elevator car in all directions, thereby obtaining acceleration and vibration monitoring data.

[0010] Furthermore, the brake action monitoring subsystem includes a controller and two micro switches. The micro switches are respectively installed on the brake body or near it, and the two micro switches are connected to the controller. A pair of auxiliary contacts are taken from the brake contactor and connected to the controller to judge whether the logic of the brake contactor action and the brake action is correct, that is, to obtain the brake logic judgment data.

[0011] Furthermore, in the brake operation monitoring subsystem, the temperature of the brake is monitored in real time by four external temperature sensors. When the data analysis and processing subsystem performs analysis, the maximum and minimum values ​​of the temperature data obtained by the four temperature sensors are deleted, and the average of the remaining two temperature values ​​is used to obtain the final temperature value for analysis. The final temperature value is the temperature monitoring data.

[0012] Furthermore, in the insufficient braking force monitoring subsystem, the acceleration and vibration signals in the vertical direction are monitored by the acceleration and vibration sensors in the acceleration and vibration monitoring subsystem, and the camera installed in the car monitors the door opening and closing process for analysis and judgment. When the acceleration exceeds the safety threshold T1, the vibration signal deviates from the baseline B1, or the door opening and closing time monitored by the camera does not meet the safety range t2-t3, it is judged that the braking force is insufficient, that is, the monitoring data of whether the braking force is insufficient is obtained, that is, the braking force monitoring data.

[0013] Furthermore, in the counterweight overtravel distance monitoring subsystem, when the car is level on the top floor, the distance measurement function is triggered, and the laser ranging sensor is used to measure the distance L between the measurement reference plane and the top surface of the buffer, which is compared with the alarm threshold s3 to determine whether the distance L is within the range of the alarm threshold s3, thereby obtaining the counterweight overtravel distance monitoring data, wherein the range of the alarm threshold s3 is (L+αs1, L+αs2), s1 and s2 are the basic counterweight overtravel distances, and α is the proportional coefficient.

[0014] Furthermore, in the improved ResNet-12 network, the original ResNet-12 network is improved by introducing a detail enhancement attention module before the second batch normalization layer and the third batch normalization layer of the original ResNet-12 network.

[0015] Furthermore, the detail enhancement attention module includes a detail enhancement convolutional layer and a content-guided attention mechanism module. The detail enhancement convolutional layer extracts features by deploying multiple ordinary convolutions and difference convolutions in parallel. The content-guided attention mechanism module generates a spatial importance map of a set channel in a coarse-to-fine manner and fuses channel attention and spatial attention.

[0016] Compared with the existing technology, the present invention has the following advantages: the elevator self-diagnosis fault warning system based on information fusion continuously and uninterruptedly measures the acceleration and deceleration and vibration in all directions during the operation of the elevator, the action of the brake, the temperature rise of the brake, the insufficient braking force and the counterweight overtravel distance, etc., and conducts self-learning training to establish a model. When the monitored data shows an abnormal deviation from the comparison result of the established model, it is immediately sent to relevant personnel for targeted investigation, so as to timely discover and eliminate the safety hazards of the elevator, and solve the problems of weak sense of responsibility of safety management personnel, insufficient level of maintenance personnel, and difficulty in discovering elevator safety hazards in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 1 is a schematic diagram of the basic functions of each subsystem in the elevator self-diagnosis fault warning system based on information fusion in an embodiment of the present invention;

[0018] Figure 2 This is a schematic diagram of the communication connection between each monitoring subsystem and the data analysis and processing subsystem in the information collection subsystem in an embodiment of the present invention;

[0019] Figure 3 This is a schematic diagram of a braking force deficiency determination process in an embodiment of the present invention;

[0020] Figure 4 Schematic diagram of the improved ResNet-12 network structure in an embodiment of the present invention;

[0021] Figure 5 2 is a schematic diagram of the structure of the detail enhancement attention module in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process. However, the protection scope of the present invention is not limited to the following embodiment.

[0023] This embodiment provides a technical solution: an elevator self-diagnosis fault warning system based on information fusion, including an information acquisition subsystem and a data analysis and processing subsystem. The basic functions of each subsystem are as follows: Figure 1 The functions of the information acquisition subsystem are mainly: data acquisition and transmission, and the functions of the data analysis and processing subsystem are mainly: monitoring data collection, adaptive learning, data comparison operation and information transmission.

[0024] The following is a further description of the above-mentioned information collection subsystem and data analysis and processing subsystem:

[0025] 1. Information collection subsystem

[0026] 1.1 Acceleration and vibration monitoring subsystem

[0027] To address the issue of malfunctioning of the safety clamp caused by the lining of the elevator car guide shoe falling into the safety clamp body, analysis revealed that the horizontal vibration data of the elevator car would change significantly after the guide shoe fell off. Therefore, acceleration and vibration sensors were installed at appropriate locations on the elevator car floor to measure acceleration and vibration data in all directions (to obtain acceleration and vibration monitoring data). The system uses a MEMS triaxial digital acceleration sensor installed on the elevator car floor. Its support device uses strong magnetic adsorption, which makes it easy to install and remove and has good adaptability. The vibration sensor system includes an electronic level, the DigPas DWL-8500XY model. A signal line follows the car's accompanying cable to connect to the data analysis and processing subsystem.

[0028] 1.2. Brake action monitoring subsystem

[0029] For older elevators that don't have brake action monitoring installed, a microswitch can be added to the brake. The brake contactor's auxiliary contacts can be used to perform logical judgments, and the judgment results are input into the data analysis and processing subsystem of this fault warning system. Specifically, the system includes a controller and two microswitches. Each microswitches is installed on or near the brake body, where they can easily capture the brake's action status. The two microswitches are connected to the controller, and a pair of auxiliary contacts from the brake contactor are connected to the controller. The controller determines whether the logic of the brake contactor and brake action is correct. The controller then outputs the logical judgment data (brake logic monitoring data) to the signal analysis and processing subsystem.

[0030] To ensure reliable data transmission, all functional units and data analysis and processing units are connected by wires. Figure 2 shown.

[0031] 1.3. Brake operation monitoring subsystem

[0032] To address the issue of brake-operated operation due to insufficient brake travel when the brake is open, an external temperature sensor monitors the brake temperature in real time and issues an alarm if the brake temperature rises too high. This system includes four digital temperature sensors, which transmit the brake temperature data acquired by these sensors to the data analysis and processing subsystem in real time. The maximum and minimum temperature values ​​are removed, and the remaining two temperature values ​​are averaged to obtain the final temperature value (temperature monitoring data), preventing errors caused by sensor failure.

[0033] 1.4. Braking force deficiency monitoring subsystem

[0034] Regarding the problem of unexpected movement caused by insufficient braking force, there are usually early signs of insufficient braking force, such as vertical acceleration in a very short time after leveling (under normal circumstances, there will be a door opening and closing time T after leveling). Specifically, it is composed of the acceleration and vibration sensors mentioned in Article 1.1, and the camera installed in the car. The acceleration and vibration sensors are used to monitor the acceleration and vibration signals in the vertical direction, and the camera installed in the car monitors the door opening and closing process for analysis and judgment. As long as the acceleration exceeds the safety threshold T1, the vibration signal deviates from the baseline B1, or the door opening and closing time monitored by the camera does not meet the safety range t2-t3, as long as any of these conditions is not met, it will prompt that the braking force is insufficient, which is safer and more accurate. Figure 3 As shown, the monitoring data of whether the braking force is insufficient, that is, the braking force monitoring data, is finally obtained.

[0035] 1.5. Counterweight overtravel distance monitoring subsystem

[0036] To address potential risks such as insufficient overhead clearance and guide rail travel due to the elongation or shortening of the suspension wire rope, which can cause the counterweight overtravel distance to fall outside the range specified by the permanent markings when the car is leveled at the top floor, a laser distance sensor can automatically measure the counterweight overtravel distance when the car is leveled at the top floor and compare it with a set value to achieve automatic judgment. Specifically, this system includes a laser distance sensor and its fixing device. The laser distance sensor is fixed to the base of the buffer (counterweight buffer) through the fixing device. The distance L between the laser distance sensor's measurement reference surface and the buffer's top surface is measured. The counterweight overtravel distance ranges from αs1 to αs2, and the alarm threshold is set to s3, which has a range of (L+αs1,L+αs2). Here, s1 and s2 are the basic counterweight overtravel distances. The final counterweight overtravel distance range changes dynamically in direct proportion to the floor height; α is the proportional coefficient: the range is smaller for lower floors and larger for higher floors, ensuring maximum elevator safety. When the car is leveled on the top floor, the distance measurement function is automatically triggered, and the measured distance L is compared with the alarm threshold s3. If there is an abnormality (the distance L is not within the range of the alarm threshold s3), an alarm message is issued.

[0037] 2. Digital analysis and processing subsystem

[0038] The system includes parameter setting and display, data storage and analysis, and information transmission. The parameter setting and display function allows the operator to set key elevator system parameters, such as operating speed, acceleration, deceleration, and door opening and closing times, and displays the current status of these parameters in real time. By setting parameters, the elevator's performance can be adjusted to suit different usage requirements or special circumstances. The display function provides transparency, allowing the operator to understand the elevator's operating status at all times and detect anomalies promptly. The data storage and analysis function collects various data during elevator operation, such as fault records, operating hours, and usage frequency, and stores this data for subsequent analysis. The data storage provides a record of historical faults and operating patterns, while the analysis function identifies fault patterns, predicts potential problems, and optimizes elevator maintenance plans. This is crucial for improving elevator reliability and reducing downtime. The information transmission function allows the system to automatically send notifications or reports to relevant personnel or systems, such as maintenance teams, property management, or remote monitoring centers, when an anomaly or fault is detected. This information transmission function ensures that faults and anomalies are quickly notified to the relevant personnel, enabling timely response, minimizing the impact on users, and improving problem resolution efficiency.

[0039] Data storage and analysis uses the method of receiving and storing externally transmitted data and status. When the comparison result exceeds the set threshold, an alarm message is output and sent to relevant personnel for timely disposal. The elevator operation data is continuously learned through the algorithm, and finally the elevator fault warning is realized. Here, the five kinds of original data mentioned above, namely acceleration and vibration monitoring data, braking logic monitoring data, temperature monitoring data, braking force monitoring data, and counterweight overtravel distance monitoring data, are mainly composed of a five-dimensional array and input into the deep network. Through continuous learning through the deep network, the elevator fault warning is finally realized to determine whether the elevator has a fault. The deep network here adopts the improved ResNet-12 network, such as Figure 4 shown.

[0040] The improvement is to incorporate the detail enhancement attention module (DEAB), such as Figure 5 DEAB consists of a detail-enhanced convolutional layer (DEConv) and a content-guided attention mechanism (CGA) module. The DEConv layer extracts features by deploying multiple ordinary convolutions and differential convolutions in parallel, enhancing representation and generalization capabilities without introducing additional parameters or computational costs. The CGA module generates channel-specific spatial importance maps in a coarse-to-fine manner and fuses channel attention with spatial attention to achieve information interaction and efficient gradient flow.

[0041] In summary, the elevator self-diagnosis fault warning system based on information fusion in the above-mentioned embodiment establishes a model by continuously and uninterruptedly measuring the acceleration and deceleration and vibration in all directions during the operation of the elevator, the action of the brake, the temperature rise of the brake, the insufficient braking force and the counterweight overtravel distance, etc., and conducting self-learning training. When the monitored data shows an abnormal deviation from the comparison result of the established model, it is immediately sent to relevant personnel for targeted investigation, so as to timely discover and eliminate the safety hazards of the elevator, and solve the problems of weak sense of responsibility of safety management personnel, insufficient level of maintenance personnel, and difficulty in discovering elevator safety hazards in advance.

[0042] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. An elevator self-diagnosis fault warning system based on information fusion, characterized in that: Including information collection subsystem and data analysis and processing subsystem; The information collection subsystem includes an acceleration and vibration monitoring subsystem, a brake action monitoring subsystem, a brake with brake operation monitoring subsystem, a braking force deficiency monitoring subsystem and a counterweight overtravel distance monitoring subsystem; Among them, the acceleration and vibration monitoring subsystem is used to obtain the acceleration and vibration monitoring data of the elevator car in all directions, the brake action monitoring subsystem is used to obtain the judgment data of whether the logic of the brake contactor action and the brake action is correct, that is, the brake logic monitoring data, and the brake operation monitoring subsystem is used to obtain the temperature monitoring data of the brake; the insufficient braking force monitoring subsystem is used to obtain the acceleration and vibration monitoring data of the elevator car in the vertical direction within a preset time after the elevator car is leveled and the door opening and closing time data after leveling, and judge whether the braking force is insufficient based on the acceleration and vibration monitoring data and the door opening and closing time data to obtain the braking force monitoring data; the counterweight overtravel distance monitoring subsystem is used for the counterweight overtravel distance monitoring data when the elevator car is leveled on the top floor; The data analysis and processing subsystem includes an analysis subsystem, which is used to input a five-dimensional array consisting of acceleration and vibration monitoring data, braking logic monitoring data, temperature monitoring data, braking force monitoring data and counterweight overtravel distance monitoring data into a deep network, and continuously learn through the deep network to ultimately achieve elevator fault warning and determine whether the elevator has a fault; wherein the deep network adopts an improved ResNet-12 network.

2. The elevator self-diagnosis fault warning system based on information fusion according to claim 1 is characterized in that: In the acceleration and vibration monitoring subsystem, acceleration and vibration sensors are installed at set positions on the elevator car bottom to measure the acceleration and vibration of the elevator car in all directions, thereby obtaining acceleration and vibration monitoring data.

3. The elevator self-diagnosis fault warning system based on information fusion according to claim 2 is characterized in that: The brake action monitoring subsystem includes a controller and two micro switches. The micro switches are respectively installed on the brake body or near it, and the two micro switches are connected to the controller. A pair of auxiliary contacts are taken from the brake contactor and connected to the controller to determine whether the logic of the brake contactor action and the brake action is correct, so as to obtain the brake logic judgment data.

4. The elevator self-diagnosis fault warning system based on information fusion according to claim 3 is characterized in that: In the brake operation monitoring subsystem, the temperature of the brake is monitored in real time by four external temperature sensors. When the data analysis and processing subsystem performs analysis, the maximum and minimum values ​​of the temperature data obtained by the four temperature sensors are deleted, and the average of the remaining two temperature values ​​is used to obtain the final temperature value for analysis. The final temperature value is the temperature monitoring data.

5. The elevator self-diagnosis fault warning system based on information fusion according to claim 4 is characterized in that: In the insufficient braking force monitoring subsystem, the acceleration and vibration signals in the vertical direction are monitored by the acceleration and vibration sensors in the acceleration and vibration monitoring subsystem, and the camera installed in the car monitors the door opening and closing process for analysis and judgment. When the acceleration exceeds the safety threshold T1, the vibration signal deviates from the baseline B1, or the door opening and closing time monitored by the camera does not meet the safety range t2-t3, it is judged that the braking force is insufficient, that is, the monitoring data of whether the braking force is insufficient is obtained, that is, the braking force monitoring data.

6. The elevator self-diagnosis fault warning system based on information fusion according to claim 5 is characterized in that: In the counterweight over-travel distance monitoring subsystem, when the car is leveled on the top floor, the distance measurement function is triggered, and the laser distance measuring sensor is used to measure the distance L between the measurement reference plane and the top surface of the buffer, which is compared with the alarm threshold s3 to determine whether the distance L is within the range of the alarm threshold s3, thereby obtaining the counterweight over-travel distance monitoring data, wherein the range of the alarm threshold s3 is (L+αs1, L+αs2), s1 and s2 are the basic counterweight over-travel distances, and α is the proportional coefficient.

7. An elevator self-diagnosis fault warning system based on information fusion according to claim 1 or 6, characterized in that: In the improved ResNet-12 network, the original ResNet-12 network is improved by introducing a detail enhancement attention module before the second batch normalization layer and the third batch normalization layer of the original ResNet-12 network.

8. The elevator self-diagnosis fault warning system based on information fusion according to claim 7 is characterized in that: The detail enhancement attention module includes a detail enhancement convolution layer and a content-guided attention mechanism module. The detail enhancement convolution layer extracts features by deploying multiple ordinary convolutions and difference convolutions in parallel. The content-guided attention mechanism module generates a spatial importance map of a set channel in a coarse-to-fine manner and fuses channel attention and spatial attention.

Images