A real-time diagnosis method and system for motor speed sensor fault

Abstract

The present invention discloses a real-time fault diagnosis method and system for a motor speed sensor, which belongs to the technical field of motor detection. The method comprises: obtaining motor speed information based on a speed sensor, and performing signal processing on the obtained motor speed information to obtain a preprocessing signal; performing feature extraction according to the obtained preprocessing signal to obtain a corresponding fault type; constructing an RBF neural network model, inputting the real-time collected preprocessing signal into the RBF neural network model to be trained, and outputting the current fault diagnosis result. The present invention preprocesses the motor speed information collected by the speed sensor, effectively filters out noise interference, improves signal quality, and uses feature extraction technology to mine feature information closely related to the fault type from the preprocessing signal, providing a reliable basis for subsequent fault diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of motor detection, and in particular to a real-time fault diagnosis method and system for a motor speed sensor. Background Art

[0002] In the motor control system, the speed sensor is a key component, and its performance directly affects the stability and safety of the system. However, most of the existing motor speed sensor fault diagnosis methods rely on traditional threshold judgment or simple pattern recognition. These methods often have low diagnostic accuracy when facing complex and changeable working conditions, and cannot reflect the development trend of the fault in real time, which easily leads to false alarms or missed alarms.

[0003] In order to overcome the deficiencies of the prior art, the present invention proposes a real-time fault diagnosis method for a motor speed sensor. Summary of the invention

[0004] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a real-time diagnosis method and system for motor speed sensor faults, which preprocesses the motor speed information collected by the speed sensor, effectively filters out noise interference, improves signal quality, and uses feature extraction technology to mine feature information closely related to the fault type from the preprocessed signal, thereby providing a reliable basis for subsequent fault diagnosis.

[0005] A first aspect of the present invention provides a method for real-time diagnosis of motor speed sensor faults, the method comprising:

[0006] Based on the speed sensor, the motor speed information is obtained, and the obtained motor speed information is signal processed to obtain a preprocessed signal;

[0007] According to the obtained preprocessed signal, feature extraction is performed to obtain the corresponding fault type;

[0008] Construct an RBF neural network model, input the real-time collected pre-processed signal into the trained RBF neural network model, and output the current fault diagnosis result;

[0009] When outputting the current fault diagnosis result, the fault probability within the time window is calculated and a probability threshold is set. If the fault probability is greater than or equal to the probability threshold, an alarm is triggered.

[0010] Furthermore, the failure probability within the time window is calculated and the probability threshold is set, including:

[0011] Set a dynamic time window and accumulate the failure probability output by the RBF neural network within the dynamic time window;

[0012] The average failure probability within the dynamic time window is calculated and judged based on the preset probability threshold; if the average failure probability is greater than or equal to the probability threshold, it is judged as a sensor failure and an alarm is triggered. If the average failure probability is less than the probability threshold, the sensor is judged to be normal and no alarm is triggered.

[0013] Furthermore, a dynamic time window is set, including:

[0014] Based on the historical fault diagnosis results of the motor, the time window module is activated; the time window module includes a timer, which is used to check the fault diagnosis results within each set cycle at the end of the cycle and calculate the frequency of the fault diagnosis results within the cycle. When the frequency is greater than the set value, the dynamic time window is shortened.

[0015] Further, based on the speed sensor, the motor speed information is obtained, and the obtained motor speed information is signal processed to obtain a preprocessed signal, which includes:

[0016] De-noise and perform data conversion on motor speed information;

[0017] The denoising process is to remove the noise data caused by environmental factors; the data conversion is to convert the rotation speed data from the original unit to the analysis unit.

[0018] Furthermore, the construction of the RBF neural network model includes:

[0019] Train the RBF neural network based on known fault types;

[0020] During training, the Gaussian function is used as the activation function of the hidden layer.

[0021] Furthermore, when an alarm is triggered, the fault diagnosis results are classified and prioritized, and the alarm information is pushed through multi-channel alarm notifications.

[0022] A second aspect of the present invention provides a real-time fault diagnosis system for a motor speed sensor, the system comprising:

[0023] An information acquisition module, wherein the information acquisition module acquires motor speed information through a speed sensor, and performs signal processing on the acquired motor speed information to obtain a preprocessing signal;

[0024] A feature extraction module, wherein the feature extraction module performs feature extraction according to the obtained preprocessed signal to obtain a corresponding fault type;

[0025] RBF neural network module, wherein the RBF neural network module is used to construct an RBF neural network model;

[0026] The fault output module is used to input the pre-processed signal collected in real time into the RBF neural network model to be trained, and output the current fault diagnosis result.

[0027] The beneficial effects of the present invention are:

[0028] The method first pre-processes the motor speed information collected by the speed sensor to effectively filter out noise interference and improve signal quality. Then, feature extraction technology is used to mine feature information closely related to the fault type from the pre-processed signal, providing a reliable basis for subsequent fault diagnosis.

[0029] At the same time, a RBF neural network model was constructed. By inputting the pre-processed signal collected in real time into the trained RBF neural network model, the current fault diagnosis result can be quickly output, realizing real-time monitoring and diagnosis of faults.

[0030] In addition, the present invention introduces a fault probability calculation mechanism within a time window and sets a probability threshold. When the calculated fault probability is greater than or equal to the threshold, the system will immediately trigger an alarm, thereby effectively avoiding the problem of false alarms or missed alarms caused by diagnostic errors at a single moment, and significantly improving the accuracy and reliability of fault diagnosis.

[0031] In summary, the real-time diagnosis method of motor speed sensor faults of the present invention shows significant advantages in diagnostic accuracy, real-time performance and reliability compared with the prior art, and provides a strong guarantee for the safe and stable operation of the motor control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] Figure 1 A schematic flow chart of a method for real-time diagnosis of motor speed sensor faults provided by the present invention.

[0034] Figure 2 This is a schematic diagram of the real-time fault diagnosis system for a motor speed sensor provided by the present invention. DETAILED DESCRIPTION

[0035] 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.

[0036] Embodiment 1

[0037] like Figure 1 As shown, the present invention provides a real-time diagnosis method for motor speed sensor fault, the method comprising:

[0038] S1: Based on the speed sensor, the motor speed information is obtained, and the obtained motor speed information is signal processed to obtain a preprocessed signal;

[0039] Specifically, the speed sensor is set on the motor shaft to ensure that it can obtain the motor's speed information in real time and accurately. During the acquisition, the output signal of the speed sensor, such as a pulse signal or an analog voltage signal, is obtained, the collected original signal is denoised, and data conversion is performed;

[0040] Among them, denoising is to remove the noise data caused by environmental factors, and data conversion refers to converting the speed data from the original unit (such as revolutions per minute) to a unit that is more suitable for analysis (such as radians per second).

[0041] S2: Extract features based on the obtained preprocessed signal to obtain the corresponding fault type;

[0042] It is mainly used to extract characteristic parameters related to motor speed from the preprocessed signal, such as frequency, amplitude, phase, etc.

[0043] S3: construct an RBF neural network model, input the pre-processed signal collected in real time into the RBF neural network model to be trained, and output the current fault diagnosis result;

[0044] Specifically, a three-layer RBF neural network model is constructed, including an input layer, a hidden layer, and an output layer, and a Gaussian function is used as the activation function of the hidden layer, which is formed as follows:

[0045]

[0046] in, is the input feature vector, is the center vector of the Gaussian function, is the width parameter of the Gaussian function.

[0047] Before building the RBF neural network model, it is necessary to train the RBF neural network based on the known fault types; then the feature parameters collected and preprocessed in real time are input into the trained RBF neural network model. The RBF neural network model calculates and analyzes the input feature parameters and outputs the current fault diagnosis result, which can take a probability value between 0 and 1 to indicate the possibility of the fault.

[0048] Furthermore, when outputting the current fault diagnosis result, the fault probability within the time window is calculated, and a probability threshold (such as 0.5) is set to determine whether the fault probability output by the RBF neural network reaches the alarm level. Once it is determined to be a sensor failure, the alarm mechanism is immediately triggered to notify the operator to pay attention.

[0049] Among them, the failure probability calculation within the time window is performed, and the probability threshold is set, specifically:

[0050] Set a dynamic time window and accumulate the failure probability output by the RBF neural network within the dynamic time window;

[0051] The average failure probability within the dynamic time window is calculated and judged based on the preset probability threshold; if the average failure probability is greater than or equal to the probability threshold, it is judged as a sensor failure and an alarm is triggered. If the average failure probability is less than the probability threshold, the sensor is judged to be normal and no alarm is triggered.

[0052] Furthermore, when setting the dynamic time window, the time window module is activated based on the historical fault diagnosis results of the motor; the time window module includes a timer, and the timer is used to check the fault diagnosis results within each set cycle at the end of the cycle, and calculate the frequency of the fault diagnosis results within the cycle. When the frequency is greater than the set value, the dynamic time window is shortened.

[0053] In this embodiment, the time window is set to a specific time (such as 15 seconds), during which the fault probability output by the RBF neural network model is accumulated, and the average fault probability is calculated, that is, the sum of all fault probability values ​​in the window divided by the number of samples in the window; at the same time, the above-mentioned probability threshold is a set value (such as 0.6).

[0054] S4: When an alarm is triggered, the fault diagnosis results are classified and prioritized, and the alarm information is pushed through multi-channel alarm notifications;

[0055] Specifically, they can be classified according to factors such as nature, scope of impact, and urgency so that processing personnel can respond more efficiently.

[0056] Embodiment 2:

[0057] Based on the first embodiment, Figure 2 As shown, this embodiment provides a real-time fault diagnosis system for a motor speed sensor, the system comprising:

[0058] The information acquisition module 10 acquires the motor speed information through a speed sensor, and performs signal processing on the acquired motor speed information to obtain a preprocessing signal.

[0059] The feature extraction module 20 performs feature extraction according to the obtained preprocessed signal to obtain a corresponding fault type.

[0060] The RBF neural network module 30 is used to construct a RBF neural network model.

[0061] The time window module 40 is used to calculate the fault probability within the time window and set the probability threshold when outputting the current fault diagnosis result. If the fault probability is greater than or equal to the probability threshold, an alarm is triggered.

[0062] The fault output module 50 is used to input the pre-processed signal collected in real time into the RBF neural network model to be trained, and output the current fault diagnosis result.

[0063] Furthermore, the time window module 40 is further configured to perform the following steps:

[0064] Set a dynamic time window and accumulate the failure probability output by the RBF neural network within the dynamic time window;

[0065] The average failure probability within the dynamic time window is calculated and judged based on the preset probability threshold; if the average failure probability is greater than or equal to the probability threshold, it is judged as a sensor failure and an alarm is triggered. If the average failure probability is less than the probability threshold, the sensor is judged to be normal and no alarm is triggered.

[0066] Among them, a dynamic time window is set, and a time window module is activated based on the historical fault diagnosis results of the motor; the time window module includes a timer, and the timer is used to check the fault diagnosis results within each set cycle at the end of the cycle, and calculate the frequency of the fault diagnosis results within the cycle. When the frequency is greater than the set value, the dynamic time window is shortened.

[0067] Furthermore, the information acquisition module 10 is also used to perform denoising and data conversion on the motor speed information. The denoising is to remove the noise data caused by environmental factors; and the data conversion is to convert the speed data from the original unit to the analysis unit.

[0068] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A real-time diagnosis method for motor speed sensor fault, characterized in that: The method comprises: Based on the speed sensor, the motor speed information is obtained, and the obtained motor speed information is signal processed to obtain a preprocessed signal; According to the obtained preprocessed signal, feature extraction is performed to obtain the corresponding fault type; Construct an RBF neural network model, input the real-time collected pre-processed signal into the trained RBF neural network model, and output the current fault diagnosis result; When outputting the current fault diagnosis result, the fault probability within the time window is calculated and a probability threshold is set. If the fault probability is greater than or equal to the probability threshold, an alarm is triggered. Calculate the failure probability within the time window and set the probability threshold, including: Set a dynamic time window and accumulate the failure probability output by the RBF neural network within the dynamic time window; The average failure probability within the dynamic time window is calculated and judged based on the preset probability threshold; if the average failure probability is greater than or equal to the probability threshold, it is judged as a sensor failure and an alarm is triggered. If the average failure probability is less than the probability threshold, the sensor is judged to be normal and no alarm is triggered.

2. A motor speed sensor fault real-time diagnosis method as claimed in claim 1, characterized in that: Set dynamic time windows, including: Based on the historical fault diagnosis results of the motor, the time window module is activated; the time window module includes a timer, which is used to check the fault diagnosis results within each set cycle at the end of the cycle and calculate the frequency of the fault diagnosis results within the cycle. When the frequency is greater than the set value, the dynamic time window is shortened.

3. A motor speed sensor fault real-time diagnosis method as claimed in claim 1, characterized in that: Based on the speed sensor, the motor speed information is obtained, and the obtained motor speed information is signal processed to obtain a preprocessed signal, which includes: De-noise and perform data conversion on motor speed information; The denoising process is to remove the noise data caused by environmental factors; the data conversion is to convert the rotation speed data from the original unit to the analysis unit.

4. A motor speed sensor fault real-time diagnosis method as claimed in claim 1, characterized in that: Before building the RBF neural network model, the following steps are required: Train the RBF neural network based on known fault types; During training, the Gaussian function is used as the activation function of the hidden layer.

5. A motor speed sensor fault real-time diagnosis method as claimed in claim 1, characterized in that: The method further comprises: When an alarm is triggered, the fault diagnosis results are classified and prioritized, and the alarm information is pushed through multi-channel alarm notifications.

6. A real-time diagnosis system for motor speed sensor fault, characterized in that: For executing the method according to any one of claims 1 to 5, the system comprises: An information acquisition module, wherein the information acquisition module acquires motor speed information through a speed sensor, and performs signal processing on the acquired motor speed information to obtain a preprocessing signal; A feature extraction module, wherein the feature extraction module performs feature extraction according to the obtained preprocessed signal to obtain a corresponding fault type; RBF neural network module, wherein the RBF neural network module is used to construct an RBF neural network model; A time window module, which is used to calculate the fault probability within the time window and set the probability threshold when outputting the current fault diagnosis result; The fault output module is used to input the pre-processed signal collected in real time into the RBF neural network model to be trained, and output the current fault diagnosis result.

Images