Method and device for predicting residual life of rolling bearing of general wharf equipment
Through the signal reconstruction of the vibration signal of the rolling bearings of the dock device, the bearing degradation trend consistency is improved, and the reconstruction signal training life predictive model is used to solve the problem of low accuracy in the existing technology and achieve high accuracy life forecasting. Essence
Patent Information
- Application Number
- CN202411365080.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-29
- Publication Date
- 2025-01-21
AI Technical Summary
In the prior art, due to the inconsistent degradation trends of bearings of different models, the prediction accuracy of the remaining service life of the bearing is low.
By obtaining the vibration signal of the rolling bearings, the average root value sequence is determined, and the signal reconstruction is performed to enhance the consistency of the bearing degradation trend. Then, the reconstruction signal is used to train the rolling bearings life prediction model to obtain a well -trained model.
The prediction accuracy of the rolling bearing life prediction model is significantly improved, and high-precision prediction of rolling bearing life is achieved.
Smart Images
Figure CN119337086A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of equipment status monitoring and fault diagnosis, and in particular to a method and a device for predicting the remaining life of a rolling bearing of general terminal equipment. Background Art
[0002] In port equipment, rolling bearings, as one of the key components of rotating machinery, are known as the "hub of industry" because of their key role in industrial operations. However, rolling bearings often face harsh conditions when working, including high-speed operation, heavy loads, high-temperature environments, dust, and alternating loads. These factors make rolling bearings the most vulnerable part of the mechanical system. Once a rolling bearing fails, it may cause serious economic losses and even pose a threat to the safety of operators. In view of these risks, it is crucial to ensure the safety of equipment and avoid catastrophic accidents. Therefore, accurate prediction of the remaining service life of rolling bearings has become a topic of widespread concern in academia and industry.
[0003] However, due to the inconsistent degradation trends of different bearings of the same model, when predicting the remaining service life, the parameters of the prediction model training set and the test set do not match, resulting in low model training accuracy. Therefore, improving the consistency of bearing degradation trends is crucial to improving the accuracy of the prediction model. Summary of the invention
[0004] In view of this, it is necessary to provide a method and device for predicting the remaining life of rolling bearings of general terminal equipment to solve the technical problem in the prior art of low accuracy in predicting the remaining service life of bearings due to inconsistent degradation trends of different bearings of the same model.
[0005] In order to solve the above problems, the present invention provides a method for predicting the remaining life of a rolling bearing of a general terminal equipment, comprising:
[0006] Acquire a vibration signal of the rolling bearing, and determine a root mean square value sequence of the vibration signal based on the time series;
[0007] Reconstructing the root mean square value sequence of the vibration signal to obtain a reconstructed signal;
[0008] The reconstructed signal is used to train the rolling bearing life prediction model, and a fully trained rolling bearing life prediction model is obtained;
[0009] The reconstructed signal is input into the well-trained rolling bearing life prediction model to obtain the predicted value of rolling bearing life.
[0010] In a possible implementation, performing signal reconstruction on a root mean square value sequence of the vibration signal to obtain a reconstructed signal includes:
[0011] The preset seven-point rolling smoothing method is used to smooth the RMS sequence to obtain a smooth signal;
[0012] Using a preset abnormal value replacement to replace the abnormal value of the smoothed signal to obtain a replacement signal;
[0013] The replacement signal is accumulated to obtain the reconstructed signal.
[0014] In a possible implementation, the preset seven-point rolling smoothing method is used to smooth the RMS sequence, which can be expressed by the following formula:
[0015]
[0016] Among them, x n is the RMS value data point, N is the number of sequences, are the smoothed data points.
[0017] In a possible implementation, replacing the abnormal value of the smoothed signal with a preset abnormal value to obtain a replacement signal includes:
[0018] Differences are made between the data points in the smoothed signal to obtain a signal degradation difference sequence;
[0019] According to the preset 3sigma principle, determine the standard value interval of the signal degradation difference;
[0020] According to the size relationship between each degradation difference value in the signal degradation difference value sequence and the standard value interval, determining whether each degradation difference value in the signal degradation difference value sequence is an abnormal value;
[0021] According to the difference consistency principle, the outliers are replaced to obtain the replacement signal.
[0022] In a possible implementation, the accumulating the replacement signal to obtain the reconstructed signal includes:
[0023] Calculate the sum of the 1st value to the nth value in the replacement signal to obtain the nth value of the reconstructed signal sequence.
[0024] In a possible implementation, the rolling bearing life prediction model includes a gated neural unit and an attention mechanism module, and the reconstructed signal is used to train the rolling bearing life prediction model, including:
[0025] The reconstructed signal is input into the gated neural unit network, and the hidden state of the reconstructed signal is output;
[0026] Based on the attention mechanism module, the hidden state is weighted and the target weight is obtained;
[0027] Multiplying the hidden layer in the gated neural network by the target weight to obtain the target hidden state of the gated neural network;
[0028] According to the target hidden state, the rolling bearing life prediction model is trained to obtain a fully trained rolling bearing life prediction model.
[0029] In a possible implementation, the root mean square value of the vibration signal determined by the vibration signal can be expressed by the following formula:
[0030]
[0031] Where, RMS is the root mean square value of the vibration signal, x i is the amplitude of the time series at time i, N is the number of x in the time series i Total number.
[0032] In a second aspect, the present invention further provides a device for predicting the remaining life of a rolling bearing of a general terminal equipment, comprising:
[0033] An acquisition module, used for acquiring a vibration signal of a rolling bearing, and determining a root mean square value sequence of the vibration signal based on a time series;
[0034] A reconstruction module, used for performing signal reconstruction on the root mean square value sequence of the vibration signal to obtain a reconstructed signal;
[0035] A model training module is used to train the rolling bearing life prediction model using the reconstructed signal to obtain a fully trained rolling bearing life prediction model;
[0036] The life prediction module is used to input the reconstructed signal into a well-trained rolling bearing life prediction model to obtain a predicted value of the rolling bearing life.
[0037] In a third aspect, the present invention further provides an electronic device, comprising: a processor and a memory;
[0038] The memory stores a computer-readable program executable by the processor;
[0039] When the processor executes the computer-readable program, the steps in the method for predicting the remaining life of a rolling bearing of a general terminal equipment as described above are implemented.
[0040] In a fourth aspect, the present invention also provides a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the method for predicting the remaining life of rolling bearings of general terminal equipment as described above.
[0041] The beneficial effects of the present invention are as follows: first, the vibration signal of the rolling bearing is obtained, and the root mean square value sequence of the vibration signal is determined based on the time series; then, the root mean square value sequence of the vibration signal is reconstructed to obtain a reconstructed signal; thereby achieving the purpose of improving the consistency of the bearing degradation trend by signal reconstruction, and using the reconstructed signal to train the rolling bearing life prediction model to obtain a fully trained rolling bearing life prediction model; the prediction accuracy of the model is significantly improved, and finally the reconstructed signal is input into the fully trained rolling bearing life prediction model to obtain the rolling bearing life prediction value, thereby achieving the purpose of high-precision prediction of the rolling bearing life. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 A method flow chart of an embodiment of a method for predicting the remaining life of a rolling bearing of a universal terminal equipment provided by the present invention;
[0043] Figure 2 for Figure 1 A method flow chart of an embodiment of step S102;
[0044] Figure 3 A schematic diagram of abnormal fluctuations in a vibration signal degradation curve in the remaining life prediction method for rolling bearings of universal terminal equipment provided by the present invention;
[0045] Figure 4 A schematic diagram of a rolling bearing life prediction model in the method for predicting the remaining life of a rolling bearing for universal terminal equipment provided by the present invention;
[0046] Figure 5 It is a schematic diagram of an embodiment of a device for predicting the remaining life of a rolling bearing of a universal terminal equipment provided by the present invention;
[0047] Figure 6 It is a schematic diagram of the operating environment of an electronic device provided by the present invention. DETAILED DESCRIPTION
[0048] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.
[0049] A specific embodiment of the present invention discloses a method for predicting the remaining life of a rolling bearing of a general terminal equipment. Figure 1 ,include:
[0050] S101, acquiring a vibration signal of a rolling bearing, and determining a root mean square value sequence of the vibration signal based on a time series;
[0051] S102, performing signal reconstruction on a root mean square value sequence of the vibration signal to obtain a reconstructed signal;
[0052] S103, using the reconstructed signal to train the rolling bearing life prediction model to obtain a fully trained rolling bearing life prediction model;
[0053] S104, inputting the reconstructed signal into a fully trained rolling bearing life prediction model to obtain a predicted value of the rolling bearing life.
[0054] In this embodiment, the vibration signal of the rolling bearing is first obtained, and the root mean square value sequence of the vibration signal is determined based on the time series; then the root mean square value sequence of the vibration signal is reconstructed to obtain a reconstructed signal; thereby achieving the purpose of improving the consistency of the bearing degradation trend by means of signal reconstruction, and using the reconstructed signal to train the rolling bearing life prediction model to obtain a fully trained rolling bearing life prediction model; the prediction accuracy of the model is significantly improved, and finally the reconstructed signal is input into the fully trained rolling bearing life prediction model to obtain the rolling bearing life prediction value, thereby achieving the purpose of high-precision prediction of the rolling bearing life.
[0055] In step S101, before predicting the bearing life, an input signal needs to be selected. Common input signals include vibration signals, oil signals, sound signals, etc. Vibration signals are widely used in rolling bearing fault diagnosis and remaining service life prediction because they are easy to extract and contain a large amount of fault information. The root mean square, as one of the time domain characteristics of the vibration signal, is highly sensitive to the bearing life degradation label curve. Therefore, the root mean square is selected as the input signal.
[0056] It should be noted that the root mean square (RMS) value of a vibration signal is usually a single value, but it can be expressed as a sequence that varies with time. In practical applications, the root mean square value of a vibration signal can be calculated at different time points. Usually, the root mean square calculation is performed on the signal within a time window to obtain an instantaneous root mean square value. In this way, the root mean square value will change with the passage of time to form a continuous time series. In this embodiment, by obtaining a root mean square value sequence that varies with time, the changing trend of the signal is monitored and analyzed in real time, thereby realizing the prediction of the remaining service life of the rolling bearing.
[0057] Among them, the calculation formula of the root mean square degradation curve RMS of the vibration signal of the rolling bearing is as follows:
[0058]
[0059] Among them, x i is the amplitude of the time series at time i, N is the number of x in the time series i Total number.
[0060] In step S102, reconstruction refers to recovering or regenerating a complete signal or some representation thereof from known signal data or features. In this embodiment, the root mean square value sequence of the vibration signal is reconstructed so that the rolling bearing degradation trend shows strong consistency. By using a model with strong trend consistency to train the rolling bearing remaining life prediction model, the prediction accuracy of the rolling bearing remaining life prediction model can be effectively improved.
[0061] In some embodiments, the root mean square value sequence of the vibration signal is reconstructed to obtain a reconstructed signal. Figure 2 ,include:
[0062] S201, smoothing the RMS sequence using a preset seven-point rolling smoothing method to obtain a smoothed signal;
[0063] S202, replacing the abnormal value of the smoothed signal with a preset abnormal value replacement to obtain a replacement signal;
[0064] S203: Perform accumulation processing on the replacement signal to obtain a reconstructed signal.
[0065] In this embodiment, after extracting the RMS degradation curve of the bearing vibration signal, in order to improve the consistency of its degradation trend, it is necessary to reconstruct the signal.
[0066] It should be noted that since the RMS degradation curve contains many burrs, it needs to be smoothed. The seven-point rolling smoothing method is used to smooth it. The calculation formula is as follows:
[0067]
[0068] Among them, x n is the RMS value data point, N is the number of sequences, are the smoothed data points.
[0069] It should be noted that in the process of predicting the remaining life of rolling bearings of general terminal equipment, due to the influence of external noise and other equipment coupling factors, the input signal degradation curve will show abnormal fluctuations, such as Figure 3 As shown. The appearance of outliers will cause the degradation curve to lose its original degradation information. If it is not processed and input into the prediction model for training, it will cause the model training parameters to be distorted. Therefore, restoring the signal degradation curve is crucial to improving the prediction accuracy of the model. The 3sigma principle has the effect of eliminating random errors. In order to restore the true degradation trend of the signal degradation curve, an outlier replacement method is proposed based on the 3sigma principle. The method steps are as follows:
[0070] Assume that the original sequence of signal degradation curves is X = [x1, x2, …, xn , calculate the difference sequence Y of the signal degradation curve as shown in the following formula:
[0071] Y = [y1, y2, …, y n-1
[0072] where y n is the absolute value of the amplitude difference between adjacent points of the signal degradation curve, as shown in the following formula:
[0073] y n = |x n+1 - x n |
[0074] According to the 3sigma principle, abnormal fluctuations are small probability events. For a normal degradation curve, each time point of the difference sequence should be within the standard interval μ - 3σ < y i < μ + 3σ (1 ≤ i < n - 1), where μ and σ are the mean and standard deviation of the difference sequence Y respectively. When y i exceeds this range, it is considered an outlier and should be replaced. To restore the true degradation trend as much as possible, the replacement value should ensure the same difference as the previous time point. Therefore, the outlier replacement formula is as follows:
[0075] x i+1 = x i + y i
[0076] where y i < μ - 3σ ∪ y i > μ + 3σ.
[0077] It should be noted that due to the influence of various factors in the bearing degradation stage, the root mean square degradation curve shows low trend consistency. Therefore, to improve the trend consistency of the degradation curve, the root mean square degradation curve after outlier replacement is accumulated, and the method steps are as follows:
[0078] Let the sequence after outlier replacement be X′ = (x′1, x′2, … x′ n ).
[0079] where n is the sequence length.
[0080] Perform an accumulation process on it, and the accumulation sequence X′ A is as shown in the following formula:
[0081]
[0082] It should be noted that the root mean square degradation curve shows high trend consistency after signal reconstruction, laying a foundation for improving the prediction accuracy of the subsequent model.
[0083] In some embodiments, the rolling bearing life prediction model includes a gated neural unit and an attention mechanism module, and the reconstructed signal is used to train the rolling bearing life prediction model, including:
[0084] The reconstructed signal is input into the gated neural unit network, and the hidden state of the reconstructed signal is output;
[0085] Based on the attention mechanism module, the hidden state is weighted and the target weight is obtained;
[0086] Multiplying the hidden layer in the gated neural network by the target weight to obtain the target hidden state of the gated neural network;
[0087] According to the target hidden state, the rolling bearing life prediction model is trained to obtain a fully trained rolling bearing life prediction model.
[0088] In this embodiment, the gated recurrent unit (GRU) is particularly sensitive to time series compared to other artificial neural networks, and avoids the problems of gradient vanishing and gradient explosion. It has high computational efficiency and is widely used in the prediction of the remaining useful life of bearings. In the process of predicting the remaining useful life of bearings, the contribution of each moment of the time series to the prediction result is different. Therefore, different weights need to be assigned to the contribution at different moments of the time series. Attention is a machine learning algorithm that draws on the principle of human eyes processing key information, identifies key information and assigns it a larger weight, and vice versa.
[0089] It should be noted that the GRU model is combined with the Attention mechanism to construct a new bearing RUL prediction model GRUA. The schematic diagram is as follows: Figure 4 shown.
[0090] in It means adding the same positions of matrices. Indicates matrix multiplication at the same position.
[0091] Assume that the input time series X = (x1, x2, ... x T ), after the GRU model outputs the hidden state H t (1≤t≤T), the calculation formula is as follows:
[0092] Z t =σ(W Z x t +U Z H t-1 )
[0093] R t =σ(W R x t+U R H t-1 )
[0094] Hh=tanh(W H x t +U H (R t ⊙H t-1 ))
[0095] H t =(1-Z t )hH t-1 +Z t ⊙h t
[0096] Among them, x t represents the input at each moment, σ(*) and tanh(*) are activation functions, Z t represents the update gate output, R t Represents the reset gate output, W Z , W R , W H , U Z , U R , U H Represents the weight matrix.
[0097] After the GRU model pre-trains the time series, its hidden state H is t Redistribute weights, weights at each moment ω i The calculation formula is as follows:
[0098]
[0099] Among them, ω A is the trainable weight matrix.
[0100] Multiply the hidden layer by the weights at each moment to get the final output H o , the calculation formula is as follows:
[0101]
[0102] Furthermore, when using the GRUA model for prediction, it is necessary to first select a certain number of training set data of bearings to train its model parameters until convergence, and then select the test set bearings to verify its prediction accuracy.
[0103] Based on the above-mentioned method for predicting the remaining life of rolling bearings of general terminal equipment, the embodiment of the present invention also provides a device for predicting the remaining life of rolling bearings of general terminal equipment, see Figure 5 ,include:
[0104] An acquisition module 510 is used to acquire a vibration signal of a rolling bearing and determine a root mean square value sequence of the vibration signal based on a time series;
[0105] A reconstruction module 520, configured to perform signal reconstruction on the root mean square value sequence of the vibration signal to obtain a reconstructed signal;
[0106] A model training module 530 is used to train the rolling bearing life prediction model using the reconstructed signal to obtain a fully trained rolling bearing life prediction model;
[0107] The life prediction module 540 is used to input the reconstructed signal into a well-trained rolling bearing life prediction model to obtain a predicted value of the rolling bearing life.
[0108] like Figure 6 As shown, based on the above-mentioned method for predicting the remaining life of rolling bearings of general terminal equipment, the present invention also provides an electronic device, which can be a computing electronic device such as a mobile terminal, a desktop computer, a notebook, a palm computer, and a server. The electronic device includes a processor 610, a memory 620, and a display 630. Figure 6 Only some components of the electronic device are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.
[0109] In some embodiments, the memory 620 may be an internal storage unit of the electronic device, such as a hard disk or memory of the electronic device. In other embodiments, the memory 620 may also be an external storage electronic device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the memory 620 may also include both an internal storage unit of the electronic device and an external storage electronic device. The memory 620 is used to store application software and various types of data installed in the electronic device, such as program codes installed in the electronic device. The memory 620 may also be used to temporarily store data that has been output or is to be output. In one embodiment, the memory 620 stores a universal terminal equipment rolling bearing remaining life prediction program 640, which can be executed by the processor 610, thereby realizing the universal terminal equipment rolling bearing remaining life prediction method of each embodiment of the present application.
[0110] In some embodiments, the processor 610 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 620, such as executing a method for predicting the remaining life of rolling bearings of general terminal equipment.
[0111] In some embodiments, the display 630 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 630 is used to display information of the rolling bearing remaining life prediction electronic device of the universal terminal equipment and to display a visual user interface. The components 610-630 of the electronic device communicate with each other via a system bus.
[0112] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.
[0113] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.
Claims
1. A method for predicting the remaining life of rolling bearings of general terminal equipment, characterized in that: include: Acquire a vibration signal of the rolling bearing, and determine a root mean square value sequence of the vibration signal based on the time series; Reconstructing the root mean square value sequence of the vibration signal to obtain a reconstructed signal; The reconstructed signal is used to train the rolling bearing life prediction model, and a fully trained rolling bearing life prediction model is obtained; The reconstructed signal is input into the well-trained rolling bearing life prediction model to obtain the predicted value of rolling bearing life.
2. The method for predicting the remaining life of rolling bearings of general terminal equipment according to claim 1 is characterized in that: Performing signal reconstruction on the root mean square value sequence of the vibration signal to obtain a reconstructed signal includes: The preset seven-point rolling smoothing method is used to smooth the RMS sequence to obtain a smooth signal; Using a preset abnormal value replacement to replace the abnormal value of the smoothed signal to obtain a replacement signal; The replacement signal is accumulated to obtain the reconstructed signal.
3. The method for predicting the remaining life of rolling bearings of general terminal equipment according to claim 2 is characterized in that: The preset seven-point rolling smoothing method is used to smooth the RMS sequence, which can be expressed by the following formula: Among them, x n is the RMS value data point, N is the number of sequences, are the smoothed data points.
4. The method for predicting the remaining life of rolling bearings of general terminal equipment according to claim 2 is characterized in that: The step of replacing the abnormal value of the smoothed signal with a preset abnormal value to obtain a replacement signal includes: Differences are made between the data points in the smoothed signal to obtain a signal degradation difference sequence; According to the preset 3sigma principle, determine the standard value interval of the signal degradation difference; According to the size relationship between each degradation difference value in the signal degradation difference value sequence and the standard value interval, determining whether each degradation difference value in the signal degradation difference value sequence is an abnormal value; According to the difference consistency principle, the outliers are replaced to obtain the replacement signal.
5. The method for predicting the remaining life of rolling bearings of general terminal equipment according to claim 2 is characterized in that: The step of performing accumulation processing on the replacement signal to obtain the reconstructed signal includes: Calculate the sum of the 1st value to the nth value in the replacement signal to obtain the nth value of the reconstructed signal sequence.
6. The method for predicting the remaining life of rolling bearings of general terminal equipment according to claim 1 is characterized in that: The rolling bearing life prediction model includes a gated neural unit and an attention mechanism module, and the reconstructed signal is used to train the rolling bearing life prediction model, including: The reconstructed signal is input into the gated neural unit network, and the hidden state of the reconstructed signal is output; Based on the attention mechanism module, the hidden state is weighted and the target weight is obtained; Multiplying the hidden layer in the gated neural network by the target weight to obtain the target hidden state of the gated neural network; According to the target hidden state, the rolling bearing life prediction model is trained to obtain a fully trained rolling bearing life prediction model.
7. The method for predicting the remaining life of rolling bearings of general terminal equipment according to claim 1 is characterized in that: The vibration signal determines the root mean square value of the vibration signal which can be expressed by the following formula: Where, RMS is the root mean square value of the vibration signal, x i is the amplitude of the time series at time i, N is the number of x in the time series i Total number.
8. A device for predicting the remaining life of rolling bearings of general terminal equipment, characterized in that: include: An acquisition module, used for acquiring a vibration signal of a rolling bearing, and determining a root mean square value sequence of the vibration signal based on a time series; A reconstruction module, used for performing signal reconstruction on the root mean square value sequence of the vibration signal to obtain a reconstructed signal; A model training module is used to train the rolling bearing life prediction model using the reconstructed signal to obtain a fully trained rolling bearing life prediction model; The life prediction module is used to input the reconstructed signal into a well-trained rolling bearing life prediction model to obtain a predicted value of the rolling bearing life.
9. An electronic device, characterized in that: include: Processor and memory; The memory stores a computer-readable program executable by the processor; When the processor executes the computer-readable program, the steps in the method for predicting the remaining life of a rolling bearing of a universal terminal equipment as described in any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the method for predicting the remaining life of a rolling bearing of a universal terminal equipment as described in any one of claims 1-7.