Escalator fault prediction system and method based on vibration algorithm

Abstract

The invention discloses an escalator fault prediction system and method based on a vibration algorithm. The escalator fault prediction system comprises a data acquisition module, a cloud server and a monitoring center, the data acquisition module comprises a motor vibration sensor, a speed reducer vibration sensor, a main machine base vibration sensor, an oil temperature sensor, a main drive vibration sensor and a handrail belt temperature sensor, wherein the motor vibration sensor, the speed reducer vibration sensor, the main machine base vibration sensor and the oil temperature sensor are installed on an escalator main machine; the main drive vibration sensor is installed on a chain wheel on the upper portion of an escalator; the step chain extension position sensor and the step chain wheel vibration sensor are arranged on a chain wheel at the lower part of the escalator; the sensors are respectively connected with a signal input end of a data processing module of the cloud server through the Ethernet, a signal output end of the data processing module is connected with an intelligent diagnosis module, and the intelligent diagnosis module of the cloud server is connected with a visual display module of a monitoring center through 485 communication. According to the invention, comprehensive monitoring and accurate diagnosis of the key transmission part of the escalator can be realized.

Description

Technical Field

[0001] The present invention relates to the field of escalator safety, and particularly to an escalator fault prediction system and method based on a vibration algorithm. Background Art

[0002] With the rapid development of urban rail transit, escalators, as important public transportation facilities, the reliability of their operation is directly related to the safety and travel experience of passengers. However, the traditional manual inspection method is inefficient and it is difficult to detect potential faults in a timely manner, resulting in an increased risk of sudden equipment failures. In recent years, equipment health management technologies based on sensor technology and big data analysis have gradually emerged. However, existing systems often rely only on a single data source for analysis, resulting in high false positive rates and insufficient diagnostic accuracy. To solve the above problems, the present invention proposes an escalator fault prediction system and method based on a vibration algorithm. This system integrates various sensor data such as vibration, temperature, and current, and uses advanced signal processing and machine learning algorithms to achieve comprehensive monitoring and accurate diagnosis of key drive components of escalators. Summary of the Invention

[0003] The purpose of the present invention is to provide an escalator fault prediction system and method based on a vibration algorithm, aiming to detect potential faults in advance and optimize maintenance strategies through real-time monitoring and data analysis of key components of escalators.

[0004] The technical solution of the present invention is as follows:

[0005] An escalator fault prediction system based on a vibration algorithm, comprising a data acquisition module, a cloud server, and a monitoring center;

[0006] The data acquisition module includes:

[0007] An electric motor vibration sensor, a speed reducer vibration sensor, a main machine base vibration sensor, and an oil temperature sensor installed on the escalator main machine;

[0008] A main drive vibration sensor installed on the upper sprocket of the escalator;

[0009] An armrest belt temperature sensor installed on the escalator armrest belt;

[0010] A step chain elongation position sensor and a step sprocket vibration sensor installed on the lower sprocket of the escalator;

[0011] The cloud server includes a data processing module and an intelligent diagnosis module. The motor vibration sensor, the reducer vibration sensor, the main machine base vibration sensor, the oil temperature sensor, the main drive vibration sensor, the handrail belt temperature sensor, the step chain elongation position sensor, and the step sprocket vibration sensor are respectively connected to the signal input end of the data processing module through Ethernet. The signal output end of the data processing module is connected to the intelligent diagnosis module. The intelligent diagnosis module is connected to the visualization display module of the monitoring center through 485 communication.

[0012] Furthermore, the data processing module includes:

[0013] Data cleaning to remove noise and outliers;

[0014] Data fusion to fuse multi-source data of vibration, temperature, and current to generate a comprehensive feature vector.

[0015] Furthermore, the intelligent diagnosis module includes:

[0016] Fault model establishment to build a model library of various typical faults based on historical data and expert experience;

[0017] Trend analysis to identify the change trend of the equipment operation status by comparing the current data with the historical data;

[0018] Self-learning function. When an unmatched special fault occurs, the system automatically updates the knowledge base to achieve continuous optimization;

[0019] Fault diagnosis to classify and predict the equipment status using machine learning algorithms.

[0020] Furthermore, the visualization display module includes:

[0021] Interface design to provide a graphical interface to display the equipment distribution map, the monitoring map of a single piece of equipment, and the content of alarm information;

[0022] Display function to show the distribution of station equipment, the operation status of a single piece of equipment, and key parameters;

[0023] Alarm function to push alarm information in real time and support historical alarm query;

[0024] Statistics function to generate an equipment status statistical report, including indicators such as monthly reliability and average daily repair time;

[0025] Mobile support with a supporting mobile phone APP to facilitate users to view the equipment status anytime and anywhere.

[0026] An escalator fault prediction method based on a vibration algorithm includes the following steps:

[0027] S1. The data acquisition module collects relevant data of the escalator main machine, upper sprocket, handrail belt, and lower sprocket;

[0028] S2. The data collected by the data acquisition module is uploaded to the data processing module of the cloud server through Ethernet for data cleaning to remove noise and outliers in the data;

[0029] S3. After data cleaning, the vibration, temperature, and current multi-source data are fused to generate a comprehensive feature vector;

[0030] S4. The comprehensive feature vector is then transmitted to the intelligent diagnosis module to analyze the states of the escalator main machine part, upper sprocket part, handrail belt part, and lower sprocket part;

[0031] S5. Finally, it is output to the visualization display module of the monitoring center through 485 communication to realize the status monitoring of the escalator main machine, upper sprocket, handrail belt, and lower sprocket.

[0032] Furthermore, the data cleaning in step S2 specifically includes the following steps:

[0033] S21. Noise removal

[0034] Use a low-pass filter or band-pass filter to remove high-frequency noise;

[0035] Apply wavelet transform technology to decompose and reconstruct the signal, removing the noise component while retaining the useful signal;

[0036] For the current signal, use the moving average method to smooth the data points with large fluctuations;

[0037] S22. Outlier detection and processing

[0038] Statistical method: By calculating the mean and standard deviation, the data points outside a certain range are regarded as outliers and removed;

[0039] Box plot method: Use quartiles to calculate the upper and lower bounds, and the data points outside the range are regarded as outliers;

[0040] Time series analysis: For continuously collected data, if a certain point is too different from the data before and after, it is marked as an outlier;

[0041] S23. Missing value processing

[0042] Interpolation method: Perform linear interpolation or polynomial interpolation according to the values of adjacent data points to fill in the missing values;

[0043] Mean filling: Replace the missing values with the mean of other normal data in the same time period;

[0044] Model prediction: Establish a prediction model based on historical data to estimate missing values;

[0045] S24. Data standardization and normalization

[0046] Standardization: Convert the data into a standard normal distribution with a mean of 0 and a standard deviation of 1;

[0047]

[0048] Among them, χ is the original data, μ is the mean, and σ is the standard deviation;

[0049] Normalization: Map the data to the interval [0, 1] for subsequent algorithm processing;

[0050]

[0051] Among them, χ min and χ max are the minimum and maximum values of the data respectively;

[0052] S25. Data consistency check

[0053] Check the correlation between data from different sensors;

[0054] Use redundant sensors for cross - verification to ensure data consistency and reliability;

[0055] Feature extraction: Extract characteristic parameters of time - domain waveforms, frequency spectra, envelope spectra, and power spectra from the original data.

[0056] Furthermore, the data fusion in step S3 specifically includes the following steps:

[0057] S31. Data acquisition and standardization

[0058] Collect vibration, temperature, and current data from each sensor and perform standardization processing on each type of data;

[0059] For each data point Xi, the formula for its standardized value Zi is as follows:

[0060]

[0061] Among them, Xi is the original data point, μ is the mean of this type of data, and σ is the standard deviation of this type of data;

[0062] S32. Feature extraction

[0063] Extract features from the standardized data to obtain a vibration feature vector V, a temperature feature vector T, and a current feature vector I;

[0064] S33. Weight assignment

[0065] According to the influence degree of each feature on the device state, a weight is assigned to each feature. The weight of the vibration feature is Wv, the weight of the temperature feature is Wt, and the weight of the current feature is Wi, and Wv + Wt + Wi = 1;

[0066] S34. Generation of comprehensive feature vector

[0067] The features are weighted and summed according to their weights to generate a comprehensive feature vector Z. The calculation formula of the comprehensive feature vector Z is as follows:

[0068] Z = Wv·V + Wt·T + Wi·I.

[0069] Compared with the prior art, the beneficial effects of the present invention are as follows: The escalator fault prediction system based on vibration algorithm provided by the present invention collects multi-source data such as vibration, temperature, and current of key components of the escalator, combines data analysis and fault diagnosis algorithms, realizes the health status monitoring and fault prediction of the transmission components of the escalator. The system has functions of real-time alarm, trend analysis, and self-learning, provides a visual interface, supports cloud deployment and secondary development, and can effectively discover potential faults, optimize the maintenance cycle, and improve the operation reliability of the escalator through the system. Description of the drawings

[0070] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments or the description of the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0071] Figure 1 It is a system block diagram of an escalator fault prediction system based on vibration algorithm provided by the present invention. Detailed implementation manners

[0072] In order to make the purpose, technical solutions and advantages of the present invention clearer, the following further details the present invention with reference to the drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, and are not used to limit the present invention.

[0073] In order to illustrate the technical solutions described in the present invention, the following will be described through specific embodiments.

[0074] Embodiment

[0075] Please refer to Figure 1 , this embodiment provides an escalator fault prediction system based on vibration algorithm, including a data acquisition module, a cloud server, and a monitoring center.

[0076] Wherein:

[0077] The data acquisition module includes:

[0078] An electric motor vibration sensor, a speed reducer vibration sensor, a main machine base vibration sensor, and an oil temperature sensor installed on the main machine of the escalator;

[0079] A main drive vibration sensor installed on the upper sprocket of the escalator;

[0080] An armrest belt temperature sensor installed on the armrest belt of the escalator;

[0081] A step chain elongation position sensor and a step sprocket vibration sensor installed on the lower sprocket of the escalator.

[0082] The cloud server includes a data processing module and an intelligent diagnosis module. The electric motor vibration sensor, the speed reducer vibration sensor, the main machine base vibration sensor, the oil temperature sensor, the main drive vibration sensor, the armrest belt temperature sensor, the step chain elongation position sensor, and the step sprocket vibration sensor are respectively connected to the signal input end of the data processing module through Ethernet. The signal output end of the data processing module is connected to the intelligent diagnosis module, and the intelligent diagnosis module is connected to the visualization display module of the monitoring center through 485 communication.

[0083] Specifically:

[0084] The data processing module includes:

[0085] Data cleaning to remove noise and outliers;

[0086] Data fusion to fuse multi-source data of vibration, temperature, and current to generate a comprehensive feature vector.

[0087] The intelligent diagnosis module includes:

[0088] Fault model establishment to build a model library of various typical faults based on historical data and expert experience;

[0089] Trend analysis to identify the change trend of the equipment operation status by comparing the current data with the historical data;

[0090] Self-learning function. When an unmatched special fault occurs, the system automatically updates the knowledge base to achieve continuous optimization;

[0091] Fault diagnosis to classify and predict the equipment status using machine learning algorithms.

[0092] The visualization display module includes:

[0093] Interface design, providing a graphical interface to display the device distribution map, single-device monitoring map, and alarm information content;

[0094] Display function, showing the distribution of station devices, the operating status of single devices, and key parameters;

[0095] Alarm function, pushing alarm information in real time and supporting historical alarm query;

[0096] Statistics function, generating a device status statistical report, including indicators such as monthly reliability and average daily repair time;

[0097] Mobile support, with a supporting mobile phone APP to facilitate users to view the device status anytime and anywhere.

[0098] An escalator fault prediction method based on the above system specifically includes the following steps:

[0099] S1. The data acquisition module collects relevant data of the escalator main machine, upper sprocket, handrail belt, and lower sprocket;

[0100] S2. The data collected by the data acquisition module is uploaded to the data processing module of the cloud server through Ethernet for data cleaning to remove noise and outliers in the data, ensuring the quality of the data;

[0101] This step S2 specifically includes the following steps:

[0102] S21. Noise removal

[0103] Use a low-pass filter or a band-pass filter to remove high-frequency noise;

[0104] Apply wavelet transform technology to decompose and reconstruct the signal, removing the noise component while retaining the useful signal;

[0105] For current signals, use the moving average method to smooth data points with large fluctuations;

[0106] S22. Outlier detection and processing

[0107] Statistical method: By calculating the mean and standard deviation, data points outside a certain range are regarded as outliers and removed;

[0108] Box plot method: Use quartiles to calculate the upper and lower bounds, and data points outside the range are regarded as outliers;

[0109] Time series analysis: For continuously collected data, if a certain point differs too much from the previous and subsequent data, it is marked as an outlier;

[0110] S23. Missing value processing

[0111] Interpolation method: Perform linear interpolation or polynomial interpolation based on the values of adjacent data points to fill in the missing values;

[0112] Mean filling: Replace the missing values with the mean of other normal data in the same time period;

[0113] Model prediction: Establish a prediction model based on historical data to estimate the missing values;

[0114] S24. Data standardization and normalization

[0115] Standardization: Convert the data into a standard normal distribution with a mean of 0 and a standard deviation of 1;

[0116]

[0117] Among them, χ is the original data, μ is the mean, and σ is the standard deviation;

[0118] Normalization: Map the data to the interval [0, 1] for subsequent algorithm processing;

[0119]

[0120] Among them, χ min and χ max are the minimum and maximum values of the data respectively;

[0121] S25. Data consistency check

[0122] Check the correlation between data from different sensors. For example, when the motor current increases, the vibration amplitude usually also increases. If the change trends of the two are inconsistent, the accuracy of the data needs to be further verified;

[0123] Use redundant sensors for cross - verification to ensure data consistency and reliability;

[0124] Feature extraction: Extract the characteristic parameters of time - domain waveforms, frequency spectra, envelope spectra, and power spectra from the original data;

[0125] S3. After data cleaning, fuse the vibration, temperature, and current multi - source data to generate a comprehensive feature vector;

[0126] This step S3 specifically includes the following steps:

[0127] S31. Data acquisition and standardization

[0128] Collect vibration, temperature, and current data from each sensor and perform standardization processing on each type of data;

[0129] For each data point Xi, the formula for its standardized value Zi is as follows:

[0130]

[0131] Among them, Xi is the original data point, μ is the mean of this type of data, and σ is the standard deviation of this type of data;

[0132] S32. Feature extraction

[0133] Feature extraction is performed from the standardized data to obtain a vibration feature vector V, a temperature feature vector T, and an electric current feature vector I;

[0134] S33. Weight assignment

[0135] According to the influence degree of each feature on the equipment state, a weight is assigned to each feature. The weight of the vibration feature is Wv, the weight of the temperature feature is Wt, and the weight of the electric current feature is Wi, and Wv + Wt + Wi = 1;

[0136] S34. Generation of the comprehensive feature vector

[0137] The features are weighted and summed according to their weights to generate a comprehensive feature vector Z. The calculation formula for the comprehensive feature vector Z is as follows:

[0138] Z = Wv·V + Wt·T + Wi·I;

[0139] Among them:

[0140] V is the vibration feature vector, including F v and E v etc.;

[0141] T is the temperature feature vector, including T avg and T rate etc.;

[0142] I is the temperature feature vector, including I rms and I peak etc.;

[0143] If all features are concatenated into a column vector, then there is:

[0144] Z = [Wv·F v , Wv·E v , Wt·T avg , Wt·T rate , Wi·I rms , Wi·I peak T ;

[0145] S4. The comprehensive feature vector is then transmitted to the intelligent diagnosis module to analyze the states of the main machine part, upper sprocket part, handrail part, and lower sprocket part of the escalator;

[0146] ​S5. Finally, it is output to the visualization display module of the monitoring center through 485 communication to realize the status monitoring of the main machine, upper sprocket, handrail belt and lower sprocket of the escalator.

[0147] The escalator fault prediction system has the following technical characteristics:

[0148] (1) Multi-source data fusion: By integrating multiple data sources such as vibration, temperature, and current, the diagnostic accuracy is improved;

[0149] (2) Self-learning ability: By continuously updating the fault model library, it can adapt to the changes in the equipment operating conditions;

[0150] (3) Real-time and reliability: It supports high-frequency data collection and real-time analysis, which can ensure the timeliness of fault warning;

[0151] (4) Support for cloud deployment: Facilitate large-scale application and remote monitoring;

[0152] (5) Openness and scalability: The system design follows open standards and supports the flexible expansion of new devices and measurement points.

[0153] In summary, the escalator fault prediction system based on vibration algorithm provided in this embodiment has the characteristics of high efficiency, accuracy and reliability, which can effectively reduce the failure rate of the escalator, extend the service life of the equipment, and provide a strong guarantee for the safe operation of urban rail transit.

[0154] The above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An escalator fault prediction system based on a vibration algorithm, characterized in that: It includes a data acquisition module, a cloud server, and a monitoring center; The data acquisition module includes: A motor vibration sensor, a reducer vibration sensor, a main machine base vibration sensor, and an oil temperature sensor installed on the escalator main machine; A main drive vibration sensor installed on the upper sprocket of the escalator; An armrest belt temperature sensor installed on the escalator armrest belt; A step chain elongation position sensor and a step sprocket vibration sensor installed on the lower sprocket of the escalator; The cloud server includes a data processing module and an intelligent diagnosis module. The motor vibration sensor, the reducer vibration sensor, the main machine base vibration sensor, the oil temperature sensor, the main drive vibration sensor, the armrest belt temperature sensor, the step chain elongation position sensor, and the step sprocket vibration sensor are respectively connected to the signal input end of the data processing module through Ethernet. The signal output end of the data processing module is connected to the intelligent diagnosis module. The intelligent diagnosis module and the visualization display module of the monitoring center are connected through 485 communication.

2. The escalator fault prediction system based on a vibration algorithm according to claim 1, wherein, The data processing module includes: Data cleaning to remove noise and outliers; Data fusion to fuse multi-source data of vibration, temperature, and current to generate a comprehensive feature vector.

3. The escalator fault prediction system based on a vibration algorithm according to claim 2, characterized in that The intelligent diagnosis module includes: Fault model establishment to build a model library of various typical faults based on historical data and expert experience; Trend analysis to identify the change trend of the equipment operation status by comparing current data with historical data; Self-learning function. When an unmatched special fault occurs, the system automatically updates the knowledge base to achieve continuous optimization; Fault diagnosis to classify and predict the equipment status using machine learning algorithms.

4. The escalator fault prediction system based on a vibration algorithm according to claim 3, characterized in that, The visualization display module includes: Interface design to provide a graphical interface to display the equipment distribution map, single equipment monitoring map, and alarm information content; Display function to display the station equipment distribution, single equipment operation status, and key parameters; Alarm function to push alarm information in real time and support historical alarm query; Statistics function to generate an equipment status statistical report, including indicators such as monthly reliability and average daily repair time; Mobile support with a supporting mobile phone APP to facilitate users to view the equipment status anytime and anywhere.

5. The escalator fault prediction method of an escalator fault prediction system based on a vibration algorithm according to claim 4, characterized in that, It includes the following steps: S1. The data acquisition module collects relevant data of the escalator main machine, upper sprocket, armrest belt, and lower sprocket; S2. The data collected by the data acquisition module is uploaded to the data processing module of the cloud server through Ethernet for data cleaning to remove noise and outliers in the data; S3. After data cleaning, multi-source data of vibration, temperature, and current are fused to generate a comprehensive feature vector; S4. The comprehensive feature vector is then transmitted to the intelligent diagnosis module to analyze the status of the main machine part, upper sprocket part, armrest belt part, and lower sprocket part of the escalator; S5. Finally, it is output to the visualization display module of the monitoring center through 485 communication to achieve status monitoring of the escalator main machine, upper sprocket, armrest belt, and lower sprocket.

6. The escalator fault prediction method based on a vibration algorithm according to claim 5, wherein The data cleaning in step S2 specifically includes the following steps: S21. Noise removal Use a low-pass filter or a band-pass filter to remove high-frequency noise; The wavelet transform technique is applied to decompose and reconstruct the signal, removing the noise components while retaining the useful signal; For the current signal, the moving average method is used to smooth the data points with large fluctuations; S22. Outlier detection and processing Statistical method: By calculating the mean and standard deviation, the data points outside a certain range are regarded as outliers and removed; Box plot method: Using quartiles to calculate the upper and lower bounds, the data points outside the range are regarded as outliers; Time series analysis: For continuously collected data, if a certain point has a large difference from the previous and subsequent data, it is marked as an outlier; S23. Missing value processing Interpolation method: Linear interpolation or polynomial interpolation is performed according to the values of adjacent data points to fill in the missing values; Mean filling: Replace the missing values with the mean of other normal data in the same time period; Model prediction: Based on historical data, a prediction model is established to estimate the missing values; S24. Data standardization and normalization Standardization: Convert the data into a standard normal distribution with a mean of 0 and a standard deviation of 1; where χ is the original data, μ is the mean, and σ is the standard deviation; Normalization: Map the data to the interval [0, 1] for subsequent algorithm processing; where χ min and χ max are the minimum and maximum values of the data, respectively; S25. Data consistency check Check the correlation between data from different sensors; Use redundant sensors for cross - verification to ensure the consistency and reliability of the data; Feature extraction: Extract the characteristic parameters of the time - domain waveform, frequency spectrum, envelope spectrum, and power spectrum from the original data.

7. A method for predicting escalator faults based on a vibration algorithm according to claim 5, characterized in that, The data fusion in step S3 specifically includes the following steps: S31. Data acquisition and standardization Collect vibration, temperature, and current data from each sensor, and perform standardization processing on each type of data; For each data point Xi, the formula for its standardized value Zi is as follows: where Xi is the original data point, μ is the mean of this type of data, and σ is the standard deviation of this type of data; S32. Feature extraction Perform feature extraction from the standardized data to obtain the vibration feature vector V, temperature feature vector T, and current feature vector I; S33. Weight assignment According to the influence degree of each feature on the equipment state, assign a weight to each feature. The weight of the vibration feature is Wv, the weight of the temperature feature is Wt, and the weight of the current feature is Wi, and Wv + Wt+Wi = 1; S34. Generation of comprehensive feature vector Weighted sum the features according to their weights to generate the comprehensive feature vector Z. The formula for the comprehensive feature vector Z is as follows: Z = Wv·V+Wt·T+Wi·I.

Images