Elevator door system fault diagnosis method and system based on multi-source signals

Abstract

The invention discloses an elevator door system fault diagnosis method based on a multi-source signal, and the method comprises the steps: S1, installing a vibration signal sensor and a sound signal sensor at an elevator door system, and collecting a vibration signal and a sound signal; s2, speed signal extraction is conducted on the collected vibration signals and sound signals, the threshold value range of the speed signals generated by the elevator under the normal working condition is set, then data obtained after data within the threshold value range is removed according to the extracted speed signals is processed, and corresponding fault types are analyzed; the processing result and the corresponding fault type form a training data set, and the training data set is imported into a CNN for model training to obtain an elevator door system fault diagnosis model; and S3, using the model obtained by training to diagnose the type of a mechanical fault occurring in the elevator door system. According to the method, multi-source signal cross processing is adopted, so that the accuracy of signal processing is ensured, and meanwhile, faults can be specifically analyzed. Meanwhile, the system is composed of a complete system, and the correct rate of fault diagnosis is guaranteed.

Description

Technical Field

[0001] The invention belongs to the technical field of fault diagnosis of elevator door systems, and relates to a method and system for fault diagnosis of elevator door systems based on multi-source signals. Background Art

[0002] With more and more elevators being installed now, more than half of the elevator accidents are caused by the failure of the elevator door system. Therefore, being able to diagnose the fault of the elevator door system in advance is crucial to life safety. In the prior art, only a single sensor is used to identify the fault of the elevator door system. However, when facing a specific fault, the single signal is used for fault diagnosis. It can only generally point out the general direction and cannot point out the problem in detail. At the same time, because a single signal is used for processing, the accuracy of fault diagnosis cannot be guaranteed. Summary of the invention

[0003] Purpose of the invention: The purpose of the present invention is to provide a method and system for elevator door system fault diagnosis based on multi-source signals, so as to solve the problems that the prior art uses single signal processing, cannot ensure the accuracy of fault diagnosis, and requires a large amount of data calculation, which requires a large amount of calculation and has low efficiency.

[0004] Technical solution: The method for diagnosing elevator door system faults based on multi-source signals of the present invention comprises:

[0005] S1. Install a vibration signal sensor and a sound signal sensor in the elevator door system to collect vibration signals and sound signals;

[0006] S2. Extract speed signals from the vibration signals and sound signals collected above, set the threshold range of the speed signals generated by the elevator under normal working conditions, and then remove the data within the threshold range according to the extracted speed signals, process the remaining data and analyze the corresponding fault types, and form a training data set with the processing results and the corresponding fault types, which are imported into CNN for model training to obtain an elevator door system fault diagnosis model;

[0007] S3. Use the elevator door system fault diagnosis model of step S2 to diagnose the type of mechanical fault occurring in the elevator door system.

[0008] Furthermore, the vibration signal sensor of S1 includes a first vibration signal sensor and a second vibration signal sensor. The first vibration signal sensor is installed at the door leaf of the car door to collect the vibration signal generated by the door leaf during the movement of the car door; the second vibration signal sensor is installed at the car door motor to collect the vibration signal at the motor; the sound sensor signal is installed at the top guide rail of the car door to collect the sound signal.

[0009] Furthermore, different elevator devices have different threshold ranges set.

[0010] Furthermore, the step S2 specifically includes the following steps:

[0011] S21, extracting the speed signal from the vibration signal and the sound signal by combining ridge extraction with time-frequency analysis technology, and after determining that the speed signal exceeds the threshold range, selecting the decomposition layer number for the data exceeding the range to perform Morlet wavelet packet decomposition processing;

[0012] S22, after decomposition, wavelet packet coefficients of different frequency sub-bands are obtained, wherein the wavelet packet coefficients contain information of the signal at different frequencies and time scales;

[0013] S23, calculating the spectral kurtosis value of each sub-band wavelet packet coefficient, and selecting the sub-band whose spectral kurtosis value exceeds the set value as the sensitive frequency band by analyzing the distribution of the spectral kurtosis;

[0014] S24, taking the sensitive frequency band as data, importing it into the input layer of CNN and then performing model training through the convolution layer and the pooling layer, and finally obtaining the training result as the system for elevator door system fault diagnosis.

[0015] Furthermore, the function of the Morlet wavelet packet decomposition is:

[0016] where ω 0 is the center frequency, and the number of decomposition layers of the decomposition processing is N=4.

[0017] Furthermore, the method for determining the fault type of the training data set in step S2 includes the following contents:

[0018] When the amplitude of the speed signal extracted from the vibration signal collected by the first vibration signal sensor reaches more than 3 times the average value under normal working conditions, the variance of the speed signal extracted from the vibration signal collected by the second vibration signal sensor exceeds the threshold range, and the sound signal collected by the sound signal sensor has an instantaneous peak, it is determined that the cause of the fault is that the door knife collides with the door wheel, causing the locking arm to disengage;

[0019] When the amplitude of the speed signal extracted from the vibration signal collected by the first vibration signal sensor reaches more than 3 times the average value under normal working conditions, the variance of the speed signal extracted from the vibration signal collected by the second vibration signal sensor exceeds the threshold range, and the sound signal collected by the sound signal sensor does not show an instantaneous peak, it is determined that the cause of the fault is that there is a foreign object blocking the door guide rail;

[0020] When the first vibration signal sensor collects a continuous vibration signal, the root mean square of the sound signal collected by the sound signal sensor reaches 1.5 to 2 times that of the normal working condition, and the duration exceeds 4.9 seconds, it is determined that the cause of the fault is deformation of the door rail or door leaf;

[0021] When the effective value of the speed signal extracted from the vibration signal collected by the first vibration signal sensor is ≤1.8mm / s and the time in this state exceeds 4.9s without intermittent interruption, and the root mean square of the sound signal collected by the sound signal sensor reaches 3 times that of the normal working condition and lasts for more than 4.9s, it is judged that the cause of the fault is severe wear of the door pulley;

[0022] When the variance of the speed signal extracted from the vibration signal collected by the first vibration signal sensor is ≤0.2 (mm / s)2, and there is no significant abnormal frequency component in the spectrum, but the amplitude is lower than the normal value under normal working conditions, or the first vibration signal sensor does not collect vibration signals, the amplitude of the speed signal extracted from the vibration signal collected by the second vibration signal sensor reaches more than 3 times the average value under normal working conditions, and the LAeq of the sound signal collected by the sound signal sensor is ≥75dB and the duration is more than 5s, it is judged that the cause of the fault is the slippage of the door machine belt;

[0023] When the first vibration signal sensor does not collect a vibration signal, the second vibration signal sensor does not collect a vibration signal, and the sound signal sensor does not collect a sound signal, it is determined that the cause of the fault is that the door machine motor is damaged.

[0024] The present invention provides an elevator door system fault diagnosis system based on multi-source signals, comprising:

[0025] A signal acquisition module, the signal acquisition module is used to collect vibration signals and sound signals, including a first vibration signal sensor, a second vibration signal sensor, and a sound signal sensor; the first vibration signal sensor is installed at the car door leaf, and is used to collect vibration signals generated by the door leaf during the movement of the car door; the second vibration signal sensor is installed at the car door motor, and is used to collect vibration signals at the motor; the sound sensor signal is installed at the top guide rail of the car door, and is used to collect sound signals;

[0026] A data processing module, used for processing the collected signals;

[0027] Model training module, used to train the CNN model;

[0028] The fault diagnosis module is used to input the collected signals into the elevator door system fault diagnosis model and obtain the fault type of the elevator door system.

[0029] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: 1. The present invention uses multi-source signal cross processing to ensure the accuracy of signal processing and specific analysis of faults. At the same time, the present invention has a complete system composition to ensure the accuracy of fault diagnosis;

[0030] 2. The present invention obtains an elevator door system fault diagnosis system by training the CNN model. In the process of diagnosing elevator door faults, only simple feature extraction processing is required after data collection. The fault type diagnosis result can be quickly obtained by introducing the model, thereby improving the analysis and judgment efficiency and reducing the computing requirements, thereby achieving the purpose of cost saving. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the installation position of the sensor in the system of the present invention.

[0032] Figure 2 Flow chart of the diagnostic method of the present invention. DETAILED DESCRIPTION

[0033] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0034] like Figure 1 As shown, the method and system for diagnosing elevator door system faults based on multi-source signals of this embodiment use multi-source signals to monitor the elevator door system. When a fault occurs in the elevator door system, the fault is diagnosed to identify the source of the fault and the type of the fault. A first vibration signal sensor is installed at the door leaf of the car door to collect the vibration signal generated by the door leaf during the movement of the car door; a second vibration signal sensor is installed at the motor of the car door to collect the vibration signal of the motor; the sound sensor signal is installed at the top guide rail of the car door to collect the sound signal.

[0035] like Figure 2 FIG. 1 is a flow chart of a method for diagnosing a fault of an elevator door system based on multi-source signals of the present invention, comprising the following steps:

[0036] S1. Install a vibration signal sensor and a sound signal sensor in the elevator door system to collect vibration signals and sound signals;

[0037] S2. Extract speed signals from the vibration signals and sound signals collected above, set the threshold range of the speed signals generated by the elevator under normal working conditions, and then remove the data within the threshold range according to the extracted speed signals, process the remaining data and analyze the corresponding fault types, and use the processing results and the corresponding fault types to form a training data set, which is imported into CNN for model training to obtain an elevator door system fault diagnosis model;

[0038] S3. Use the elevator door system fault diagnosis model of step S2 to diagnose the type of mechanical fault occurring in the elevator door system.

[0039] The signal processing method for step S2 comprises the following steps:

[0040] S21, extracting the speed signal from the vibration signal and the sound signal by combining ridge extraction with time-frequency analysis technology, and after determining that the speed signal exceeds or is less than a threshold range, selecting the number of decomposition layers for the data exceeding the range to perform Morlet wavelet packet decomposition processing;

[0041] S22, after decomposition, wavelet packet coefficients of different frequency sub-bands are obtained, and these coefficients contain information of the signal at different frequencies and time scales;

[0042] S23. Calculate the spectral kurtosis value of each sub-band wavelet packet coefficient, and select the sub-band whose spectral kurtosis value exceeds the set value as the sensitive frequency band by analyzing the distribution of the spectral kurtosis. The calculation formula of the spectral kurtosis value is: θ=μ+3σ, where θ refers to the threshold, μ refers to the average value of the sub-band spectral kurtosis, and σ refers to the standard deviation of the sub-band spectral kurtosis. For different working environments, θ=k×θ, (k=0.8~1.2). In this embodiment, the set value is 4, that is, the sub-band whose spectral kurtosis value exceeds 4 is selected as the sensitive frequency band. It should be noted that the set value is not a fixed value, but is determined by the threshold according to different equipment and environments;

[0043] S24, taking the sensitive frequency band as data, importing it into the input layer of CNN and then performing model training through the convolution layer and the pooling layer, and finally obtaining the training result as the system for elevator door system fault diagnosis.

[0044] Wherein, the Morlet wavelet packet basis function of step S21 is:

[0045] where ω 0 is the center frequency, and the number of decomposition layers of the decomposition processing is N=4.

[0046] The method for determining the fault type of the training data set in step S2 includes the following contents:

[0047] When the amplitude of the speed signal extracted from the vibration signal collected by the first vibration signal sensor reaches more than 3 times the average value under normal working conditions, the variance of the speed signal extracted from the vibration signal collected by the second vibration signal sensor exceeds the threshold range, and the sound signal collected by the sound signal sensor has an instantaneous peak, it is determined that the cause of the fault is that the door knife collides with the door wheel, causing the locking arm to disengage;

[0048] When the amplitude of the speed signal extracted from the vibration signal collected by the first vibration signal sensor reaches more than 3 times the average value under normal working conditions, the variance of the speed signal extracted from the vibration signal collected by the second vibration signal sensor exceeds the threshold range, and the sound signal collected by the sound signal sensor does not show an instantaneous peak, it is determined that the cause of the fault is that there is a foreign object blocking the door guide rail;

[0049] When the first vibration signal sensor collects a continuous vibration signal, the root mean square of the sound signal collected by the sound signal sensor reaches 1.5 to 2 times that of the normal working condition, and the duration exceeds 4.9 seconds, it is determined that the cause of the fault is deformation of the door rail or door leaf;

[0050] When the effective value of the speed signal extracted from the vibration signal collected by the first vibration signal sensor is ≤1.8mm / s and the time in this state exceeds 4.9s without intermittent interruption, and the root mean square of the sound signal collected by the sound signal sensor reaches 3 times that of the normal working condition and lasts for more than 4.9s, it is judged that the cause of the fault is severe wear of the door pulley;

[0051] When the variance of the speed signal extracted from the vibration signal collected by the first vibration signal sensor is ≤0.2 (mm / s)2, and there is no significant abnormal frequency component in the spectrum, but the amplitude is lower than the normal value under normal working conditions, or the first vibration signal sensor does not collect vibration signals, the amplitude of the speed signal extracted from the vibration signal collected by the second vibration signal sensor reaches more than 3 times the average value under normal working conditions, and the LAeq of the sound signal collected by the sound signal sensor is ≥75dB and the duration is more than 5s, it is judged that the cause of the fault is the slippage of the door machine belt;

[0052] When the first vibration signal sensor does not collect a vibration signal, the second vibration signal sensor does not collect a vibration signal, and the sound signal sensor does not collect a sound signal, it is determined that the cause of the fault is that the door machine motor is damaged.

[0053] The present invention provides an elevator door system fault diagnosis system based on multi-source signals, comprising:

[0054] A signal acquisition module, the signal acquisition module is used to collect vibration signals and sound signals, including a first vibration signal sensor, a second vibration signal sensor, and a sound signal sensor; the first vibration signal sensor is installed at the car door leaf, and is used to collect vibration signals generated by the door leaf during the movement of the car door; the second vibration signal sensor is installed at the car door motor, and is used to collect vibration signals at the motor; the sound sensor signal is installed at the top guide rail of the car door, and is used to collect sound signals;

[0055] A data processing module, used for processing the collected signals;

[0056] Model training module, used to train the CNN model;

[0057] The fault diagnosis module is used to input the collected signals into the elevator door system fault diagnosis model and obtain the fault type of the elevator door system.

[0058] The present invention uses multi-source signal cross processing to ensure the accuracy of signal processing and can analyze the fault in detail. At the same time, the present invention has a complete system composition to ensure the accuracy of fault diagnosis.

[0059] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for fault diagnosis of elevator door system based on multi-source signals, characterized in that: The method comprises: S1. Install a vibration signal sensor and a sound signal sensor in the elevator door system to collect vibration signals and sound signals; S2. Extract speed signals from the vibration signals and sound signals collected above, set the threshold range of the speed signals generated by the elevator under normal working conditions, and then remove the data within the threshold range according to the extracted speed signals, process the remaining data and analyze the corresponding fault types, and form a training data set with the processing results and the corresponding fault types, which are imported into CNN for model training to obtain an elevator door system fault diagnosis model; S3. Use the elevator door system fault diagnosis model of step S2 to diagnose the type of mechanical fault occurring in the elevator door system.

2. The method for elevator door system fault diagnosis based on multi-source signals according to claim 1, characterized in that: The vibration signal sensor of S1 includes a first vibration signal sensor and a second vibration signal sensor. The first vibration signal sensor is installed at the door leaf of the car door to collect the vibration signal generated by the door leaf during the movement of the car door; the second vibration signal sensor is installed at the car door motor to collect the vibration signal at the motor; the sound sensor signal is installed at the top guide rail of the car door to collect the sound signal.

3. The method for elevator door system fault diagnosis based on multi-source signals according to claim 1, characterized in that: Different elevator devices have different threshold ranges set.

4. The method for elevator door system fault diagnosis based on multi-source signals according to claim 1, characterized in that: The step S2 specifically includes the following steps: S21, extracting the speed signal from the vibration signal and the sound signal by combining ridge extraction with time-frequency analysis technology, and after determining that the speed signal exceeds the threshold range, selecting the decomposition layer number for the data exceeding the range to perform Morlet wavelet packet decomposition processing; S22, after decomposition, wavelet packet coefficients of different frequency sub-bands are obtained, wherein the wavelet packet coefficients contain information of the signal at different frequencies and time scales; S23, calculating the spectral kurtosis value of each sub-band wavelet packet coefficient, and selecting the sub-band whose spectral kurtosis value exceeds the set value as the sensitive frequency band by analyzing the distribution of the spectral kurtosis; S24, taking the sensitive frequency band as data, importing it into the input layer of CNN and then performing model training through the convolution layer and the pooling layer, and finally obtaining the training result as the system for elevator door system fault diagnosis.

5. The method for elevator door system fault diagnosis based on multi-source signals according to claim 4, characterized in that: The function of the Morlet wavelet packet decomposition is: Wherein ω0 is the center frequency, and the number of decomposition layers of the decomposition process is N=4.

6. The method for elevator door system fault diagnosis based on multi-source signals according to claim 1, characterized in that: The method for determining the fault type of the training data set in step S2 includes the following contents: When the amplitude of the speed signal extracted from the vibration signal collected by the first vibration signal sensor reaches more than 3 times the average value under normal working conditions, the variance of the speed signal extracted from the vibration signal collected by the second vibration signal sensor exceeds the threshold range, and the sound signal collected by the sound signal sensor has an instantaneous peak, it is determined that the cause of the fault is that the door knife collides with the door wheel, causing the locking arm to disengage; When the amplitude of the speed signal extracted from the vibration signal collected by the first vibration signal sensor reaches more than 3 times the average value under normal working conditions, the variance of the speed signal extracted from the vibration signal collected by the second vibration signal sensor exceeds the threshold range, and the sound signal collected by the sound signal sensor does not show an instantaneous peak, it is determined that the cause of the fault is that there is a foreign object blocking the door guide rail; When the first vibration signal sensor collects a continuous vibration signal, the root mean square of the sound signal collected by the sound signal sensor reaches 1.5 to 2 times that of the normal working condition, and the duration exceeds 4.9 seconds, it is determined that the cause of the fault is deformation of the door rail or door leaf; When the effective value of the speed signal extracted from the vibration signal collected by the first vibration signal sensor is ≤1.8mm / s and the time in this state exceeds 4.9s without intermittent interruption, and the root mean square of the sound signal collected by the sound signal sensor reaches 3 times that of the normal working condition and lasts for more than 4.9s, it is judged that the cause of the fault is severe wear of the door pulley; When the variance of the speed signal extracted from the vibration signal collected by the first vibration signal sensor is ≤0.2 (mm / s)2, and there is no significant abnormal frequency component in the spectrum, but the amplitude is lower than the normal value under normal working conditions, or the first vibration signal sensor does not collect vibration signals, the amplitude of the speed signal extracted from the vibration signal collected by the second vibration signal sensor reaches more than 3 times the average value under normal working conditions, and the LAeq of the sound signal collected by the sound signal sensor is ≥75dB and the duration is more than 5s, it is judged that the cause of the fault is the slippage of the door machine belt; When the first vibration signal sensor does not collect a vibration signal, the second vibration signal sensor does not collect a vibration signal, and the sound signal sensor does not collect a sound signal, it is determined that the cause of the fault is that the door machine motor is damaged.

7. An elevator door system fault diagnosis system based on multi-source signals, characterized in that: include: A signal acquisition module, the signal acquisition module is used to collect vibration signals and sound signals, including a first vibration signal sensor, a second vibration signal sensor, and a sound signal sensor; the first vibration signal sensor is installed at the car door leaf, and is used to collect vibration signals generated by the door leaf during the movement of the car door; the second vibration signal sensor is installed at the car door motor, and is used to collect vibration signals at the motor; the sound sensor signal is installed at the top guide rail of the car door, and is used to collect sound signals; A data processing module, used for processing the collected signals; Model training module, used to train the CNN model; The fault diagnosis module is used to input the collected signals into the elevator door system fault diagnosis model and obtain the fault type of the elevator door system.

Images