Turnout switch machine fault diagnosis and prediction model method based on artificial intelligence

Abstract

The invention discloses a turnout switch machine fault diagnosis and prediction model method based on artificial intelligence. The method comprises the following steps: step 1, data acquisition; step 2, data preprocessing; step 3, feature extraction; step 4, model training; 5, verifying the model; 6, deploying the model; step 7, fault diagnosis and prediction; an artificial intelligence algorithm is introduced to automatically analyze data and identify faults, safe operation of equipment is ensured, instant fault response is realized through real-time monitoring and rapid data processing, the equipment safety is improved, an automatic modeling tool is adopted to simplify the construction process, non-professional personnel operation is facilitated, system deployment is accelerated, and the system reliability is improved. The method combines expert experience and data driving to form a hybrid model, improves the adaptability and accuracy of the model, integrates multi-sensor data to form a unified monitoring platform, enhances the comprehensiveness and accuracy of fault diagnosis, and provides powerful guarantee for the safety and efficiency of railway transportation.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a turnout machine fault diagnosis and prediction model method based on artificial intelligence. Background Art

[0002] Turnout machines are crucial equipment in the railway transportation system, responsible for controlling the running direction of trains at the turnouts. With the continuous increase in railway transportation demand, the complexity and scale of the railway network are also expanding. The reliability and safety of turnout machines directly affect the operating efficiency and safety of trains. Therefore, ensuring the normal operation of turnout machines is an important task in railway operation management.

[0003] In the prior art, the fault diagnosis methods for turnout machines mainly include the following:

[0004] 1. The diagnosis method based on artificial experience has a large workload, is greatly affected by subjectivity, and cannot respond quickly to faults; 2. The diagnosis method based on rules relies on expert experience and rule bases, and diagnoses by setting a series of fault judgment rules. The disadvantage of this method is that the formulation and updating of rules requires a lot of professional knowledge, and it is difficult to adapt to complex and changeable fault conditions; 3. The diagnosis method based on models establishes a mathematical model of the turnout machine, compares the output of the model with the actual data, and judges the fault. This method improves the accuracy of diagnosis to a certain extent, but the establishment and maintenance of the model requires high professional skills, and the sensitivity to the model assumptions makes it limited in practical application; 4. The diagnosis method based on data-driven analysis can improve the accuracy and efficiency of fault diagnosis to a certain extent by analyzing historical data, extracting features and performing pattern recognition. However, the existing data-driven methods still face problems such as insufficient data volume, difficulty in feature selection, and poor model generalization ability. The scarcity of data and the complexity of feature extraction easily lead to insufficient accuracy and generalization ability of the model, which cannot effectively deal with changeable fault conditions. Therefore, a fault diagnosis and prediction model method for turnout machines based on artificial intelligence is proposed. Summary of the invention

[0005] The purpose of the present invention is to provide a turnout machine fault diagnosis and prediction model method based on artificial intelligence to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for fault diagnosis and prediction model of turnout machine based on artificial intelligence, comprising the following steps: step one, data collection; step two, data preprocessing; step three, feature extraction; step four, model training; step five, model verification; step six, model deployment; step seven, fault diagnosis and prediction;

[0007] In the above step 1, a data acquisition module is installed on the turnout machine to collect sensor data of the turnout machine in normal operation and fault state, the collected data is annotated, a multi-category fault data set is constructed, and the multi-category fault data set is transmitted to the data processing center through the data transmission module;

[0008] In the above step 2, the data preprocessing module preprocesses the multi-category fault data set obtained in step 1;

[0009] In the above step 3, the feature extraction module extracts features from the multi-category fault data set preprocessed in step 2 to obtain a feature set;

[0010] In the above step 4, the feature set obtained in step 3 is divided into a training set, a validation set and a test set in a ratio of 7:2:1. The model is iteratively trained using the training set, and the model is evaluated using the validation set to monitor the performance of the model. After the training is completed, a fault diagnosis and prediction model is obtained;

[0011] In the above step 5, the fault diagnosis and prediction model trained in step 4 is tested using the test set to evaluate its accuracy and robustness, and parameters are tuned according to the evaluation results;

[0012] In the above step 6, the fault diagnosis and prediction model after parameter tuning in step 5 is deployed to the artificial intelligence module;

[0013] In the above step seven, the data collected in real time by the data acquisition module is input into the artificial intelligence module after preprocessing and feature extraction. The artificial intelligence module performs fault diagnosis and prediction, outputs fault reports and warning information, and displays them through the interactive interface module.

[0014] Preferably, in step 1, the multi-category fault data set specifically includes mechanical faults, software faults, electrical faults and environmental faults.

[0015] Preferably, in the step one, the data processing center includes a data storage module, a data preprocessing module, a feature extraction module, an artificial intelligence module, a maintenance solution library, a feedback optimization module and an interactive interface module, and the data storage module establishes a data connection with the data preprocessing module, the data preprocessing module establishes a data connection with the feature extraction module, the feature extraction module establishes a data connection with the artificial intelligence module, the artificial intelligence module establishes data connections with the maintenance solution library, the feedback optimization module and the interactive interface module respectively, the data storage module is used to store the acquired data, the data preprocessing module is used to preprocess the data, the feature extraction module is used to extract feature data, the artificial intelligence module is used for fault diagnosis and prediction, the maintenance solution library is used to provide maintenance suggestions, the feedback optimization module is used to collect the difference information between the actual fault situation and the model prediction results, and optimize the model, and the interactive interface module is used for human-computer interaction to view the equipment status, fault prediction information and historical data analysis results.

[0016] Preferably, in step one, the data acquisition module includes a temperature sensor, a vibration sensor, a pressure sensor, a current sensor and a voltage sensor.

[0017] Preferably, in step 2, the preprocessing specifically includes data cleaning, denoising, normalization and standardization.

[0018] Preferably, in step three, the feature extraction module uses one or more of Fourier transform, wavelet transform, and convolutional neural network to extract features, and uses principal component analysis to select features.

[0019] Preferably, in step four, the model includes a fault diagnosis model and a fault prediction model, the fault diagnosis model is a combination of one or more of a random forest, a support vector machine and a deep neural network, and the fault prediction model uses a long short-term memory network.

[0020] Preferably, in step 4, the model uses binary cross entropy as the loss function to judge the fault state and the normal state, and uses the Adam optimizer to dynamically adjust the learning rate to accelerate convergence. The binary cross entropy is specifically:

[0021] The probability of the true label y and the model prediction for a sample Binary cross entropy is defined as:

[0022]

[0023] Where y is 0 or 1. The range is [0,1];

[0024] For N samples, the average loss of binary cross entropy is:

[0025]

[0026] Preferably, in step 4, the evaluation indicators specifically include accuracy, precision, recall rate and F1 value.

[0027] Preferably, in step seven, the fault report includes the fault type, fault location and recommended maintenance measures.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention introduces artificial intelligence algorithms to automatically analyze data and identify faults, thereby ensuring safe operation of equipment, achieving instantaneous response to faults through real-time monitoring and rapid data processing, improving equipment safety, using automated modeling tools to simplify the construction process, facilitating operation by non-professionals, and accelerating system deployment, combining expert experience with data-driven to form a hybrid model, improving model adaptability and accuracy, integrating multi-sensor data to form a unified monitoring platform, enhancing the comprehensiveness and accuracy of fault diagnosis, and providing a strong guarantee for the safety and efficiency of railway transportation. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a flow chart of the steps of the present invention;

[0030] Figure 2 is a flow chart of the method of the present invention;

[0031] Figure 3 It is a data preprocessing flow chart of the present invention;

[0032] Figure 4 It is a schematic diagram of feature extraction of the present invention;

[0033] Figure 5 The model construction and training flow chart of the present invention;

[0034] Figure 6 It is a flow chart of fault diagnosis and prediction of the present invention;

[0035] Figure 7 This is a schematic diagram of the fault types and maintenance suggestions of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] Please see attached Figure 1 -Attached Figure 7The present invention provides an embodiment: a fault diagnosis and prediction model method for a turnout machine based on artificial intelligence, comprising the following steps: step one, data collection; step two, data preprocessing; step three, feature extraction; step four, model training; step five, model verification; step six, model deployment; step seven, fault diagnosis and prediction;

[0038] In the above step one, a data acquisition module is installed on the turnout machine to collect sensor data of the turnout machine in normal operation and fault state, the collected data is annotated, a multi-category fault data set is constructed, and the multi-category fault data set is transmitted to the data processing center through the data transmission module; wherein the multi-category fault data set specifically includes mechanical faults, software faults, electrical faults and environmental faults; the data processing center includes a data storage module, a data preprocessing module, a feature extraction module, an artificial intelligence module, a maintenance solution library, a feedback optimization module and an interactive interface module, and the data storage module establishes a data connection with the data preprocessing module, the data preprocessing module establishes a data connection with the feature extraction module, and the feature extraction module The module establishes data connection with the artificial intelligence module, and the artificial intelligence module establishes data connection with the maintenance solution library, feedback optimization module and interactive interface module respectively. The data storage module is used to store the acquired data, the data preprocessing module is used to preprocess the data, the feature extraction module is used to extract feature data, the artificial intelligence module is used for fault diagnosis and prediction, the maintenance solution library is used to provide maintenance suggestions, the feedback optimization module is used to collect the difference information between the actual fault situation and the model prediction results, and optimize the model. The interactive interface module is used for human-computer interaction to view the equipment status, fault prediction information and historical data analysis results; the data acquisition module includes temperature sensors, vibration sensors, pressure sensors, current sensors and voltage sensors;

[0039] In the above step 2, the data preprocessing module preprocesses the multi-category fault data set obtained in step 1, specifically including data cleaning, denoising, normalization and standardization;

[0040] In the above step 3, the feature extraction module extracts features from the multi-category fault data set preprocessed in step 2 to obtain a feature set; specifically, the feature extraction module uses one or more of Fourier transform, wavelet transform, and convolutional neural network to extract features, and uses principal component analysis to select features;

[0041] In the above step 4, the feature set obtained in step 3 is divided into a training set, a validation set and a test set in a ratio of 7:2:1. The model is iteratively trained using the training set, and the model is evaluated using the validation set. The accuracy, precision, recall and F1 value are used as evaluation indicators to monitor the performance of the model. After the training is completed, a fault diagnosis and prediction model is obtained; wherein, the model includes a fault diagnosis model and a fault prediction model. The fault diagnosis model is a combination of one or more of a random forest, a support vector machine and a deep neural network. The fault prediction model uses a long short-term memory network. The model uses binary cross entropy as a loss function to judge the fault state and the normal state. The Adam optimizer is used to dynamically adjust the learning rate to accelerate convergence. The binary cross entropy is specifically:

[0042] The probability of the true label y and the model prediction for a sample Binary cross entropy is defined as:

[0043]

[0044] Among them, y takes the value of 0 or 1. The range is [0,1];

[0045] For N samples, the average loss of binary cross entropy is:

[0046]

[0047] In the above step 5, the fault diagnosis and prediction model trained in step 4 is tested using the test set to evaluate its accuracy and robustness, and parameters are tuned according to the evaluation results;

[0048] In the above step 6, the fault diagnosis and prediction model after parameter tuning in step 5 is deployed to the artificial intelligence module;

[0049] In the above step seven, the data collected in real time by the data acquisition module is input into the artificial intelligence module after preprocessing and feature extraction. The artificial intelligence module performs fault diagnosis and prediction, outputs fault report and warning information, and displays it through the interactive interface module. The fault report includes the fault type, fault location and recommended maintenance measures.

[0050] Based on the above, the advantages of the present invention are that when the present invention is used, by introducing artificial intelligence algorithms, historical data and real-time data are automatically analyzed, fault modes are identified, the accuracy and reliability of fault diagnosis are improved, and the safe operation of equipment is ensured; through real-time monitoring and rapid data processing technology, analysis and response are performed at the moment of fault occurrence, and alarms are issued in time to improve equipment safety and reduce potential accident risks; by adopting automated modeling tools and algorithms, the model building process is simplified so that non-professionals can also quickly establish and adjust models, lower technical barriers, and accelerate system deployment and application; by combining expert experience with data-driven methods, a hybrid model is formed to overcome the limitations of data-driven methods and improve model adaptability and accuracy; by integrating multiple sensor data to form a unified monitoring platform, the equipment status is comprehensively evaluated, data utilization efficiency is improved, and the comprehensiveness and accuracy of fault diagnosis are enhanced.

[0051] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A fault diagnosis and prediction model method for a turnout machine based on artificial intelligence, comprising the following steps: Step 1: data collection; Step 2: data preprocessing; Step 3: feature extraction; Step 4: model training; Step 5: model verification; Step 6: model deployment; Step 7: fault diagnosis and prediction; It is characterized by: In the above step 1, a data acquisition module is installed on the turnout machine to collect sensor data of the turnout machine in normal operation and fault state, the collected data is annotated, a multi-category fault data set is constructed, and the multi-category fault data set is transmitted to the data processing center through the data transmission module; In the above step 2, the data preprocessing module preprocesses the multi-category fault data set obtained in step 1; In the above step 3, the feature extraction module extracts features from the multi-category fault data set preprocessed in step 2 to obtain a feature set; In the above step 4, the feature set obtained in step 3 is divided into a training set, a validation set and a test set in a ratio of 7:2:

1. The model is iteratively trained using the training set, and the model is evaluated using the validation set to monitor the performance of the model. After the training is completed, a fault diagnosis and prediction model is obtained; In the above step 5, the fault diagnosis and prediction model trained in step 4 is tested using the test set to evaluate its accuracy and robustness, and parameters are tuned according to the evaluation results; In the above step 6, the fault diagnosis and prediction model after parameter tuning in step 5 is deployed to the artificial intelligence module; In the above step seven, the data collected in real time by the data acquisition module is input into the artificial intelligence module after preprocessing and feature extraction. The artificial intelligence module performs fault diagnosis and prediction, outputs fault reports and warning information, and displays them through the interactive interface module.

2. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1 is characterized in that: In the step 1, the multi-category fault data set specifically includes mechanical faults, software faults, electrical faults and environmental faults.

3. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1 is characterized in that: In the step one, the data processing center includes a data storage module, a data preprocessing module, a feature extraction module, an artificial intelligence module, a maintenance solution library, a feedback optimization module and an interactive interface module, and the data storage module establishes a data connection with the data preprocessing module, the data preprocessing module establishes a data connection with the feature extraction module, the feature extraction module establishes a data connection with the artificial intelligence module, and the artificial intelligence module establishes data connections with the maintenance solution library, the feedback optimization module and the interactive interface module respectively. The data storage module is used to store the acquired data, the data preprocessing module is used to preprocess the data, the feature extraction module is used to extract feature data, the artificial intelligence module is used for fault diagnosis and prediction, the maintenance solution library is used to provide maintenance suggestions, the feedback optimization module is used to collect the difference information between the actual fault situation and the model prediction results, and optimize the model, and the interactive interface module is used for human-computer interaction to view the equipment status, fault prediction information and historical data analysis results.

4. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1 is characterized in that: In the step 1, the data acquisition module includes a temperature sensor, a vibration sensor, a pressure sensor, a current sensor and a voltage sensor.

5. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1 is characterized in that: In the step 2, the preprocessing specifically includes data cleaning, denoising, normalization and standardization.

6. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1 is characterized in that: In the step three, the feature extraction module uses one or more of Fourier transform, wavelet transform, and convolutional neural network to extract features, and uses principal component analysis to select features.

7. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1 is characterized in that: In the step 4, the model includes a fault diagnosis model and a fault prediction model. The fault diagnosis model is a combination of one or more of a random forest, a support vector machine, and a deep neural network, and the fault prediction model uses a long short-term memory network.

8. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1 is characterized by: In step 4, the model uses binary cross entropy as the loss function to judge the fault state and the normal state, and uses the Adam optimizer to dynamically adjust the learning rate to accelerate convergence. The binary cross entropy is specifically: The probability of the true label y and the model prediction for a sample Binary cross entropy is defined as: Where y is 0 or 1. The range is [0,1]; For N samples, the average loss of binary cross entropy is:

9. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1, characterized in that: In the step 4, the evaluation indicators specifically include accuracy, precision, recall rate and F1 value.

10. The method for fault diagnosis and prediction model of turnout machine based on artificial intelligence according to claim 1, characterized in that: In step seven, the fault report includes the fault type, fault location and recommended maintenance measures.

Images