Intelligent elevator maintenance management method and system based on speed curve comparison
By using elevator speed curve comparison technology combined with machine learning, real-time diagnosis and predictive maintenance of elevator faults can be achieved, solving the problems of high maintenance costs and insufficient data in traditional elevators, and improving diagnostic accuracy and coverage.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ZHEJIANG XINCHENG BUILDING TECHNOLOGY CO LTD
- Filing Date
- 2026-03-02
- Publication Date
- 2026-04-24
AI Technical Summary
Existing intelligent elevator maintenance systems struggle to achieve effective and reliable fault prediction and diagnosis while controlling costs. High sensor installation costs and insufficient data support make it difficult for the system to achieve true fault tracing and proactive early warning.
By collecting the actual operating speed curve of the elevator and the speed curve of the traction machine encoder, comparing and analyzing them, an anomaly dataset is generated. Combined with maintenance interaction information, a sample dataset is formed, which is then used for machine learning training to achieve real-time fault diagnosis and predictive maintenance.
Real-time fault diagnosis and prediction can be achieved without the need for high-cost sensors, reducing hardware costs, improving diagnostic accuracy and fault coverage, and expanding the dimensions of elevator health status monitoring.
Smart Images
Figure CN121913397A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of elevator maintenance, management and fault diagnosis technology, and in particular to an intelligent elevator maintenance management method and system based on speed curve comparison. Background Technology
[0002] The increasing production and usage of elevators has led to a continuous expansion of elevator maintenance needs. Because traditional maintenance methods relying on manual inspections are insufficient to meet the demands for comprehensive coverage and accurate diagnosis, there is a growing trend towards intelligent maintenance.
[0003] Currently, intelligent elevator maintenance mainly focuses on three technological directions: first, an IoT-based monitoring model, which uses sensors installed on key components to collect real-time operational data; second, an AI and big data-based fault prediction model, which uses algorithms to analyze historical data to identify fault patterns; and third, an intelligent diagnostic model combining IoT and AI. Independent IoT monitoring can only reflect existing anomalies and lacks fault prediction capabilities; AI algorithms require high-quality, multi-dimensional real-time data support, and their prediction accuracy is easily limited. Furthermore, the IoT-AI combined model requires the installation of sensors on elevator components, but currently commonly used sensors (such as vibration, temperature, and load sensors) are mostly limited to condition monitoring and cannot provide continuous physical quantity data with precise measurement characteristics, making it difficult for the system to achieve true fault tracing and proactive early warning. Using sensors with precise measurement functions, however, will face high production costs and subsequent system maintenance costs.
[0004] Therefore, how to build an effective, reliable, and feasible intelligent elevator maintenance system while controlling costs is an urgent technical problem to be solved. Summary of the Invention
[0005] To address the aforementioned technical problems, this invention provides an intelligent elevator maintenance management method and system based on speed curve comparison. Through multimodal data acquisition and information processing, it achieves real-time diagnosis, predictive maintenance, and maintenance management of elevator faults; through physical sensor data acquisition and speed curve comparison analysis, it realizes a transformation from scheduled repairs to predictive diagnosis-based maintenance. The technical solution adopted by this invention is as follows: An intelligent elevator maintenance management method based on speed curve comparison includes the following steps: Data acquisition includes real-time collection of elevator actual operating speed curves, elevator traction machine encoder speed curves, and maintenance interaction information. Speed curve monitoring involves continuously comparing the two speed curves collected in real time and monitoring the comparison data. If the actual operating speed curve of the elevator is inconsistent with the speed curve of the elevator traction machine encoder, an abnormal event is determined to have occurred. The two curves for that time period are marked and stored to generate the first abnormal dataset. Information processing includes sending maintenance task instructions based on a first abnormal dataset, automatically recording fault characteristic information based on collected maintenance interaction information, and generating a second interaction dataset; determining whether there is a time correlation between the first abnormal dataset and the second interaction dataset, and if there is a time correlation, merging the first abnormal dataset and the second interaction dataset to generate a sample dataset. Receive instructions and processing feedback, conduct on-site verification and processing feedback based on the maintenance task instructions, determine the root cause of the fault, bind the root cause of the fault to the sample dataset, form a machine learning training sample with causal orientation and store it; Model updates and database maintenance utilize historically accumulated training samples to continuously train and optimize the fault diagnosis model, enabling intelligent prediction and diagnosis of fault causes for subsequent new abnormal events.
[0006] Furthermore, the step of obtaining the actual operating speed curve of the elevator is to monitor the linear velocity of the elevator speed governor and generate a speed curve; the step of obtaining the speed curve of the elevator traction machine encoder is to monitor the angular velocity of the elevator traction machine through a rotary encoder and generate a speed curve.
[0007] Furthermore, the comparison between the actual elevator operating speed curve and the elevator traction machine encoder speed curve includes a curve integrity comparison and a difference comparison; the integrity comparison: monitors whether the speed curve is a complete curve, the complete curve includes an acceleration segment, a constant speed segment, and a deceleration segment; when a sudden drop is detected in the curve, it is determined that the curve has lost its integrity; the difference comparison: monitors whether there are graphical differences in the speed curve in the acceleration segment, constant speed segment, or deceleration segment, and whether these differences can be captured during the comparison process.
[0008] Furthermore, it also includes comparative correlation of speed curves, which includes self-comparison and mutual comparison; the self-comparison is based on the speed curve monitored by integrity comparison. If the speed curve loses its integrity, the speed curve is compared with the samples in the sample dataset, and the most similar sample curve is directly called to obtain the associated fault cause; the mutual comparison is based on the speed curve monitored by difference comparison, which determines whether there is a speed curve deviation between the actual elevator operating speed curve and the elevator traction machine encoder speed curve, and obtains the fault cause based on the speed curve deviation.
[0009] Furthermore, the first abnormal dataset includes information on the abnormal time period, the actual operating speed curve segment of the elevator, the speed curve segment of the elevator traction machine encoder, the location of the elevator car, and the direction of elevator car operation; the second interactive dataset includes information on the fault time, fault location, and fault phenomenon.
[0010] Furthermore, the maintenance task instructions are sent to the mobile terminal of the maintenance personnel; the maintenance task includes on-site fault diagnosis, fault troubleshooting, and checking the elevator system status and fault codes.
[0011] Furthermore, it also includes inferring abnormal elevator conditions based on elevator operation characteristics such as the number of elevator runs, operating range, and operating direction monitored; when the elevator operation characteristics deviate significantly from historical benchmarks, a system warning is generated.
[0012] An intelligent elevator maintenance and management system based on speed curve comparison includes: A multimodal data acquisition module is used to acquire interactive information, the actual operating speed curve of the elevator, and the speed curve of the elevator traction machine encoder. The speed monitoring module is used to monitor the comparison data between the actual operating speed curve of the elevator and the speed curve of the elevator traction machine encoder. An information processing module is used for anomaly detection and the generation of a first anomaly dataset and a second interactive dataset. A data fusion and task engine module, which is used to generate sample datasets and maintenance task instructions; A sample library module, which stores training samples and historical diagnostic cases; The intelligent diagnosis and learning engine module trains and optimizes the model based on the sample information of the sample library module. The maintenance interaction platform module is used to send task instructions and receive on-site feedback.
[0013] Furthermore, the speed monitoring module includes a first speed acquisition unit and a second speed acquisition unit; the first speed acquisition unit is connected to the speed governor encoder signal of the elevator and is used to monitor the speed governor speed in real time and generate the actual operating speed curve of the elevator; the second speed acquisition unit is connected to the rotary encoder signal of the elevator traction machine and is used to monitor the pulse signal of the rotary encoder of the traction machine in real time and generate the encoder speed curve of the elevator traction machine.
[0014] Furthermore, it also includes a non-associated information module, which is used to collect and analyze information on the elevator's operating characteristics.
[0015] The beneficial effects of this invention are: By monitoring the elevator's speed and continuously comparing the actual elevator speed curve with the elevator traction machine encoder speed curve, the system identifies ongoing and impending explicit and implicit faults, collecting a large number of samples to build a model. Then, through task assignment and feedback to maintenance personnel, the model is precisely refined. This eliminates the need for additional sensors with precise measurement capabilities to provide a continuous data source, reducing hardware costs and achieving true fault diagnosis, prediction, and tracing.
[0016] By merging the first abnormal dataset and the second interactive dataset based on time correlation to generate a sample dataset, the root causes of the faults discovered during on-site inspection are bound to the sample dataset to form training samples. These high-quality and labeled training samples can drive the fault diagnosis model to continuously learn and optimize, thereby achieving a continuous improvement in diagnostic accuracy and an expansion of fault coverage.
[0017] By monitoring the elevator's operating characteristics, we can identify hidden faults that do not directly affect speed but indicate elevator abnormalities, thereby increasing the breadth of the system's absolute fault perception and expanding the dimensions of elevator health status monitoring. Attached Figure Description
[0018] Figure 1 This is a structural block diagram of the present invention.
[0019] Figure 2 This is the speed curve diagram of the elevator traction machine encoder of the present invention (complete speed curve diagram).
[0020] Figure 3 This is the emergency stop fault speed curve diagram of the present invention.
[0021] Figure 4 This is a speed curve diagram of the speed limiter of the present invention. Detailed Implementation
[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will now be described in further detail with reference to the accompanying drawings and the following embodiments, so that the public can better understand the implementation method of this invention. The specific implementation scheme of this invention is as follows: Example 1:
[0023] During normal elevator operation, the actual operating status of the elevator is collected in real time and the actual operating speed curve of the elevator is generated.
[0024] An intelligent elevator maintenance management method based on speed curve comparison includes the following steps: Real-time acquisition of elevator actual operating speed curves, elevator traction machine encoder speed curves, and maintenance interaction information.
[0025] Based on speed monitoring, the two speed curves collected in real time are continuously compared and the comparison data is monitored. If the actual operating speed curve of the elevator is inconsistent with the speed curve of the elevator traction machine encoder, an abnormal event is determined to have occurred. The two curves for that time period are marked and stored to generate the first abnormal dataset.
[0026] Specifically, during normal elevator operation, the comparison between the actual elevator speed curve and the elevator traction machine encoder speed curve includes a comparison of the integrity of the curves and a comparison of their differences.
[0027] The integrity comparison involves monitoring whether the speed curve is a complete curve, which includes acceleration, constant speed, and deceleration sections. If a sudden drop is detected in the curve, it is determined that the curve has lost its integrity. Figure 2-3 As shown, the Figure 2 This represents a complete speed curve. When the speed curve suddenly loses its integrity, and the elevator experiences a sudden and drastic change (i.e., an emergency stop), the speed curve will become... Figure 3 Represents the velocity curve, the Figure 3 It is a sudden stop speed curve, in which the elevator stops abruptly after running at a constant speed for a period of time. Figure 3 Elevator malfunctions associated with speed curves include door lock system malfunctions, elevator system malfunctions, inverter malfunctions, and vibrations during operation that trigger safety circuits.
[0028] It should be noted that, Figure 2 The curve pattern is only one of the ways in which a fault can cause curve changes, and sudden drops may occur in the acceleration, constant speed, and deceleration phases; each curve pattern that drops suddenly at different times corresponds to a different fault cause.
[0029] The difference comparison involves monitoring the speed curves for graphical differences during acceleration, constant speed, or deceleration, ensuring these differences are captured during the comparison process. As elevator operating time increases, any change in the shape of any speed curve, if captured during the comparison, indicates a change in the elevator's performance or structure, prompting maintenance personnel to pay attention or conduct an investigation. Figure 4 As shown, the velocity curve exhibits anomalies in the uniform speed range, and the associated causes of elevator malfunctions include structural failures of elevator components, mechanical coupling issues, and problems with the operating posture.
[0030] When the friction between the wire rope and the traction sheave changes, wear occurs in the wire rope traction sheave, and the groove shape of the traction sheave rope changes, leading to a "rope skipping" phenomenon. This type of fault falls under the category of structural faults in the elevator components. When the balance coefficient disrupts the mechanical balance of the elevator system, the elevator's unbalanced state will eventually lead to a "slippage" phenomenon. This type of fault falls under the category of mechanical faults. When the guide shoes wear and the elevator car itself is unstable, abnormal intermittent disconnections may occur in the elevator door system or safety devices. This type of fault falls under the category of problems related to component coupling and operating posture.
[0031] Based on the first abnormal dataset, a maintenance task instruction is sent. Based on the collected maintenance interaction information, fault feature information is automatically recorded by capturing keywords to generate a second interaction dataset. It is determined whether there is a time correlation between the first abnormal dataset and the second interaction dataset. If there is a time correlation, the first abnormal dataset and the second interaction dataset are merged to generate a sample dataset.
[0032] Specifically, the first abnormal dataset includes information on the abnormal time period, segments of the elevator's actual operating speed curve, segments of the elevator traction machine encoder speed curve, the location of the elevator car, and the direction of the elevator car's movement. The second interactive dataset includes information on the fault time, fault location, and fault symptoms.
[0033] When the speed monitoring module detects an inconsistency between the actual operating speed curve of the elevator and the speed curve of the elevator traction machine encoder, the speed curve comparison and analysis module analyzes the inconsistency and uploads the associated data to generate the first abnormal dataset. In the following period, if maintenance personnel receive a fault alarm call, the multimodal data acquisition module captures keywords through the telephone voice and automatically records them to generate the second interactive dataset. The data fusion and task engine module merges the two datasets based on their temporal correlation to generate a sample dataset.
[0034] It should be noted that the maintenance task instructions are sent to the mobile terminal of the maintenance personnel; the maintenance task includes on-site fault diagnosis, fault troubleshooting, and checking the elevator system status and fault codes.
[0035] Based on the maintenance task instructions, on-site verification and processing feedback are conducted to determine the root cause of the fault. The root cause of the fault is then bound to the sample dataset to form causal machine learning training samples, which are then stored. Specifically, after receiving the maintenance task instructions on the terminal, maintenance personnel need to conduct on-site verification of the fault cause and provide feedback on the determined fault cause through the maintenance interaction platform module. The fault cause data will be bound to the sample dataset to form training samples, which are then stored in the sample library module.
[0036] The fault diagnosis model is continuously trained and optimized using historically accumulated training samples, enabling intelligent prediction and diagnosis of fault causes for subsequent new anomalies. Specifically, the intelligent diagnosis and learning engine module continuously trains and optimizes a large number of training samples stored in the sample library module. This achieves increasingly accurate identification of fault causes based on monitored samples.
[0037] It should be noted that the steps for obtaining the actual elevator operating speed curve are to monitor the speed of the elevator speed governor and generate a speed curve; the steps for obtaining the elevator traction machine encoder speed curve are to monitor the angular velocity of the elevator traction machine through a rotary encoder and generate a speed curve. The elevator traction machine encoder speed curve reflects the operating state of the elevator and its components under the current operating conditions, while the actual elevator operating speed curve reflects the true speed of the elevator (the speed governor's speed).
[0038] Compared to monitoring the actual operating speed of an elevator by detecting the linear velocity of the elevator car or the linear velocity of the counterweight movement using sensors, measuring the angular velocity of the speed governor is a simple, easy-to-implement, and low-cost method based on relatively simple measurement technology and the principle of fewer measurement points.
[0039] It should be explained that there are two reasons why the speed governor directly reflects the actual speed of the elevator: First, both swings of the speed governor wire rope are fixed on the car, the car drags the wire rope, and the position of the wire rope relative to the outer edge of the speed governor wheel is fixed; second, the speed governor wheel is driven to rotate by the car, which is a driven operation, and there will be no slippage of the wire rope under normal elevator operation.
[0040] The comparison of the two curves described in the above steps includes two modes: By comparing the actual operating speed curve of the elevator with the speed curve of the elevator traction machine encoder, and by capturing anomalies in the curves, faults can be found, enabling the prediction of faults before they occur.
[0041] The current abnormal curve is compared with the historical abnormal curve. By comparing the subtle changes of the current abnormal curve with the historical abnormal curve, the higher the similarity, the higher the correlation of the fault, the higher the accuracy of the system diagnosis, the more accurate the maintenance task instructions are issued, the higher the efficiency of maintenance personnel verification, and the higher the overall maintenance operation efficiency.
[0042] It should be noted that the speed curve comparison process includes comparative correlation, which includes self-comparison and mutual comparison. The self-comparison is based on the speed curve monitored by integrity comparison (the speed curve monitored by integrity comparison includes the actual elevator operating speed curve and the elevator traction machine encoder speed curve). If the speed curve loses its integrity, the speed curve is compared with the samples in the sample dataset, and the most similar sample curve is directly called to obtain the associated fault cause. The mutual comparison is based on the speed curve monitored by difference comparison to determine whether there is a speed curve deviation (manifested as a difference in curve linearity) between the actual elevator operating speed curve and the elevator traction machine encoder speed curve, and the fault cause is obtained based on the speed curve deviation.
[0043] It should be noted that the elevator intelligent maintenance management method based on speed curve comparison also includes inferring abnormal elevator states based on elevator operation characteristics such as the number of elevator runs, operating range, and operating direction. These elevator operation characteristics are obtained through the non-associated information module. When the elevator operation characteristics deviate significantly from historical benchmarks, a system warning is generated. The elevator operation characteristics, including the number of elevator runs, operating range, and operating direction, are also monitored using speed curves.
[0044] Specifically, taking two parallel elevators or a grouped elevator arranged side-by-side as an example, by monitoring the elevator speed, the average number of operating cycles of one elevator per day can be determined. If, on a certain day or during a certain period, the number of cycles of one elevator increases while the number of cycles of the other elevator decreases significantly, this change in operating characteristics does not necessarily indicate a fault, but it should alert maintenance personnel. For some latent elevator faults (encoder faults, communication faults, or door operator faults), the characteristics include not responding to external calls and the elevator not moving, leading users to choose the adjacent elevator, resulting in a change in the operating frequency of the two elevators. This design increases the breadth of the system's absolute fault detection and expands the dimensions of elevator health status monitoring. Example 2:
[0045] An intelligent elevator maintenance and management system based on speed curve comparison includes: A multimodal data acquisition module is used to acquire interactive information, the actual operating speed curve of the elevator, and the speed curve of the elevator traction machine encoder. The speed monitoring module is used to monitor the comparison data between the actual operating speed curve of the elevator and the speed curve of the elevator traction machine encoder. An information processing module is used for anomaly detection and the generation of a first anomaly dataset and a second interactive dataset. A data fusion and task engine module, which is used to generate sample datasets and maintenance task instructions; A sample library module, which stores training samples and historical diagnostic cases; The intelligent diagnosis and learning engine module trains and optimizes the model based on the sample information of the sample library module. The maintenance interaction platform module is used to send task instructions and receive on-site feedback.
[0046] It should be noted that the speed monitoring module includes a first speed acquisition unit and a second speed acquisition unit; the first speed acquisition unit is connected to the encoder signal of the elevator speed governor and is used to monitor the speed governor speed in real time and generate the actual operating speed curve of the elevator; the second speed acquisition unit is connected to the rotary encoder signal of the elevator traction machine and is used to monitor the pulse signal of the rotary encoder of the traction machine in real time and generate the encoder speed curve of the elevator traction machine.
[0047] It should be noted that it also includes a non-associated information module, which is used to collect and analyze information on the elevator's operating characteristics.
[0048] In the description of the invention, it should be understood that although the invention has been described with respect to a limited number of embodiments, those skilled in the art should understand from the above description that other embodiments may be conceived within the scope of the invention described herein.
Claims
1. An intelligent elevator maintenance management method based on speed curve comparison, characterized in that, Includes the following steps: Data acquisition includes real-time collection of elevator actual operating speed curves, elevator traction machine encoder speed curves, and maintenance interaction information. Speed curve monitoring involves continuously comparing the two speed curves collected in real time and monitoring the comparison data. If the actual operating speed curve of the elevator is inconsistent with the speed curve of the elevator traction machine encoder, an abnormal event is determined to have occurred. The two curves for that time period are marked and stored to generate the first abnormal dataset. Information processing includes sending maintenance task instructions based on the first abnormal dataset, automatically recording fault characteristic information based on the collected maintenance interaction information, and generating a second interaction dataset. Determine whether there is a temporal correlation between the first abnormal dataset and the second interactive dataset. If there is a temporal correlation, merge the first abnormal dataset and the second interactive dataset to generate a sample dataset. Receive instructions and process feedback, conduct on-site verification and processing feedback based on the maintenance task instructions, and determine the root cause of the fault; The root cause of the failure is bound to the sample dataset to form a machine learning training sample with causal orientation and then stored. Model updates and database maintenance utilize historically accumulated training samples to continuously train and optimize the fault diagnosis model, enabling intelligent prediction and diagnosis of fault causes for subsequent new abnormal events.
2. The elevator intelligent maintenance management method based on speed curve comparison according to claim 1, characterized in that: The steps for obtaining the actual operating speed curve of the elevator are to monitor the speed of the elevator speed limiter and generate the speed curve. The steps for obtaining the speed curve of the elevator traction machine encoder are as follows: monitoring the angular velocity of the elevator traction machine through a rotary encoder and generating a speed curve.
3. The elevator intelligent maintenance management method based on speed curve comparison according to claim 1, characterized in that: The comparison between the actual operating speed curve of the elevator and the speed curve of the elevator traction machine encoder includes a comparison of the integrity of the curves and a comparison of their differences. The integrity comparison involves monitoring whether the speed curve is a complete curve, which includes an acceleration segment, a constant speed segment, and a deceleration segment; if a sudden drop is detected in the curve, it is determined that the curve has lost its integrity. The difference comparison: whether there are graphical differences in the speed curve during the acceleration, constant speed, or deceleration phases, and whether these differences can be captured during the comparison process.
4. The elevator intelligent maintenance management method based on speed curve comparison according to claim 3, characterized in that: It also includes comparative correlation of velocity curves, which includes self-comparison and mutual comparison; The self-comparison is based on the speed curve monitored by integrity comparison. If the speed curve loses its integrity, the speed curve is compared with the samples in the sample dataset, and the most similar sample curve is directly called to obtain the associated fault cause. The comparison is based on the speed curve monitoring of the difference comparison to determine whether there is a speed curve deviation between the actual operating speed curve of the elevator and the speed curve of the elevator traction machine encoder, and to obtain the cause of the fault based on the speed curve deviation.
5. The elevator intelligent maintenance management method based on speed curve comparison according to claim 1, characterized in that: The first abnormal dataset includes information on the abnormal time period, the actual operating speed curve segment of the elevator, the speed curve segment of the elevator traction machine encoder, the location of the elevator car, and the direction of elevator car operation; the second interactive dataset includes information on the fault time, fault location, and fault phenomenon.
6. The elevator intelligent maintenance management method based on speed curve comparison according to claim 1, characterized in that: The maintenance task instructions are sent to the mobile terminal of the maintenance personnel; the maintenance task includes on-site fault diagnosis, fault troubleshooting, and inspection of elevator system status and fault codes.
7. The elevator intelligent maintenance management method based on speed curve comparison according to claim 1, characterized in that: It also includes inferring abnormal elevator conditions based on elevator operation characteristics such as the number of elevator runs, operating range, and operating direction; when the elevator operation characteristics deviate significantly from historical benchmarks, a system warning is generated.
8. An intelligent elevator maintenance and management system based on speed curve comparison, characterized in that, include: A multimodal data acquisition module is used to acquire interactive information, the actual operating speed curve of the elevator, and the speed curve of the elevator traction machine encoder. The speed monitoring module is used to monitor the comparison data between the actual operating speed curve of the elevator and the speed curve of the elevator traction machine encoder. An information processing module is used for anomaly detection and the generation of a first anomaly dataset and a second interactive dataset. A data fusion and task engine module, which is used to generate sample datasets and maintenance task instructions; A sample library module, which stores training samples and historical diagnostic cases; The intelligent diagnosis and learning engine module trains and optimizes the model based on the sample information of the sample library module. The maintenance interaction platform module is used to send task instructions and receive on-site feedback.
9. The elevator intelligent maintenance management system based on speed curve comparison according to claim 8, characterized in that: The speed monitoring module includes a first speed acquisition unit and a second speed acquisition unit; The first speed acquisition unit is connected to the speed governor encoder signal of the elevator and is used to monitor the speed governor speed in real time and generate the actual operating speed curve of the elevator. The second speed acquisition unit is connected to the rotary encoder signal of the elevator traction machine, and is used to monitor the rotary encoder of the traction machine in real time and generate the speed curve of the elevator traction machine encoder.
10. The elevator intelligent maintenance management system based on speed curve comparison according to claim 8, characterized in that: It also includes a non-associated information module, which is used to collect and analyze information on the elevator's operating characteristics.