A method, system and electronic equipment for predicting motor bearing wear

Abstract

The present disclosure relates to the field of motor technology, and in particular to a method, system and electronic device for predicting the degree of wear of motor bearings. The method comprises: determining a set of auxiliary variables, wherein the auxiliary variables in the set of auxiliary variables are used to predict each bearing wear variable in the set of bearing wear variables; taking the prediction of each bearing wear variable as a subtask, constructing a first multi-task learning support vector machine model, and training the first multi-task learning support vector machine model to obtain a second multi-task learning support vector machine model; monitoring the auxiliary variable set of the motor to be tested, inputting the auxiliary variable data set obtained by monitoring into the second multi-task learning support vector machine model, and obtaining the bearing wear variable data set corresponding to the motor to be tested. The present disclosure using the above scheme can predict the degree of wear of motor bearings in real time, and the prediction accuracy is high.

Description

Technical Field

[0001] This disclosure relates to the field of motor technology, and in particular to a method, system and electronic device for predicting the wear degree of motor bearings. Background Technology

[0002] During motor operation, bearings are one of the core components, and their wear directly affects the motor's operational stability and safety. Bearing wear can lead to increased vibration, increased energy consumption, and even motor failure or explosion risks. Directly detecting bearing wear requires stopping the machine and disassembling it. However, explosion-proof motors are equipped with robust explosion-proof housings and are encapsulated with a casting process, making disassembly extremely difficult and preventing real-time monitoring. Summary of the Invention

[0003] This disclosure aims to at least partially address one of the technical problems in the related art.

[0004] Therefore, the first objective of this disclosure is to propose a method for predicting the wear degree of motor bearings, so as to predict the wear degree of motor bearings in real time with high accuracy.

[0005] The second objective of this disclosure is to propose a system for predicting the wear degree of motor bearings.

[0006] The third objective of this disclosure is to propose an electronic device.

[0007] The fourth objective of this disclosure is to provide a computer-readable storage medium.

[0008] The fifth objective of this disclosure is to provide a computer program product.

[0009] To achieve the above objectives, the first aspect of this disclosure provides a method for predicting the wear degree of motor bearings, comprising:

[0010] A set of auxiliary variables is determined, wherein the auxiliary variables in the set of auxiliary variables are used to predict each bearing wear variable in the set of bearing wear variables;

[0011] The prediction of each bearing wear variable is taken as a subtask, a first multi-task learning support vector machine model is constructed, and the first multi-task learning support vector machine model is trained to obtain a second multi-task learning support vector machine model.

[0012] The auxiliary variable set of the motor under test is monitored, and the monitored auxiliary variable data set is input into the second multi-task learning support vector machine model to obtain the bearing wear variable data set corresponding to the motor under test.

[0013] Optionally, constructing the first multi-task learning support vector machine model includes:

[0014] Construct a model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set to obtain a third multi-task learning support vector machine model;

[0015] The first multi-task learning support vector machine model is obtained by weighting each input variable in the third multi-task learning support vector machine model.

[0016] Optionally, constructing a model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set includes:

[0017] In the time dimension, each auxiliary variable in the set of auxiliary variables is expanded based on the data expansion parameter to obtain the expanded set of auxiliary variables.

[0018] Construct a model between each of the bearing wear variables and the set of auxiliary variables after the data dimension expansion.

[0019] Optionally, constructing a model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set includes:

[0020] Determine the data feedback step size, and determine the bearing wear history output variable set corresponding to each bearing wear variable in the bearing wear variable set;

[0021] Construct a model relating each of the bearing wear variables to the set of auxiliary variables and the set of bearing wear history output variables corresponding to each of the bearing wear variables.

[0022] Optionally, training the first multi-task learning support vector machine model includes:

[0023] Obtain a sample dataset and train the model parameters of the first multi-task learning support vector machine model based on the sample dataset, wherein the model parameters include the weights corresponding to each input variable.

[0024] Optionally, inputting the monitored auxiliary variable data set into the second multi-task learning support vector machine model includes:

[0025] The auxiliary variable data set obtained from monitoring is preprocessed to obtain a preprocessed auxiliary variable data set, wherein the data preprocessing includes noise reduction processing, data alignment processing, and data standardization processing;

[0026] The preprocessed auxiliary variable data set is input into the second multi-task learning support vector machine model.

[0027] Optionally, determining the set of auxiliary variables includes:

[0028] A set of bearing wear variables is determined, and a correlation analysis is performed on the bearing wear variables in the set of bearing wear variables to analyze the variables related to the bearing wear variables during motor operation, thereby obtaining an initial set of auxiliary variables.

[0029] The measurability of the initial auxiliary variables in the initial auxiliary variable set is evaluated, and the initial auxiliary variables that meet the measurability evaluation requirements are selected from the initial auxiliary variable set to obtain the measurable auxiliary variable set;

[0030] The number and type of auxiliary variables to be used are determined from the set of measurable auxiliary variables, thus obtaining the set of auxiliary variables.

[0031] Optionally, the set of auxiliary variables includes voltage, current, temperature, speed, cumulative running time, single running time, and vibration amplitude. Monitoring the set of auxiliary variables of the motor under test includes:

[0032] A temperature sensor is installed on the motor under test to monitor the temperature of the motor under test;

[0033] An acceleration sensor is installed on the motor under test to monitor the vibration amplitude of the motor under test;

[0034] An encoder is installed on the motor under test to monitor the rotational speed of the motor under test;

[0035] A current sensor is installed on the power supply of the motor under test to monitor the current of the motor under test;

[0036] A voltage sensor is installed on the power supply to monitor the voltage of the motor under test;

[0037] Based on the clock unit in the control module of the motor under test, the cumulative running time and single running time of the motor under test are monitored.

[0038] To achieve the above objectives, a second aspect of this disclosure provides a system for predicting the wear degree of motor bearings, comprising:

[0039] A set determination unit is used to determine an auxiliary variable set, wherein the auxiliary variables in the auxiliary variable set are used to predict each bearing wear variable in the bearing wear variable set;

[0040] The model determination unit is used to construct a first multi-task learning support vector machine model by taking the prediction of each bearing wear variable as a sub-task, and to train the first multi-task learning support vector machine model to obtain a second multi-task learning support vector machine model.

[0041] The data prediction unit is used to monitor the set of auxiliary variables of the motor under test, and input the monitored set of auxiliary variable data into the second multi-task learning support vector machine model to obtain the set of bearing wear variable data corresponding to the motor under test.

[0042] To achieve the above objectives, a third aspect of this disclosure provides an electronic device comprising:

[0043] Memory, used to store executable program code;

[0044] A processor for calling and running the executable program code from the memory, causing the electronic device to perform the method shown in any of the first aspects above.

[0045] To achieve the above objectives, a fourth aspect of this disclosure provides a computer-readable storage medium storing a computer program that, when executed, implements the method shown in any of the first aspects above.

[0046] To achieve the above objectives, a fifth aspect of this disclosure provides a computer program product including a computer program that, when executed by a processor, implements the method shown in any of the first aspects above.

[0047] In summary, the method, system, and electronic equipment provided in this disclosure can predict the wear degree of motor bearings with high accuracy by using soft measurement technology and a multi-task learning support vector machine model.

[0048] Additional aspects and advantages of this disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this disclosure. Attached Figure Description

[0049] The above and / or additional aspects and advantages of this disclosure will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, in which:

[0050] Figure 1 This is a flowchart illustrating a method for predicting the wear degree of motor bearings provided in an embodiment of this disclosure.

[0051] Figure 2 A schematic diagram of the architecture of a first multi-task learning support vector machine model provided in an embodiment of this disclosure;

[0052] Figure 3 This is a schematic diagram illustrating the monitoring of an auxiliary variable set provided in an embodiment of this disclosure;

[0053] Figure 4 This is a schematic flowchart illustrating a method for predicting the wear level of an electric motor bearing, as provided in an embodiment of this disclosure.

[0054] Figure 5 This is a schematic diagram of the structure of a motor bearing wear prediction system provided in an embodiment of this disclosure. Detailed Implementation

[0055] Embodiments of this disclosure are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0056] The present disclosure will now be described in detail with reference to specific embodiments.

[0057] In the first embodiment, such as Figure 1 As shown, Figure 1 This is a flowchart illustrating a method for predicting the wear level of motor bearings according to an embodiment of this disclosure. This method can be implemented using a computer program and can run on a system for predicting the wear level of motor bearings. The computer program can be integrated into an application or run as a standalone utility application.

[0058] The method for predicting the wear of motor bearings can be executed by electronic equipment.

[0059] For example, this method for predicting the wear of motor bearings includes the following steps:

[0060] S101, Determine the set of auxiliary variables;

[0061] Among them, the auxiliary variables in the auxiliary variable set are used to predict each bearing wear variable in the bearing wear variable set.

[0062] It's important to note that soft measurement technology not only applies the concepts of automatic control but also integrates with computer software. It uses easily measurable variables as auxiliary variables as inputs to construct mathematical relationships to predict variables that are difficult to measure directly (called dominant variables). In other words, computer software replaces sensors. Compared to direct measurement, soft measurement technology can quickly calculate the values ​​of dominant variables based on a model, and it is also easy to maintain and update. Therefore, using soft measurement technology to predict bearing wear variables using easily measurable auxiliary variables has significant engineering implications.

[0063] S102, the prediction of each bearing wear variable is taken as a subtask, the first multi-task learning support vector machine model is constructed, and the first multi-task learning support vector machine model is trained to obtain the second multi-task learning support vector machine model.

[0064] It's important to note that the goal of multi-task learning is to improve the learning efficiency of prediction, regression, or classification problems by leveraging the correlations between different tasks. In practice, multiple tasks are often interconnected and mutually restrictive, such as face recognition, dog breed identification, and age prediction. These all involve modeling facial or head features, but single-task learning can only learn for a specific task. This training method does not consider the relationships between tasks. Multi-task learning, on the other hand, can uncover the relationships between different tasks to some extent, thereby improving the learning efficiency of each task. Applying the concept of multi-task learning to soft sensor modeling can significantly improve the accuracy of mathematical models.

[0065] Secondly, ordinary Least-Squares Support Vector Regression Machines (LSSVR) can only handle single-output regression. For multi-output models, multiple independent LSSVRs are often trained, but this ignores the potential correlation between different outputs. Therefore, based on the idea of ​​multi-task learning, multi-task learning LSSVR is adopted, namely Multi-output Least-Squares Support Vector Regression Machines (MLSSVR), or simply Multi-task Learning Support Vector Machine (SVM). This method fully considers the possible relationships between different outputs, extending the single-output model to multi-output models, which theoretically improves the prediction accuracy of the regression model.

[0066] In other words, by introducing a multi-task learning support vector machine model into the soft prediction of bearing wear variables, the accuracy of bearing wear variable prediction can be greatly improved.

[0067] S103, monitor the set of auxiliary variables of the motor under test, and input the monitored set of auxiliary variable data into the second multi-task learning support vector machine model to obtain the set of bearing wear variable data corresponding to the motor under test.

[0068] In summary, the method provided in this embodiment, based on soft measurement technology and a multi-task learning support vector machine model, can predict the wear degree of motor bearings with high accuracy.

[0069] Another embodiment of this disclosure provides a method for predicting the wear degree of motor bearings. This method can be performed by an electronic device.

[0070] For example, the method for predicting the wear of motor bearings may include the following steps:

[0071] S201, Determine the set of auxiliary variables;

[0072] It's important to note that soft sensing modeling presupposes the identification of variables, including the selection of dominant and auxiliary variables. The selection of dominant variables can be determined based on the technological process and the soft sensing task. After identifying the dominant variables, auxiliary variables—those highly correlated with the dominant variables—are identified based on the industrial process. These auxiliary variables must also be easily measurable, either directly or indirectly. The number and type of auxiliary variables, as well as the type of monitoring points, are interconnected and mutually restrictive. In practice, they are also influenced by many external factors such as ease of implementation, maintenance costs, and operational stability.

[0073] In this embodiment, the dominant variable is the bearing wear variable, and different bearing wear variables can be selected for different types of motors. For example, for explosion-proof motors, directly measuring the wear degree of the explosion-proof motor requires stopping and disassembling, which is extremely difficult and cannot be monitored in real time. Therefore, the bearing wear variables can be the percentage of wear depth and the remaining service life of the bearing. These two variables directly reflect the wear degree and health status of the bearing.

[0074] Next, based on the selected bearing wear variables, mechanistic analysis methods can be used to determine the set of auxiliary variables. For example, correlation analysis can be performed on the bearing wear variables to analyze variables related to bearing wear during motor operation, thus obtaining an initial set of auxiliary variables. Then, the measurability of the initial auxiliary variables in the initial set is assessed, and those initial auxiliary variables that meet the measurability assessment requirements are selected, resulting in a set of measurable auxiliary variables. Finally, the number and type of auxiliary variables to be used are determined from the set of measurable auxiliary variables, thus obtaining the final set of auxiliary variables.

[0075] In some embodiments, the number and type of auxiliary variables to be used can be determined from the set of measurable auxiliary variables by combining external factors and based on actual needs and measurement conditions, thus obtaining a set of auxiliary variables.

[0076] Taking a scenario as an example, during correlation analysis of bearing wear variables, we can analyze variables such as current, voltage, temperature, speed, and vibration. Changes in these variables may be affected by bearing wear. When bearing wear is severe, the motor's operating resistance increases, leading to a rise in operating current and voltage. Simultaneously, as the current in the circuit increases, and given that the explosion-proof motor operates in a sealed explosion-proof enclosure with poor heat dissipation, the temperature inside the enclosure rises, further affecting bearing wear. Motor speed also shows a strong correlation with the degree of bearing wear; high speeds increase the contact surface temperature and inevitably lead to mechanical wear. When the motor bearing experiences a certain amount of wear, a gap will form between it and the structural components. The motor will vibrate during high-speed operation; the greater the wear, the more pronounced the vibration and the greater the friction, which in turn leads to increased temperature. Finally, the single-run duration and cumulative run duration of the motor can also reflect the degree of wear. Therefore, voltage, current, temperature, speed, cumulative run duration, single-run duration, vibration amplitude, and bearing wear variables are correlated to some extent.

[0077] Secondly, current and voltage can be measured using Hall effect current sensors and patch voltage sensors; the internal cavity temperature can be measured using temperature sensors installed within the cavity; rotational speed can be measured using an encoder; vibration amplitude can be measured using an accelerometer; and cumulative runtime and single runtime can be recorded using the controller clock. Therefore, voltage, current, temperature, rotational speed, cumulative runtime, single runtime, vibration amplitude, and bearing wear variables are all initial auxiliary variables that meet the requirements for measurability assessment.

[0078] Furthermore, considering external factors such as ease of implementation, maintenance costs, and operational stability, the sensors used for monitoring voltage, current, temperature, rotational speed, cumulative operating time, single operating time, and vibration amplitude are easy to install and provide stable and reliable measurement data, ensuring the accuracy and reliability of the soft measurement model. Therefore, these six variables—voltage, current, temperature, rotational speed, cumulative operating time, single operating time, and vibration amplitude—can be selected as auxiliary variables in the auxiliary variable set.

[0079] S202, the prediction of each bearing wear variable is taken as a subtask, and a model is constructed between each bearing wear variable in the bearing wear variable set and the auxiliary variable set to obtain the third multi-task learning support vector machine model;

[0080] It should be noted that, in order to study certain relationships that may exist between different outputs, we can assume... w i It can be represented as w i = w 0+ vi ,in, , , , w 0 indicates information shared across different outputs. v i This indicates that different outputs carry their own unique information. The objective function of the multi-task-based regression model is:

[0081] (1)

[0082] The constraints are:

[0083] (2)

[0084] in, , , λ , γ These are two regularization parameters. .

[0085] make The `trace` command is used to calculate the sum of the diagonal elements of a matrix. Using the `trace` command, the sum of squared terms can be obtained, which is the square of the error in least squares support vector regression. The equation using the Lagrange multiplier method is as follows:

[0086] (3)

[0087] By solving the system of equations in formula (3), we can calculate:

[0088] (4)

[0089] In other words, w 0 is v 1, v 2,… v m A linear combination, i.e. w i Too v 1, v 2,… v m The linear combination of these problems can be transformed into an optimization problem involving only the linear combination of these problems. v and b The problem with these two vectors is as follows:

[0090] (5)

[0091] Similar to the single-output LSSVR model, simplifying equation (5) yields the following formula:

[0092] (6)

[0093] in, Positive definite matrix , , , , , .

[0094] By solving formula (6), we obtain... and b Then, the mapping function can be obtained:

[0095] (7)

[0096] Since formula (7) is very complex to calculate, the following method is adopted for calculation, similar to the LSSVR model. First, the identity transformation is performed as follows:

[0097] (8)

[0098] in, Then by , Seeking , Then calculate Finally, solve... , This algorithm reduces the amount of computation, greatly improves operational efficiency, and is also easy to implement in computer programming.

[0099] Through the above derivation, it is found that the ultimate goal of multi-task learning is to find... α and b This allows us to obtain the model's parameters, which are then used to perform regression analysis on the prediction dataset. α Values ​​equal to zero have no effect on the model. The core of multi-task modeling lies in considering the influence of the dissimilar parts between different output vectors, while single-task modeling does not consider the relationship between output vectors. This is the advantage of multi-task modeling over single-task modeling.

[0100] According to some embodiments, during the dynamic motion process, the value of the dominant variable at time k is not only affected by the auxiliary variable at time k, but also related to the auxiliary variable at past times. Therefore, in the time dimension, the auxiliary variable set can be expanded based on the data expansion parameter to obtain the expanded auxiliary variable set; and a model can be constructed between each bearing wear variable and the expanded auxiliary variable set.

[0101] In other words, for a model with input x and output y, we construct a model between y(k) and x(k), x(k-1), x(k-2)...x(km). Here, m is the data dimensionality expansion parameter, representing the first m time steps. Dimension expansion can solve the transformation problem from time-dynamic modeling to spatial static modeling, and can improve the prediction accuracy of the model to a certain extent.

[0102] For example, when m=3, construct y(k)= f (x(k),x(k-1),x(k-2),x(k-3)). The original input data matrix is ​​1000*6, and after dimension expansion, the input matrix becomes 1000*24. When m=4, the original input data matrix is ​​1000*6, and after dimension expansion, the input matrix becomes 1000*30.

[0103] It should be noted that data dimensionality expansion can increase the accuracy of soft measurement models in some cases. However, since it introduces many input nodes, the number of parameters increases, which may lead to overfitting and reduce the model's prediction accuracy. Therefore, it is necessary to flexibly adjust the size of the data dimensionality expansion parameters during training to achieve the best possible prediction accuracy.

[0104] According to some embodiments, bearing wear is a slow process; unless rigid fracture occurs, the wear gradually increases, and there is a strong relationship between the wear level at the next moment and the wear level at the previous moment. Therefore, historical wear prediction data can be introduced to improve the accuracy of prediction.

[0105] In other words, the data feedback step size can be determined, and the set of bearing wear history output variables corresponding to each bearing wear variable in the bearing wear variable set can be determined; a model can be constructed between each bearing wear variable and the set of auxiliary variables, as well as the set of bearing wear history output variables corresponding to each bearing wear variable.

[0106] The data feedback step size represents the number of steps backward from the current output. The size of the data feedback step size can be flexibly adjusted during training to achieve the best possible prediction accuracy.

[0107] S203, weight each input variable in the third multi-task learning support vector machine model to obtain the first multi-task learning support vector machine model;

[0108] It's important to note that different inputs have different predictive effects on the output variable. In the third multi-task learning support vector machine model, each input variable has a weight of 1, which ignores the influence of different input variables on each output variable. Therefore, assigning a certain weight to each input can improve the accuracy of prediction.

[0109] In some embodiments, Figure 2 This is a schematic diagram of the architecture of a first multi-task learning support vector machine model provided in an embodiment of this disclosure. Figure 2 As shown, the bearing wear depth percentage y1(k) and remaining service life y2(k) are selected as the dominant output variables. The auxiliary variable set after data expansion and the two sets of historical bearing wear output variables, namely historical dominant variable 1 input and historical dominant variable 2 input, are selected. The data expansion parameter is m and the data feedback step size is n.

[0110] Since this embodiment selects six auxiliary variables—voltage, current, temperature, rotational speed, cumulative running time, single running time, and vibration amplitude—and further expands their data dimensions, in order to simplify the model structure diagram, Figure 2 The text only shows the input of one of the auxiliary variables, namely x(k), x(k-1), x(k-2)…x(km), where α0, α1, α2, …α m These are their respective weights.

[0111] For the historical output variable of bearing wear, the assigned weight λ can be understood as a forgetting factor. By continuously adjusting the weights of historical prediction data and errors, the model can pay more attention to recent prediction biases, thus quickly adapting to changes in bearing wear. By dynamically adjusting the forgetting factor, the model can better adapt to the gradual increase in bearing wear, avoiding prediction lag caused by over-reliance on historical data, thereby improving prediction accuracy.

[0112] S204. Obtain the sample dataset and train the model parameters of the first multi-task learning support vector machine model based on the sample dataset to obtain the second multi-task learning support vector machine model.

[0113] The model parameters include the weights corresponding to each input variable.

[0114] According to some implementations, when training the weights corresponding to each input variable, the weight of a certain input variable can be reduced or the input variable can be directly deleted, while the weights of other input variables remain unchanged. The change in prediction accuracy of the two output variables is then observed. If the prediction accuracy increases significantly, it indicates that the input variable has a greater effect on the prediction, meaning it needs to be assigned a larger weight. Conversely, if the prediction accuracy decreases significantly, it indicates that the input has a smaller effect on the prediction, meaning it needs to be assigned a smaller weight.

[0115] In some embodiments, the metric used to measure the deviation between the predicted value and the actual observed value can be the root mean square error (RMSE).

[0116] S205, monitors the set of auxiliary variables of the motor under test;

[0117] According to some embodiments, when predicting six auxiliary variables—voltage, current, temperature, speed, cumulative running time, single running time, and vibration amplitude—a temperature sensor can be installed on the motor under test to monitor its temperature; an acceleration sensor can be installed on the motor under test to monitor its vibration amplitude; an encoder can be installed on the motor under test to monitor its speed; a current sensor can be installed on the power supply of the motor under test to monitor its current; a voltage sensor can be installed on the power supply to monitor its voltage; and the cumulative running time and single running time of the motor under test can be monitored based on the clock unit in the control module of the motor under test.

[0118] To give an example from a scenario, Figure 3 This is a schematic diagram illustrating the monitoring of a set of auxiliary variables provided in an embodiment of this disclosure. For example... Figure 3 As shown, the motor bearing wear prediction method provided in this embodiment is applied to an exoskeleton robot. The exoskeleton robot includes an explosion-proof joint motor, an explosion-proof battery, and a control board. The control board monitors the operating temperature, vibration, and speed of the explosion-proof joint motor through temperature sensors, acceleration sensors, and encoders installed on the explosion-proof joint motor. It monitors the current and voltage of the explosion-proof joint motor through current sensors and voltage sensors installed on the explosion-proof battery. It records the running time through its internal clock and monitors the cumulative running time and the single running time.

[0119] S206, Perform data preprocessing on the auxiliary variable data set obtained from monitoring to obtain the preprocessed auxiliary variable data set;

[0120] According to some embodiments, data preprocessing includes, but is not limited to, noise reduction, data alignment, and data standardization.

[0121] In some embodiments, noise reduction processing includes, but is not limited to, wavelet denoising of the vibration signal to eliminate environmental interference and electromagnetic interference.

[0122] In some embodiments, data alignment processing includes, but is not limited to, ensuring the synchronization of current, vibration, and rotational speed data. This can be achieved by synchronously sampling them.

[0123] In some embodiments, data standardization processing includes, but is not limited to, Z-score standardization of current and vibration characteristics to eliminate the influence of dimensions.

[0124] In some embodiments, Kalman filtering can also be applied to the acceleration data during data preprocessing to remove erroneous data and ensure the authenticity of the acceleration signal.

[0125] S207. Input the preprocessed auxiliary variable data set into the second multi-task learning support vector machine model to obtain the bearing wear variable data set corresponding to the motor under test.

[0126] To give an example from a scenario, Figure 4 This is a schematic flowchart illustrating a method for predicting the wear level of a motor bearing, as provided in an embodiment of this disclosure. Figure 4 As shown, applying the motor bearing wear prediction method provided in this embodiment to an exoskeleton robot allows for the acquisition of auxiliary variable signals, data preprocessing, and prediction output via the exoskeleton robot's main controller. During the prediction process, a feedback mechanism can be used, where the historical bearing wear output variable is used as input to dynamically train the second multi-task learning support vector machine model, dynamically adjusting the weights of the input variables.

[0127] In summary, the method provided in this embodiment, which uses soft measurement technology to predict bearing wear using easily measurable variables, has significant engineering implications and can be applied to fields such as exoskeleton robots, wearable devices, and explosion-proof modifications. Furthermore, by using data expansion, output feedback, and variable weighting, the accuracy of the prediction is greatly improved.

[0128] To achieve the above embodiments, this disclosure also proposes a motor bearing wear prediction system.

[0129] For example, Figure 5 This is a schematic diagram of the structure of a motor bearing wear prediction system provided in an embodiment of this disclosure. Figure 5 As shown, the motor bearing wear prediction system 500 includes:

[0130] The set determination unit 501 is used to determine the set of auxiliary variables, wherein the auxiliary variables in the set of auxiliary variables are used to predict each bearing wear variable in the set of bearing wear variables;

[0131] The model determination unit 502 is used to construct a first multi-task learning support vector machine model by taking the prediction of each bearing wear variable as a sub-task, and to train the first multi-task learning support vector machine model to obtain a second multi-task learning support vector machine model.

[0132] The data prediction unit 503 is used to monitor the set of auxiliary variables of the motor under test, and input the monitored set of auxiliary variable data into the second multi-task learning support vector machine model to obtain the set of bearing wear variable data corresponding to the motor under test.

[0133] Optionally, when the model determination unit 502 is used to construct the first multi-task learning support vector machine model, it is specifically used for:

[0134] Construct a model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set to obtain the third multi-task learning support vector machine model;

[0135] By weighting each input variable in the third multi-task learning support vector machine model, the first multi-task learning support vector machine model is obtained.

[0136] Optionally, when the model determination unit 502 is used to construct a model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set, it is specifically used for:

[0137] In the time dimension, the auxiliary variable set is expanded based on the data expansion parameter to obtain the expanded auxiliary variable set.

[0138] Construct a model between each bearing wear variable and the set of auxiliary variables after data dimension expansion.

[0139] Optionally, when the model determination unit 502 is used to construct a model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set, it is specifically used for:

[0140] Determine the data feedback step size and determine the set of bearing wear history output variables corresponding to each bearing wear variable in the bearing wear variable set;

[0141] Construct a model relating each bearing wear variable to a set of auxiliary variables and the set of bearing wear history output variables corresponding to each bearing wear variable.

[0142] Optionally, when training the first multi-task learning support vector machine model, the model determination unit 502 is specifically used for:

[0143] Obtain a sample dataset and train the model parameters of the first multi-task learning support vector machine model based on the sample dataset. The model parameters include the weights corresponding to each input variable.

[0144] Optionally, when the data prediction unit 503 inputs the monitored auxiliary variable data set into the second multi-task learning support vector machine model, it is specifically used for:

[0145] The auxiliary variable dataset obtained from monitoring is preprocessed to obtain a preprocessed auxiliary variable dataset. The data preprocessing includes noise reduction, data alignment, and data standardization.

[0146] The preprocessed auxiliary variable data set is input into the second multi-task learning support vector machine model.

[0147] Optionally, when the set determination unit 501 is used to determine the set of auxiliary variables, it is specifically used for:

[0148] Determine the set of bearing wear variables, and conduct correlation analysis on the bearing wear variables in the set of bearing wear variables. Analyze the variables related to bearing wear variables during motor operation to obtain the initial set of auxiliary variables.

[0149] The measurability of the initial auxiliary variables in the initial auxiliary variable set is evaluated, and the initial auxiliary variables that meet the measurability evaluation requirements are selected from the initial auxiliary variable set to obtain the measurable auxiliary variable set;

[0150] The number and type of auxiliary variables to be used are determined from the set of measurable auxiliary variables, thus obtaining the set of auxiliary variables.

[0151] Optionally, the set of auxiliary variables includes voltage, current, temperature, speed, cumulative running time, single running time, and vibration amplitude. The data prediction unit 503 is used to monitor the set of auxiliary variables of the motor under test, specifically for:

[0152] A temperature sensor is installed on the motor under test to monitor its temperature.

[0153] An acceleration sensor is installed on the motor under test to monitor the vibration amplitude of the motor under test;

[0154] An encoder is installed on the motor under test to monitor its speed.

[0155] A current sensor is installed on the power supply of the motor under test to monitor the current of the motor under test.

[0156] A voltage sensor is installed on the power supply to monitor the voltage of the motor under test;

[0157] Based on the clock unit in the control module of the motor under test, the cumulative running time and single running time of the motor under test are monitored.

[0158] It should be noted that the foregoing explanation of the embodiment of the method for predicting the wear degree of motor bearings also applies to the motor bearing wear degree prediction system of this embodiment, and will not be repeated here.

[0159] In summary, the system provided in this disclosure, based on soft measurement technology and a multi-task learning support vector machine model, can predict the wear degree of motor bearings with high accuracy.

[0160] To implement the above embodiments, this disclosure also proposes an electronic device, including: a processor and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method provided in the foregoing embodiments.

[0161] To implement the above embodiments, this disclosure also proposes a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the methods provided in the foregoing embodiments.

[0162] To implement the above embodiments, this disclosure also proposes a computer program product, including a computer program that, when executed by a processor, implements the methods provided in the foregoing embodiments.

[0163] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in this disclosure all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0164] It should be noted that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. Furthermore, such collection / sharing should only be conducted after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes authorization of relevant user information before the user uses the function. In addition, any necessary steps must be taken to protect and safeguard access to such personal information data and ensure that others with access to personal information data comply with their privacy policies and procedures.

[0165] This disclosure is intended to provide implementation schemes for users to selectively prevent the use or access to their personal information data. Specifically, this disclosure is intended to provide hardware and / or software to prevent or block access to such personal information data. Once personal information data is no longer needed, risks can be minimized by restricting data collection and deleting data. Furthermore, where applicable, such personal information is de-identified to protect user privacy.

[0166] In the foregoing descriptions of the embodiments, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0167] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. Throughout the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0168] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present disclosure includes additional implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present disclosure belong.

[0169] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequential list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disks (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, fiber optic devices, and compact disc read-only memory (CDROM). Additionally, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0170] It should be understood that various parts of this disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0171] Those skilled in the art will appreciate that all or part of the steps in the method for implementing the above-mentioned embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0172] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0173] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present disclosure have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. A person of ordinary skill in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A method for predicting the wear degree of motor bearings, characterized in that, include: A set of auxiliary variables is determined, wherein the auxiliary variables in the set of auxiliary variables are used to predict each bearing wear variable in the set of bearing wear variables, and the set of auxiliary variables includes voltage, current, temperature, rotational speed, cumulative running time, single running time, and vibration amplitude; The prediction of each bearing wear variable is taken as a subtask, a first multi-task learning support vector machine model is constructed, and the first multi-task learning support vector machine model is trained to obtain a second multi-task learning support vector machine model. The auxiliary variable set of the motor under test is monitored, and the monitored auxiliary variable data set is input into the second multi-task learning support vector machine model to obtain the bearing wear variable data set corresponding to the motor under test. The construction of the first multi-task learning support vector machine model includes: Construct a model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set to obtain a third multi-task learning support vector machine model; By weighting each input variable in the third multi-task learning support vector machine model, the first multi-task learning support vector machine model is obtained. The construction of the model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set includes: In the time dimension, each auxiliary variable in the set of auxiliary variables is expanded based on the data expansion parameter to obtain the expanded set of auxiliary variables. Construct a model between each of the bearing wear variables and the set of auxiliary variables after the data dimension expansion.

2. The method according to claim 1, characterized in that, The construction of the model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set includes: Determine the data feedback step size, and determine the bearing wear history output variable set corresponding to each bearing wear variable in the bearing wear variable set; Construct a model relating each of the bearing wear variables to the set of auxiliary variables and the set of bearing wear history output variables corresponding to each of the bearing wear variables.

3. The method according to claim 1, characterized in that, The training of the first multi-task learning support vector machine model includes: Obtain a sample dataset and train the model parameters of the first multi-task learning support vector machine model based on the sample dataset, wherein the model parameters include the weights corresponding to each input variable.

4. The method according to claim 1, characterized in that, The step of inputting the monitored auxiliary variable data set into the second multi-task learning support vector machine model includes: The auxiliary variable data set obtained from monitoring is preprocessed to obtain a preprocessed auxiliary variable data set, wherein the data preprocessing includes noise reduction processing, data alignment processing, and data standardization processing; The preprocessed auxiliary variable data set is input into the second multi-task learning support vector machine model.

5. The method according to claim 1, characterized in that, The determination of the set of auxiliary variables includes: Determine the set of bearing wear variables, and perform correlation analysis on the bearing wear variables in the set of bearing wear variables to analyze the variables related to the bearing wear variables during motor operation, and obtain an initial set of auxiliary variables; The measurability of the initial auxiliary variables in the initial auxiliary variable set is evaluated, and the initial auxiliary variables that meet the measurability evaluation requirements are selected from the initial auxiliary variable set to obtain the measurable auxiliary variable set; The number and type of auxiliary variables to be used are determined from the set of measurable auxiliary variables, thus obtaining the set of auxiliary variables.

6. The method according to claim 1, characterized in that... The monitoring of the set of auxiliary variables of the motor under test includes: A temperature sensor is installed on the motor under test to monitor the temperature of the motor under test; An acceleration sensor is installed on the motor under test to monitor the vibration amplitude of the motor under test; An encoder is installed on the motor under test to monitor the rotational speed of the motor under test; A current sensor is installed on the power supply of the motor under test to monitor the current of the motor under test; A voltage sensor is installed on the power supply to monitor the voltage of the motor under test; Based on the clock unit in the control module of the motor under test, the cumulative running time and single running time of the motor under test are monitored.

7. A system for predicting the wear degree of motor bearings, characterized in that, include: A set determination unit is used to determine an auxiliary variable set, wherein the auxiliary variables in the auxiliary variable set are used to predict each bearing wear variable in the bearing wear variable set, and the auxiliary variable set includes voltage, current, temperature, rotational speed, cumulative running time, single running time, and vibration amplitude; The model determination unit is used to construct a first multi-task learning support vector machine model by taking the prediction of each bearing wear variable as a sub-task, and to train the first multi-task learning support vector machine model to obtain a second multi-task learning support vector machine model. The data prediction unit is used to monitor the set of auxiliary variables of the motor under test, and input the monitored set of auxiliary variable data into the second multi-task learning support vector machine model to obtain the set of bearing wear variable data corresponding to the motor under test. The model determination unit is also used to construct a model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set, to obtain a third multi-task learning support vector machine model; By weighting each input variable in the third multi-task learning support vector machine model, the first multi-task learning support vector machine model is obtained. The construction of the model between each bearing wear variable in the bearing wear variable set and the auxiliary variable set includes: In the time dimension, each auxiliary variable in the set of auxiliary variables is expanded based on the data expansion parameter to obtain the expanded set of auxiliary variables. Construct a model between each of the bearing wear variables and the set of auxiliary variables after the data dimension expansion.

8. An electronic device, characterized in that, The electronic device includes: Memory, used to store executable program code; A processor for calling and running the executable program code from the memory, causing the electronic device to perform the method as described in any one of claims 1 to 6.

Images