Escalator or moving sidewalk fault prediction and analysis system based on deep learning

Abstract

The invention discloses an escalator or moving sidewalk fault prediction and analysis system based on deep learning. The system comprises an Internet of Things sensor network, a data acquisition module, a deep learning model module, a fault analysis module and a user interaction interface and early warning module. The deep learning model module performs fault mode identification on the operation state of the escalator or the moving sidewalk by using a deep learning model based on the collected multi-dimensional data, and predicts potential faults and development trends thereof; the fault analysis module is used for carrying out deep analysis on fault data when the deep learning model module predicts a potential fault, and providing a detailed fault analysis report and maintenance suggestion for maintenance personnel; according to the invention, the deep learning algorithm is applied to the fault prediction of the escalator and the moving sidewalk, the deep mining and intelligent analysis of the equipment operation data are realized through the construction of the neural network model, and the accuracy of fault prediction is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of escalator and moving walk fault monitoring, and more specifically, to a fault prediction and analysis system for escalators or moving walks based on deep learning. Background Art

[0002] Traditional escalator and moving walk fault detection mainly relies on manual inspections and regular maintenance. This method is not only time-consuming and laborious but also difficult to accurately predict faults before they occur. With the rapid development of big data and artificial intelligence technologies, especially the advantages of deep learning algorithms in data processing and pattern recognition, fault prediction for escalators and moving walks has become possible. Summary of the Invention

[0003] An object of the present invention is to overcome the deficiencies of the above-mentioned prior art and provide a fault prediction and analysis system for escalators or moving walks based on deep learning. This system collects equipment operation data through a data acquisition module, uses a deep learning model to intelligently analyze the data, predicts potential faults, and provides a detailed fault analysis report, improving the reliability and safety of the equipment and reducing maintenance costs.

[0004] To achieve the above object, the present invention adopts the following technical solutions:

[0005] A fault prediction and analysis system for escalators or moving walks based on deep learning, comprising:

[0006] An Internet of Things sensor network, including a plurality of sensors deployed on various key components of the escalator or moving walk, for real-time monitoring of the operating states of the various key components;

[0007] A data acquisition module, which acquires multi-dimensional data of the escalator or moving walk through the Internet of Things sensor network and transmits the multi-dimensional data to a cloud server or an edge computing device;

[0008] A deep learning model module, which, based on the collected multi-dimensional data, uses a deep learning model to perform fault mode recognition on the operating state of the escalator or moving walk and predicts potential faults and their development trends;

[0009] A fault analysis module, which, when the deep learning model module predicts a potential fault, deeply analyzes the fault data and provides a detailed fault analysis report and maintenance suggestions for maintenance personnel;

[0010] A user interface and warning module, which is used to display the operating state of the escalator or moving walk, the fault prediction result, and the fault analysis report, and automatically sends a warning message to relevant personnel when a potential fault is predicted.

[0011] Further, the deep learning model module includes:

[0012] Model selection: Select an appropriate deep learning model according to the requirements of fault prediction.

[0013] Model training: Use historical fault data and extracted features to train the deep learning model so that it can learn the mapping relationship between faults and features.

[0014] Model prediction: Input the collected multi-dimensional data into the deep learning model for fault prediction, and the deep learning model outputs the probability of fault occurrence and the fault type.

[0015] Model optimization: Continuously train and optimize the deep learning model using newly collected multi-dimensional data.

[0016] Further, the deep learning model collects deep learning algorithms, including but not limited to convolutional neural networks, recurrent neural networks, and long short-term memory networks.

[0017] Further, the deep learning model module uses transfer learning technology to fine-tune the pre-trained deep learning model.

[0018] Further, the multi-dimensional data collected by the data acquisition module needs to be pre-processed, including data cleaning, feature extraction, and normalization.

[0019] Further, the data acquisition module also includes an image recognition component and a sound recognition component, which are used to capture image and sound information in the area of the escalator or moving walkway to assist in identifying potential faults.

[0020] Further, the fault analysis module includes:

[0021] Fault type identification: Identify the fault type according to the prediction result of the deep learning model and in combination with the characteristics of key components.

[0022] Fault cause analysis: Based on the relevant characteristics when a fault occurs, obtain the contribution degree of the fault cause, and locate the cause of the fault by analyzing the contribution degree of the fault cause.

[0023] Fault impact assessment: According to the predicted fault type and severity, evaluate the impact of the fault on the operation of the escalator or moving walkway and passenger safety, and provide corresponding fault analysis reports and maintenance suggestions.

[0024] Further, the fault analysis module also includes:

[0025] Key component life prediction, based on data before and after the maintenance of escalators or moving walks, uses significance analysis technology to extract features with large changes before and after maintenance and fits a degradation curve, and uses the degradation curve to predict the remaining life of key components and adjust the maintenance cycle.

[0026] Furthermore, it also includes a remote monitoring and management platform module. Through network connection, it allows managers to remotely access and control the system, view the real-time operating status, fault prediction results and analysis reports, and conduct fault diagnosis, maintenance arrangement and system upgrade.

[0027] The beneficial effects of the present invention are as follows:

[0028] 1. The present invention applies deep learning algorithms to the fault prediction of escalators and moving walks. By constructing a neural network model, it realizes the in-depth mining and intelligent analysis of equipment operation data, and significantly improves the accuracy of fault prediction.

[0029] 2. The present invention not only collects traditional physical parameters such as vibration and temperature, but also integrates multi-source data such as image recognition and sound recognition. Through deep learning algorithms, it conducts fusion analysis on multi-source data, so as to more comprehensively capture the equipment operating status and improve the accuracy and reliability of fault prediction.

[0030] 3. The present invention can real-time monitor equipment operation data, use deep learning models for rapid analysis, immediately trigger an early warning mechanism once abnormalities are found, and at the same time generate a detailed fault analysis report to provide timely and accurate guidance for maintenance personnel.

[0031] 4. In the present invention, the deep learning model has the ability of self-learning and self-adaptation, and can continuously optimize the prediction algorithm with the addition of new data to improve the prediction accuracy; at the same time, the system also has a self-diagnosis function, which can identify and correct errors in the model to ensure long-term stable operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a schematic structural diagram of a fault prediction and analysis system for escalators or moving walks based on deep learning in this embodiment;

[0033] Figure 2 is a line chart showing the development trend of the vibration characteristics of the escalator motor in this embodiment;

[0034] Figure 3 is a line chart comparing the historical data of the motor vibration characteristics and operating conditions data in this embodiment;

[0035] Figure 4 is a mapping relationship diagram between the motor characteristics and the operating data in this embodiment.

[0036] Reference numerals: Internet of Things sensor network 1, data acquisition module 2, deep learning model module 3, fault analysis module 4, user interaction interface and warning module 5, remote monitoring and management platform module 6. Detailed implementation manners

[0037] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0038] Embodiment: An escalator or moving walkway fault prediction and analysis system based on deep learning, as Figure 1 shown, includes an Internet of Things sensor network 1, a data acquisition module 2, a deep learning model module 3, a fault analysis module 4, and a user interaction interface and warning module 5; wherein, the Internet of Things sensor network 1 includes a plurality of sensors deployed on various key components of the escalator or moving walkway, such as vibration sensors, temperature sensors, current sensors, voltage sensors, and the sensors are communicatively connected to the Internet of Things for real-time monitoring of the operating states of various key components.

[0039] The data acquisition module 2 acquires multi-dimensional data of the escalator or moving walkway through the Internet of Things sensor network 1, and transmits the multi-dimensional data to a cloud server or edge computing device wirelessly or wiredly to ensure the real-time and integrity of the data; the multi-dimensional data includes but is not limited to vibration data, temperature data, current data, voltage data, etc.

[0040] Furthermore, the data acquisition module 2 further includes an image recognition component and a sound recognition component for capturing image and sound information in the area of the escalator or moving walkway to assist in identifying potential faults.

[0041] Furthermore, the multi-dimensional data acquired by the data acquisition module 2 needs to be preprocessed, including data cleaning, feature extraction, and normalization processing. Specifically, the acquired raw data is cleaned, denoised, and other processing operations are performed to improve the data quality and analysis efficiency, and then features related to faults, such as vibration frequency, amplitude, etc., are extracted from the processed data.

[0042] The deep learning model module 3, based on the collected multi-dimensional data, uses a deep learning model to identify fault modes of the operating state of the escalator or moving walkway, and predicts potential faults and their development trends, as Figure 2As shown, it is a line chart of the development trend of the vibration characteristics of the escalator motor. The deep learning model uses deep learning algorithms such as convolutional neural network CNN, recurrent neural network RNN, and long short-term memory network LSTM. The deep learning model module 3 uses transfer learning technology to fine-tune the pre-trained deep learning model to improve the efficiency and accuracy of model training.

[0043] Furthermore, the deep learning model module 3 includes model selection, model training, model prediction, and model optimization; among them, model selection includes selecting a suitable deep learning module according to the requirements of fault prediction, such as convolutional neural network CNN, recurrent neural network RNN, and long short-term memory network LSTM; model training includes using historical fault data and extracted features (such as Figure 3 As shown, it is a line chart comparing the historical data of the motor vibration characteristics and operating conditions data) to train the deep learning model so that it can learn the mapping relationship between faults and features (such as Figure 4 As shown, it is a mapping relationship diagram between the motor characteristics and operating data); model prediction includes inputting the collected multi-dimensional data into the deep learning model for fault prediction, and the deep learning model outputs the probability of the occurrence of the fault and the fault type; model optimization includes continuously training and optimizing the deep learning model using newly collected multi-dimensional data to improve the accuracy of fault prediction.

[0044] The deep learning module has the ability of self-learning and self-adaptation, and can continuously optimize the prediction algorithm with the addition of new data to improve the prediction accuracy; at the same time, it also has a self-diagnosis function, which can identify and correct the faults in the model to ensure long-term stable operation.

[0045] When the deep learning model module 3 predicts a potential fault, the fault analysis module 4 is activated to deeply analyze the fault data, including fault type identification, fault cause analysis, and fault impact assessment, and provide a detailed fault analysis report and maintenance suggestions for maintenance personnel. Specifically:

[0046] Fault type identification is: according to the prediction result of the deep learning model, combined with the characteristics of key components, identify the fault type;

[0047] Fault cause analysis is: based on the characteristics related to the occurrence of the fault, obtain the fault cause contribution degree, and locate the cause of the fault by analyzing the fault cause contribution degree, so as to provide guidance for fault repair, that is, by analyzing, find the feature value or operating condition parameter with the strongest correlation when the fault occurs, so as to obtain the fault cause contribution degree of the specific components of the escalator, locate the cause of the fault, accumulate the escalator fault cause library, and avoid the next fault of the escalator from the source.

[0048] The fault impact assessment is as follows: Based on the predicted fault type and severity, the impact of the fault on the operation of the escalator or moving walkway and passenger safety is evaluated, and the corresponding risk level is provided. At the same time, the fault impact assessment also has the function of generating maintenance suggestions. It can automatically generate a fault analysis report and maintenance suggestions according to the fault type and possible causes to guide the maintenance personnel to handle the fault.

[0049] The fault analysis report includes the fault type, severity, possible causes, and recommended maintenance measures.

[0050] Furthermore, the fault analysis module 4 also includes the remaining life prediction of key components. The remaining life prediction of key components is based on the data before and after the maintenance of the escalator or moving walkway. The significance analysis technology is used to extract the features with large changes before and after maintenance and fit the degradation curve. The degradation curve is used to predict the remaining life of key components and adjust the maintenance cycle.

[0051] The user interface and warning module 5 is designed in the form of a touch screen or a mobile application, providing multiple language options, and supporting the management personnel to input commands through gestures, voice, and text. The user interface and warning module 5 is used to display the operation status of the escalator or moving walkway, the fault prediction results, and the fault analysis report, allowing the management personnel to view and make decisions in real time. At the same time, a warning system is set up. When a potential fault is predicted, a warning message is automatically sent to the relevant personnel to ensure that the fault can be handled in a timely manner.

[0052] The above-mentioned fault prediction and analysis system also includes a remote monitoring and management platform module 6. Through network connection, it allows the management personnel to remotely access and control the system, view the real-time operation status, fault prediction results, and analysis reports, and conduct fault diagnosis, maintenance arrangement, and system upgrade.

[0053] The following is an illustration with the following case:

[0054] A large shopping mall has multiple escalators. In the past, the traditional maintenance method was adopted, that is, the maintenance personnel conducted manual inspections at fixed intervals (once a month). However, there are many problems with this method. For example, during a shopping mall promotion event, the number of people flow increased significantly, and the operation duration and frequency of the escalators far exceeded normal. However, since it was not yet time for the inspection, the maintenance personnel failed to detect in time that the temperature of the drive motor of one of the escalators had risen abnormally. Eventually, the motor failed, and the escalator had to stop operating, causing great inconvenience to the normal operation of the shopping mall and the passage of customers.

[0055] After this system is installed in the shopping mall, corresponding sensors are installed on key components such as the drive motor, step chain, and handrail belt of each escalator. Among them, a temperature sensor and a vibration sensor are installed on the drive motor, a vibration sensor and a pressure sensor are installed on the step chain, and a displacement sensor is installed on the handrail belt, etc. These sensors collect the device operation parameters in real time according to the preset sampling frequency (such as once every 5 minutes), and transmit the data to the data processing center through the Internet of Things communication module (using 4G communication method).

[0056] When the temperature sensor of the drive motor of a certain escalator detects that the motor temperature rises rapidly within a short period of time and exceeds the set warning threshold (such as 85°C), the system immediately issues a warning prompt. At this time, there is still a certain time margin before the motor actually fails, enabling the maintenance personnel to receive the notice in time and take corresponding measures.

[0057] After receiving the warning information, the data processing center transmits the relevant data to the cloud server. The deep learning model module 3 uses the trained deep learning model to analyze the current data. By comparing the historical failure data and the current motor operation data patterns, it accurately determines that the motor has an overheating failure, and the severity of the failure is moderate. At the same time, based on the long-term operation data of this escalator in the device operation status database (such as daily operation duration, start-stop times, motor cumulative operation time, etc.), the failure analysis module 4 predicts that if the motor is not maintained in time within the next week, it may cause the motor to be completely damaged and require replacement of motor components. It also predicts that the best maintenance time node is the morning of the second day after receiving the warning, when the traffic flow in the shopping mall is relatively small, which is conducive to carrying out maintenance work.

[0058] Based on the above failure diagnosis and maintenance prediction results, the failure analysis module 4 automatically generates a detailed maintenance task list. The task list specifies that the maintenance item is to cool down the drive motor, check the wear of the internal components of the motor and replace them as appropriate. The responsible person is Li Mou, a senior maintenance personnel designated by the shopping mall. The estimated completion time is from 9:00 to 11:00 in the morning of the next day. The required maintenance tools include professional temperature detection equipment, motor disassembly tools, etc.

[0059] Through the mobile application of the cloud server, the maintenance task is accurately pushed to Li Mou. After receiving the task notice, Li Mou views the task details through the mobile application, including the device location, failure situation, maintenance requirements, etc. During the maintenance process, Li Mou uploads the work progress in real time, such as the completed motor cooling treatment and new problems found such as slight wear of the motor bearing. The system evaluates the work quality of Li Mou in real time according to the information uploaded by Li Mou, confirms that his maintenance operation is standardized, and completes all maintenance items within the specified time, enabling the escalator to resume normal operation, effectively avoiding further damage to the motor and greater impact on the operation of the shopping mall.

[0060] The above are only the preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the concept of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art, several improvements and refinements made without departing from the principle of the present invention should also be regarded as within the protection scope of the present invention.

Claims

1. An escalator or moving walkway fault prediction and analysis system based on deep learning, characterized in that Including: The Internet of Things sensor network (1), including multiple sensors deployed on each key component of the escalator or moving walkway, for real-time monitoring of the operating status of each key component; The data acquisition module (2), which acquires multi-dimensional data of the escalator or moving walkway through the Internet of Things sensor network (1) and transmits the multi-dimensional data to the cloud server or edge computing device; The deep learning model module (3), which, based on the collected multi-dimensional data, uses the deep learning model to identify the fault mode of the operating status of the escalator or moving walkway, and predicts potential faults and their development trends; The fault analysis module (4), which, when the deep learning model module (3) predicts a potential fault, deeply analyzes the fault data and provides a detailed fault analysis report and maintenance suggestions for maintenance personnel; The user interaction interface and warning module (5), which is used to display the operating status of the escalator or moving walkway, the fault prediction result, and the fault analysis report, and automatically sends a warning message to relevant personnel when a potential fault is predicted.

2. The automatic escalator or moving walkway fault prediction and analysis system based on deep learning according to claim 1, wherein The deep learning model module (3) includes: Model selection, selecting the corresponding deep learning model according to the requirements of fault prediction; Model training, using historical fault data and extracted features to train the deep learning model so that it can learn the mapping relationship between faults and features; Model prediction, inputting the collected multi-dimensional data into the deep learning model for fault prediction, and the deep learning model outputs the probability of the occurrence of the fault and the fault type; Model optimization, continuously training and optimizing the deep learning model using newly acquired multi-dimensional data.

3. The automatic escalator or moving walkway fault prediction and analysis system based on deep learning according to claim 1, characterized in that, The deep learning model collects deep learning algorithms, including but not limited to convolutional neural networks, recurrent neural networks, and long short-term memory networks.

4. The automatic escalator or moving walkway fault prediction and analysis system based on deep learning according to claim 1, characterized in that, The deep learning model module (3) adopts transfer learning technology to fine-tune the pre-trained deep learning model.

5. The automatic escalator or moving walkway fault prediction and analysis system based on deep learning according to claim 1, characterized in that, The multi-dimensional data collected by the data acquisition module (2) needs to be pre-processed, including data cleaning, feature extraction, and normalization processing.

6. The automatic escalator or moving walkway fault prediction and analysis system based on deep learning according to claim 1, wherein, The data acquisition module (2) also includes an image recognition component and a sound recognition component, which are used to capture the image and sound information in the area of the escalator or moving walkway to assist in identifying potential faults.

7. The automatic escalator or moving walkway fault prediction and analysis system based on deep learning according to claim 1, characterized in that The fault analysis module (4) includes: Fault type identification, identifying the fault type according to the result predicted by the deep learning model and combining the characteristics of the key components; Fault cause analysis, obtaining the fault cause contribution degree based on the characteristics related to the occurrence of the fault, and locating the cause of the fault by analyzing the fault cause contribution degree; Fault impact assessment, evaluating the impact of the fault on the operation of the escalator or moving walkway and passenger safety according to the predicted fault type and severity, and providing corresponding fault analysis reports and maintenance suggestions.

8. The automatic escalator or moving walkway fault prediction and analysis system based on deep learning according to claim 7, characterized in that, The fault analysis module (4) also includes: Remaining life prediction of key components, based on the data before and after the maintenance of the escalator or moving walkway, using significance analysis technology to extract the features with large changes before and after maintenance and fit the degradation curve, and using the degradation curve to predict the remaining life of the key components and adjust the maintenance cycle.

9. The automatic escalator or moving walkway fault prediction and analysis system based on deep learning according to claim 1, characterized in that It also includes a remote monitoring and management platform module (6), which, through network connection, allows managers to remotely access and control the system, view the real-time operating status, fault prediction results, and analysis reports, and perform fault diagnosis, maintenance scheduling, and system upgrades.

Images