Motor fault diagnosis method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of motor diagnosis, and discloses a motor fault diagnosis method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining motor vibration noise big data; correlation analysis is carried out on the vibration noise big data, and parameters causing motor vibration noise faults and parameters of qualified motor vibration noise are screened out; obtaining a vibration noise data set containing the fault data set and the qualified data set; training a preset AI model based on the vibration noise data set to obtain an AI vibration noise diagnosis model; and acquiring a real-time vibration noise data set of the motor, and performing motor vibration noise fault diagnosis based on the real-time vibration noise data set and the AI vibration noise diagnosis model, so that a root cause of the motor vibration noise fault and a method for solving the problem can be directly given.

Description

Technical Field

[0001] The present invention relates to the technical field of motor diagnosis, and in particular to a motor fault diagnosis method, device, computer equipment and storage medium. Background Art

[0002] Existing motor NVH (Noise, Vibration, Harshness) end-of-line testing software's fault diagnosis methods can only extract fault characteristics from raw data, providing specialized terms such as fault frequency and fault order anomaly. However, it cannot identify the abnormal parameters that cause NVH issues in the motor (abnormal parameters include incoming material quality parameters and process support parameters). This diagnostic approach poses significant challenges to motor production line personnel and wastes the production line's big data resources.

[0003] The current existing technology, publication number CN112881014B, discloses an off-line NVH test system for a transmission assembly. The system collects the vibration acceleration signal, sound pressure signal and speed signal of the transmission system for analysis to obtain the sound pressure signal order spectrum and the vibration signal order spectrum. The analysis results are compared with the preset evaluation threshold to determine whether the NVH of the transmission assembly is qualified. The invention uses conventional signal processing methods to perform NVH analysis on the collected signals and determines the NVH status of the motor according to the set threshold. It can only test and diagnose specific regular NVH, and the detection effect is poor for special fault types. The diagnostic results given only stay at whether the NVH status of the motor is OK or NG. It is impossible to give targeted improvement suggestions from the production and manufacturing level, and end-to-end diagnostic effects cannot be achieved.

[0004] Publication number CN110132394A proposed a Labview-based software and hardware system to build a motor NVH evaluation system. The system is simple, with a single function and a simple detection method. It is only suitable for specific work scenarios. The test results cannot provide the root cause of the NVH problem of the motor, and engineers need to conduct further experimental verification and analysis, which cannot effectively reduce human resource costs.

[0005] Publication number CN117310484A discloses a motor noise detection system. This system primarily uses a bone conduction probe sensor to collect motor vibration data and transmits the noise data to a detection circuit assembly for identification, determining product conformity, and identifying the type of defective products. The NVH detection system then extracts data features and uses these features to intercept motors exhibiting abnormal NVH performance. This invention, therefore, only proposes a specific implementation plan for the detection sensor and does not provide any description of the detection method.

[0006] In summary, some existing technologies do not propose effective solutions for detection methods, while some proposed detection methods but all require the pre-extraction of raw data features, and the extracted data features are only effective for specific scenarios. These data features can only provide feedback on NVH problems in the motor, but cannot directly reflect the specific influencing parameters that cause NVH problems in the motor and provide reasonable improvement suggestions. They cannot achieve end-to-end diagnostic effects and cannot fundamentally solve the NVH problems in the motor. Summary of the Invention

[0007] In view of this, the present invention provides a motor fault diagnosis method, device, computer equipment and storage medium to solve the problem in the prior art that only possible fault judgments can be given through data feature extraction, but the fault parameters of the motor causing NVH problems cannot be directly given.

[0008] In a first aspect, the present invention provides a motor fault diagnosis method, the method comprising:

[0009] Obtain big data on motor vibration and noise;

[0010] Conduct correlation analysis on vibration noise big data to screen out the first key parameter that causes motor vibration noise failure and the second key parameter of qualified motor vibration noise;

[0011] The first key parameter is used as a fault data set and the second key parameter is used as a qualified data set, and the fault data set and the qualified data set are aggregated to obtain a vibration noise data set;

[0012] The preset AI model is trained based on the vibration noise dataset to obtain the AI ​​vibration noise diagnosis model;

[0013] Obtain the real-time vibration and noise dataset of the motor, and perform motor vibration and noise fault diagnosis based on the real-time vibration and noise dataset and the AI ​​vibration and noise diagnostic model.

[0014] The present invention provides a motor fault diagnosis method, which obtains big data on vibration and noise during the motor production process, covers motor data in different states, provides a broad and rich source of information for subsequent analysis, and helps to fully understand the vibration and noise characteristics of the motor. It performs correlation analysis on the incoming material quality parameters and the production process parameters, and can accurately screen out the first key parameter that causes the fault and the second key parameter of the corresponding qualified motor. This meticulous classification and screening makes it more accurate to grasp the root cause of the problem and the factors affecting the normal state, and provides high-quality data for subsequent model training. The key parameters after labeling and classification are used as a fault data set and a qualified data set, and then summarized to obtain a vibration and noise data set. This construction method enables the data set to have clear category distinctions, can better adapt to AI model training, and improve the accuracy of the model in identifying different states. The preset AI model is trained based on the constructed vibration and noise data set to obtain an AI model specifically for motor vibration and noise diagnosis. The model is used to diagnose the real-time vibration and noise data sets of actual mass-produced motors, realizing an intelligent and efficient fault diagnosis process. It can quickly and accurately judge motor vibration and noise faults, and can specifically judge the problem parameters of motor vibration and noise faults, which helps to timely discover production problems and take measures to improve product quality and production efficiency. It can directly provide the root cause of motor vibration and noise faults and solutions to the problems, improve motor production efficiency, reduce labor costs, and achieve closed-loop problem solving. It solves the problem that the existing technology can only give possible fault judgments through data feature extraction, and cannot directly give the fault parameters of the motor that cause NVH problems.

[0015] In an optional embodiment, the vibration noise big data includes first vibration noise big data of a motor with a vibration noise fault, second vibration noise big data of a qualified motor, first expert experience big data of a motor with a vibration noise fault that has not been improved after repair, and second expert experience big data of a motor with a vibration noise fault that has been qualified after repair;

[0016] Correlation analysis is performed on the motor material quality parameters and manufacturing process parameters in the vibration noise big data to screen out the first key parameters that lead to motor vibration noise failure and the second key parameters of qualified motor vibration noise, including:

[0017] Extracting incoming material quality parameters and production manufacturing process parameters from the first vibration noise big data, the second vibration noise big data, the first expert experience big data, and the second expert experience big data respectively;

[0018] Performing a correlation analysis on the incoming material quality parameters in the first vibration noise big data and the incoming material quality parameters in the first expert experience big data to obtain the incoming material quality unqualified parameters. At the same time, performing a correlation analysis on the production and manufacturing process parameters in the first vibration noise big data and the production and manufacturing process parameters in the first expert experience big data to obtain the production and manufacturing process unqualified parameters. The incoming material quality unqualified parameters and the production license process unqualified parameters are summarized to obtain the first key parameter.

[0019] A correlation analysis is performed on the incoming material quality parameters in the second vibration noise big data and the incoming material quality parameters in the second expert experience big data to obtain the incoming material quality qualified parameters. At the same time, a correlation analysis is performed on the production and manufacturing process parameters in the second vibration noise big data and the production and manufacturing process parameters in the second expert experience big data to obtain the production and manufacturing process qualified parameters. The incoming material quality qualified parameters and the production license process qualified parameters are summarized to obtain the second key parameters.

[0020] The present invention provides a motor fault diagnosis method. Its vibration and noise big data encompasses qualified motors, faulty motors, and expert experience data from different repair results, providing ample samples for comprehensive analysis. By performing correlation analysis on incoming material quality parameters and manufacturing process parameters in different data types, qualified and unqualified parameters are accurately extracted. This meticulous data classification and utilization approach can deeply explore the inherent connections between motors and related parameters under different conditions, providing strong support for fault diagnosis and quality improvement. During the parameter screening process, the vibration and noise big data for qualified and faulty motors are correlated with the corresponding expert experience big data to accurately determine qualified and unqualified parameters, which are then summarized to obtain key parameters. This allows for more precise identification of the causes of motor vibration and noise failures, clearly distinguishing between incoming material quality and specific factors in the manufacturing process that lead to failure. By incorporating data from different sources and in different states into the analysis system, the motor vibration and noise problem is considered from multiple dimensions. The analysis not only focuses on the faulty motor itself but also references case studies of motors that passed and did not improve after repair, resulting in a more comprehensive analysis of motor vibration and noise failures. By obtaining key parameters through a rigorous correlation analysis process, the reliability of the fault diagnosis basis is greatly improved, laying a solid foundation for subsequent model training and practical application.

[0021] In an optional embodiment, correlation analysis is performed on the vibration noise big data to screen out the first key parameter that causes the motor vibration noise fault and the second key parameter of the qualified motor vibration noise, further comprising:

[0022] Perform one-to-one labeling and classification on the unqualified incoming material quality parameters in the first vibration noise big data in the first key parameter and the unqualified incoming material quality parameters in the first expert experience big data; at the same time, perform one-to-one labeling and classification on the unqualified production and manufacturing process parameters in the first vibration noise big data in the first key parameter and the unqualified production and manufacturing process parameters in the first expert experience big data;

[0023] The incoming material quality qualified parameters in the second vibration noise big data in the second key parameters and the incoming material quality qualified parameters in the second expert experience big data are labeled and classified one-to-one. At the same time, the production and manufacturing process qualified parameters in the second vibration noise big data in the second key parameters and the production and manufacturing process qualified parameters in the second expert experience big data are labeled and classified one-to-one.

[0024] The present invention provides a motor fault diagnosis method, which can accurately establish the correspondence between the first vibration noise big data and the qualified parameters in the first expert experience big data, as well as the connection between the second vibration noise big data and the unqualified parameters in the second expert experience big data through one-to-one labeling and classification. This makes it possible to clearly trace the source and evolution process of each parameter in subsequent analysis. When a vibration noise problem occurs, the original data and expert experience related to it can be quickly located, providing a convenient path for in-depth analysis of the cause of the fault. One-to-one precise matching labeling avoids parameter confusion and greatly improves the accuracy of the data. In a complex motor production and manufacturing data system, this method ensures that each qualified or unqualified parameter can be accurately associated with the corresponding experience data, providing a reliable data foundation for subsequent model training, fault diagnosis and other work based on these data, and reducing the risk of misjudgment due to data errors or mismatches. When providing training data for the AI ​​model, this kind of data that has been carefully labeled and classified can enable the model to better learn the difference characteristics between qualified and unqualified parameters. The model can more accurately identify the key factors that cause vibration and noise failures, thereby improving the model's diagnostic accuracy and generalization ability, enabling the trained AI model to more reliably diagnose and predict motor vibration and noise failures in practical applications.

[0025] In an optional embodiment, before training the preset AI model based on the vibration noise dataset, the motor fault diagnosis method further includes:

[0026] The vibration noise dataset is divided into a training dataset and a test dataset according to a preset ratio.

[0027] In an optional embodiment, a preset AI model is trained and tested based on a vibration noise dataset to obtain an AI vibration noise diagnostic model, including:

[0028] Input the training data set into the preset AI model for training to obtain the AI ​​model to be tested;

[0029] The test data set is input into the AI ​​model to be tested for test verification to obtain the test results. When the accuracy of the test results meets the preset test threshold, the AI ​​model is used as the AI ​​vibration noise diagnosis model.

[0030] The present invention provides a motor fault diagnosis method, which inputs a training data set into a preset AI model for training, allowing the model to fully learn the characteristics and patterns contained in the qualified and unqualified parameters in the vibration noise data set. A large amount of diverse data input prompts the model to continuously optimize its own parameters, thereby having stronger pattern recognition capabilities. After verification with a test data set, the model training effect can be objectively evaluated. When the accuracy of the test results meets the preset test threshold, it means that the model has been able to accurately capture the key factors that cause vibration noise faults, effectively improving the accuracy of model diagnosis and providing a reliable basis for fault diagnosis in actual motor production.

[0031] In an optional embodiment, motor vibration noise fault diagnosis is performed based on a real-time vibration noise dataset and an AI vibration noise diagnostic model, including:

[0032] The real-time vibration noise data set is input into the AI ​​vibration noise diagnosis model for vibration noise fault diagnosis, and the qualified state of the motor without vibration noise fault and the unqualified state with vibration noise fault are obtained.

[0033] This invention provides a motor fault diagnosis method that promptly inputs real-time vibration and noise data sets into an AI-powered vibration and noise diagnostic model, enabling continuous, real-time monitoring of the motor's operating status. If a motor vibration or noise fault occurs, the model quickly responds and outputs a diagnostic result indicating a failure. This allows businesses to identify problems immediately and take swift action, effectively preventing further deterioration and reducing production losses due to downtime for repairs.

[0034] In an optional implementation, the motor fault diagnosis method further includes:

[0035] When an unqualified state with a vibration noise fault is obtained, the fault parameters of the motor generating the vibration noise fault are analyzed, and the fault parameters are compared with the qualified motor vibration noise parameters in a preset database to obtain the change difference between the fault parameters and the qualified motor vibration noise parameters.

[0036] The present invention provides a motor fault diagnosis method that analyzes the fault parameters that produce vibration and noise faults, and compares them with qualified parameters in a preset database to obtain a change difference, which can clearly identify the specific parameter deviation that causes the motor fault. Compared with simply judging that the motor is in an unqualified state, this method goes further into the core of the fault, allowing technicians to quickly focus on the problem, providing a precise direction for the subsequent formulation of targeted repair plans, and greatly improving the efficiency of fault detection and resolution. The change difference presents the degree to which the fault parameters deviate from the qualified standards in the form of specific numerical values. This allows enterprises to have a quantitative understanding of the severity of the fault, and is no longer limited to vague fault judgments.

[0037] In a second aspect, the present invention provides a motor fault diagnosis device, the device comprising:

[0038] Vibration and noise big data acquisition module, used to obtain motor vibration and noise big data;

[0039] A correlation analysis module is used to perform correlation analysis on vibration noise big data to screen out the first key parameter that causes motor vibration noise failure and the second key parameter of qualified motor vibration noise;

[0040] a vibration and noise data set determination module, configured to use the first key parameter as a fault data set and the second key parameter as a qualified data set, and to aggregate the fault data set and the qualified data set to obtain a vibration and noise data set;

[0041] The training and testing module is used to train and test the preset AI model based on the vibration and noise dataset to obtain the AI ​​vibration and noise diagnostic model;

[0042] The fault diagnosis module is used to obtain the real-time vibration and noise data set of the motor and perform motor vibration and noise fault diagnosis based on the real-time vibration and noise data set and the AI ​​vibration and noise diagnosis model.

[0043] In a third aspect, the present invention provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the motor fault diagnosis method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0044] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the motor fault diagnosis method of the first aspect or any corresponding embodiment thereof.

[0045] In a fifth aspect, the present invention provides a computer program product, comprising computer instructions for causing a computer to execute the motor fault diagnosis method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 is a flow chart of a motor fault diagnosis method according to an embodiment of the present invention;

[0048] Figure 2 is a flow chart of another motor fault diagnosis method according to an embodiment of the present invention;

[0049] Figure 3 is a flow chart of another motor fault diagnosis method according to an embodiment of the present invention;

[0050] Figure 4 is a flow chart of another motor fault diagnosis method according to an embodiment of the present invention;

[0051] Figure 5 is a schematic diagram of the operation process of the motor fault diagnosis method according to an embodiment of the present invention;

[0052] Figure 6 is a structural block diagram of a motor fault diagnosis device according to an embodiment of the present invention;

[0053] Figure 7 Schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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 those skilled in the art without making creative work are within the scope of protection of the present invention. In addition, the terms "first", "second" and "third" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance.

[0055] According to an embodiment of the present invention, an embodiment of a motor fault diagnosis method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0056] In this embodiment, a motor fault diagnosis method is provided, which can be used in the above-mentioned computer equipment. Figure 1 FIG. 1 is a flow chart of a motor fault diagnosis method according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0057] Step S101: Obtaining motor vibration noise big data.

[0058] Specifically, motor vibration and noise big data refers to the collection, analysis and processing of various data related to motor vibration and noise. Vibration acceleration, noise spectrum, speed and other parameters are collected in real time at key parts of the motor (stator, rotor, bearings, etc.) through high-precision sensors (such as MEMS microphones, accelerometers, laser vibrometers). The sampling frequency is usually higher than the motor operating frequency to capture high-frequency details.

[0059] Furthermore, high-precision vibration and noise sensors are installed at key locations on the motor production line. Accelerometers can be used as vibration sensors, with their range and sensitivity appropriately set based on the motor's vibration characteristics to ensure accurate capture of subtle vibration changes. Directional or omnidirectional microphones are used as noise sensors, with their gain and frequency response adjusted based on the ambient noise level in the production workshop and the intensity of the motor noise. The microphones should be mounted at an appropriate distance from the motor to effectively capture motor noise while minimizing ambient noise interference. For example, they should be mounted horizontally or vertically at a distance of 1 meter from the motor. Data collection is performed throughout the entire motor production process, from the beginning of motor component processing to the testing phase after finished product assembly. During component processing, such as core stamping and winding, data is collected for a longer period of time, such as 30 seconds, after each key assembly step to comprehensively document the impact of assembly on motor vibration and noise. During motor performance testing, vibration and noise data are collected at a high frequency, in the millisecond range, under simulated operating conditions, such as varying loads and speeds. Because the motor's operating state is complex, high-frequency acquisition can capture subtle dynamic changes.

[0060] In addition to real-time data collected directly from sensors, other relevant data from the production process is also integrated. This includes manufacturing process parameter data, such as core lamination pressure and time, winding speed and turn accuracy, and the installation position accuracy and assembly torque of various components during motor assembly. Fluctuations in these process parameters can cause changes in the motor's structural stability, generating vibration and noise.

[0061] The motor vibration and noise big data in this embodiment can also include vibration and noise big data from qualified motors, vibration and noise big data from motors with NVH issues, and expert experience big data from motors with NVH issues that have been improved after repair. The motor vibration and noise big data is collected in the form of vibration and noise time domain signals and is structured and stored using database management software.

[0062] Step S102 , performing correlation analysis on the vibration noise big data, screening out the first key parameter that causes the motor vibration noise fault and the second key parameter that is qualified for the motor vibration noise.

[0063] For example, correlation analysis is performed on the motor incoming material quality parameters and production manufacturing process parameters in the vibration noise big data. The motor incoming material quality parameters are the actual incoming material size parameters of the motor, including: cylindricity, coaxiality, length, verticality, etc. related to the stator, rotor, housing, shaft and other components; the motor production manufacturing process parameters include: core stacking pressure, rotor dynamic balance, bolt tightening torque, assembly alignment, etc.

[0064] A preliminary analysis of incoming motor material quality parameters, such as the perpendicularity of components like the stator, rotor, housing, and shaft, as well as manufacturing process parameters like core stacking pressure and rotor dynamic balance, was conducted using the Pearson correlation coefficient method. This method calculates the linear correlation between two sets of data to produce a correlation coefficient ranging from -1 to 1. For example, the correlation coefficient between stator and rotor perpendicularity and motor vibration noise was calculated. A coefficient close to 1 or -1 indicates a strong linear correlation; a coefficient close to 0 indicates a weak correlation.

[0065] A faulty motor vibration and noise dataset was established, including the incoming material quality and manufacturing process parameters of motors with obvious vibration and noise faults during production. A correlation coefficient threshold was set, such as parameters with an absolute value of the Pearson correlation coefficient greater than 0.6 or an absolute value of the Spearman rank correlation coefficient greater than 0.7 as potential key parameters. For example, if the Pearson correlation coefficient between the stator and rotor perpendicularity and the faulty motor vibration and noise is 0.75, exceeding the set threshold, then this parameter is included in the list of potential first key parameters. For example, in multiple faulty motors, it was found that when the core stacking pressure was below a certain value, the vibration and noise increased significantly, and this parameter performed outstandingly in the correlation coefficient analysis. After comprehensive judgment, the core stacking pressure was determined to be one of the first key parameters causing the fault.

[0066] A qualified motor vibration and noise data set was constructed, encompassing relevant parameters for motors without vibration or noise failures during production. Using the same correlation analysis method described above, the correlation coefficients between incoming material quality and manufacturing process parameters and qualified motor vibration and noise were calculated. Different thresholds were set, such as the absolute value of the Pearson correlation coefficient between 0.2 and 0.5. Potential parameters were evaluated from the perspectives of production stability and quality consistency. After carefully screening potential parameters, secondary key parameters, such as stator and rotor perpendicularity, were identified as crucial for maintaining acceptable motor vibration and noise levels.

[0067] In step S103 , the first key parameter is used as a fault data set and the second key parameter is used as a qualified data set, and the fault data set and the qualified data set are aggregated to obtain a vibration noise data set.

[0068] Specifically, the first key parameter was organized and systematically integrated into a fault dataset. A qualified dataset was constructed for the second key parameter. It can be seen that the first and second key parameters are one or more of the motor's incoming material quality parameters and manufacturing process parameters. The fault dataset and qualified dataset were combined to form a vibration and noise dataset.

[0069] Step S104: training a preset AI model based on the vibration noise data set to obtain an AI vibration noise diagnostic model.

[0070] Specifically, the preset AI models include deep learning network models, machine learning models, etc., and the AI ​​model parameters include the number of neural network layers, the number of neurons, activation functions, learning rates, weight vectors, etc. The structures of deep learning network models and machine learning models and the functions of each structure can be found in mature related technologies and will not be repeated here.

[0071] For the convenience of training and testing, the vibration noise data set can be divided into a training data set and a test training data set. The training data set is input into the preset AI model for training to obtain a trained AI model. The accuracy of the trained AI model is then tested using the test data set, and the trained AI model with the load test threshold is used as the AI ​​vibration noise diagnostic model.

[0072] Step S105 , obtaining a real-time vibration noise data set of the motor, and performing motor vibration noise fault diagnosis based on the real-time vibration noise data set and the AI ​​vibration noise diagnosis model.

[0073] Specifically, the AI ​​model is applied to the motor NVH offline detection system, and the vibration and noise data collected in real time by the detection system is input into the AI ​​vibration and noise diagnosis model for intelligent data diagnosis to determine qualified motors and motors with NVH problems.

[0074] The motor fault diagnosis method provided in this embodiment, by acquiring big data on vibration and noise during the motor production and manufacturing process, covers motor data in different states, provides a broad and rich source of information for subsequent analysis, helps to fully understand the vibration and noise characteristics of the motor, and performs correlation analysis on the incoming material quality parameters and the production and manufacturing process parameters respectively, and can accurately screen out the first key parameter that causes the fault and the second key parameter of the corresponding qualified motor. This meticulous classification and screening makes it more accurate to grasp the root cause of the problem and the factors affecting the normal state, and provides high-quality data for subsequent model training. The key parameters after labeling and classification are used as a fault data set and a qualified data set, and then summarized to obtain a vibration and noise data set. This construction method enables the data set to have clear category distinctions, can better adapt to AI model training, and improve the accuracy of the model in identifying different states. The preset AI model is trained based on the constructed vibration and noise data set to obtain an AI model specifically for motor vibration and noise diagnosis. The model is used to diagnose the real-time vibration and noise data sets of actual mass-produced motors, realizing an intelligent and efficient fault diagnosis process. It can quickly and accurately judge motor vibration and noise faults, and can specifically judge the problem parameters of motor vibration and noise faults, which helps to timely discover production problems and take measures to improve product quality and production efficiency. It can directly provide the root cause of motor vibration and noise faults and solutions to the problems, improve motor production efficiency, reduce labor costs, and achieve closed-loop problem solving. It solves the problem that the existing technology can only give possible fault judgments through data feature extraction, and cannot directly give the fault parameters of the motor that cause NVH problems.

[0075] In this embodiment, a motor fault diagnosis method is provided, which can be used in the above-mentioned computer equipment. Figure 2 FIG. 1 is a flow chart of a motor fault diagnosis method according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0076] Step S201: Obtain motor vibration noise big data. Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0077] Step S202 , performing correlation analysis on the vibration noise big data, screening out the first key parameter that causes the motor vibration noise fault and the second key parameter of the qualified motor vibration noise.

[0078] Specifically, the vibration noise big data includes first vibration noise big data of a vibration noise fault, second vibration noise big data of a qualified motor, first expert experience big data of a vibration noise fault motor that has not improved after repair, and second expert experience big data of a vibration noise fault motor that has passed repair; wherein the expert experience big data includes: the motor production serial number, the motor time domain signal, the motor NVH status OK or NG, the specific reasons and optimization measures leading to the motor NVH status NG, such as the motor rotor cylindricity size problem, the motor housing and stator cylindricity problem, etc. The above step S202 includes:

[0079] Step S2021 , extracting incoming material quality parameters and production manufacturing process parameters from the first vibration noise big data, the second vibration noise big data, the first expert experience big data, and the second expert experience big data respectively.

[0080] Step S2022, perform correlation analysis on the incoming material quality parameters in the first vibration noise big data and the incoming material quality parameters in the first expert experience big data to obtain the incoming material quality unqualified parameters, and at the same time perform correlation analysis on the production and manufacturing process parameters in the first vibration noise big data and the production and manufacturing process parameters in the first expert experience big data to obtain the production and manufacturing process unqualified parameters, and summarize the incoming material quality unqualified parameters and the production license process unqualified parameters to obtain the first key parameters.

[0081] In an optional embodiment, the correlation analysis algorithm uses the Pearson correlation coefficient method, and the above step S2022 includes:

[0082] The first set of correlation analysis (finding unqualified parameters):

[0083] Correlation Analysis of Incoming Material Quality Parameters: The Spearman Rank Correlation Coefficient method was used to analyze incoming material quality parameters from the first vibration noise big data and the first expert experience big data. Because parameter relationships can be more complex in fault conditions, the Spearman Rank Correlation Coefficient method can better capture nonlinear relationships.

[0084] Correlation analysis of manufacturing process parameters: Using the Spearman rank correlation coefficient method, we calculated the correlation between the manufacturing process parameters in the first vibration noise big data and the corresponding parameters in the first expert experience big data. For example, we analyzed the rank correlation between the winding speed in the two data sets.

[0085] Determine unqualified parameters: Set an appropriate Spearman rank correlation coefficient threshold. For example, if the absolute value is greater than 0.65, the parameter is determined to be unqualified for incoming material quality or manufacturing process. Parameters that meet the criteria are screened for subsequent summary.

[0086] Step S2023, perform correlation analysis on the incoming material quality parameters in the second vibration noise big data and the incoming material quality parameters in the second expert experience big data to obtain the incoming material quality qualified parameters, and at the same time perform correlation analysis on the production and manufacturing process parameters in the second vibration noise big data and the production and manufacturing process parameters in the second expert experience big data to obtain the production and manufacturing process qualified parameters, and summarize the incoming material quality qualified parameters and the production license process qualified parameters to obtain the second key parameters.

[0087] The second set of correlation analysis (to find qualified parameters):

[0088] Correlation Analysis of Incoming Material Quality Parameters: Using the Pearson correlation coefficient method, we calculated the correlation coefficient between incoming material quality parameters in the second vibration and noise big data set and those in the second expert experience big data set. Correlation Analysis of Manufacturing Process Parameters: Also using the Pearson correlation coefficient method, we analyzed the correlation between manufacturing process parameters in the second vibration and noise big data set and the corresponding parameters in the second expert experience big data set. For example, we calculated the correlation coefficient between the core stacking time in the two data sets to determine its impact on the qualified motor manufacturing process.

[0089] Determine qualified parameters: Set a correlation coefficient threshold. For example, parameters with an absolute value of the Pearson correlation coefficient greater than 0.7 are considered qualified parameters for incoming material quality or manufacturing process. Parameters that meet the criteria are screened out for subsequent summary.

[0090] In step S2024, the unqualified incoming material quality parameters in the first vibration noise big data in the first key parameter and the unqualified incoming material quality parameters in the first expert experience big data are labeled and classified one-to-one. At the same time, the unqualified production and manufacturing process parameters in the first vibration noise big data in the first key parameter and the unqualified production and manufacturing process parameters in the first expert experience big data are labeled and classified one-to-one.

[0091] Labeling of parameters indicating substandard incoming material quality: Within the first key parameter set, the substandard incoming material quality parameters in the first vibration and noise big data set are accurately matched one-to-one with the corresponding parameters in the first expert experience big data set, based on parameter name. Each matching pair of parameters is then labeled in detail based on the specific cause and mechanism of the motor vibration and noise failure. For example, if the stator's verticality is substandard, causing increased electromagnetic vibration and noise in the motor, the label might be "Stator Verticality - Substandard Verticality Leading to Electromagnetic Vibration and Noise Failure - Substandard."

[0092] Labeling of unqualified manufacturing process parameters: A one-to-one match is performed between the first vibration and noise big data in the first key parameter and the first expert experience big data for unqualified manufacturing process parameters, and these parameters are labeled accordingly. For example, if improper tension control during winding leads to uneven winding turns, causing mechanical vibration and noise in the motor, the label might be "Winding Tension - Uneven Turns Leading to Mechanical Vibration and Noise Fault - Unqualified."

[0093] Step S2025, perform one-to-one labeling and classification on the incoming material quality qualified parameters in the second vibration noise big data in the second key parameters and the incoming material quality qualified parameters in the second expert experience big data, and at the same time perform one-to-one labeling and classification on the production and manufacturing process qualified parameters in the second vibration noise big data in the second key parameters and the production and manufacturing process qualified parameters in the second expert experience big data.

[0094] The labeling of incoming material quality parameters involves a one-to-one match between the incoming material quality parameters in the second vibration and noise big data and the corresponding parameters in the second expert experience big data within the second key parameter set. Labeling is performed based on the parameters' role in ensuring the proper operation and performance of qualified motors.

[0095] Labeling of qualified manufacturing process parameters: A one-to-one match is performed between the qualified manufacturing process parameters in the second key parameter, the second vibration and noise big data, and the second expert experience big data. For example, precise pressure and time control during the core lamination process ensures a tight bond between the cores and effectively reduces the vibration and noise of the motor. This label could be "Core lamination pressure and time control - key parameters for ensuring tight bond between the cores and reducing vibration and noise - qualified."

[0096] Step S203: The first key parameter is used as a fault data set and the second key parameter is used as a qualified data set, and the fault data set and the qualified data set are aggregated to obtain a vibration noise data set. Figure 1 Step S103 of the illustrated embodiment will not be described in detail here.

[0097] Step S204: Train the preset AI model based on the vibration noise data set to obtain an AI vibration noise diagnosis model. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0098] Step S205: Acquire the real-time vibration noise data set of the motor, and perform motor vibration noise fault diagnosis based on the real-time vibration noise data set and the AI ​​vibration noise diagnosis model. Figure 1 Step S105 of the illustrated embodiment will not be described in detail here.

[0099] The motor fault diagnosis method provided in this embodiment utilizes vibration and noise big data encompassing qualified motors, faulty motors, and expert experience data from different repair results, providing ample samples for comprehensive analysis. By performing correlation analysis on incoming material quality parameters and manufacturing process parameters within different data types, qualified and unqualified parameters are accurately extracted. This meticulous data classification and utilization approach allows for in-depth exploration of the inherent connections between motors in different states and related parameters, providing strong support for fault diagnosis and quality improvement. During the parameter screening process, correlation analysis is performed on the vibration and noise big data for qualified and faulty motors, along with the corresponding expert experience big data, to accurately determine qualified and unqualified parameters, which are then aggregated to form key parameters. This allows for more precise identification of the causes of motor vibration and noise failures, enabling clear distinction between incoming material quality and specific factors contributing to the failure within the manufacturing process. By incorporating data from different sources and states into the analysis system, the motor vibration and noise problem is considered from multiple perspectives. This approach not only focuses on the faulty motor itself but also references case studies of motors that passed and failed repairs, resulting in a more comprehensive analysis of motor vibration and noise failures.

[0100] Through a rigorous correlation analysis process, key parameters are derived, significantly improving the reliability of fault diagnosis and laying a solid foundation for subsequent model training and practical application. Through one-to-one labeling and classification, a precise correspondence is established between the first vibration and noise big data and the qualified parameters in the first expert experience big data, as well as between the second vibration and noise big data and the unqualified parameters in the second expert experience big data. This enables subsequent analysis to clearly trace the source and evolution of each parameter. When vibration and noise issues arise, the relevant original data and expert experience can be quickly located, providing a convenient path for in-depth analysis of the fault cause. Precise one-to-one labeling avoids parameter confusion and significantly improves data accuracy. Within the complex motor manufacturing data system, this approach ensures that each qualified or unqualified parameter is accurately associated with the corresponding empirical data, providing a reliable data foundation for subsequent model training and fault diagnosis based on this data, reducing the risk of misjudgment due to data errors or mismatches. When providing training data for AI models, this carefully labeled and classified data allows the model to better learn the distinguishing characteristics between qualified and unqualified parameters. The model can more accurately identify the key factors that cause vibration and noise failures, thereby improving the model's diagnostic accuracy and generalization ability, enabling the trained AI model to more reliably diagnose and predict motor vibration and noise failures in practical applications.

[0101] In this embodiment, a motor fault diagnosis method is provided, which can be used in the above-mentioned computer equipment. Figure 3 FIG. 1 is a flow chart of a motor fault diagnosis method according to an embodiment of the present invention. Figure 3 As shown, the process includes the following steps:

[0102] Step S301: Obtain motor vibration noise big data. Figure 2 Step S201 of the illustrated embodiment will not be described in detail here.

[0103] Step S302: perform correlation analysis on the vibration noise big data to screen out the first key parameter that causes the motor vibration noise fault and the second key parameter that is qualified for the motor vibration noise. Figure 2 Step S202 of the illustrated embodiment will not be described in detail here.

[0104] Step S303: The first key parameter is used as a fault data set and the second key parameter is used as a qualified data set, and the fault data set and the qualified data set are aggregated to obtain a vibration noise data set. Figure 2 Step S203 of the illustrated embodiment will not be described in detail here.

[0105] Step S304: Divide the vibration noise dataset into a training dataset and a test dataset according to a preset ratio.

[0106] Specifically, the preset ratio is set according to actual conditions and is not specifically limited here. For example, the training data set and the test data set are set according to 7:3.

[0107] Step S305: Training and testing the preset AI model based on the vibration noise data set to obtain an AI vibration noise diagnostic model.

[0108] Specifically, the above step S305 includes:

[0109] Step S3051: Input the training data set into the preset AI model for training to obtain the AI ​​model to be tested.

[0110] Specifically, if Figure 5 As shown, the actual collected motor vibration noise time domain signal is marked as OK for qualified motors and NG for motors with NVH problems. At the same time, the cause of the failure of the NG motor is described. The cause of the failure should be specific to the incoming material quality parameters and production process parameters of the components. Finally, the labeled and classified vibration noise dataset is divided into the training dataset and test dataset required for AI model training. For example, when the preset AI model adopts a deep learning network model, the training dataset is input into the deep learning network model for training, and the training results are output, that is, an iterative training process is completed, and it is iterated in sequence until the iteration termination condition is met. The iteration is terminated. The iteration termination condition can be set to a preset algebra or a set loss function threshold. The trained AI model is used as the AI ​​model to be tested.

[0111] Step S3052: Input the test data set into the AI ​​model to be tested for testing and verification to obtain the test results. When the accuracy of the test results meets the preset test threshold, the AI ​​model is used as the AI ​​vibration noise diagnosis model.

[0112] Specifically, the preset test threshold can be set based on actual conditions and is not specifically limited here. It can be set to a diagnostic accuracy of 99% or greater. When the test data set is input into the AI ​​model to be tested for verification and test results are obtained, the AI ​​model corresponding to the test result being 99% or greater is used as the AI ​​vibration noise diagnostic model.

[0113] Step S306 , obtaining a real-time vibration noise data set of the motor, and performing motor vibration noise fault diagnosis based on the real-time vibration noise data set and the AI ​​vibration noise diagnostic model.

[0114] Specifically, the real-time vibration noise data set is input into the AI ​​vibration noise diagnosis model to perform vibration noise fault diagnosis, and obtain the qualified state of the motor without vibration noise fault and the unqualified state with vibration noise fault.

[0115] Furthermore, the AI ​​model is applied to the motor NVH offline detection system, and the vibration and noise data collected by the detection system in real time are input into the neural network model for data intelligent diagnosis, and finally the motor NVH is determined to be qualified or unqualified. Figure 5 As shown, the motor NVH is in an unqualified state, including: abnormal rotor dynamic balance, abnormal stator paint dripping, abnormal rotor core stacking pressure, abnormal stator verticality size, abnormal rotor core stacking, abnormal bolt tightening torque, abnormal casing coaxiality size and abnormal motor assembly alignment, etc.

[0116] Step S307, when an unqualified state of vibration noise fault is obtained, the fault parameters of the motor generating the vibration noise fault are analyzed, and the fault parameters are compared with the qualified motor vibration noise parameters in the preset database to obtain the change difference between the fault parameters and the qualified motor vibration noise parameters.

[0117] Specifically, when the diagnosed motor is in an unqualified state, the specific reasons that cause the motor to produce NVH problems must be given: such as incoming material quality parameters or process parameters of the manufacturing process, and the problem parameters must be compared with the OK motor parameters in the database, and the specific changes in the parameters must be given.

[0118] Furthermore, the system connects to a pre-set database of qualified motor vibration and noise parameters, which contains a large number of standard parameters for various motor models under various operating conditions. Based on key information such as the faulty motor's model and production batch, the system performs a precise search in the database to obtain the corresponding qualified motor parameter set.

[0119] The determined fault parameters are matched one by one with the qualified parameters retrieved from the database to ensure the consistency and comparability of the compared parameters. For example, the pressure change curve during the core stacking process is matched using the same time node and pressure measurement unit. According to the type and characteristics of the parameter, select an appropriate method to calculate the change difference. For numerical parameters, directly subtract the standard value of the qualified motor from the measured value of the faulty motor to obtain the change difference. For parameters with range attributes, such as the content range of a certain component of the core material, calculate the difference between the actual content of the component in the faulty motor and the boundary value of the qualified range to determine the degree of deviation from the qualified range. For curve parameters, such as the core stacking pressure curve, a curve fitting error algorithm is used to calculate the fitting error between the faulty motor pressure curve and the qualified motor standard pressure curve as the change difference between the two.

[0120] The motor fault diagnosis method provided in this embodiment uses a training dataset to input a preset AI model for training, enabling the model to fully learn the characteristics and patterns inherent in the qualified and unqualified parameters in the vibration and noise dataset. This large and diverse data input enables the model to continuously optimize its parameters, thereby enhancing its pattern recognition capabilities. Subsequently, validation with a test dataset allows for objective evaluation of the model training results. When the accuracy of the test results meets the preset test threshold, it indicates that the model has accurately captured the key factors leading to vibration and noise failures, effectively improving the accuracy of the model diagnosis and providing a reliable basis for fault diagnosis in actual motor production. Promptly inputting real-time vibration and noise datasets into the AI ​​vibration and noise diagnosis model enables uninterrupted, real-time monitoring of the motor's operating status. Once a vibration and noise failure occurs in the motor, the model reacts quickly and outputs a diagnosis result indicating an unqualified condition. This enables companies to detect problems immediately and take swift action, effectively preventing further deterioration of the failure and reducing production losses caused by downtime for maintenance. By analyzing the fault parameters causing the vibration and noise failure and comparing them with qualified parameters in a preset database to determine the difference in change, the specific parameter deviations that caused the motor failure can be clearly identified. Compared to simply determining that the motor is unqualified, this method goes deeper into the core of the fault, allowing technicians to quickly focus on the problem and provide precise guidance for developing targeted repair plans, greatly improving the efficiency of troubleshooting and resolution. The variation difference provides a specific numerical value indicating the degree to which the fault parameter deviates from the qualified standard. This allows companies to gain a quantitative understanding of the severity of the fault, eliminating the need for vague fault diagnosis.

[0121] As one or more specific application embodiments of the present invention, combined with Figure 4 The motor fault diagnosis method provided by the present invention is further described in detail. Figure 4 The specific process steps are as follows:

[0122] Step S1, (1) collects big data on vibration and noise without NVH problems in the motor manufacturing process, (2) collects big data on vibration and noise with NVH problems in the motor manufacturing process, (3) collects big data on expert experience of motors with NVH problems that have been improved after repair, and uses database management software to perform structured storage of big data, where the expert experience big data is the experience big data of solving motor NVH problems by improving incoming material quality parameters and production process parameters.

[0123] Among them, the collected big data without NVH faults and big data with NVH faults are the vibration noise time domain signals collected by the test equipment during NVH offline testing.

[0124] Step S2: Use big data analysis methods to perform correlation analysis on the incoming material quality parameters and production manufacturing process parameters in the vibration noise big data, retain the main parameters that cause NVH failures, and use the vibration noise fault data set (including NVH fault big data and fault big data in expert experience big data) and qualified fault set (including non-NVH fault big data and qualified big data in expert experience big data) as the vibration noise data set, and label and classify the incoming material quality parameters and production manufacturing process parameters in the vibration noise data set, and divide the data set into the training data set and test data set required by the AI ​​model.

[0125] Among them, the incoming material quality parameters are the actual incoming material size parameters of the motor, including: cylindricity, coaxiality, length, verticality, etc. related to components such as the stator, rotor, housing, and shaft; manufacturing process parameters include: core stacking pressure, rotor dynamic balance, bolt tightening torque, assembly alignment, etc.

[0126] Step S21: Using the big data analysis method to perform correlation analysis on the incoming material quality parameter and production process parameter data set is as follows:

[0127] The actual incoming material quality parameters of the components are correlated with the manufacturing process parameters. Completely independent parameters are retained, and mutually correlated parameters are discarded. At the same time, the parameters are marked with the motor NVH problems caused by the parameters.

[0128] Step S22: Label and classify the vibration noise fault dataset and the repair expert experience dataset, and divide the dataset into the training dataset and test dataset required by the AI ​​model, specifically:

[0129] The actual motor vibration and noise time domain signals collected are marked as OK for qualified motors and NG for motors with NVH problems. At the same time, the fault causes of NG motors are described, and the fault causes are specific to the incoming material quality parameters and process manufacturing parameters of the components. Finally, the labeled and classified large data sets are divided into training data sets and test data sets required for AI model training.

[0130] In step S3, an AI model is used to train the classified and labeled vibration noise dataset, and the trained AI model is tested and verified using a test dataset to ensure that the model's diagnostic accuracy is above 99%, thereby determining the specific parameters of the AI ​​model and using the AI ​​model that passes the test as the AI ​​vibration noise diagnostic model.

[0131] AI models include deep learning network models, machine learning models, etc. AI model parameters include the number of neural network layers, the number of neurons, activation functions, learning rates, weight vectors, etc.

[0132] In step S4, the AI ​​vibration and noise diagnostic model is applied to the actual mass-produced motor NVH off-line inspection to perform motor NVH fault diagnosis, directly determine the key parameters that cause the motor NVH problem, and provide improvement measures.

[0133] Specifically, the AI ​​vibration and noise diagnostic model is applied to the actual NVH end-of-line testing of mass-produced motors to diagnose motor NVH faults. This directly identifies the key parameters that cause motor NVH problems and provides improvement measures, specifically:

[0134] The AI ​​vibration and noise diagnosis model is applied to the motor NVH offline detection system. The vibration and noise data collected by the detection system in real time is input into the neural network model for intelligent data diagnosis. Finally, it is determined whether the motor NVH is qualified or unqualified. When the diagnosed motor is unqualified, the specific reasons that cause the motor NVH problem need to be given: such as the quality parameters of the incoming materials or the process parameters of the manufacturing process, and the problem parameters are compared with the OK motor parameters in the database, and the specific changes in the parameters are given.

[0135] This embodiment also provides a motor fault diagnosis device for implementing the above-mentioned embodiments and preferred embodiments. Details already described will not be repeated. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0136] This embodiment provides a motor fault diagnosis device, such as Figure 6 Shown, including:

[0137] The vibration and noise big data acquisition module 601 is used to acquire motor vibration and noise big data.

[0138] The correlation analysis module 602 is used to perform correlation analysis on the vibration noise big data to screen out the first key parameter that causes the motor vibration noise fault and the second key parameter of the qualified motor vibration noise.

[0139] The vibration and noise data set determination module 603 is configured to use the first key parameter as a fault data set and the second key parameter as a qualified data set, and aggregate the fault data set and the qualified data set to obtain a vibration and noise data set.

[0140] The training and testing module 604 is used to train and test the preset AI model based on the vibration and noise data set to obtain an AI vibration and noise diagnostic model.

[0141] The fault diagnosis module 605 is used to obtain a real-time vibration and noise data set of the motor, and perform motor vibration and noise fault diagnosis based on the real-time vibration and noise data set and the AI ​​vibration and noise diagnosis model.

[0142] In some optional implementations, the vibration noise big data includes first vibration noise big data of motors with vibration noise faults, second vibration noise big data of qualified motors, first expert experience big data of motors with vibration noise faults that have not been improved after rework, and second expert experience big data of motors with vibration noise faults that have been qualified after rework; the correlation analysis module 602 includes:

[0143] The incoming material quality parameter and production manufacturing process parameter extraction unit is used to extract the incoming material quality parameters and production manufacturing process parameters from the first vibration noise big data, the second vibration noise big data, the first expert experience big data and the second expert experience big data respectively.

[0144] The first correlation analysis unit is used to perform a correlation analysis on the incoming material quality parameters in the first vibration noise big data and the incoming material quality parameters in the first expert experience big data to obtain the incoming material quality unqualified parameters, and at the same time perform a correlation analysis on the production and manufacturing process parameters in the first vibration noise big data and the production and manufacturing process parameters in the first expert experience big data to obtain the production and manufacturing process unqualified parameters, and summarize the incoming material quality unqualified parameters and the production license process unqualified parameters to obtain the first key parameters.

[0145] The second correlation analysis unit is used to perform a correlation analysis on the incoming material quality parameters in the second vibration noise big data and the incoming material quality parameters in the second expert experience big data to obtain the incoming material quality qualified parameters, and at the same time perform a correlation analysis on the production and manufacturing process parameters in the second vibration noise big data and the production and manufacturing process parameters in the second expert experience big data to obtain the production and manufacturing process qualified parameters, and summarize the incoming material quality qualified parameters and the production license process qualified parameters to obtain the first key parameters.

[0146] The first labeling unit is used to perform one-to-one labeling and classification on the unqualified incoming material quality parameters in the first vibration noise big data in the first key parameter and the unqualified incoming material quality parameters in the first expert experience big data, and at the same time, perform one-to-one labeling and classification on the unqualified production and manufacturing process parameters in the first vibration noise big data in the first key parameter and the unqualified production and manufacturing process parameters in the first expert experience big data.

[0147] The second labeling unit is used to perform one-to-one labeling and classification on the incoming material quality qualified parameters in the second vibration noise big data in the second key parameters and the incoming material quality qualified parameters in the second expert experience big data, and at the same time, perform one-to-one labeling and classification on the production and manufacturing process qualified parameters in the second vibration noise big data in the second key parameters and the production and manufacturing process qualified parameters in the second expert experience big data.

[0148] In some optional implementations, the motor fault diagnosis device further includes:

[0149] The data set division module is used to divide the vibration noise data set into a training data set and a test data set according to a preset ratio.

[0150] In some optional implementations, the training and testing module 604 includes:

[0151] The training unit is used to input the training data set into the preset AI model for training to obtain the AI ​​model to be tested.

[0152] The testing unit is used to input the test data set into the AI ​​model to be tested for test verification to obtain the test results. When the accuracy of the test results meets the preset test threshold, the AI ​​model is used as the AI ​​vibration noise diagnosis model.

[0153] In some optional implementations, the fault diagnosis module 605 includes:

[0154] The fault diagnosis unit is used to input the real-time vibration noise data set into the AI ​​vibration noise diagnosis model to perform vibration noise fault diagnosis, and obtain the qualified state of the motor without vibration noise fault and the unqualified state with vibration noise fault.

[0155] In some optional implementations, the motor fault diagnosis device further includes:

[0156] The change difference calculation module is used to analyze the fault parameters of the motor generating the vibration noise fault when an unqualified state of vibration noise fault is obtained, and compare the fault parameters with the qualified motor vibration noise parameters in a preset database to obtain the change difference between the fault parameters and the qualified motor vibration noise parameters.

[0157] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0158] The motor fault diagnosis device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0159] The embodiment of the present invention also provides a computer device having the above Figure 6 The motor fault diagnosis device shown.

[0160] See also Figure 7 , Figure 7 is a structural diagram of a computer device provided by an optional embodiment of the present invention, such as Figure 7 As shown, the computer device includes: one or more processors 10, memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in the memory or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Equally, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 7 A processor 10 is taken as an example.

[0161] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0162] The memory 20 stores instructions that can be executed by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0163] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0164] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0165] The computer device further includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 7 The bus connection is taken as an example.

[0166] The input device 30 can receive input digital or character information and generate key signal input related to user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a trackpad, a touch pad, an indicator stick, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 can include a display device, an auxiliary lighting device (e.g., an LED), and a tactile feedback device (e.g., a vibration motor). The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display, and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0167] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0168] A portion of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the form in which the computer program instruction exists in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc. Accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium that can be accessed by the computer.

[0169] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A motor fault diagnosis method, characterized in that: The method comprises: Obtain big data on motor vibration and noise; Performing correlation analysis on the vibration noise big data to screen out a first key parameter that causes motor vibration noise failure and a second key parameter that is a qualified motor vibration noise; The first key parameter is used as a fault data set and the second key parameter is used as a qualified data set, and the fault data set and the qualified data set are aggregated to obtain a vibration noise data set; Training and testing a preset AI model based on the vibration noise dataset to obtain an AI vibration noise diagnostic model; A real-time vibration and noise data set of the motor is obtained, and motor vibration and noise fault diagnosis is performed based on the real-time vibration and noise data set and an AI vibration and noise diagnostic model.

2. The method according to claim 1, characterized in that The vibration noise big data includes first vibration noise big data of motors with vibration noise faults, second vibration noise big data of qualified motors, first expert experience big data of motors with vibration noise faults that have not been improved after repair, and second expert experience big data of motors with vibration noise faults that have been qualified after repair; Performing correlation analysis on the vibration noise big data to screen out the first key parameter that causes the motor vibration noise fault and the second key parameter of the qualified motor vibration noise, including: Extracting incoming material quality parameters and production manufacturing process parameters from the first vibration noise big data, the second vibration noise big data, the first expert experience big data, and the second expert experience big data respectively; Performing a correlation analysis on the incoming material quality parameters in the first vibration noise big data and the incoming material quality parameters in the first expert experience big data to obtain an incoming material quality unqualified parameter; performing a correlation analysis on the production and manufacturing process parameters in the first vibration noise big data and the production and manufacturing process parameters in the first expert experience big data to obtain a production and manufacturing process unqualified parameter; and summarizing the incoming material quality unqualified parameter and the production license process unqualified parameter to obtain a first key parameter; A correlation analysis is performed on the incoming material quality parameters in the second vibration noise big data and the incoming material quality parameters in the second expert experience big data to obtain the incoming material quality qualified parameters. At the same time, a correlation analysis is performed on the production and manufacturing process parameters in the second vibration noise big data and the production and manufacturing process parameters in the second expert experience big data to obtain the production and manufacturing process qualified parameters. The incoming material quality qualified parameters and the production license process qualified parameters are summarized to obtain the second key parameters.

3. The method according to claim 2, characterized in that Correlation analysis is performed on the vibration noise big data to screen out the first key parameter that causes motor vibration noise failure and the second key parameter of qualified motor vibration noise, including: Perform one-to-one labeling and classification on the unqualified incoming material quality parameters in the first vibration noise big data among the first key parameters and the unqualified incoming material quality parameters in the first expert experience big data; and perform one-to-one labeling and classification on the unqualified production and manufacturing process parameters in the first vibration noise big data among the first key parameters and the unqualified production and manufacturing process parameters in the first expert experience big data; The incoming material quality qualified parameters in the second vibration noise big data among the second key parameters are labeled and classified one-to-one with the incoming material quality qualified parameters in the second expert experience big data. At the same time, the production and manufacturing process qualified parameters in the second vibration noise big data among the second key parameters are labeled and classified one-to-one with the production and manufacturing process qualified parameters in the second expert experience big data.

4. The method according to claim 1, wherein Before training the preset AI model based on the vibration noise dataset, the method further includes: The vibration noise data set is divided into a training data set and a test data set according to a preset ratio.

5. The method according to claim 4, characterized in that The method of training and testing a preset AI model based on the vibration noise dataset to obtain an AI vibration noise diagnostic model includes: Input the training data set into a preset AI model for training to obtain an AI model to be tested; The test data set is input into the AI ​​model to be tested for test verification to obtain a test result. When the accuracy of the test result meets the preset test threshold, the AI ​​model is used as the AI ​​vibration noise diagnosis model.

6. The method according to claim 1, characterized in that The motor vibration noise fault diagnosis based on the real-time vibration noise data set and the AI ​​vibration noise diagnostic model includes: The real-time vibration noise data set is input into the AI ​​vibration noise diagnosis model to perform vibration noise fault diagnosis, and a qualified state of the motor without vibration noise fault and an unqualified state with vibration noise fault are obtained.

7. The method according to claim 6, characterized in that The method further comprises: When an unqualified state with a vibration noise fault is obtained, the fault parameters of the motor generating the vibration noise fault are analyzed, and the fault parameters are compared with the qualified motor vibration noise parameters in a preset database to obtain the change difference between the fault parameters and the qualified motor vibration noise parameters.

8. A motor fault diagnosis device, characterized in that: The device comprises: Vibration and noise big data acquisition module, used to obtain motor vibration and noise big data; A correlation analysis module is used to perform correlation analysis on the vibration noise big data to screen out a first key parameter that causes motor vibration noise failure and a second key parameter that is qualified for motor vibration noise; a vibration and noise data set determination module, configured to use the first key parameter as a fault data set and the second key parameter as a qualified data set, and to aggregate the fault data set and the qualified data set to obtain a vibration and noise data set; A training and testing module, configured to train and test a preset AI model based on the vibration and noise dataset to obtain an AI vibration and noise diagnostic model; The fault diagnosis module is used to obtain a real-time vibration and noise data set of the motor and perform motor vibration and noise fault diagnosis based on the real-time vibration and noise data set and an AI vibration and noise diagnosis model.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the motor fault diagnosis method according to any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the motor fault diagnosis method according to any one of claims 1 to 7.

Images