Fault diagnosis method and device, equipment and storage medium

Abstract

The invention discloses a fault diagnosis method and device, equipment and a storage medium, and relates to the technical field of fault diagnosis, and the fault diagnosis method comprises the steps: carrying out the data selection of the module operation data of a drive module according to a fault impact factor corresponding to the drive module on target equipment, and determining the target operation data; and inputting the module operation characteristics corresponding to the target operation data into a fault diagnosis model for fault diagnosis, determining the current fault information of the driving module and the predicted service life of the module, and further obtaining a fault diagnosis result of the driving module. Through the above mode, redundant data during fault diagnosis is greatly reduced, the diagnosis efficiency is improved, fault diagnosis is performed on the driving module through the fault diagnosis model, the diagnosis accuracy is further improved, and the obtained fault diagnosis result provides an important guiding effect for guaranteeing normal operation of equipment and improving production efficiency. And the problem of long downtime of equipment is avoided.

Description

Technical Field

[0001] This application relates to the technical field of fault diagnosis, and particularly to a fault diagnosis method, device, equipment, and storage medium. Background Art

[0002] With the rapid development of modern technology, various modules are widely used in fields such as electronic devices and industrial production. However, various faults may occur during the operation of the modules, and timely and accurate fault diagnosis is of great significance for ensuring the normal operation of the equipment and improving production efficiency. Traditional fault diagnosis methods often rely on manual experience or simple rule judgments, resulting in problems such as low diagnosis efficiency, low accuracy, and long equipment downtime.

[0003] The above content is only used to assist in understanding the technical solution of this application, and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of this application is to provide a fault diagnosis method, device, equipment, and storage medium, aiming to solve the technical problems of low diagnosis efficiency and low diagnosis accuracy when the prior art performs fault diagnosis on modules.

[0005] To achieve the above object, this application proposes a fault diagnosis method, and the method includes:

[0006] Select data from the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device to determine the target operation data of the drive module;

[0007] Input the module operation characteristics corresponding to the target operation data into the fault diagnosis model, and perform fault diagnosis on the drive module through the fault diagnosis model to determine the current fault information and module predicted life of the drive module;

[0008] Obtain the fault diagnosis result of the drive module according to the current fault information and the module predicted life.

[0009] In an embodiment, before the step of selecting data from the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device to determine the target operation data of the drive module, it further includes:

[0010] Collect the initial operation data of the drive module during the operation of the target device according to a preset acquisition period;

[0011] Perform data analysis on the initial operation data within the preset acquisition period to determine at least one acceleration data point existing in the initial operation data and a deceleration data point adjacent to the acceleration data point;

[0012] Determine at least one module motion segment existing in the initial operation data according to the acceleration data point and the deceleration data point adjacent to the acceleration data point;

[0013] Select data segments from the initial operation data according to the module motion segment to determine the module operation data of the drive module.

[0014] In an embodiment, before the step of performing fault diagnosis on the drive module according to the module operation characteristics corresponding to the target operation data and the fault diagnosis model to determine the current fault information and the module predicted life of the drive module, it further includes:

[0015] Obtain the operation feature set of the drive module, where the operation feature set is composed of operation feature columns of the drive module in multiple operation states;

[0016] Perform feature selection on the operation feature set according to the target feature columns corresponding to each operation feature column to determine multiple groups of first training features;

[0017] Train a machine learning model according to multiple groups of first training features and the training operation states corresponding to each first training feature to determine a basic diagnosis model;

[0018] Perform model optimization on the first diagnosis model according to multiple groups of second training features and the training operation states corresponding to each second training feature to obtain a derivative diagnosis model, where the second training features are obtained by numerically adjusting the key feature parameters in the first training features;

[0019] Perform model integration on the basic diagnosis model and the derivative diagnosis model to obtain the fault diagnosis model of the drive module.

[0020] In an embodiment, before the step of obtaining the operation feature set of the drive module, it further includes:

[0021] Perform data preprocessing on the operation source data of the drive module in each operation state to obtain the target source data in each operation state;

[0022] Perform feature transformation on the target source data in each operation state to determine the original feature data in each operation state;

[0023] Perform random generation according to the original feature data in each operation state to obtain operation feature columns in multiple operation states;

[0024] Construct the operation feature set of the drive module according to the operation feature columns in each operation state.

[0025] In one embodiment, the step of performing feature selection on the operation feature set according to the target feature columns corresponding to each operation feature column to determine multiple groups of first training features includes:

[0026] Calculate the information gain between each operation feature column in the operation feature set and the target feature column corresponding to each operation feature column respectively, and obtain the information gain corresponding to each operation feature column;

[0027] Calculate the mean value of the information gain corresponding to each operation feature column to determine the average gain;

[0028] Based on the comparison relationship between the average gain and the information gain corresponding to each operation feature column, screen multiple groups of operation feature columns to determine multiple groups of training feature columns;

[0029] Calculate the importance of multiple groups of training feature columns, and screen multiple groups of training feature columns according to the calculation result of the importance to determine multiple groups of first training features.

[0030] In one embodiment, the step of calculating the importance of multiple groups of training feature columns, screening multiple groups of training feature columns according to the calculation result of the importance, and determining multiple groups of first training features includes:

[0031] Calculate the importance of each training feature column through multiple importance evaluation models to determine the importance score of each training feature column under each importance evaluation model;

[0032] Calculate the mean value according to the importance score of each training feature column under each importance evaluation model to determine the average score of each training feature column;

[0033] Perform normalization processing on the average score of each training feature column to obtain the target importance score of each training feature column;

[0034] Sort the target importance scores of each training feature column, and perform feature screening on multiple groups of training feature columns according to the sorting result to determine multiple groups of first training features.

[0035] In one embodiment, before the step of performing fault diagnosis on the drive module according to the module operation features and the fault diagnosis model corresponding to the target operation data to determine the current fault information and the module predicted life of the drive module, it further includes:

[0036] Perform feature classification on the target operation data to determine the first type of data and the second type of data;

[0037] Perform outlier processing on the first type of data to obtain the first processed data;

[0038] Perform feature scaling on the first processed data to determine the first operation feature;

[0039] Perform encoding conversion on the second type of data to determine the second operation characteristic;

[0040] Perform feature summarization based on the first operation characteristic and the second operation characteristic to obtain the module operation characteristic corresponding to the target operation data.

[0041] In addition, to achieve the above object, the present application also provides a fault diagnosis device, which includes: a selection module for selecting the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device to determine the target operation data of the drive module;

[0042] A diagnosis module for performing fault diagnosis on the drive module according to the module operation characteristic corresponding to the target operation data and a fault diagnosis model to determine the current fault information and the module predicted life of the drive module;

[0043] A processing module for obtaining the fault diagnosis result of the drive module according to the current fault information and the module predicted life.

[0044] In addition, to achieve the above object, the present application also provides a fault diagnosis device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program is configured to implement the steps of the fault diagnosis method as described above.

[0045] In addition, to achieve the above object, the present application also provides a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium, and when the computer program is executed by a processor, the steps of the fault diagnosis method as described above are implemented.

[0046] In addition, to achieve the above object, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the fault diagnosis method as described above are implemented.

[0047] The present application provides a fault diagnosis method. According to the fault impact factors corresponding to the drive module on the target device, the present application selects data from the module operation data of the drive module to determine the target operation data of the drive module; inputs the module operation characteristics corresponding to the target operation data into a fault diagnosis model, diagnoses the fault of the drive module through the fault diagnosis model, and determines the current fault information and the predicted module life of the drive module; obtains the fault diagnosis result of the drive module according to the current fault information and the predicted module life. By the above method, the amount of redundant data during fault diagnosis is greatly reduced, the diagnosis efficiency is improved. At the same time, the fault of the drive module is diagnosed through the fault diagnosis model, further improving the diagnosis accuracy. The obtained fault diagnosis result provides an important guiding role for ensuring the normal operation of the device and improving production efficiency, avoiding the problem of long device downtime, and reducing the device maintenance cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings are incorporated herein and form a part of this specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0049] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative efforts.

[0050] Figure 1 It is a schematic flowchart provided for Embodiment 1 of the fault diagnosis method of the present application;

[0051] Figure 2 It is a schematic diagram of the module operation mechanism provided for Embodiment 1 of the present application;

[0052] Figure 3 It is a schematic flowchart provided for Embodiment 2 of the fault diagnosis method of the present application;

[0053] Figure 4 It is a schematic diagram of data interception provided for Embodiment 2 of the present application;

[0054] Figure 5 It is a schematic flowchart provided for Embodiment 3 of the fault diagnosis method of the present application;

[0055] Figure 6 It is a schematic flowchart of the brief process of the fault diagnosis method provided for Embodiment 3 of the present application;

[0056] Figure 7 It is a schematic diagram of the module structure of the fault diagnosis device in the embodiment of the present application;

[0057] Figure 8 It is a schematic diagram of the device structure of the hardware operating environment involved in the fault diagnosis method in the embodiments of the present application.

[0058] The implementation, functional characteristics and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. Detailed implementation manners

[0059] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0060] To better understand the technical solutions of the present application, the following will be described in detail in conjunction with the drawings in the specification and specific implementation manners.

[0061] The main solution of the embodiments of the present application is: select data from the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device to determine the target operation data of the drive module; input the module operation characteristics corresponding to the target operation data into the fault diagnosis model, and perform fault diagnosis on the drive module through the fault diagnosis model to determine the current fault information and module predicted life of the drive module; obtain the fault diagnosis result of the drive module according to the current fault information and the module predicted life.

[0062] With the rapid development of modern technology, various modules are widely used in fields such as electronic devices and industrial production. However, various faults may occur during the operation of the module. Timely and accurately diagnosing faults is of great significance for ensuring the normal operation of the device and improving production efficiency. Traditional fault diagnosis methods often rely on manual experience or simple rule judgment, and have problems such as low diagnosis efficiency, low accuracy, and long device downtime.

[0063] The present application greatly reduces the amount of redundant data during fault diagnosis, improves the diagnosis efficiency. At the same time, the fault diagnosis model is used to perform fault diagnosis on the drive module, further improving the diagnosis accuracy. The obtained fault diagnosis result provides an important guiding role for ensuring the normal operation of the device and improving production efficiency, avoiding the problem of long device downtime, and reducing the device maintenance cost.

[0064] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device, a fault diagnosis device, etc. that can implement the above functions. The following takes the fault diagnosis device as an example to illustrate this embodiment and the following embodiments.

[0065] Based on this, the embodiments of the present application provide a fault diagnosis method, referring toFigure 1 , Figure 1 is a schematic flowchart of the first embodiment of the fault diagnosis method of this application.

[0066] In this embodiment, the fault diagnosis method includes steps S10 to S30:

[0067] Step S10: Select data from the module operation data of the drive module according to the fault impact factors corresponding to the drive module on the target device, and determine the target operation data of the drive module.

[0068] It should be noted that in this embodiment, the target device can be a device with a drive module in various fields. The drive module refers to a module that realizes various mechanical movements through motor drive, and the mechanical movements include but are not limited to linear motion, rotation and other movements.

[0069] It can be understood that the fault impact factor is a parameter that affects the associated motor after the module fails. For example, when the drive module jams, the running resistance of the module increases, which in turn affects parameters such as the real-time current and real-time speed of the associated motor. Since different drive modules use different types of motors, such as stepper motors, servo motors, DC motors or AC motors, etc.; at the same time, different drive modules may exhibit different motion forms. Therefore, there may be multiple drive modules on the target device, and the fault impact factors corresponding to each drive module are different. The fault impact factors corresponding to each drive module are determined by analyzing the operation data of a large number of the same type of drive modules.

[0070] In this embodiment, the fault impact factors of drive module A include but are not limited to X1 motor parameters, X2 module parameters, and X3 other parameters. Among them, X1 motor parameters include but are not limited to x1 real-time motor current and x2 real-time motor speed; X2 module parameters include but are not limited to the power of x3 motor, the load of x4 module, the lead of x5 module, the running angle of x6 module (the angle between the module running surface and the horizontal plane), the maximum speed of x7 module, and the maximum acceleration of x8 module; X3 other parameters include but are not limited to lubricating oil, dust, and chemical corrosion. In general, the X3 other parameters can be ignored, so the relevant data of the X3 other parameters are not collected.

[0071] It should be understood that the real-time current of the x1 motor and the real-time speed of the x2 motor can be recorded by the motor driver of the PCL (Programmable Logic Controller) and uploaded to the digital system in real time for storage; the power of the x3 motor, the load of the x4 module, and the lead of the x5 module are generally rated values, and the PCL can upload all the above values to the digital system for storage; the operating angle of the x6 module is also generally a rated value, and the range of the operating angle is (0, 90°) (mainly horizontal 0° and vertical 90°) which is the rated value, and the PLC can upload this value to the digital system for storage; the maximum speed of the x7 module and the maximum acceleration of the x8 module are parameters issued by the PLC and can be directly uploaded to the digital system for storage.

[0072] In a specific implementation, the module operation data includes the motion data collected in real time when the driving module completes at least one acceleration and deceleration process within one acquisition period. When the target device is running, the module operation data of the driving module is obtained in real time, and in the module operation data, the specific values of the fault influence factors corresponding to the driving module are extracted, so as to obtain the target operation data of the driving module. In this embodiment, the target operation data refers to the specific values of the fault influence factors corresponding to the driving module during this motion process.

[0073] In a feasible implementation manner, before step S10, steps A11 to A14 may be included:

[0074] Step A11, collect the initial operation data of the driving module during the operation of the target device according to a preset acquisition period.

[0075] It should be noted that the preset acquisition period refers to the time interval for collecting data once in each period / beat, and the preset acquisition period is preset to ensure the regularity of data collection. The initial operation data refers to the original data collected from the driving module within the preset acquisition period, including but not limited to position, speed, acceleration, real-time current of the motor, power of the motor, etc.

[0076] Step A12, perform data analysis on the initial operation data within the preset acquisition period to determine at least one acceleration data point and a deceleration data point adjacent to the acceleration data point existing in the initial operation data.

[0077] It should be noted that by analyzing the initial operation data to identify at least one data point in the acceleration phase and its adjacent data points in the deceleration phase, in the above manner, the specific moments and positions of the acceleration and deceleration processes experienced by the drive module during task execution can be accurately located. In this embodiment, the acceleration data points refer to the data in the acceleration phase, and the deceleration data points refer to the data points in the deceleration phase. The deceleration data points adjacent to the acceleration data points can be located before or after the acceleration data points, and this embodiment does not limit this.

[0078] Step A13, determine at least one module motion segment existing in the initial operation data according to the acceleration data points and the deceleration data points adjacent to the acceleration data points.

[0079] It should be noted that based on the identified at least one acceleration data point and its adjacent deceleration data points, at least one module motion segment existing in the initial operation data is identified. In this embodiment, the module motion segment refers to the data segment in the initial operation data from the acceleration data point to the adjacent deceleration data point. This data segment represents the motion process of the drive module in one cycle, including the acceleration and deceleration processes. The starting point of the data segment is the acceleration data point, and the ending point of the data segment is the adjacent deceleration data point; or, the starting point of the data segment is the adjacent deceleration data point, and the ending point of the data segment is the acceleration data point.

[0080] Step A14, perform data segment selection on the initial operation data according to the module motion segment to determine the module operation data of the drive module.

[0081] It should be noted that after obtaining the module motion segment, it is intercepted from the initial operation data. If there are multiple module motion segments, one module motion segment A→B can be randomly selected and used as the module operation data of the drive module; if there is only one module motion segment, it is directly used as the module operation data of the drive module.

[0082] It can be understood that since the operation of the drive module is a periodic operation, its operation trajectory is complex and the differences are large. The data to be collected is large and its related regularity is poor. If all are recorded, a large amount of useless data will be imported, resulting in an overly large and irregular AI model data, causing the failure of AI model creation or overfitting, affecting the diagnostic efficiency and accuracy. Through this embodiment, targeted data collection is carried out based on the module fault mechanism, reducing the data collection volume and model processing volume, and improving the feasibility of the model.

[0083] Step S20, input the module operation characteristics corresponding to the target operation data into the fault diagnosis model, and perform fault diagnosis on the drive module through the fault diagnosis model to determine the current fault information and module predicted life of the drive module.

[0084] It should be noted that the fault diagnosis model is an AI (Artificial Intelligence Model) model that can predict the fault types, fault locations, and remaining service life of all drive modules on the target device. In this embodiment, the AI model is obtained by first training a machine learning model with the historical operation data of one drive module to obtain a benchmark model, and then optimizing the benchmark model with the historical operation data of other drive modules. The machine learning model is a multi-classification AI model, including but not limited to any one of LogisticRegression (logistic regression), RandomForest (random forest), XGBoost (extreme gradient boosting), AdaBoost (adaptive boosting), Lightgbm (lightweight gradient boosting machine), and other classification models.

[0085] It can be understood that data preprocessing and feature transformation are performed on the target operation data to obtain the module operation characteristics corresponding to the target operation data.

[0086] In a specific implementation, the module operation characteristics corresponding to the target operation data of the drive module are input into the fault diagnosis model. The fault diagnosis model performs fault diagnosis on the drive module based on the module operation characteristics to determine whether the drive module has a fault, the fault type, fault location, and remaining service life of the module if a fault occurs. In this embodiment, the current fault information includes the result of whether the drive module has a fault, the fault type, and fault location if a fault occurs, etc.

[0087] In a feasible implementation manner, before step S20, steps B11 to B15 may be included:

[0088] Step B11: Classify the features of the target operation data to determine the first type of data and the second type of data.

[0089] Step B12: Process the outliers of the first type of data to obtain the first processed data.

[0090] Step B13: Perform feature scaling on the first processed data to determine the first operation characteristics.

[0091] Step B14: Perform encoding conversion on the second type of data to determine the second operation characteristics.

[0092] Step B15: Summarize the features according to the first operation characteristics and the second operation characteristics to obtain the module operation characteristics corresponding to the target operation data.

[0093] It should be noted that after obtaining the target operation data of the driving module, the target operation data is classified by data characteristics to determine the categorical feature data and numerical feature data therein. In this embodiment, the first type of data refers to numerical feature data, and the second type of data refers to categorical feature data.

[0094] It can be understood that the following processing is performed on the first type of data to obtain the first processed data. The specific processing process includes: 1. Invalid value processing: According to the current curve graph, identify the invalid data in the end part of the data and delete it. 2. Missing value processing: Identify the missing values and fill the missing values with the median of all non-missing values in this column of features. 3. Outlier processing: Identify the maximum and minimum values among them and delete them; at the same time, the six-sigma principle can be used to identify outliers. When a data point exceeds the mean plus or minus six times the standard deviation of the overall data set, it will be determined as an outlier and deleted.

[0095] In a specific implementation, the following three methods can be used to perform feature scaling on the first processed data to obtain the first operation feature corresponding to the first type of data. The specific methods are as follows: 1. Standardization strategy (standard_scaling): Convert the first processed data into a normal distribution with a mean of 0 and a standard deviation of 1. The calculation method is to subtract the mean from each data point and then divide by the standard deviation. 2. Normalization strategy (minmax_scaling): Scale the first processed data to between 0 and 1. 3. Leave-alone strategy (leave_alone), this method will not change the original data. In this embodiment, the leave-alone strategy is selected to perform feature scaling on the first processed data to obtain the first operation feature.

[0096] It should be noted that in this embodiment, the following two methods can be used to perform encoding conversion on the second type of data to obtain the second operation feature corresponding to the second type. The specific methods are as follows: 1. One-hot encoding: Convert the second type of data into a binary matrix. Each category is represented by a binary vector, and the length of the vector is equal to the total number of categories, where only one position is 1 and the rest are 0. 2. Label Encoder: Convert the second type of data into a unique integer value, and each category is mapped to an integer label. In this embodiment, the label encoding strategy is selected to perform conversion on the second type of data to obtain the second operation feature.

[0097] It can be understood that after obtaining the first operation feature corresponding to the first type of data and the second operation feature corresponding to the second type of data, the two are summarized to obtain the module operation feature corresponding to the target operation data.

[0098] Step S30: Obtain the fault diagnosis result of the drive module according to the current fault information and the predicted life of the module.

[0099] It should be noted that the current fault information and the predicted life of the module are summarized to obtain the fault diagnosis result of the drive module. Based on the fault diagnosis result, early warnings are given to the target device and sent to the digital system. When a fault occurs in the drive module, corresponding repair / compensation strategies are formulated according to the type of the drive fault and pushed to the target device or the maintenance personnel. When the target device can adjust according to the repair / compensation strategy, it is pushed to the target device, and the adjustment record and the fault diagnosis result are sent to the maintenance personnel. When the target device cannot handle it, the repair / compensation strategy and the fault diagnosis result are pushed to the maintenance personnel, and the maintenance personnel will handle the fault according to the pushed information.

[0100] In this embodiment, when the drive module is normal, the fault diagnosis result is: the module is operating normally, and the predicted life of the module is x hours. When the drive module is abnormal, the fault diagnosis result is: Fault type: XX; Fault reason: XX; Predicted life of the module: x hours; Repair / compensation strategy: The module needs maintenance, for example: the module needs to add lubricating oil, the connection between the load and the module is loose, the coupling is abnormal, etc.

[0101] It can be understood that the fault diagnosis process is described by taking the drive module A as an example. The operating mechanism of the drive module A is as Figure 2 shown. The load 3 of the drive module is controlled by the PCL, drives the motor to rotate through the driver 4, and then drives the module slider to move through the coupling, that is, the linear motion of the load 3 is completed. The load completes the controllable reciprocating linear motion on the ball screw under the drive of the motor. Among them, the motor: the power device - provides the power for rotational motion. The coupling: the connecting component - completes the connection between the motor and the ball screw. The ball screw: the motion device - completes the horizontal forward and backward movement of the slider under the drive of the motor. After the fault diagnosis model diagnoses the fault based on the module operation characteristics corresponding to the target operation data of the drive module A, it is determined that the fault of the drive module is caused by factors such as lack of lubricating oil, ball detachment, and bead explosion, which hinder the operation of the module. At this time, the servo motor performs current compensation according to the signal feedback of the motor rotation to meet the requirements of the target operation trajectory. When it is determined that although the motion trajectory of the drive module can meet the usage requirements during operation, the module has entered the wear period, and if it cannot be maintained and repaired in time to improve the module life, it will be damaged rapidly in this environment. At this time, the fault diagnosis result and the maintenance strategy will be pushed to the maintenance personnel.

[0102] This embodiment provides a fault diagnosis method. In this embodiment, data selection is performed on the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device to determine the target operation data of the drive module; the module operation characteristics corresponding to the target operation data are input into the fault diagnosis model, and the drive module is fault-diagnosed through the fault diagnosis model to determine the current fault information and the predicted module life of the drive module; the fault diagnosis result of the drive module is obtained according to the current fault information and the predicted module life. Through the above method, the amount of redundant data during fault diagnosis is greatly reduced, the diagnosis efficiency is improved, and at the same time, the fault diagnosis of the drive module is performed through the fault diagnosis model, further improving the diagnosis accuracy. The obtained fault diagnosis result provides an important guiding role for ensuring the normal operation of the equipment and improving the production efficiency, avoiding the problem of long equipment downtime, and reducing the equipment maintenance cost.

[0103] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar content as in the above-mentioned first embodiment can be referred to the above introduction and will not be repeated hereinafter. On this basis, please refer to Figure 3 , before step S20, the fault diagnosis method further includes steps S21 to S25:

[0104] Step S21, obtaining an operation feature set of the drive module, where the operation feature set is composed of operation feature columns of the drive module in multiple operation states.

[0105] It should be noted that historical operation data of a large number of drive modules of the same type are obtained; or, historical operation data of the drive module in various operation states. The operation states include normal operation states and the states corresponding to the drive module during operation under various fault types.

[0106] It can be understood that by analyzing the historical operation data in various operation states, the operation feature columns in each operation state are obtained. One operation state includes multiple groups of operation feature columns; the operation feature columns contain data points of the drive module in multiple dimensions in this operation state, there are multiple feature measurement values, and each data point reflects the performance of the drive module at a certain moment.

[0107] In a specific implementation, the operation feature set will include operation feature columns in multiple operation states, and the operation feature columns in multiple operation states are summarized to construct the operation feature set of the drive module.

[0108] In a feasible implementation manner, before step S21, steps C11 to C14 may be included:

[0109] Step C11: Perform data preprocessing on the operation source data of the drive module in each operation state to obtain the target source data in each operation state.

[0110] It should be noted that the digital system will obtain the historical operation data of the drive module in various operation states. The historical operation data includes, but is not limited to, raw data such as position, speed, acceleration, real-time motor current, and motor power, which reflects the working conditions of the drive module in different operation states. The historical operation data is the historical operation data of the same type of drive module or the historical operation data of this drive module.

[0111] It should be understood that the historical operation data segments in each operation state are cropped to obtain the operation source data corresponding to the operation process of the drive module in one cycle in each operation state. In this embodiment, the operation source data includes the acceleration and deceleration processes. For example, Figure 4 As shown, the operation source data takes the intermediate working data in the historical operation data, where the blue line represents the real-time current of the drive module in the historical operation process, and the red line represents the real-time speed of the drive module in the historical operation process. Since the X2 module parameters of the drive module are generally rated values under normal circumstances, no processing is performed on the multiple parameters included in the X2 module parameters.

[0112] It can be understood that since the file uploaded by PCL to the digital system is a csv (Comma-Separated Values) file, the csv library of python is used to read the operation source data in each operation state, and the quotes existing in the operation source data in each operation state are removed through the replacement function (such as the replace function in Python), and the missing values are filled with 0. Then, the strings of the operation source data in each operation state are converted into a list format, so as to obtain the list data corresponding to the operation source data in each operation state.

[0113] In a specific implementation, data exploration is performed on the list data corresponding to the operation source data in each operation state, and data volume reduction is performed on the operation source data in each operation state. The specific strategy is as follows: explore each feature column in the list data. If the duplicate data in this column exceeds a certain proportion (for example, 95%) or all the data is duplicate, then this feature column is filtered; or if the number of null values in this column is too large, then this feature column is filtered. Finally, the remaining feature columns of the operation source data in this operation state are used as the target source data in each operation state.

[0114] For example, in the normal operating state, there are 2,379 columns of operating source data, and the proportion of duplicate data in the data exceeds 95%. (567 - 570) are not selected as feature columns. All the data in the data are duplicate data, and (571 - 600) are not selected as feature columns. There are too many null values in the data, and (601 - 2,396) are not selected as feature columns. After processing according to the above three steps, the data information is as follows: there are XX rows in total, 562 columns remaining, and the data volume is reduced by 76.5%.

[0115] Step C12: Perform feature transformation on the target source data in each operating state to determine the original feature data in each operating state.

[0116] It should be noted that after obtaining the target source data in each operating state, the invalid values, missing values, and abnormal values in the target source data in each operating state are processed separately. After processing, the categorical feature data in the target source data in each operating state is encoded and transformed, and the numerical feature data is feature-scaled to obtain the original feature data in each operating state. In this embodiment, the encoding and transformation methods include, but are not limited to, the one_hot method and the label_encoder method. The feature-scaling methods include, but are not limited to, the standardization strategy, the normalization strategy, and the no-processing strategy, etc.

[0117] Step C13: Randomly generate based on the original feature data in each operating state to obtain the operating feature columns in multiple operating states.

[0118] It should be noted that the original feature data in each operating state is randomly generated, for example, by randomly combining the original feature data or other random sampling methods to obtain the operating feature columns in each operating state. In this embodiment, one operating state includes multiple groups of operating feature columns.

[0119] Step C14: Construct the operating feature set of the drive module according to the operating feature columns in each operating state.

[0120] It should be noted that the operating feature columns in multiple operating states are summarized, and a certain number of operating feature columns are selected from them to construct the operating feature set of the drive module. In this embodiment, 100 groups of operating feature columns can be selected from the operating feature columns in multiple operating states to construct the operating feature set.

[0121] Step S22: Perform feature selection on the operating feature set according to the target feature columns corresponding to each operating feature column to determine multiple groups of first training features.

[0122] It should be noted that the target feature column refers to the standard operating feature column corresponding to each operating state, and the target feature column can be used as the label column during fault diagnosis training. Feature screening is performed on the operating feature set through the target feature columns corresponding to each group of operating feature columns, and the most predictive features are selected from them to reduce redundant information, improve the efficiency and effectiveness of the model, and thus obtain multiple groups of first training features. In this embodiment, there is a corresponding operating state label for each group of training features.

[0123] Step S23: Train the machine learning model according to multiple groups of first training features and the training operating states corresponding to each first training feature to determine the basic diagnosis model.

[0124] It should be noted that the training operating state refers to the operating state label corresponding to each group of first training features. The machine learning model is trained through multiple groups of first training features and the training operating states corresponding to each first training feature, and the trained machine learning model is the basic diagnosis model. In this embodiment, the machine learning model refers to a multi-classification AI model, including but not limited to any one of LogisticRegression (logical regression), RandomForest (random forest), XGBoost (extreme gradient boosting), AdaBoost (adaptive boosting), Lightgbm (lightweight gradient boosting machine), and other classification models.

[0125] Step S24: Optimize the first diagnosis model according to multiple groups of second training features and the training operating states corresponding to each second training feature to obtain a derivative diagnosis model, where the second training features are obtained by numerically adjusting the key feature parameters in the first training features.

[0126] It should be noted that since many of the parameters in the X2 module parameters among the fault impact factors of each drive module are rated values, during the training of the basic diagnosis model, there are differences in the X1 motor parameters and X3 other parameters under various operating states, but many of the parameters in the X2 module parameters are a set of rated values. Therefore, a key feature parameter in multiple sets of first training features is numerically adjusted to obtain new X2 module parameters, and the operating feature columns under various operating states are obtained for these module parameters. Then, the operating feature columns under various operating states are processed to obtain multiple sets of second training features and their corresponding training operating states. For example, among multiple sets of first training features, the X2 module parameters include the power of the x3 motor, the load of the x4 module, the lead of the x5 module, the operating angle of the x6 module (the angle between the module operating surface and the horizontal plane), the maximum speed of the x7 module, and the maximum acceleration of the x8 module. x3 - x8 are all rated values. By adjusting x3, new x'3, x4 - x8 are obtained, and thus multiple sets of second training features under these module parameters are obtained. The multiple sets of second training features and the corresponding training operating states are input into the basic diagnosis model for further training and optimization of the model. This process can help the model learn more diverse patterns and boundary conditions, improve its generalization ability and robustness, and thus obtain a derivative diagnosis model under these module parameters.

[0127] It can be understood that by repeating the above process and sequentially adjusting other parameters in the X2 module parameters, the basic diagnosis model is trained based on the new second training features and their corresponding training operating states to obtain derivative diagnosis models under multiple sets of module parameters.

[0128] Step S25: Integrate the basic diagnosis model and the derivative diagnosis models to obtain the fault diagnosis model of the drive module.

[0129] It should be noted that integrating the basic diagnosis model and multiple derivative diagnosis models to obtain an integrated diagnosis model, and the integration methods include but are not limited to: 1. Weighted average: Assign a weight to each model and perform weighted averaging according to the performance or importance of the model. 2. Voting method: For classification tasks, the majority voting method can be used to determine the final output. 3. Stacking method: Use a meta - model to combine the prediction results of multiple models.

[0130] It can be understood that when evaluating the integrated diagnostic model using the validation set, the evaluation metrics include but are not limited to accuracy, recall, and other metrics. If the integrated diagnostic model meets the standard, it will be used as the fault diagnosis model for the drive module. In this embodiment, the operating characteristic sets of other drive modules can be used to optimize the fault diagnosis model, so that the fault diagnosis model can diagnose faults for all drive modules existing on the target device. At the same time, the data where the predicted results and the actual results recorded by the digital system do not match is regularly pushed into the fault diagnosis model for model optimization, so as to achieve the self-optimization of the AI algorithm, make the fault diagnosis method have the ability of self-learning and self-adaptation, and be able to continuously adapt to the changes of the module and new fault modes.

[0131] This embodiment provides a fault diagnosis method. In this embodiment, by obtaining the operating characteristic set of the drive module, the operating characteristic set is composed of the operating characteristics of the drive module in multiple operating states; feature selection is performed on the operating characteristic set according to the target characteristic columns corresponding to each operating characteristic column to determine multiple groups of first training features; the machine learning model is trained according to multiple groups of first training features and the training operating states corresponding to each first training feature to determine the basic diagnosis model; the first diagnosis model is optimized according to multiple groups of second training features and the training operating states corresponding to each second training feature to obtain the derivative diagnosis model, and the second training features are obtained by numerically adjusting the key feature parameters in the first training features; the basic diagnosis model and the derivative diagnosis model are integrated to obtain the fault diagnosis model of the drive module. Through the above method, the accuracy and diagnostic efficiency of the model are improved.

[0132] Based on the first embodiment and / or the second embodiment of the present application, in the third embodiment of the present application, the same or similar content as in the above-mentioned first and second embodiments can be referred to the above introduction and will not be repeated hereinafter. On this basis, please refer to Figure 5 , step S22, the fault diagnosis method further includes steps S01 to S04:

[0133] Step S01, calculate the information gain between each operating characteristic column in the operating characteristic set and the target characteristic column corresponding to each operating characteristic column respectively, and obtain the information gain corresponding to each operating characteristic column.

[0134] It should be noted that information gain is an index to measure the predictive ability of a feature for the target column. The greater the information gain, the greater the contribution to the target column. Therefore, by calculating the information gain between each group of running feature columns and the corresponding target feature columns respectively, the information gain corresponding to each group of running feature columns can be obtained. In this embodiment, there is a corresponding target feature column for each running state, and the running state of each group of running feature columns is known. Therefore, the target feature columns corresponding to each group of running feature columns can be determined based on the above mapping relationship.

[0135] Step S02, calculate the mean of the information gains corresponding to each running feature column to determine the average gain.

[0136] It should be noted that by calculating the mean of the information gains corresponding to all running feature columns, the average gain is obtained, and this value represents the average level of the predictive ability of the running feature columns for the target column.

[0137] Step S03, screen multiple groups of running feature columns based on the comparison relationship between the average gain and the information gains corresponding to each running feature column to determine multiple groups of training feature columns.

[0138] It should be noted that the magnitudes of the information gains corresponding to each running feature column and the average gain are compared respectively to obtain multiple comparison relationships. Based on the comparison relationship between the average gain and the information gains corresponding to each running feature column, the columns in the running feature columns whose information gains are greater than the average gain are screened out, and this group of running feature columns is used as the training feature columns.

[0139] Step S04, calculate the importance of multiple groups of training feature columns, and screen multiple groups of training feature columns according to the calculation results of the importance to determine multiple groups of first training features.

[0140] It should be noted that the feature importance of each group of training feature columns is calculated through multiple trained importance evaluation models (such as Random Forest, XGBoost, LightGBM, etc.) to obtain the importance scores of each group of training feature columns. Based on the importance scores of each group of training feature columns, multiple groups of training feature columns are screened, and the training feature columns with a certain proportion (such as 95%) of the top importance are retained, so as to obtain multiple groups of first training features.

[0141] In a feasible implementation manner, step S04 may include steps D11 to D14:

[0142] Step D11, calculate the importance of each training feature column through multiple importance evaluation models to determine the importance scores of each training feature column under each importance evaluation model.

[0143] It should be noted that in this embodiment, there are multiple importance evaluation models, which are models that can calculate feature importance and are trained based on random forest, XGBoost, and LightGBM respectively. They are trained with a large number of sample features and their corresponding importance labels. Multiple groups of training feature columns are respectively input into multiple importance evaluation models. For each group of training feature columns, the importance scores of the training feature columns obtained under multiple importance evaluation models are calculated, which reflects the feature importance of the training feature columns under the importance evaluation models.

[0144] Step D12: Calculate the mean value based on the importance scores of each training feature column under each importance evaluation model to determine the average score of each training feature column.

[0145] It should be noted that for each group of training feature columns: calculate the average value of the importance scores of the training feature columns under multiple importance evaluation models, so as to obtain the average score of the training feature columns.

[0146] Step D13: Normalize the average scores of each training feature column to obtain the target importance scores of each training feature column.

[0147] It should be noted that normalize the average scores of all training feature columns so that all scores fall within the interval [0, 1], thereby obtaining the target importance scores of each training feature column. In this embodiment, the normalization process is specifically: divide the average score of each training feature column by the maximum score value to obtain the final target importance score.

[0148] Step D14: Sort the target importance scores of each training feature column, and perform feature screening on multiple groups of training feature columns according to the sorting result to determine multiple groups of first training features.

[0149] It should be noted that sort the target importance scores of each training feature column. The sorting method can be ascending order or descending order. In this embodiment, the descending order sorting method is selected. Based on the sorting result, screen multiple groups of training feature columns, and retain a certain proportion (such as 95%) of the training feature columns with the highest importance, thereby obtaining multiple groups of first training features.

[0150] This embodiment provides a fault diagnosis method. In this embodiment, the information gain between each operation feature column in the operation feature set and the corresponding target feature column is calculated respectively to obtain the information gain corresponding to each operation feature column; the mean value of the information gain corresponding to each operation feature column is calculated to determine the average gain; based on the comparison relationship between the average gain and the information gain corresponding to each operation feature column, multiple groups of operation feature columns are screened to determine multiple groups of training feature columns; the importance of multiple groups of training feature columns is calculated, and multiple groups of training feature columns are screened according to the calculation result of the importance to determine multiple groups of first training features. Through the above method, the most predictive features are selected, redundant information is reduced, and the model training efficiency and training effect are improved.

[0151] Exemplarily, to help understand the implementation process of the fault diagnosis method obtained by combining the above Embodiment 1 and Embodiment 2, please refer to Figure 6 , Figure 6 which provides a schematic diagram of the brief process of a fault diagnosis method. Specifically: 1. Data acquisition: The IOT system (Internet of Things system) completes data acquisition and outputs it to the fault diagnosis model. 2. Fault diagnosis: 2.1 The fault diagnosis model (i.e., the AI model) completes model parameter inference based on the acquired information and outputs a conclusion. The IOT system sends the model diagnosis data to the device to complete the push of module replacement or maintenance information. 2.2 The process of establishing the AI model: 2.2.1 Data cleaning; 2.2.2 Data exploration; 2.2.3 Feature engineering: processing invalid values, missing values, and outliers; 2.2.4 Feature transformation; 2.2.5 Feature selection; 2.2.6 Model training; 2.2.7 Model testing. 3. Self-optimization of the AI model: According to the data recorded by the IOT system where the deduction does not match the actual conclusion, it is regularly pushed into the AI model to optimize the existing AI model. The inconsistent data is added to the establishment of the AI model, and the specific steps for regenerating the model are repeated the same as 2.2.

[0152] The method of this embodiment finally obtained has the following beneficial effects: 1. Improve the diagnosis efficiency and accuracy, and can quickly and accurately diagnose the fault type and location of the module. 2. Reduce the maintenance cost and reduce the production loss and maintenance cost caused by faults. 3. Have self-learning and self-adaptive capabilities, and can continuously adapt to the changes of the module and new fault modes. 4. Realize real-time monitoring and diagnosis of the module, and improve the reliability and stability of the device. 5. Targeted acquisition of data based on the module fault mechanism reduces the data acquisition volume and model processing volume, and improves the feasibility and promotion. 6. The investment related to the model is low, which is an upgrade based on the existing digital acquisition, with low investment and high return.

[0153] It should be noted that the above examples are only for understanding the present application and do not constitute a limitation on the fault diagnosis method of the present application. Based on this technical concept, more forms of simple transformations are within the protection scope of the present application.

[0154] The present application also provides a fault diagnosis device. Please refer to Figure 7 , the fault diagnosis device includes:

[0155] A selection module 10, configured to perform data selection on the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device, and determine the target operation data of the drive module.

[0156] A diagnosis module 20, configured to perform fault diagnosis on the drive module according to the module operation characteristics corresponding to the target operation data and a fault diagnosis model, and determine the current fault information and module predicted life of the drive module.

[0157] A processing module 30, configured to obtain the fault diagnosis result of the drive module according to the current fault information and the module predicted life.

[0158] Optionally, the selection module 10 is further configured to:

[0159] Collect the initial operation data of the drive module during the operation of the target device according to a preset acquisition period; perform data analysis on the initial operation data within the preset acquisition period to determine at least one acceleration data point existing in the initial operation data and a deceleration data point adjacent to the acceleration data point; determine at least one module motion segment existing in the initial operation data according to the acceleration data point and the deceleration data point adjacent to the acceleration data point; perform data segment selection on the initial operation data according to the module motion segment to determine the module operation data of the drive module.

[0160] Optionally, the diagnosis module 20 is further configured to:

[0161] Obtain an operation feature set of the drive module, where the operation feature set is composed of operation features of the drive module in multiple operation states; perform feature selection on the operation feature set according to target feature columns corresponding to each operation feature column to determine multiple groups of first training features; train a machine learning model according to multiple groups of first training features and training operation states corresponding to each first training feature to determine a basic diagnosis model; perform model optimization on the first diagnosis model according to multiple groups of second training features and training operation states corresponding to each second training feature to obtain a derivative diagnosis model, where the second training features are obtained by numerically adjusting key feature parameters in the first training features; perform model integration on the basic diagnosis model and the derivative diagnosis model to obtain the fault diagnosis model of the drive module.

[0162] Optionally, the diagnostic module 20 is further configured to:

[0163] Perform data preprocessing on the operation source data of the drive module in each operation state to obtain target source data in each operation state; perform feature transformation on the target source data in each operation state to determine the original feature data in each operation state; perform random generation according to the original feature data in each operation state to obtain operation feature columns in multiple operation states; construct an operation feature set of the drive module according to the operation feature columns in each operation state.

[0164] Optionally, the diagnostic module 20 is further configured to:

[0165] Calculate the information gain between each operation feature column in the operation feature set and the corresponding target feature column respectively to obtain the information gain corresponding to each operation feature column; perform a mean calculation on the information gain corresponding to each operation feature column to determine the average gain; screen multiple groups of operation feature columns based on the comparison relationship between the average gain and the information gain corresponding to each operation feature column to determine multiple groups of training feature columns; perform importance calculation on multiple groups of training feature columns, and screen multiple groups of training feature columns according to the importance calculation results to determine multiple groups of first training features.

[0166] Optionally, the diagnostic module 20 is further configured to:

[0167] Perform importance calculation on each training feature column through multiple importance evaluation models to determine the importance score of each training feature column under each importance evaluation model; perform a mean calculation according to the importance score of each training feature column under each importance evaluation model to determine the average score of each training feature column; perform normalization processing on the average score of each training feature column to obtain the target importance score of each training feature column; sort the target importance scores of each training feature column, and perform feature screening on multiple groups of training feature columns according to the sorting results to determine multiple groups of first training features.

[0168] Optionally, the diagnostic module 20 is further configured to:

[0169] Perform feature classification on the target operation data to determine first-type data and second-type data; perform outlier processing on the first-type data to obtain first-processed data; perform feature scaling on the first-processed data to determine first operation features; perform encoding conversion on the second-type data to determine second operation features; perform feature summarization according to the first operation features and the second operation features to obtain the module operation features corresponding to the target operation data.

[0170] The fault diagnosis device provided by the present application adopts the fault diagnosis method in the above embodiment, and can solve the technical problems of low diagnosis efficiency and low diagnosis accuracy in the prior art when diagnosing faults of modules. Compared with the prior art, the beneficial effects of the fault diagnosis device provided by the present application are the same as those of the fault diagnosis method provided by the above embodiment, and other technical features in the fault diagnosis device are the same as the features disclosed in the method of the above embodiment, which will not be elaborated here.

[0171] The present application provides a fault diagnosis device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the fault diagnosis method in the first embodiment above.

[0172] Refer to the following Figure 8 , which shows a schematic structural diagram of a fault diagnosis device suitable for implementing the embodiments of the present application. The fault diagnosis device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 8 The fault diagnosis device shown is only an example and should not impose any limitation on the functions and usage scope of the embodiments of the present application.

[0173] As shown in Figure 8As shown in the figure, the fault diagnosis device may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which may perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. In the RAM 1004, various programs and data required for the operation of the fault diagnosis device are also stored. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems may be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 may allow the fault diagnosis device to communicate with other devices wirelessly or wiredly to exchange data. Although the figure shows a fault diagnosis device having various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or had alternatively.

[0174] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program codes for executing the methods shown in the flowcharts. In such an embodiment, the computer program may be downloaded and installed from a network through the communication device, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above functions defined in the methods of the embodiments disclosed in the present application are executed.

[0175] The fault diagnosis device provided by the present application adopts the fault diagnosis method in the above embodiment, and can solve the technical problems of low diagnosis efficiency and low diagnosis accuracy rate in the prior art when diagnosing faults of modules. Compared with the prior art, the beneficial effects of the fault diagnosis device provided by the present application are the same as those of the fault diagnosis method provided by the above embodiment, and other technical features in the fault diagnosis device are the same as those disclosed in the method of the previous embodiment, and will not be elaborated here.

[0176] It should be understood that each part disclosed in this application can be implemented by hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in a suitable manner in any one or more embodiments or examples.

[0177] As described above, this is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0178] This application provides a computer-readable storage medium with computer-readable program instructions (i.e., computer programs) stored thereon, and the computer-readable program instructions are used to execute the fault diagnosis method in the above embodiments.

[0179] The computer-readable storage medium provided by this application can be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this embodiment, the computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or combined with an instruction execution system, device, or device. The program code contained on the computer-readable storage medium can be transmitted by any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0180] The above computer-readable storage medium can be included in the fault diagnosis device; it can also exist separately and not be assembled into the fault diagnosis device.

[0181] The above computer-readable storage medium carries one or more programs, which, when executed by a fault diagnosis device, cause the fault diagnosis device to: perform data selection on the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device to determine the target operation data of the drive module; input the module operation characteristics corresponding to the target operation data into a fault diagnosis model, and perform fault diagnosis on the drive module through the fault diagnosis model to determine the current fault information and the predicted module life of the drive module; and obtain a fault diagnosis result of the drive module according to the current fault information and the predicted module life.

[0182] Computer program code for performing the operations of the present application may be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., by connecting through an Internet service provider using the Internet).

[0183] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that, in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system for performing the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0184] The modules involved in the embodiments of the present application can be implemented in software or in hardware. In some cases, the name of the module does not constitute a limitation on the unit itself.

[0185] The readable storage medium provided by the present application is a computer-readable storage medium, and the computer-readable storage medium stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned fault diagnosis method, which can solve the technical problems of low diagnosis efficiency and low diagnosis accuracy rate in the prior art when diagnosing faults in modules. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as those of the fault diagnosis method provided by the above embodiments, and will not be elaborated here.

[0186] The present application further provides a computer program product, including a computer program, and the steps of the above-mentioned fault diagnosis method are implemented when the computer program is executed by a processor.

[0187] The computer program product provided by the present application can solve the technical problems of low diagnosis efficiency and low diagnosis accuracy rate in the prior art when diagnosing faults in modules. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as those of the fault diagnosis method provided by the above embodiments, and will not be elaborated here.

[0188] The above are only some embodiments of the present application, and do not limit the patent scope of the present application. All equivalent structural transformations made under the technical concept of the present application by using the content of the specification and drawings of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A fault diagnosis method, characterized in that, The method includes: Select data from the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device, and determine the target operation data of the drive module; Input the module operation characteristics corresponding to the target operation data into the fault diagnosis model, and perform fault diagnosis on the drive module through the fault diagnosis model to determine the current fault information and the predicted module life of the drive module; Obtain the fault diagnosis result of the drive module according to the current fault information and the predicted module life.

2. The method according to claim 1, characterized in that, Before the step of selecting data from the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device and determining the target operation data of the drive module, it further includes: Collect the initial operation data of the drive module during the operation of the target device according to a preset collection period; Perform data analysis on the initial operation data within the preset collection period to determine at least one acceleration data point existing in the initial operation data and the deceleration data point adjacent to the acceleration data point; Determine at least one module motion segment existing in the initial operation data according to the acceleration data point and the deceleration data point adjacent to the acceleration data point; Select data segments from the initial operation data according to the module motion segment to determine the module operation data of the drive module.

3. The method according to claim 1, wherein Before the step of performing fault diagnosis on the drive module according to the module operation characteristics corresponding to the target operation data and the fault diagnosis model to determine the current fault information and the predicted module life of the drive module, it further includes: Obtain the operation feature set of the drive module, and the operation feature set is composed of operation feature columns of the drive module in multiple operation states; Perform feature selection on the operation feature set according to the target feature columns corresponding to each operation feature column to determine multiple groups of first training features; Train the machine learning model according to multiple groups of first training features and the training operation states corresponding to each first training feature to determine the basic diagnosis model; Optimize the first diagnosis model according to multiple groups of second training features and the training operation states corresponding to each second training feature to obtain the derivative diagnosis model, and the second training features are obtained by numerically adjusting the key feature parameters in the first training features; Perform model integration on the basic diagnosis model and the derivative diagnosis model to obtain the fault diagnosis model of the drive module.

4. The method according to claim 3, wherein Before the step of obtaining the operation feature set of the drive module, it further includes: Perform data preprocessing on the operation source data of the drive module in each operation state to obtain the target source data in each operation state; Perform feature transformation on the target source data in each operation state to determine the original feature data in each operation state; Randomly generate according to the original feature data in each operation state to obtain operation feature columns in multiple operation states; Construct the operation feature set of the drive module according to the operation feature columns in each operation state.

5. The method according to claim 3, wherein The step of performing feature selection on the operation feature set according to the target feature columns corresponding to each operation feature column to determine multiple groups of first training features includes: Calculate the information gain between each operation feature column in the operation feature set and the corresponding target feature column, and obtain the information gain corresponding to each operation feature column; Calculate the mean value of the information gain corresponding to each operation feature column to determine the average gain; Based on the comparison relationship between the average gain and the information gain corresponding to each operation feature column, screen multiple groups of operation feature columns to determine multiple groups of training feature columns; Calculate the importance of multiple groups of training feature columns, and screen multiple groups of training feature columns according to the calculation result of the importance to determine multiple groups of first training features.

6. The method according to claim 5, characterized in that, The step of calculating the importance of multiple groups of training feature columns and screening multiple groups of training feature columns according to the calculation result of the importance to determine multiple groups of first training features includes: Calculate the importance of each training feature column through multiple importance evaluation models to determine the importance score of each training feature column under each importance evaluation model; Calculate the mean value according to the importance scores of each training feature column under each importance evaluation model to determine the average score of each training feature column; Perform normalization processing on the average scores of each training feature column to obtain the target importance score of each training feature column; Sort the target importance scores of each training feature column, and screen the multiple groups of training feature columns according to the sorting result to determine multiple groups of first training features.

7. The method according to any one of claims 1 to 6, characterized in that, Before the step of performing fault diagnosis on the drive module according to the module operation characteristics and the fault diagnosis model corresponding to the target operation data to determine the current fault information and the module predicted life of the drive module, it further includes: Classify the features of the target operation data to determine the first type of data and the second type of data; Perform outlier processing on the first type of data to obtain the first processed data; Perform feature scaling on the first processed data to determine the first operation feature; Perform encoding conversion on the second type of data to determine the second operation feature; Perform feature aggregation according to the first operation feature and the second operation feature to obtain the module operation characteristics corresponding to the target operation data.

8. A fault diagnosis device, characterized in that, The fault diagnosis device includes: A selection module, configured to select the module operation data of the drive module according to the fault impact factor corresponding to the drive module on the target device to determine the target operation data of the drive module; A diagnosis module, configured to perform fault diagnosis on the drive module according to the module operation characteristics and the fault diagnosis model corresponding to the target operation data to determine the current fault information and the module predicted life of the drive module; A processing module, configured to obtain the fault diagnosis result of the drive module according to the current fault information and the module predicted life.

9. A fault diagnosis device, characterized in that, The device includes: a memory, a processor, and a fault diagnosis program stored on the memory and executable on the processor, and the fault diagnosis program is configured to implement the steps of the fault diagnosis method according to any one of claims 1 to 7.

10. A storage medium, characterized in that, A fault diagnosis program is stored on the storage medium, and when the fault diagnosis program is executed by the processor, it implements the steps of the fault diagnosis method according to any one of claims 1 to 7.

Images