Motor fault diagnosis method and system

Abstract

The present invention belongs to the technical field of fault diagnosis, and discloses a motor fault diagnosis method and system, comprising: collecting signals of a motor to be diagnosed to obtain corresponding motor operation signals; performing signal processing on the obtained motor operation signals to obtain enhanced operation signals, and performing feature extraction on the obtained enhanced operation signals to obtain corresponding feature information; inputting the obtained feature information into a pre-built fault diagnosis model to obtain corresponding fault prediction results; and performing diagnostic verification on the fault prediction results to obtain corresponding fault diagnosis results based on the verification results; the present invention achieves higher diagnostic accuracy, faster response speed and lower false alarm rate, which is of great significance for ensuring the normal operation of the motor and extending its service life.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and more specifically, to a motor fault diagnosis method and system. Background Art

[0002] Compared with existing technologies, traditional motor maintenance relies more on regular inspections and manual experience judgment. This method is not only time-consuming and labor-intensive, but also often fails to detect potential problems in a timely manner, leading to sudden failures. In addition, excessive maintenance will also cause a waste of resources and increase unnecessary costs. At the same time, it is too dependent on specific types of sensors or complex data processing processes, which limits their universality and flexibility.

[0003] In view of this, the present invention proposes a motor fault diagnosis method and system to solve the above problems. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solutions:

[0005] A motor fault diagnosis method, comprising:

[0006] Step 1: Collect signals from the motor to be diagnosed and obtain corresponding motor operation signals;

[0007] Step 2: performing signal processing on the obtained motor operation signal to obtain an enhanced operation signal, and performing feature extraction on the obtained enhanced operation signal to obtain corresponding feature information;

[0008] Step three: input the obtained characteristic information into a pre-built fault diagnosis model to obtain a corresponding fault prediction result; and perform diagnostic verification on the fault prediction result to obtain a corresponding fault diagnosis result based on the verification result.

[0009] Furthermore, the process of collecting signals from the motor to be diagnosed and obtaining corresponding motor operation signals includes:

[0010] Setting a data acquisition unit, wherein the data acquisition unit is used to collect operation signals corresponding to the operation parameters during the operation of the target motor;

[0011] The motor to be diagnosed is obtained, and is marked as a motor to be diagnosed. The signal of the corresponding motor to be diagnosed is collected based on the data acquisition unit to obtain a corresponding motor operation signal.

[0012] Furthermore, the obtained motor operation signal is processed to obtain an enhanced operation signal, which includes:

[0013] Read the motor operation signal in the collected motor operation information, and perform signal decomposition on the corresponding motor operation signal based on wavelet transform to obtain high-frequency signals and low-frequency signals at different scales corresponding to N1 decomposition layers;

[0014] Constructing a signal optimization function, and performing signal processing on the target high-frequency signal based on the signal optimization function to obtain a corresponding enhanced high-frequency signal;

[0015] At the same time, the obtained low-frequency signal is adaptively enhanced to obtain a corresponding enhanced low-frequency signal;

[0016] Based on the acquisition process of the enhanced low-frequency signal and the enhanced low-frequency signal, signal processing is performed on the low-frequency signal and the high-frequency signal in each decomposition layer to obtain the enhanced high-frequency signal and the enhanced low-frequency signal corresponding to the corresponding decomposition layer;

[0017] The enhanced high-frequency signal and the enhanced low-frequency signal of each decomposition layer are reconstructed by wavelet to obtain the enhanced decomposition signal corresponding to the corresponding decomposition layer;

[0018] The enhanced decomposition signals corresponding to each decomposition layer are weightedly fused to obtain the corresponding reconstructed operation signal;

[0019] The formula for weighted fusion of the enhanced decomposition signals corresponding to each decomposition layer is: ; In the formula, Represents the reconstructed operation signal after weighted fusion is completed; Indicates The fusion weights corresponding to the layers are ; and the number of decomposition layers The closer , the corresponding fusion weight is lower; represents the total number of decomposition layers in the corresponding wavelet transform process;

[0020] After the fusion is completed, the obtained reconstructed operation signal is fused again with the corresponding original motor operation signal to obtain the corresponding enhanced operation signal;

[0021] The formula for signal fusion is: Where, Indicates enhanced operation signal, , Represent the reconstructed operation signal and the original motor operation signal respectively; is the fusion correction coefficient and .

[0022] Furthermore, the formula of the signal optimization function is:

[0023] ; In the formula, Represents high-frequency signal fragments The function value of the corresponding signal optimization function; Represents the corresponding high-frequency signal segment The corresponding small slope coefficient is , and is an integer, Indicates the total number of high-frequency signal segments;

[0024] Represents high-frequency signal fragments The corresponding wavelet threshold is used to remove noise and retain key information; the corresponding mathematical formula is: ; In the formula, Represents high-frequency signal fragments The corresponding signal length; Represents high-frequency signal fragments The corresponding noise standard deviation, where ;

[0025] represents the adjustment factor corresponding to the corresponding signal optimization function, where ;and ;

[0026] The formula for the adaptive enhancement is: ; In the formula, It represents the enhanced low-frequency signal, that is, the enhanced low-frequency signal; represents the original low-frequency signal, Represents the impulse response of the adaptively adjusted filter set based on the original low-frequency signal.

[0027] Furthermore, the process of extracting features from the obtained enhanced operation signal and obtaining corresponding feature information includes:

[0028] Obtaining enhanced operating signals corresponding to various operating parameters of the corresponding motor to be diagnosed, and performing signal transformation on the obtained enhanced operating signals based on the Hilbert transform algorithm to obtain corresponding characteristic spectra; obtaining characteristic spectra corresponding to all enhanced operating signals, and constructing corresponding characteristic maps based thereon;

[0029] Reading the obtained enhanced operation signal, and obtaining the local maximum value and the local minimum value in the corresponding enhanced operation signal;

[0030] Then, the corresponding local maximum and local minimum are interpolated based on the cubic spline interpolation algorithm to generate the corresponding upper envelope and lower envelope; and the mean of the upper envelope and the lower envelope is obtained; the corresponding mean is subtracted from the enhanced operation signal to obtain the corresponding characteristic component, and it is determined whether the characteristic component satisfies the component function; if not, the process is repeated until the characteristic component that satisfies the component function is obtained;

[0031] Subtracting the characteristic component from the enhanced operation signal to obtain a corresponding residual signal, repeating the characteristic component acquisition process until the characteristic component can no longer be decomposed into a new characteristic component, counting all characteristic components corresponding to the corresponding enhanced operation signal to obtain a corresponding signal characteristic group;

[0032] The formula of the component function is: ; In the formula, The obtained characteristic components are shown;

[0033] Respectively obtain sample entropies of envelope signals corresponding to corresponding characteristic components, and use the sample entropies as characteristic vectors of corresponding enhanced operation signals;

[0034] The characteristic vectors and characteristic spectra corresponding to the enhanced operation signal are counted to obtain corresponding characteristic information.

[0035] Furthermore, the obtained characteristic information is input into a pre-built fault diagnosis model to obtain a corresponding output result, which includes a predicted operating state to which the corresponding characteristic information belongs and a corresponding membership probability.

[0036] Furthermore, the process of building a fault diagnosis model includes:

[0037] Construct a motor database; obtain all fault categories of the motor to be diagnosed and other motors of the same type in a faulty operating state, and obtain a number of historical operating signals corresponding to each fault category based on the motor database; at the same time, it is also necessary to obtain the historical operating signals corresponding to the normal operating state;

[0038] Based on the process of acquiring characteristic information, signal processing is performed on the obtained historical operation signal to obtain characteristic information corresponding to the corresponding historical operation signal;

[0039] Construct corresponding training data based on the feature information corresponding to the corresponding historical operation signal;

[0040] The pre-built fault diagnosis model is defined to be based on a convolutional neural network as a skeleton, which is used to obtain multi-scale feature information from the input feature information at different stages of the backbone network;

[0041] Based on the training set and test set corresponding to the training data component, the fault diagnosis model is initialized; the training set is input into the initialized fault diagnosis model for forward propagation calculation, and the error is obtained through a three-layer convolutional network; the obtained error is propagated to each convolution kernel in the corresponding fault diagnosis model by back propagation;

[0042] Then, by changing the weights corresponding to different convolution kernels, the model weights of the corresponding fault diagnosis model are continuously optimized, and the optimization process of the model weights is repeated until the loss function in the corresponding fault diagnosis model tends to be stable and the accuracy of the model training reaches the highest.

[0043] When the loss function tends to be stable, the model parameters are saved, and the fault diagnosis model is tested and verified based on the test machine. If the verification passes, the construction of the fault diagnosis model is completed. If the verification passes, it is repeated until the corresponding fault diagnosis model meets the requirements.

[0044] Furthermore, the fault prediction result is diagnosed and verified, and a process of obtaining a corresponding fault diagnosis result according to the verification result includes:

[0045] Setting a probability threshold, reading the obtained output result, and comparing the membership probability in the output result with the probability threshold;

[0046] If the membership probability is less than or equal to the probability threshold, no other operation is performed;

[0047] If the membership probability is greater than the probability threshold, the corresponding output result is read, and if the operating state in the output result is a normal operating state, no other operation is performed;

[0048] If there is a fault operation state in the output result, read the fault category corresponding to the corresponding fault operation state in the corresponding output result; and obtain the corresponding historical operation signal based on the fault category, and mark it as a verification signal;

[0049] Obtaining a sensitivity coefficient between a characteristic component corresponding to a corresponding motor operation signal and a corresponding verification signal;

[0050] Set a verification threshold and compare the obtained sensitivity coefficient with the corresponding verification threshold;

[0051] If the sensitivity coefficient is lower than the verification threshold, it indicates that the characteristic component corresponding to the corresponding motor operation signal is not similar to the characteristics of the verification signal; then a corresponding motor maintenance notification is generated, and the operating load of the motor to be diagnosed is appropriately reduced through remote control to reduce the risk of damage to the motor to be diagnosed;

[0052] If the sensitivity coefficient is not lower than the verification threshold, indicating that the characteristic component corresponding to the corresponding motor operation signal is similar to the characteristic of the verification signal, a corresponding fault category warning is generated based on the output result, and corresponding warning measures are taken based on the corresponding fault category.

[0053] Furthermore, the process of obtaining the sensitivity coefficient between the characteristic component corresponding to the corresponding motor operation signal and the corresponding verification signal includes:

[0054] Obtain the characteristic component corresponding to the motor operation signal, and obtain the correlation coefficient between the corresponding characteristic component and the corresponding verification signal; wherein the formula for obtaining the correlation coefficient is:

[0055] ; In the formula, Indicates the verification signal. Indicates the motor running signal corresponding to the feature components, among which , is a natural number, Indicates the total number of characteristic components corresponding to the motor operation signal;

[0056] , Represent the verification signal and The mean square value corresponding to the characteristic components; Indicates the verification signal and The covariance between the feature components;

[0057] Furthermore, the energy corresponding to the corresponding characteristic component is obtained and recorded as ; The corresponding mathematical formula is as follows: ;

[0058] The correlation coefficient and energy corresponding to the corresponding characteristic component are obtained, and based on them, the sensitivity coefficient between the corresponding verification signals is obtained. The formula for obtaining the sensitivity coefficient is:

[0059] ; Represents the sensitivity coefficient.

[0060] Furthermore, a motor fault diagnosis system comprises:

[0061] A data acquisition module is used to collect information data of the motor to be diagnosed and obtain corresponding motor operation signals;

[0062] A data processing module is used to perform signal processing on the collected motor operation signal to obtain a corresponding enhanced operation signal, and perform feature extraction on the enhanced operation signal to obtain corresponding feature information;

[0063] The model diagnosis module is used to construct a fault diagnosis model and input the obtained characteristic information into the constructed fault diagnosis model to obtain the corresponding fault prediction result;

[0064] The data feedback module is used to perform diagnostic verification on the obtained fault prediction results, generate corresponding fault diagnosis results based on the diagnostic verification results and provide feedback.

[0065] The technical effects and advantages of a motor fault diagnosis method and system of the present invention are as follows:

[0066] 1. By constructing a fault diagnosis model, the intelligent diagnosis of motor faults is realized, and the accuracy and efficiency of diagnosis are improved; at the same time, the present invention can process different types of motor fault signals and adapt to the motor operating status under different working conditions, and has strong adaptability and generalization ability.

[0067] 2. After obtaining the fault prediction results, further diagnostic verification is performed on the prediction results by setting probability thresholds and verification thresholds. If the verification results show that the motor has a potential fault, corresponding early warning measures will be taken according to the fault type, and the motor's operating load will be appropriately reduced through remote control to reduce the risk of damage. This verification and maintenance mechanism can ensure the accuracy and reliability of motor fault diagnosis and reduce the risk of false alarms and missed alarms. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A schematic diagram of a motor fault diagnosis method of the present invention;

[0069] Figure 2 It is a schematic diagram of a motor fault diagnosis system of the present invention. DETAILED DESCRIPTION

[0070] 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 only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0071] Example 1

[0072] See also Figure 1 As shown, a motor fault diagnosis method described in this embodiment includes:

[0073] Step 1: Collect signals from the motor to be diagnosed and obtain corresponding motor operation signals;

[0074] Step 2: performing signal processing on the obtained motor operation signal to obtain an enhanced operation signal, and performing feature extraction on the obtained enhanced operation signal to obtain corresponding feature information;

[0075] Step 3: input the obtained characteristic information into a pre-built fault diagnosis model to obtain a corresponding fault prediction result; and perform diagnostic verification on the fault prediction result to obtain a corresponding fault diagnosis result based on the verification result;

[0076] It should be further explained that, in the specific implementation process, the process of collecting signals from the motor to be diagnosed and obtaining corresponding motor operation signals includes:

[0077] Setting a data acquisition unit, wherein the data acquisition unit is used to collect operation signals corresponding to the operation parameters during the operation of the target motor;

[0078] Acquire a motor to be diagnosed, mark it as a motor to be diagnosed, and acquire signals of the corresponding motor to be diagnosed based on the data acquisition unit to obtain a corresponding motor operation signal;

[0079] The operating parameters of the motor include current parameters and vibration parameters, the corresponding motor operating signals include vibration signals and current signals, and the data acquisition unit is provided with a fixed data acquisition cycle and acquisition frequency;

[0080] It should be further explained that the process of performing signal processing on the obtained motor operation signal to obtain the enhanced operation signal includes:

[0081] Read the motor operation signal in the motor operation information collected in the current collection cycle, and perform signal decomposition on the corresponding motor operation signal based on wavelet transform to obtain high-frequency signals and low-frequency signals at different scales corresponding to N1 decomposition layers;

[0082] Read the high-frequency signal in any decomposition layer, obtain the maximum value point in the corresponding high-frequency signal, obtain the midpoint between adjacent maximum value points, and divide the corresponding high-frequency signal based on the midpoint to obtain a number of high-frequency signal segments;

[0083] Constructing a signal optimization function, and performing signal processing on the target high-frequency signal based on the signal optimization function to reduce the influence of noise on the corresponding high-frequency signal and obtain a higher quality enhanced high-frequency signal;

[0084] The signal optimization function shown is:

[0085] ; In the formula, Represents high-frequency signal fragments The function value of the corresponding signal optimization function; Represents the corresponding high-frequency signal segment The corresponding small slope coefficient is , and is an integer, Indicates the total number of high-frequency signal segments;

[0086] Represents high-frequency signal fragments The corresponding wavelet threshold is used to remove noise and retain key information; the corresponding mathematical formula is: ; In the formula, Represents high-frequency signal fragments The corresponding signal length; Represents high-frequency signal fragments The corresponding noise standard deviation, where ;

[0087] represents the adjustment factor corresponding to the corresponding signal optimization function, where ;and ;

[0088] It should be further explained that, in a specific implementation process, the process of performing signal processing on the corresponding high-frequency signal based on the signal optimization function includes:

[0089] Obtain the wavelet coefficients corresponding to each high-frequency signal segment in the corresponding high-frequency signal, and substitute them into the corresponding signal optimization function to obtain the function value corresponding to the signal optimization function corresponding to the corresponding high-frequency signal segment, which is the high-frequency signal threshold;

[0090] Obtain the wavelet coefficients of the corresponding high-frequency signal fragments, and compare them with the corresponding high-frequency signal thresholds. At the same time, obtain the corresponding adjustment factors , combined with the adjustment factor The wavelet coefficients of the corresponding high-frequency signal segments are adjusted according to the comparison results;

[0091] If the adjustment factor Approaching 0, the corresponding adjustment process is: , then the wavelet coefficients of the corresponding high-frequency signal segment are adjusted to ;in, ;like , then the wavelet coefficients of the corresponding high-frequency signal fragment are compressed to 0;

[0092] If the adjustment factor Approaching 1, the corresponding adjustment process is: , then the wavelet coefficients of the corresponding high-frequency signal fragment are retained; if , then the wavelet coefficients of the corresponding high-frequency signal fragments are set to 0;

[0093] After the adjustment is completed, the corresponding high-frequency signal segments are reconstructed based on the adjusted wavelet coefficients to obtain reconstructed high-frequency signal segments, and the reconstructed high-frequency signal segments are reorganized based on the decomposition points (i.e., the reconstructed high-frequency signal segments are reconnected based on the decomposition points); and the corresponding reconstructed high-frequency signal is obtained; the decomposition point refers to the midpoint between the adjacent maximum value points corresponding to the original high-frequency signal segment acquisition process;

[0094] The obtained reconstructed high-frequency signal is fused with the original high-frequency signal to obtain a corresponding enhanced high-frequency signal;

[0095] Read the low-frequency signal corresponding to any decomposition layer, and adaptively enhance it to obtain the corresponding enhanced low-frequency signal; wherein, in the enhancement process of the corresponding low-frequency signal, it is necessary to avoid problems such as distortion or noise amplification caused by excessive enhancement;

[0096] The formula for the adaptive enhancement is: ; In the formula, It represents the enhanced low-frequency signal, that is, the enhanced low-frequency signal; represents the original low-frequency signal, represents the impulse response of the adaptively adjusted filter set based on the original low-frequency signal;

[0097] Based on the acquisition process of the enhanced low-frequency signal and the enhanced low-frequency signal, signal processing is performed on the low-frequency signal and the high-frequency signal in each decomposition layer to obtain the enhanced high-frequency signal and the enhanced low-frequency signal corresponding to the corresponding decomposition layer;

[0098] The enhanced high-frequency signal and the enhanced low-frequency signal of each decomposition layer are reconstructed by wavelet to obtain the enhanced decomposition signal corresponding to the corresponding decomposition layer;

[0099] The enhanced decomposition signals corresponding to each decomposition layer are weightedly fused to obtain the corresponding reconstructed operation signal;

[0100] The formula for weighted fusion of the enhanced decomposition signals corresponding to each decomposition layer is:

[0101] ; In the formula, Represents the reconstructed operation signal after weighted fusion is completed; Indicates The fusion weights corresponding to the layers are ; and the number of decomposition layers The closer , the corresponding fusion weight is lower; represents the total number of decomposition layers in the corresponding wavelet transform process; Indicates The enhanced decomposition signal corresponding to each sub-level;

[0102] After the fusion is completed, the obtained reconstructed operation signal is fused again with the corresponding original motor operation signal to obtain the corresponding enhanced operation signal;

[0103] The formula for signal fusion is: Where, Indicates enhanced operation signal, , Represent the reconstructed operation signal and the original motor operation signal respectively; is the fusion correction coefficient and ;

[0104] It should be further explained that, in the specific implementation process, the process of extracting features from the obtained enhanced operation signal and obtaining corresponding feature information includes:

[0105] Obtaining enhanced operating signals corresponding to various operating parameters of the corresponding motor to be diagnosed, and performing signal transformation on the obtained enhanced operating signals based on the Hilbert transform algorithm to obtain corresponding characteristic spectrum, wherein the characteristic spectrum can be used to display the frequency characteristics corresponding to the corresponding motor to be diagnosed in the current operating state, wherein the frequency characteristics refer to frequency domain information such as the envelope, instantaneous amplitude and instantaneous phase of the signal; wherein the Hilbert transform algorithm is a prior art and will not be elaborated in detail in the present invention;

[0106] Acquire characteristic spectra corresponding to all enhanced operating signals, and construct corresponding characteristic spectra based thereon, wherein the characteristic spectra include single characteristic spectra corresponding to operating parameters (composed of characteristic spectra corresponding to any one operating parameter) and multiple characteristic spectra (composed of characteristic spectra corresponding to any two operating parameters);

[0107] Reading the obtained enhanced operation signal, and obtaining the local maximum value and the local minimum value in the corresponding enhanced operation signal;

[0108] Then, the corresponding local maximum and local minimum are interpolated based on the cubic spline interpolation algorithm to generate the corresponding upper envelope and lower envelope; and the mean of the upper envelope and the lower envelope is obtained; the corresponding mean is subtracted from the enhanced operation signal to obtain the corresponding characteristic component, and it is determined whether the characteristic component satisfies the component function; if not, the process is repeated until the characteristic component that satisfies the component function is obtained;

[0109] Subtract the characteristic component from the enhanced operation signal to obtain the corresponding residual signal, repeat the characteristic component acquisition process until the characteristic component can no longer decompose into a new characteristic component, count all the characteristic components corresponding to the corresponding enhanced operation signal, and obtain the corresponding signal feature group; wherein the formula of the component function is: ; In the formula, represents the obtained characteristic components;

[0110] Respectively obtain sample entropies of envelope signals corresponding to corresponding characteristic components, and use the sample entropies as characteristic vectors of corresponding enhanced operation signals;

[0111] Counting the feature vectors and feature spectra corresponding to the enhanced operation signal to obtain corresponding feature information;

[0112] It should be further explained that, in the specific implementation process, the process of inputting the obtained characteristic information into the pre-built fault diagnosis model to obtain the corresponding fault prediction result includes:

[0113] Inputting the obtained characteristic information into the fault diagnosis model to obtain a corresponding output result, wherein the output result includes the predicted operating state to which the corresponding characteristic information belongs and the corresponding membership probability; wherein, if the operating state of the output result is a faulty operating state, the output result includes the corresponding fault category and the membership probability of belonging to the corresponding fault category;

[0114] The process of building a fault diagnosis model includes:

[0115] Constructing a motor database, wherein the motor database stores historical operation signals corresponding to the motor to be diagnosed and other motors of the notification type as the motor to be diagnosed, and also includes historical operation states corresponding to the corresponding historical operation signals, wherein the historical operation states include normal operation states and fault operation states;

[0116] Obtain all fault categories of the motor to be diagnosed and other motors of the same type in a faulty operating state, and obtain a number of historical operating signals corresponding to each fault category based on the motor database; at the same time, it is also necessary to obtain the historical operating signals corresponding to the normal operating state;

[0117] Based on the process of acquiring characteristic information, the obtained historical operation signal is processed to obtain characteristic information corresponding to the corresponding historical operation signal, wherein each fault category in the normal operation state and the fault operation state respectively obtains a number of characteristic information;

[0118] Construct corresponding training data based on the feature information corresponding to the corresponding historical operation signal;

[0119] The pre-built fault diagnosis model is defined to be based on a convolutional neural network as a skeleton, which is used to obtain multi-scale feature information from the input feature information at different stages of the backbone network;

[0120] Based on the training set and test set corresponding to the training data component, the fault diagnosis model is initialized; the training set is input into the initialized fault diagnosis model for forward propagation calculation, and the error is obtained through a three-layer convolutional network; the obtained error is propagated to each convolution kernel in the corresponding fault diagnosis model by back propagation;

[0121] Then, by changing the weights corresponding to different convolution kernels, the model weights of the corresponding fault diagnosis model are continuously optimized, and the optimization process of the model weights is repeated until the loss function in the corresponding fault diagnosis model tends to be stable and the accuracy of the model training reaches the highest.

[0122] When the loss function tends to be stable, the model parameters are saved, and the fault diagnosis model is tested and verified based on the test machine. If the verification passes, the construction of the fault diagnosis model is completed. If the verification passes, the process is repeated until the corresponding fault diagnosis model meets the requirements.

[0123] It should be further explained that, in the specific implementation process, the fault prediction result is diagnosed and verified, and the process of obtaining the corresponding fault diagnosis result according to the verification result includes:

[0124] Setting a probability threshold and comparing the membership probability in the output result with the probability threshold;

[0125] If the membership probability is less than or equal to the probability threshold, no other operation is performed;

[0126] If the membership probability is greater than the probability threshold, the corresponding output result is read, and if the operating state in the output result is a normal operating state, no other operation is performed;

[0127] If there is a fault operation state in the output result, read the fault category corresponding to the corresponding fault operation state in the corresponding output result; and obtain the corresponding historical operation signal based on the fault category, and mark it as a verification signal;

[0128] Obtain the characteristic component corresponding to the motor operation signal, and obtain the correlation coefficient between the corresponding characteristic component and the corresponding verification signal; wherein the formula for obtaining the correlation coefficient is:

[0129] ; In the formula, Indicates the verification signal. Indicates the motor running signal corresponding to the feature components, among which , is a natural number, Indicates the total number of characteristic components corresponding to the motor operation signal;

[0130] , Represent the verification signal and The mean square value corresponding to the characteristic components;

[0131] Indicates the verification signal and The covariance between the feature components;

[0132] Furthermore, the energy corresponding to the corresponding characteristic component is obtained and recorded as ; The corresponding mathematical formula is as follows: ;

[0133] The correlation coefficient and energy corresponding to the corresponding characteristic component are obtained, and based on them, the sensitivity coefficient between the corresponding verification signals is obtained. The formula for obtaining the sensitivity coefficient is: ;

[0134] Set a verification threshold and compare the obtained sensitivity coefficient with the corresponding verification threshold;

[0135] If the sensitivity coefficient is lower than the verification threshold, it indicates that the characteristic component corresponding to the corresponding motor operation signal is not similar to the characteristics of the verification signal; then a corresponding motor maintenance notification is generated, and the operating load of the motor to be diagnosed is appropriately reduced through remote control to reduce the risk of damage to the motor to be diagnosed;

[0136] If the sensitivity coefficient is not lower than the verification threshold, indicating that the characteristic component corresponding to the corresponding motor operation signal is similar to the characteristic of the verification signal, a corresponding fault category warning is generated based on the output result, and corresponding warning measures are taken based on the corresponding fault category.

[0137] The present invention realizes comprehensive diagnosis and effective prevention of motor faults through the organic combination of multiple links such as signal acquisition, processing, feature extraction, fault prediction and verification. Compared with traditional methods, it has higher diagnostic accuracy, faster response speed and lower false alarm rate, which is of great significance for ensuring the normal operation of the motor and extending its service life.

[0138] Example 2

[0139] See also Figure 2 As shown, the part not described in detail in this embodiment is described in Example 1, which provides a motor fault diagnosis system, including:

[0140] A data acquisition module is used to collect information data of the motor to be diagnosed and obtain corresponding motor operation signals;

[0141] A data processing module is used to perform signal processing on the collected motor operation signal to obtain a corresponding enhanced operation signal, and perform feature extraction on the enhanced operation signal to obtain corresponding feature information;

[0142] The model diagnosis module is used to construct a fault diagnosis model and input the obtained characteristic information into the constructed fault diagnosis model to obtain the corresponding fault prediction result;

[0143] The data feedback module is used to perform diagnostic verification on the obtained fault prediction results, generate corresponding fault diagnosis results based on the diagnostic verification results and provide feedback.

[0144] The modules are connected to each other via wired and / or wireless means to achieve data transmission between modules.

[0145] Example 3

[0146] This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the operation mode of the motor fault diagnosis method and system provided above is implemented.

[0147] Since the electronic device introduced in this embodiment is an electronic device used to implement a motor fault diagnosis method and system in the embodiment of the present application, based on the motor fault diagnosis method and system introduced in the embodiment of the present application, the technical personnel of the field can understand the specific implementation of the electronic device of the present embodiment and its various variations, so how the electronic device implements the method in the embodiment of the present application is not described in detail here. As long as the technical personnel of the field implement the electronic device used in the motor fault diagnosis method and system in the embodiment of the present application, it belongs to the scope of protection of this application.

[0148] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0149] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A motor fault diagnosis method, characterized in that: include: Step 1: Collect signals from the motor to be diagnosed and obtain corresponding motor operation signals; Step 2: performing signal processing on the obtained motor operation signal to obtain an enhanced operation signal, and performing feature extraction on the obtained enhanced operation signal to obtain corresponding feature information; Step 3: input the obtained characteristic information into a pre-built fault diagnosis model to obtain a corresponding fault prediction result; and perform diagnostic verification on the fault prediction result to obtain a corresponding fault diagnosis result based on the verification result; The process of collecting signals from the motor to be diagnosed and obtaining the corresponding motor operation signals includes: Setting a data acquisition unit; obtaining a motor to be diagnosed, marking it as a motor to be diagnosed, and acquiring signals of the corresponding motor to be diagnosed based on the data acquisition unit to obtain a corresponding motor operation signal; The process of performing signal processing on the obtained motor operation signal to obtain an enhanced operation signal includes: Read the motor operation signal in the collected motor operation information, and decompose the corresponding motor operation signal based on wavelet transform to obtain The high-frequency and low-frequency signals at different scales corresponding to each decomposition layer; Constructing a signal optimization function, and performing signal processing on the target high-frequency signal based on the signal optimization function to obtain a corresponding enhanced high-frequency signal; At the same time, the obtained low-frequency signal is adaptively enhanced to obtain a corresponding enhanced low-frequency signal; Based on the acquisition process of the enhanced low-frequency signal and the enhanced low-frequency signal, signal processing is performed on the low-frequency signal and the high-frequency signal in each decomposition layer to obtain the enhanced high-frequency signal and the enhanced low-frequency signal corresponding to the corresponding decomposition layer; The enhanced high-frequency signal and the enhanced low-frequency signal of each decomposition layer are reconstructed by wavelet to obtain the enhanced decomposition signal corresponding to the corresponding decomposition layer; Perform weighted fusion on the enhanced decomposition signals corresponding to each decomposition layer to obtain the corresponding reconstructed operation signal; The formula for weighted fusion of the enhanced decomposition signals corresponding to each decomposition layer is: ; In the formula, Represents the reconstructed operation signal after weighted fusion is completed; Indicates The fusion weights corresponding to the layers are ; Indicates The enhanced decomposition signal corresponding to the decomposition layer; After the fusion is completed, the obtained reconstructed operation signal is fused again with the corresponding original motor operation signal to obtain the corresponding enhanced operation signal; The formula for signal fusion is: Where, Indicates enhanced operation signal, , Represent the reconstructed operation signal and the original motor operation signal respectively; is the fusion correction coefficient and .

2. A motor fault diagnosis method according to claim 1, characterized in that: The formula of the signal optimization function is: ; In the formula, Represents high-frequency signal fragments The function value of the corresponding signal optimization function; Represents the corresponding high-frequency signal segment The corresponding small slope coefficient is , Indicates the total number of high-frequency signal segments; Represents high-frequency signal fragments The corresponding wavelet threshold; wavelet threshold The formula for obtaining is: ; In the formula, Represents high-frequency signal fragments The corresponding signal length; Represents high-frequency signal fragments The corresponding noise standard deviation, noise standard deviation The formula for obtaining is: ; Indicates the adjustment factor corresponding to the corresponding signal optimization function, the adjustment factor The formula for obtaining is ;and ; The formula for the adaptive enhancement is: ; In the formula, represents the enhanced low-frequency signal; represents the original low-frequency signal, Represents the impulse response of the adaptively adjusted filter set based on the original low-frequency signal.

3. A motor fault diagnosis method according to claim 2, characterized in that: The process of performing signal processing on the corresponding high-frequency signal based on the signal optimization function includes: Obtaining a maximum point in the corresponding high-frequency signal, obtaining a midpoint between adjacent maximum points, and dividing the corresponding high-frequency signal based on the maximum point to obtain a plurality of high-frequency signal segments; Obtain the wavelet coefficients corresponding to each high-frequency signal segment in the corresponding high-frequency signal, and substitute them into the corresponding signal optimization function to obtain the function value corresponding to the signal optimization function corresponding to the corresponding high-frequency signal segment, which is recorded as the high-frequency signal threshold; Obtain the wavelet coefficients of the corresponding high-frequency signal fragments, and compare them with the corresponding high-frequency signal thresholds. At the same time, obtain the corresponding adjustment factors , combined with the adjustment factor The wavelet coefficients of the corresponding high-frequency signal segments are adjusted according to the comparison results; If the adjustment factor Approaching 0, the corresponding adjustment process is: , then the wavelet coefficients of the corresponding high-frequency signal segment are adjusted to ;in, ;like , then the wavelet coefficients of the corresponding high-frequency signal fragment are compressed to 0; If the adjustment factor Approaching 1, the corresponding adjustment process is: , then the wavelet coefficients of the corresponding high-frequency signal fragment are retained; if , then the wavelet coefficients of the corresponding high-frequency signal fragments are set to 0; After the adjustment is completed, the corresponding high-frequency signal segment is reconstructed based on the adjusted wavelet coefficient to obtain a reconstructed high-frequency signal segment, and the reconstructed high-frequency signal segment is reorganized based on the decomposition point to obtain a corresponding reconstructed high-frequency signal; the decomposition point refers to the midpoint between the adjacent maximum value points corresponding to the original high-frequency signal segment acquisition process; The obtained reconstructed high-frequency signal is fused with the original high-frequency signal to obtain a corresponding enhanced high-frequency signal.

4. A motor fault diagnosis method according to claim 2, characterized in that: The process of extracting features from the obtained enhanced operation signal and obtaining corresponding feature information includes: Obtaining enhanced operating signals corresponding to various operating parameters of the corresponding motor to be diagnosed, and performing signal transformation on the obtained enhanced operating signals based on the Hilbert transform algorithm to obtain corresponding characteristic spectra; obtaining characteristic spectra corresponding to all enhanced operating signals, and constructing corresponding characteristic maps based thereon; Reading the obtained enhanced operation signal, and obtaining the local maximum value and the local minimum value in the corresponding enhanced operation signal; Then, the corresponding local maximum and local minimum are interpolated based on the cubic spline interpolation algorithm to generate the corresponding upper envelope and lower envelope; and the mean of the upper envelope and the lower envelope is obtained; the corresponding mean is subtracted from the enhanced operation signal to obtain the corresponding characteristic component, and it is determined whether the characteristic component satisfies the component function; if not, the process is repeated until the characteristic component that satisfies the component function is obtained; Subtracting the characteristic component from the enhanced operation signal to obtain a corresponding residual signal, repeating the characteristic component acquisition process until the characteristic component can no longer be decomposed into a new characteristic component, counting all characteristic components corresponding to the corresponding enhanced operation signal to obtain a corresponding signal characteristic group; The formula of the component function is: ; In the formula, represents the obtained characteristic components; Obtaining sample entropies of envelope signals corresponding to corresponding characteristic components respectively, and using the sample entropies as characteristic vectors of corresponding enhanced operation signals; The characteristic vectors and characteristic spectra corresponding to the enhanced operation signal are counted to obtain corresponding characteristic information.

5. A motor fault diagnosis method according to claim 4, characterized in that: A fault diagnosis model is constructed, and the obtained characteristic information is input into the pre-constructed fault diagnosis model to obtain a corresponding output result, wherein the output result includes a prediction of the operating state to which the corresponding characteristic information belongs and a corresponding membership probability.

6. A motor fault diagnosis method according to claim 5, characterized in that: The process of performing diagnostic verification on the fault prediction result and obtaining a corresponding fault diagnosis result according to the verification result includes: Setting a probability threshold, reading the obtained output result, and comparing the membership probability in the output result with the probability threshold; If the membership probability is less than or equal to the probability threshold, no other operation is performed; If the membership probability is greater than the probability threshold, the corresponding output result is read, and if the operating state in the output result is a normal operating state, no other operation is performed; If there is a fault operation state in the output result, read the fault category corresponding to the corresponding fault operation state in the corresponding output result; and obtain the corresponding historical operation signal based on the fault category, and mark it as a verification signal; Obtaining a sensitivity coefficient between a characteristic component corresponding to a corresponding motor operation signal and a corresponding verification signal; Set a verification threshold and compare the obtained sensitivity coefficient with the corresponding verification threshold; If the sensitivity coefficient is lower than the verification threshold, a corresponding motor maintenance notification is generated; if the sensitivity coefficient is not lower than the verification threshold, a corresponding fault category warning is generated based on the output result, and corresponding warning measures are taken based on the corresponding fault category.

7. A motor fault diagnosis method according to claim 6, characterized in that: The process of obtaining the sensitivity coefficient between the characteristic component corresponding to the corresponding motor operation signal and the corresponding verification signal includes: Obtaining a characteristic component corresponding to the motor operation signal, and obtaining a correlation coefficient between the corresponding characteristic component and the corresponding verification signal; The formula for obtaining the correlation coefficient is: ; In the formula, Indicates the verification signal. represents the pth characteristic component corresponding to the motor operation signal, where , Indicates the total number of characteristic components corresponding to the motor operation signal; , Represent the verification signal and The mean square value corresponding to the characteristic components; Indicates the verification signal and The covariance between the feature components; Get the energy corresponding to the corresponding characteristic component The formula is: ; Obtain the correlation coefficient and energy corresponding to the corresponding characteristic component, and based on it, obtain the sensitivity coefficient between the corresponding verification signals , sensitivity coefficient The formula for obtaining is: .

8. A motor fault diagnosis system, used to implement a motor fault diagnosis method according to any one of claims 1 to 7, characterized in that: include: A data acquisition module is used to collect information data of the motor to be diagnosed and obtain corresponding motor operation signals; A data processing module is used to perform signal processing on the collected motor operation signal to obtain a corresponding enhanced operation signal, and perform feature extraction on the enhanced operation signal to obtain corresponding feature information; The model diagnosis module is used to construct a fault diagnosis model and input the obtained characteristic information into the constructed fault diagnosis model to obtain the corresponding fault prediction result; The data feedback module is used to perform diagnostic verification on the obtained fault prediction results, generate corresponding fault diagnosis results based on the diagnostic verification results and provide feedback.

Images