Main shaft bearing life prediction method

Abstract

The invention relates to the field of bearing life prediction, in particular to a main shaft bearing life prediction method, which comprises the following steps: acquiring parameters of a bearing at different historical moments, the parameters comprising a plurality of sub-parameters; and constructing a long short-term memory network model, calculating the learning rate of the long short-term memory network model, and training the long short-term memory network model by using historical parameters to obtain a prediction model so as to predict the service life of the bearing. The learning rate is obtained through self-adaption, so that the accuracy of the prediction result of the long-short-term memory network model is improved.

Description

Technical Field

[0001] The invention relates to the field of bearing life prediction, and in particular to a method for predicting the life of a spindle bearing. Background Art

[0002] Bearings are common parts of mechanical equipment, and the quality of bearings is also a key factor affecting the industrial production and manufacturing process. In high-speed, high-load operating environments, spindle bearings are prone to wear or failure, which in turn affects the operating efficiency and accuracy of the equipment. Therefore, it is necessary to predict the life of the bearing so that it can be replaced or repaired before the bearing fails, thereby avoiding downtime, improving production efficiency and equipment reliability. Long short-term memory network is a time recursive neural network used to process time series data. Its unique gating mechanism enables it to effectively capture and learn long-term dependencies in data, so it is used in the field of time series prediction.

[0003] The Chinese patent application document with publication number CN111581888A discloses a method for constructing a prediction model for the remaining service life of a wind turbine bearing, which is implemented by the following steps: step 1, collecting bearing vibration frequency data; step 2, data preprocessing; step 3, constructing a long short-term memory network and configuring network parameters and specifying training options; step 4, training the network; step 5, verifying the prediction model. By providing a method for constructing a prediction model for the remaining service life of a wind turbine bearing based on a long short-term memory network, the prediction of the remaining service life of a wind turbine bearing is achieved.

[0004] When using the long short-term memory network model to predict the service life of bearings, it is necessary to first build the long short-term memory network model, then train the long short-term network model, and finally use the trained long short-term network model to predict the life of the bearing. In the existing training process, a fixed learning rate is usually used. However, if the learning rate is too large, the model may oscillate near the optimal value or even fail to converge; if the learning rate is too small, the training process will be slow and it will easily fall into the local optimum. Ultimately, the prediction results of the long short-term memory network model are less accurate. Summary of the invention

[0005] In order to solve the problem that the accuracy of bearing life prediction is low due to the use of a fixed learning rate to train a long-term and short-term network model, the present invention provides a spindle bearing life prediction method.

[0006] The present invention provides a method for predicting the life of a spindle bearing, which adopts the following technical solution: Obtain the parameters of the bearing at different historical moments, the parameters include multiple sub-parameters; Constructing a long short-term memory network model, calculating the learning rate of the long short-term memory network model, and using historical parameters to train the long short-term memory network model to obtain a prediction model for predicting the bearing life; The learning rate is calculated as follows: divide the historical moments into the time period of each historical moment; calculate the first influencing factor and the second influencing factor of the sub-parameter of each historical moment, where the first influencing factor represents the average deviation of the sub-parameter in the time period of the corresponding historical moment, and the second influencing factor represents the fluctuation range of the sub-parameter in the time period of the corresponding historical moment; perform weighted summation of the first influencing factor and the second influencing factor to obtain the influence degree of the sub-parameter at each historical moment; perform weighted summation of the influence degrees of multiple sub-parameters at each historical moment to obtain the influence degree of the corresponding historical moment, and calculate the learning rate of each historical moment. The expression is:

[0007] In the formula, represents the learning rate corresponding to the t-th historical moment, represents the preset initial learning rate, Represents the degree of influence at the t-th historical moment.

[0008] In the process of training the LSTM network model, the learning rate is adaptively adjusted according to the degree of abnormality at different historical moments. When the degree of abnormality is high, the learning rate is reduced to avoid the model's overreaction to abnormal data. When the degree of abnormality is low, the learning rate is increased to speed up the model convergence, thereby improving the accuracy of the prediction results of the LSTM network model.

[0009] Preferably, the method of dividing the historical moments to obtain the time period of each historical moment is: taking any historical moment as the center, extending multiple moments forward and backward along the time direction to obtain the time period of the corresponding historical moment.

[0010] By dividing the historical moments into corresponding time periods, it is convenient to analyze the parameters within the corresponding time period, so as to facilitate judging whether the parameters of the corresponding historical moments are in an abnormal state.

[0011] Preferably, the method further comprises: calculating the attention weight of each historical moment sub-parameter, expressed as:

[0012] In the formula, represents the attention weight of the sub-parameter at the t-th historical moment, represents the sub-parameter at the t-th historical moment, It represents the mean of all sub-parameters before the t-th historical moment, and norm represents the normalization function.

[0013] The attention weights can provide a preliminary understanding of the possibility that the sub-parameters in the corresponding historical moments are in an abnormal state, providing a theoretical basis for updating the parameters in the long short-term memory network model.

[0014] Preferably, the method further comprises: utilizing an STL algorithm to decompose sub-parameters of historical moments to obtain trend components of each moment.

[0015] Preferably, the expression of the first impact factor is:

[0016] In the formula, represents the first influencing factor of the sub-parameter at the t-th historical moment, represents the normalization function, represents the attention weight of the sub-parameter at the t-th historical moment, , Represent the minimum and maximum values ​​of the trend component, respectively. represents the trend component at the t-th historical moment, represents the sub-parameter of the t+ith historical moment, represents the mean of all sub-parameters before the t-th historical moment, Indicates the tth The time length between the sub-parameter of the i-th historical moment and the sub-parameter of the t-th historical moment, and n represents half of the number of remaining moments in the time period of the t-th historical moment.

[0017] The first impact factor is calculated through multiple factors, which improves the accuracy of the calculation results. The first impact factor comprehensively reflects the possibility that the parameters at the corresponding historical moment are abnormal.

[0018] Preferably, the method further includes: taking the variance of the sub-parameters of all moments before each historical moment as the overall variance of the corresponding historical moment, and taking the variance of the sub-parameters of each moment within the time period of each historical moment as the local variance of the corresponding historical moment.

[0019] Preferably, the expression of the second impact factor is:

[0020] In the formula, represents the second impact factor corresponding to the t-th historical moment, represents the overall variance corresponding to the t-th historical moment, represents the local variance corresponding to the t-th historical moment, and β represents the preset weight.

[0021] The second influencing factor is obtained by comparing the overall variance and the local variance. The ratio method is more suitable for capturing significant fluctuations in parameters.

[0022] Preferably, the expression of the influence degree of the sub-parameter is:

[0023] In the formula, represents the influence of the sub-parameter at the t-th historical moment, represents the first influencing factor of the sub-parameter at the t-th historical moment, Represents the second impact factor corresponding to the t-th historical moment.

[0024] The first influencing factor and the second influencing factor are weightedly fused to obtain the degree of influence, which comprehensively reflects the importance of the sub-parameters at the corresponding historical moment.

[0025] Preferably, the expression of the influence degree of historical moments is:

[0026] In the formula, represents the degree of influence at the t-th historical moment, Indicates the influence of the first sub-parameter at the t-th historical moment, Indicates the influence of the second sub-parameter at the t-th historical moment.

[0027] Preferably, the multiple sub-parameters are vibration frequency and temperature.

[0028] Temperature changes can reflect the lubrication status, friction and potential thermal damage of the bearing, and can evaluate the health status of the bearing. Vibration frequency can reflect the status information of the machinery, such as imbalance, misalignment, and bearing failure. By considering both temperature and vibration frequency, the operating status of the bearing can be more comprehensively evaluated, avoiding the limitations of a single data source.

[0029] The present invention has the following technical effects: In the process of training the LSTM network model, the learning rate is adaptively adjusted according to the degree of abnormality at different historical moments. When the degree of abnormality is high, the learning rate is reduced to avoid the model's overreaction to abnormal data. When the degree of abnormality is low, the learning rate is increased to speed up the model convergence, thereby improving the accuracy of the prediction results of the LSTM network model. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of a method for predicting the life of a spindle bearing of the present invention.

[0031] Figure 2 Schematic diagram of parameter weights of the present invention.

[0032] Figure 3It is a comparison chart between the prediction results of the long short-term memory network model of the present invention and the prior art. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are 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 creative work are within the scope of protection of the present invention.

[0034] The embodiment of the present invention discloses a method for predicting the life of a main shaft bearing, referring to Figure 1 , including the following steps: S1: Obtain the parameters of the bearing at different historical moments, which include multiple sub-parameters.

[0035] Obtain the parameters of the bearing at different times when it is in working condition. The multiple sub-parameters are temperature data and vibration frequency data. When the bearing is working abnormally, greater friction occurs inside the bearing or friction occurs between the bearing and external components, which causes the bearing temperature to rise. At the same time, the bearing will shake when it is abnormal. Therefore, the bearing temperature and vibration frequency can reflect the working condition of the bearing; use interpolation methods (such as linear interpolation) to fill in missing values ​​for the obtained parameters, and use sliding average or low-pass filtering to smooth the temperature data and vibration frequency data respectively.

[0036] Exemplarily, the frequency of collecting the historical temperature data and the historical vibration frequency data of the bearing is once every 1 second. The unit of the historical temperature data is ° C, and the unit of the historical vibration frequency data is Hz. The collected temperature and vibration frequency data are standardized or normalized.

[0037] S2: Divide the historical moments to obtain the time period of each historical moment.

[0038] Taking any historical moment as the center, we extend multiple moments forward and backward along the time direction to obtain the time period of the corresponding historical moment.

[0039] For example, the historical moment is , , …, , for which time, extending 5 times forward and backward along the time direction, we get Time period ( , …, , …, ), similarly, for time, extending 5 times forward and backward along the time direction, we get Time period ( , …, , …, ).

[0040] S3: Calculate the first impact factor and the second impact factor of the sub-parameter at each historical moment.

[0041] The first impact factor represents the average deviation of the sub-parameter in the time period of the corresponding historical moment, and the second impact factor represents the fluctuation range of the sub-parameter in the time period of the corresponding historical moment.

[0042] S31: Calculate the attention weight of each sub-parameter at each historical moment, the expression is:

[0043] In the formula, represents the attention weight of the sub-parameter at the t-th historical moment, represents the sub-parameter at the t-th historical moment, represents the mean of all sub-parameters before the t-th historical moment, and norm represents the normalization function. For example, for the 10th historical moment In terms of The value of , , …, The mean of the sub-parameters at time .

[0044] Combination Figure 2 As shown in the figure, the attention weight is adjusted according to the difference between the sub-parameter at the current t-th historical moment and the average value. The larger the difference, the larger the attention weight, making it easier to capture abnormal situations when training the long short-term memory network model; the smaller the difference, the smaller the attention weight, which reduces its attention to normal situations and reduces the amount of calculation. The attention weight can reflect the abnormal situation of the sub-parameter data itself to a certain extent. Specifically, the larger its value, the greater the possibility of abnormal sub-parameters at the corresponding moment. Conversely, the smaller its value, the smaller the possibility of abnormal sub-parameters at the corresponding moment.

[0045] Identifying abnormal trends in historical temperature data or vibration frequency data depends not only on the temperature data or vibration frequency data at a single time point, but also on the temperature data or vibration frequency data at surrounding time points. By comparing the difference between the current temperature data (or vibration frequency data) and the historical average temperature data (or vibration frequency data), the abnormality at a single time point is captured. It can be understood that the temperature data and vibration frequency data each correspond to a weight of attention.

[0046] S32: Use the STL algorithm to decompose the sub-parameters of the historical moments to obtain the trend components of each moment and calculate the first influencing factor.

[0047] The temperature data of historical moments are decomposed by using the STL algorithm to obtain the trend component of the temperature data at each moment, and the vibration frequency data of historical moments are decomposed by using the STL algorithm to obtain the trend component of the vibration frequency data at each moment. It should be noted that the STL algorithm is a prior art, and the calculation method of the trend component is not repeated here.

[0048] The expression of the first impact factor is:

[0049] In the formula, represents the first influencing factor of the sub-parameter at the t-th historical moment, represents the normalization function, represents the attention weight of the sub-parameter at the t-th historical moment, , Represent the minimum and maximum values ​​of the trend component, respectively. represents the trend component at the t-th historical moment, represents the sub-parameter of the t+ith historical moment, represents the mean of all sub-parameters before the t-th historical moment, Indicates the tth The time length between the sub-parameter of the i-th historical moment and the sub-parameter of the t-th historical moment, and n represents half of the number of remaining moments in the time period of the t-th historical moment.

[0050] For example, The time period of the moment is ( , …, , …, ), the number of moments (number of data points) in the time period is 11, and the value of n is 5.

[0051] in, It represents the difference between the sub-parameters corresponding to each moment in the time period where the t-th historical moment is located and the overall sub-parameters. The larger the value, the greater the probability that the t-th historical moment and the surrounding sub-parameters are abnormal, which further indicates that the bearing is in an abnormal working state.

[0052] The normalized value representing the difference between the trend component corresponding to the sub-parameter at the t-th historical moment and the maximum value of the trend component reflects the relative size of the trend component corresponding to the sub-parameter at the t-th historical moment in the overall trend component. The larger the difference, the more likely it is that the sub-parameter at the t-th historical moment is abnormal.

[0053] In summary, the first influencing factor indicates the possibility of abnormality of the sub-parameter at the corresponding historical moment. The larger its value is, the greater the possibility that the sub-parameter at the corresponding historical moment is abnormal, which further indicates that the bearing is more likely to be working abnormally at the corresponding moment.

[0054] S33: Calculate the second impact factor.

[0055] The variance of the sub-parameters at all moments before each historical moment is taken as the overall variance of the corresponding historical moment, and the variance of the sub-parameters at each moment in the time period of each historical moment is taken as the local variance of the corresponding historical moment.

[0056] Exemplary, for the 10th historical moment In terms of , , …, The variance of the moment sub-parameter is used as the historical moment The overall variance of The time period ( , …, , …, ) as the variance of the sub-parameters at each historical moment The local variance of .

[0057] The expression of the second impact factor is:

[0058] In the formula, represents the second impact factor corresponding to the t-th historical moment, represents the overall variance corresponding to the t-th historical moment, represents the local variance corresponding to the t-th historical moment, β represents the preset weight, and the value of the weight is set manually according to the situation. For example, the value of the weight is 0.5.

[0059] The second influencing factor indicates the difference between the local variance and the overall variance of the sub-parameter in the time period of the t-th historical moment. The larger its value is, the greater the possibility that the value of the sub-parameter has changed greatly near the t-th historical moment, and the greater the possibility that the bearing is abnormal.

[0060] It can be understood that, for the t-th historical moment, the temperature data corresponds to a first influencing factor and a second influencing factor, and the vibration frequency data corresponds to a first influencing factor and a second influencing factor.

[0061] S4: Perform a weighted summation of the first influencing factor and the second influencing factor to obtain the influence degree of the sub-parameter at each historical moment.

[0062] The expression of the influence degree of sub-parameters is:

[0063] In the formula, represents the influence of the sub-parameter at the t-th historical moment, represents the first influencing factor of the sub-parameter at the t-th historical moment, Represents the second impact factor corresponding to the t-th historical moment.

[0064] The impact degree comprehensively reflects the abnormal situation of the sub-parameters at the corresponding historical moment. It can be understood that the temperature data at each historical moment corresponds to an impact degree, and the vibration frequency data corresponds to an impact degree.

[0065] S5: Perform a weighted summation of the influence levels of multiple sub-parameters at each historical moment to obtain the influence level of the corresponding historical moment.

[0066] The expression of the influence of historical moments is:

[0067] In the formula, represents the degree of influence at the t-th historical moment, Indicates the influence of the first sub-parameter at the t-th historical moment, It indicates the influence of the second sub-parameter at the t-th historical moment. It can also be understood as, represents the influence of temperature at the t-th historical moment, Indicates the degree of influence of the vibration frequency at the t-th historical moment.

[0068] S6: Construct a long short-term memory network model, calculate the learning rate of the long short-term memory network model, and use historical parameters to train the long short-term memory network model to obtain a prediction model for predicting the bearing life.

[0069] The expression of the learning rate at each historical moment is: ; In the formula, represents the learning rate corresponding to the t-th historical moment, represents the preset initial learning rate, Indicates the influence degree of the tth historical moment. For example, the initial learning rate is 0.1.

[0070] Combination Figure 3As shown in the figure, when the influence of the t-th historical moment is greater, the corresponding learning rate is relatively small. A smaller learning rate can slow down the speed of parameter updating, making the long short-term memory network model more responsive to abnormal data and avoiding too fast updates, thereby improving the prediction accuracy of the spindle bearings. Conversely, when the influence of the t-th historical moment is smaller, the corresponding learning rate is larger. A larger learning rate can speed up the convergence of the model.

[0071] The sub-parameters of the bearing are obtained in real time, and the sub-parameters are input into the trained long short-term memory network model to obtain the service life of the bearing.

[0072] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting the life of a spindle bearing, characterized in that: Includes steps: Obtain the parameters of the bearing at different historical moments, the parameters include multiple sub-parameters; Constructing a long short-term memory network model, calculating the learning rate of the long short-term memory network model, and using historical parameters to train the long short-term memory network model to obtain a prediction model for predicting the bearing life; The learning rate is calculated as follows: divide the historical moments into the time period of each historical moment; calculate the first influencing factor and the second influencing factor of the sub-parameter of each historical moment, where the first influencing factor represents the average deviation of the sub-parameter in the time period of the corresponding historical moment, and the second influencing factor represents the fluctuation range of the sub-parameter in the time period of the corresponding historical moment; perform weighted summation of the first influencing factor and the second influencing factor to obtain the influence degree of the sub-parameter at each historical moment; perform weighted summation of the influence degrees of multiple sub-parameters at each historical moment to obtain the influence degree of the corresponding historical moment, and calculate the learning rate of each historical moment. The expression is: ; In the formula, represents the learning rate corresponding to the t-th historical moment, represents the preset initial learning rate, Represents the degree of influence at the t-th historical moment.

2. A method for predicting the life of a main shaft bearing according to claim 1, characterized in that: The method of dividing the historical moments to obtain the time period of each historical moment is: taking any historical moment as the center, extending multiple moments forward and backward along the time direction to obtain the time period of the corresponding historical moment.

3. A method for predicting the life of a main shaft bearing according to claim 1, characterized in that: The method also includes: calculating the attention weight of each sub-parameter at each historical moment, the expression is: In the formula, represents the attention weight of the sub-parameter at the t-th historical moment, represents the sub-parameter at the t-th historical moment, It represents the mean of all sub-parameters before the t-th historical moment, and norm represents the normalization function.

4. A method for predicting the life of a main shaft bearing according to claim 3, characterized in that: The method also includes: using the STL algorithm to decompose the sub-parameters of the historical moments to obtain the trend components of each moment.

5. A method for predicting the life of a main shaft bearing according to claim 4, characterized in that: The expression of the first impact factor is: In the formula, represents the first influencing factor of the sub-parameter at the t-th historical moment, norm represents the normalization function, represents the attention weight of the sub-parameter at the t-th historical moment, , Represent the minimum and maximum values ​​of the trend component, respectively. represents the trend component at the t-th historical moment, represents the sub-parameter of the t+ith historical moment, represents the mean of all sub-parameters before the t-th historical moment, Indicates the tth The time length between the sub-parameter of the i-th historical moment and the sub-parameter of the t-th historical moment, and n represents half of the number of remaining moments in the time period of the t-th historical moment.

6. A method for predicting the life of a main shaft bearing according to claim 1, characterized in that: The method also includes: taking the variance of the sub-parameters of all moments before each historical moment as the overall variance of the corresponding historical moment, and taking the variance of the sub-parameters of each moment in the time period of each historical moment as the local variance of the corresponding historical moment.

7. A method for predicting the life of a main shaft bearing according to claim 6, characterized in that: The expression of the second impact factor is: In the formula, represents the second impact factor corresponding to the t-th historical moment, represents the overall variance corresponding to the t-th historical moment, represents the local variance corresponding to the t-th historical moment, and β represents the preset weight.

8. A method for predicting the life of a main shaft bearing according to claim 1, characterized in that: The expression of the influence degree of sub-parameters is: In the formula, represents the influence of the sub-parameter at the t-th historical moment, represents the first influencing factor of the sub-parameter at the t-th historical moment, Represents the second impact factor corresponding to the t-th historical moment.

9. A method for predicting the life of a main shaft bearing according to claim 1, characterized in that: The expression of the influence of historical moments is: In the formula, represents the degree of influence at the t-th historical moment, Indicates the influence of the first sub-parameter at the t-th historical moment, Indicates the influence of the second sub-parameter at the t-th historical moment.

10. A method for predicting the life of a main shaft bearing according to claim 1, characterized in that: The multiple sub-parameters are vibration frequency and temperature.

Images