Elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion

Abstract

The invention discloses an elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion, and belongs to the technical field of elevator detection. Comprising various modules and is characterized in that a data acquisition module is used for acquiring various operation data of the elevator traction machine; the edge calculation module is used for carrying out primary data preprocessing; the data fusion module is used for fusing traction machine data from different sensors to obtain fused data by using a multi-source heterogeneous data fusion method; the fault diagnosis module is used for processing the fused data by using a deep learning algorithm so as to diagnose the running state of the elevator traction machine to obtain an elevator diagnosis result; the cloud platform is used for storing and managing elevator operation data; and the maintenance suggestion module is used for generating maintenance suggestions and early warning information according to the elevator diagnosis result. According to the elevator traction machine fault diagnosis method and system, multiple data sources are fused, fault diagnosis is conducted through the deep learning algorithm, comprehensive monitoring and accurate fault diagnosis of the elevator traction machine are achieved, and the safety and maintenance efficiency of an elevator are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of elevator detection, and particularly to an elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion. Background Art

[0002] With the continuous growth of the number of elevators, the safe operation and timely maintenance of elevators have become particularly important. Existing elevator traction machine state monitoring methods mostly use a single data source, such as current monitoring, vibration monitoring, etc., but these methods have problems such as incomplete monitoring and low fault diagnosis accuracy. Summary of the Invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art and provide an elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion.

[0004] The purpose of the present invention is achieved through the following technical solutions: An elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion includes a data acquisition module, the data acquisition module is connected to an edge computing module, the edge computing module is connected to a data fusion module, the data fusion module is connected to a fault diagnosis module, the fault diagnosis module is connected to a cloud platform, and the cloud platform is connected to a maintenance suggestion module; The data acquisition module includes a variety of sensors installed in different parts of the elevator traction machine for collecting a variety of operating data of the elevator traction machine; the edge computing module is used for performing primary data preprocessing on the operating data to obtain traction machine data; the data fusion module is used for using a multi-source heterogeneous data fusion method to fuse the traction machine data from different sensors to obtain fusion data; the fault diagnosis module is used for using a deep learning algorithm to process the fusion data, thereby diagnosing the operating state of the elevator traction machine to obtain an elevator diagnosis result; the cloud platform is used for storing and managing elevator operation data; the maintenance suggestion module is used for generating maintenance suggestions and warning information according to the elevator diagnosis result.

[0005] Preferably, the data acquisition module includes a current sensor, a vibration sensor, a temperature sensor, and a noise sensor; the operating data includes current data, vibration data, temperature data, and noise data.

[0006] Preferably, the primary data preprocessing includes data cleaning, feature extraction, and data conversion.

[0007] Preferably, when the data fusion module performs data fusion, it includes the following steps: Extract current features, vibration features, and temperature features from the traction machine data collected by different sensors: , , ; , ; ; wherein, I mean is the average current value, I rms is the root mean square value of the current, I peak is the peak value of the current, I i is the current sampling value, N is the number of sampling points; X rms is the root mean square value of the vibration, is the kurtosis factor, x i is the vibration sampling value, is the average value of the vibration, is the standard deviation of the vibration; T rate is the temperature rise rate, T current is the current temperature, T initial is the initial temperature, t current is the current time, t initial is the initial time; Fuse the extracted feature vectors to obtain a comprehensive feature vector: ; where F is the comprehensive feature vector, N level is the noise level; Use the Dempster - Shafer combination rule to fuse the diagnostic results from different data sources. Set the frame of discernment Θ = {minor fault, serious fault, normal}, and set the basic assignment function for each feature m 1, m 2… m n , and then fuse the basic probability assignment functions according to the Dempster - Shafer combination rule to obtain the fused probability assignment function m (A), which is used to judge the operating state of the elevator traction machine: ; wherein, K is the conflict coefficient, and B and C are subsets of different features.

[0008] Preferably, when the fault diagnosis module performs fault diagnosis, it includes the following steps: Normalize the fused data: ; wherein, x norm is the normalized fusion data, x is the original data, x min and x max are respectively the minimum and maximum values of the original data; Construct a deep learning model and use historical data for training to optimize the model parameters. During the training process, introduce the cross-entropy loss function: ; wherein is the loss value, is the true label, is the predicted probability, C is the number of classes; Input the real-time collected data into the trained deep learning model for fault diagnosis to obtain the fault type and severity.

[0009] Preferably, the deep learning model is a convolutional neural network model or a long short-term memory network model.

[0010] Preferably, the convolutional neural network model includes an input layer, a convolutional layer, a pooling layer, and a fully connected layer; the input layer is used to receive the normalized feature vector; the convolutional layer is used to extract the local correlation of features; the pooling layer is used to reduce the feature dimension and reduce the computational amount; the fully connected layer is used for classification.

[0011] Preferably, the long short-term memory network model includes an input layer, an LSTM layer, and a fully connected layer; the input layer is used to receive time series data; the LSTM layer is used to capture the long-term dependence of the time series; the fully connected layer is used for classification.

[0012] The beneficial effects of the present invention are: 1) By fusing multiple data sources and using deep learning algorithms for fault diagnosis, comprehensive monitoring and accurate fault diagnosis of the elevator traction machine are realized, improving the safety and maintenance efficiency of the elevator. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 is the principle block diagram of the elevator traction machine status monitoring and diagnosis platform based on multi-source heterogeneous data fusion; Figure 2 is the data processing flow chart; Figure 3 is the deep learning model structure diagram. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0014] Next, in combination with the embodiments, the technical solution of the present invention will be clearly and completely described. 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 skilled in the art without creative efforts belong to the scope of protection of the present invention.

[0015] Referring to Figures 1 - 3 , the present invention provides a technical solution: an elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion, including a data acquisition module, the data acquisition module is connected to an edge computing module, the edge computing module is connected to a data fusion module, the data fusion module is connected to a fault diagnosis module, the fault diagnosis module is connected to a cloud platform, and the cloud platform is connected to a maintenance suggestion module; The data acquisition module includes a variety of sensors installed in different parts of the elevator traction machine, and is used to collect a variety of operating data of the elevator traction machine; the edge computing module is used to perform primary data preprocessing on the operating data to obtain traction machine data; the data fusion module is used to use a multi-source heterogeneous data fusion method to fuse the traction machine data from different sensors to obtain fusion data; the fault diagnosis module is used to process the fusion data using a deep learning algorithm, so as to diagnose the operating state of the elevator traction machine to obtain an elevator diagnosis result; the cloud platform is used to store and manage elevator operation data; the maintenance suggestion module is used to generate maintenance suggestions and warning information according to the elevator diagnosis result.

[0016] In this embodiment, the data acquisition module: is used to collect a variety of operating data of the elevator traction machine, including but not limited to current data, vibration data, temperature data, noise data, etc. The data acquisition module includes a current sensor, a vibration sensor, a temperature sensor, a noise sensor, etc., and these sensors are respectively installed in different parts of the elevator traction machine to obtain comprehensive operating state information.

[0017] The edge computing module: performs preprocessing and preliminary analysis on the collected data. The edge computing module includes functions such as data cleaning, feature extraction, and data conversion, and can convert the original data into a format suitable for further analysis.

[0018] The data fusion module: adopts multi-source heterogeneous data fusion technology to fuse data from different sensors. The data fusion module uses methods such as Dempster-Shafer evidence theory to comprehensively process the information from different data sources, improving the reliability of the data and the accuracy of diagnosis.

[0019] Fault diagnosis module: Based on the fused data, machine learning and deep learning algorithms are used to diagnose the operating status of the elevator traction machine. The fault diagnosis module includes functions such as fault feature extraction, fault mode recognition, and fault type judgment, and can accurately identify the fault type and severity of the elevator traction machine.

[0020] Cloud platform: Used to store and manage a large amount of elevator operation data, and provide data query and analysis services. The cloud platform also supports remote monitoring and maintenance, can receive data from the edge computing module in real time, and feedback the diagnosis results to the maintenance personnel.

[0021] Maintenance suggestion module: According to the diagnosis results, generate corresponding maintenance suggestions and warning information. The maintenance suggestion module includes functions such as maintenance plan formulation, maintenance priority evaluation, and maintenance effect evaluation, to help maintenance personnel take effective maintenance measures in a timely manner.

[0022] In some embodiments, the data acquisition module includes a current sensor, a vibration sensor, a temperature sensor, and a noise sensor; the operation data includes current data, vibration data, temperature data, and noise data.

[0023] In some embodiments, the initial data preprocessing includes data cleaning, feature extraction, and data conversion.

[0024] In some embodiments, when the data fusion module performs data fusion, it includes the following steps: Extract current features, vibration features, and temperature features from the traction machine data collected by different sensors: , , ; , ; ; Among them, I mean is the current mean value, I rms is the root mean square value of the current, I peak is the peak value of the current, I i is the current sampling value, N is the number of sampling points; X rms is the root mean square value of the vibration, is the kurtosis factor, x i is the vibration sampling value, is the vibration mean value, is the vibration standard deviation; Trate is the temperature rise rate, T current is the current temperature, T initial is the initial temperature, t current is the current time, t initial is the initial time; Fuse the extracted feature vectors to obtain a comprehensive feature vector: ; where F is the comprehensive feature vector, N level is the noise level; Use the Dempster - Shafer combination rule to fuse the diagnostic results from different data sources. Set the frame of discernment Θ = {minor fault, severe fault, normal}, and set the basic assignment function for each feature m 1, m 2… m n , and then fuse the basic probability assignment functions according to the Dempster - Shafer combination rule to obtain the fused probability assignment function m (A), which is used to judge the operating state of the elevator traction machine: ; where, K is the conflict coefficient, and B and C are subsets of different features.

[0025] In this embodiment, the main function of the data fusion module is to fuse the data from different sensors to improve the reliability of the data and the accuracy of diagnosis. Extract features from the data collected by different sensors, such as the mean value, peak value, and effective value of current, the root mean square value and kurtosis factor of vibration, the temperature rise rate of temperature, etc. Fuse the extracted feature vectors to form a comprehensive feature vector. For example, combine the current feature, vibration feature, temperature feature, and noise feature into a comprehensive feature vector.

[0026] In some embodiments, when the fault diagnosis module performs fault diagnosis, it includes the following steps: Normalize the fused data: ; where, x norm is the normalized fused data, x is the original data, x min and x max are the minimum and maximum values of the original data respectively; Build a deep learning model, train it using historical data, optimize the model parameters, and introduce the cross-entropy loss function during the training process: ; where is the loss value, is the true label, is the predicted probability, C is the number of classes; Input the real-time collected data into the trained deep learning model for fault diagnosis to obtain the fault type and severity.

[0027] In this embodiment, the normalization process is to ensure that the data is suitable for the input of the deep learning model. Training the deep learning model using historical data can optimize the model parameters and improve the diagnostic accuracy of the model.

[0028] In some embodiments, the deep learning model is a convolutional neural network model or a long short-term memory network model.

[0029] In some embodiments, the convolutional neural network model includes an input layer, a convolutional layer, a pooling layer, and a fully connected layer; the input layer is used to receive the normalized feature vector; the convolutional layer is used to extract the local correlation of features; the pooling layer is used to reduce the feature dimension and computational amount; the fully connected layer is used for classification.

[0030] In some embodiments, the long short-term memory network model includes an input layer, an LSTM layer, and a fully connected layer; the input layer is used to receive time series data; the LSTM layer is used to capture the long-term dependencies of the time series; the fully connected layer is used for classification.

[0031] The following gives an application example: Case background: A commercial building has 20 elevators, with a daily passenger flow of up to tens of thousands of people. Due to the high usage frequency of the elevators, the traditional regular maintenance method is difficult to meet the requirements of real-time monitoring and fault prediction. The building management decides to introduce the elevator traction machine status monitoring and diagnosis platform based on multi-source heterogeneous data fusion of the present invention.

[0032] Implementation process: Equipment deployment: Install edge protocol gateways on each elevator in the building. The gateways collect real-time operation data of the elevators through sensors, including speed, load, vibration, temperature, etc. Data processing: The edge protocol gateways are built-in with machine learning algorithms that can perform real-time analysis on the collected data and identify abnormal patterns. For example, when the vibration frequency of the elevator exceeds the normal range, the gateway will immediately issue a warning. Fault prediction: Through the analysis of historical data, the edge protocol gateways can predict the possible types and times of elevator faults. For example, when it is detected that the load of the elevator continues to be high, the gateway will predict that the elevator motor may fail in the next few days and notify the maintenance staff in advance. Maintenance response: According to the prediction information provided by the gateway, the maintenance staff carry out maintenance and replacement in advance, avoiding the inconvenience and safety risks brought by sudden elevator shutdowns.

[0033] Implementation effects: Through the platform of the present invention, the failure rate of the elevators in this building has been significantly reduced and the maintenance efficiency has been improved. The specific effects are as follows: Reduction in failure rate: Since the introduction of the edge protocol gateways, the failure rate of the building's elevators has been reduced by 30%, and the elevator shutdown time has been reduced by 50%. Decrease in maintenance cost: Through predictive maintenance, the maintenance staff can detect and handle potential faults in advance, avoiding large-scale repairs and replacements, and the maintenance cost has decreased by 20%. Improvement in user satisfaction: The elevator operation is more stable, the user complaint rate has decreased significantly, and the overall operation efficiency of the building has been improved.

[0034] Data analysis: Calculation of reduction in failure rate: ; The original failure rate was an average of 3 faults per elevator per month, and the optimized failure rate is an average of 2.1 faults per elevator per month. Substituting into the formula for calculation: ; Calculation of reduction in maintenance time: ; The original maintenance time was an average of 2 hours per maintenance, and the optimized maintenance time is an average of 1.2 hours per maintenance. Substituting into the formula for calculation: ; Calculation of reduction in maintenance cost: Reduction in maintenance cost = (Original maintenance cost - Optimized maintenance cost) / Original maintenance cost × 100%. The original maintenance cost was an average of 2000 yuan per elevator per year, and the optimized maintenance cost is an average of 1600 yuan per elevator per year. Substituting into the formula for calculation: .

[0035] The above are only the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be changed within the scope of the concept described herein through the above teachings or the techniques or knowledge in related fields. And any changes and modifications made by those skilled in the art without departing from the spirit and scope of the present invention shall fall within the protection scope of the appended claims of the present invention.

Claims

1. An elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion, characterized in that: It includes a data acquisition module, the data acquisition module is connected to an edge computing module, the edge computing module is connected to a data fusion module, the data fusion module is connected to a fault diagnosis module, the fault diagnosis module is connected to a cloud platform, and the cloud platform is connected to a maintenance suggestion module; The data acquisition module includes a variety of sensors installed in different parts of the elevator traction machine, which are used to collect a variety of operating data of the elevator traction machine; the edge computing module is used to perform primary data preprocessing on the operating data to obtain traction machine data; the data fusion module is used to use a multi-source heterogeneous data fusion method to fuse the traction machine data from different sensors to obtain fusion data; The fault diagnosis module is used to process the fusion data using a deep learning algorithm, so as to diagnose the operating state of the elevator traction machine to obtain an elevator diagnosis result; the cloud platform is used to store and manage elevator operating data; the maintenance suggestion module is used to generate maintenance suggestions and warning information according to the elevator diagnosis result.

2. The elevator traction machine status monitoring and diagnosis platform based on multi-source heterogeneous data fusion according to claim 1, characterized in that: The data acquisition module includes a current sensor, a vibration sensor, a temperature sensor and a noise sensor; the operating data includes current data, vibration data, temperature data and noise data.

3. The elevator traction machine status monitoring and diagnosis platform based on multi-source heterogeneous data fusion according to claim 1, characterized in that: The primary data preprocessing includes data cleaning, feature extraction and data transformation.

4. The elevator traction machine status monitoring and diagnosis platform based on multi-source heterogeneous data fusion according to claim 1, characterized in that: When the data fusion module performs data fusion, it includes the following steps: Extract current features, vibration features and temperature features from the traction machine data collected by different sensors: ， ， ； ， ； ； Among them, I mean is the average current value, I rms is the root mean square value of the current, I peak is the peak value of the current, I i is the current sampling value, N is the number of sampling points; X rms is the root mean square value of the vibration, is the kurtosis factor, x i is the vibration sampling value, is the average vibration value, is the vibration standard deviation; T rate is the temperature rise rate, T current is the current temperature, T initial is the initial temperature, t current is the current time, t initial is the initial time; Fuse the extracted feature vectors to obtain a comprehensive feature vector: ; where F is the comprehensive feature vector, N level is the noise level; The diagnostic results from different data sources are fused using the Dempster-Shafer combination rule. The frame of discernment Θ = {minor fault, serious fault, normal} is set, and the basic assignment function is set for each feature. m 1. m 2… m n , and then the basic probability assignment functions are fused according to the Dempster-Shafer combination rule to obtain the fused probability assignment function m (A), which is used to judge the operating state of the elevator traction machine: ; wherein, K is the conflict coefficient, and B and C are subsets of different features.

5. The elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion according to claim 1, characterized in that: When the fault diagnosis module performs fault diagnosis, it includes the following steps: Perform normalization processing on the fusion data: ; wherein, x norm is the normalized fusion data, x is the original data, x min and x max are the minimum and maximum values of the original data respectively; Construct a deep learning model, and use historical data for training to optimize the model parameters. The cross-entropy loss function is introduced during the training process: ; where is the loss value, is the true label, is the predicted probability, C is the number of classes; Input the real-time collected data into the trained deep learning model for fault diagnosis to obtain the fault type and severity.

6. The elevator traction machine state monitoring and diagnosis platform based on multi-source heterogeneous data fusion according to claim 5, characterized in that: The deep learning model is a convolutional neural network model or a long short-term memory network model.

7. The elevator traction machine status monitoring and diagnosis platform based on multi-source heterogeneous data fusion according to claim 6, characterized in that: The convolutional neural network model includes an input layer, a convolutional layer, a pooling layer and a fully connected layer; the input layer is used to receive the normalized feature vector; the convolutional layer is used to extract the local correlation of features; the pooling layer is used to reduce the feature dimension and reduce the amount of calculation; the fully connected layer is used for classification.

8. The elevator traction machine status monitoring and diagnosis platform based on multi-source heterogeneous data fusion according to claim 6, characterized in that: The long short-term memory network model includes an input layer, an LSTM layer and a fully connected layer; the input layer is used to receive time series data; the LSTM layer is used to capture the long-term dependence of time series; the fully connected layer is used for classification.

Images