Circuit breaker mechanical fault identification method and device

Abstract

The invention discloses a circuit breaker mechanical fault identification method and device. The method comprises the following steps: acquiring SOE data and sound signal data of a circuit breaker in a preset time period; time features and action features are extracted from the SOE data, and a plurality of stages of the circuit breaker in the preset time period are identified according to the action features; extracting domain features corresponding to the stages from the sound signal data according to the time features; and inputting the domain feature corresponding to each stage into a pre-trained corresponding prediction model to obtain a fault identification result. According to the application, the original structural characteristics of the circuit breaker can be prevented from being changed in the fault identification process, and the accuracy of fault identification is improved at the same time.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method and device for identifying mechanical faults of circuit breaker. Background Art

[0002] As the main switchgear, high-voltage circuit breakers control the on and off of the lines to ensure the safe operation of the system. When a high-voltage circuit breaker fails, it will greatly affect the power quality and power supply reliability.

[0003] High-voltage circuit breaker failures are mostly caused by mechanical defects. When identifying such failures, the existing technology uses sensors to measure contact travel, opening and closing speed, opening and closing coil current, energy storage motor current and voltage, and vibration signals during the operation process to identify mechanical failures. However, during the actual operation of high-voltage circuit breakers, the above data collection will not only be affected by electromagnetic interference, but also affect the structure of the circuit breaker, making it difficult to achieve engineering application.

[0004] Therefore, how to accurately identify the mechanical failure of the circuit breaker without changing the structure of the circuit breaker is a technical problem that needs to be solved at present. Summary of the invention

[0005] The present application provides a circuit breaker mechanical fault identification method and device to solve the technical problem of how to accurately identify the circuit breaker mechanical fault without changing the circuit breaker structure.

[0006] In order to solve the above technical problems, in a first aspect, an embodiment of the present application provides a circuit breaker mechanical fault identification method, comprising:

[0007] Obtain SOE data and sound signal data of the circuit breaker within a preset time period;

[0008] Extracting time features and action features from the SOE data, and identifying the stages of the circuit breaker within the preset time period according to the action features;

[0009] Extracting domain features corresponding to each of the stages from the sound signal data according to the time features;

[0010] The domain features corresponding to each stage are input into the corresponding pre-trained prediction model to obtain the fault identification result.

[0011] Compared with the prior art, the embodiments of the present application have the following beneficial effects: since when a mechanical failure occurs, abnormal noises are often mixed into the operating sound of the circuit breaker, the mechanical failure of the circuit breaker can be accurately identified by comparing the difference between the currently collected sound signal data and the sound signal data under normal operation through the pre-trained prediction model, and since a large number of sensors do not need to be set in each component of the circuit breaker when collecting sound signals, a large number of changes to the circuit breaker structure are avoided. However, the sound characteristics of the normal operation of the circuit breaker are different at different stages, so the combination of SOE (Sequence Of Event) data can accurately reflect the characteristics of the operating events of the circuit breaker at different operating stages, and independently predict the sound signal data at different stages, which effectively improves the accuracy of circuit breaker fault identification.

[0012] In some embodiments of the first aspect of the present application, extracting time features and action features from the SOE data, and identifying the stages of the circuit breaker in the preset time period according to the action features, includes:

[0013] Extracting the time characteristics and action characteristics of the events occurring in the circuit breaker within the preset time period from the SOE data;

[0014] The plurality of stages of the circuit breaker within the preset time period are determined through the action characteristics.

[0015] Compared with the prior art, the above embodiment has the following beneficial effects: since the SOE data generated by the circuit breaker at each stage includes the unique action characteristics of each stage, it is possible to first confirm which stages the circuit breaker has passed through within a preset time period based on the action characteristics, and perform independent fault identification for different stages, thereby improving the accuracy of subsequent fault identification based on sound signal data.

[0016] In some embodiments of the first aspect of the present application, after acquiring the sound signal data, the method further includes:

[0017] Performing noise reduction processing on the sound signal data by spectrum subtraction to obtain first power spectrum data;

[0018] Soft threshold processing is performed on the first power spectrum data to obtain the pre-processed sound signal data.

[0019] Compared with the prior art, the above embodiment has the following beneficial effects: since the circuit breaker will cause a lot of noise in the sound signal data during operation, the noise is mainly composed of the noise of other operating equipment. Therefore, in the case of known noise, the noise interference of other operating equipment can be effectively reduced by spectrum subtraction. In order to further reduce the small amount of residual noise caused by electromagnetic interference, the sound signal data is refined and denoised by a soft threshold processing method, the key features in the sound signal data are retained, and the accuracy of subsequent fault identification is improved. In addition, the above method has low computational requirements and can be used for real-time fault detection.

[0020] In some embodiments of the first aspect of the present application, extracting domain features corresponding to each of the stages from the sound signal data includes:

[0021] The domain features include: time domain features, frequency domain features and cepstrum domain features;

[0022] Determine the time interval corresponding to each of the stages through the time feature, divide the sound signal data according to the time interval, and obtain the sound signal data segments corresponding to each stage;

[0023] Extracting the time domain features by analyzing the time waveform of the sound signal data segment that has not been preprocessed and that appears over time;

[0024] Processing the preprocessed sound signal data segment by inverse short-time Fourier transform to obtain the frequency domain feature;

[0025] The cepstrum domain features are obtained by performing short-time Fourier transform on the frequency domain features, processing them with a Mel filter, and then performing logarithmic operation and discrete cosine transform.

[0026] Compared with the prior art, the above embodiment has the following beneficial effects: through the time domain features, the waveform changes of the sound signal in time can be directly reflected, so as to capture the information such as short-time energy changes and amplitude characteristics; through the inverse short-time Fourier transform to obtain the frequency domain features, the denoised spectrum signal is converted into the time domain signal, so as to more accurately identify the vibration and noise patterns in the circuit breaker operation; and the cepstrum domain features can more effectively distinguish the periodic components and capture the nonlinear characteristics in the complex mechanical vibration signal by further transforming the frequency domain features. Through the multi-dimensional and staged sound signal data processing method, it is ensured that no matter what operating state or external environmental conditions the circuit breaker is in, the key features can be effectively extracted and fault diagnosis can be performed, so as to more accurately reflect the current operating state of the circuit breaker.

[0027] In some embodiments of the first aspect of the present application, the prediction model is obtained by setting the circuit breaker in different mechanical states for actual operation, collecting analog sound signal data of corresponding stages, and then training a support vector machine based on the analog sound signal data.

[0028] Compared with the prior art, the above embodiment has the following beneficial effects: by setting up a circuit breaker in a real environment, the circuit breaker can simulate sound signal data under different mechanical states, thereby improving the authenticity and diversity of training data and improving the accuracy of subsequent prediction models.

[0029] In some embodiments of the first aspect of the present application, the analog sound signal data and the sound signal data are both collected and acquired through evenly distributed multi-channel audio collection devices.

[0030] Compared with the prior art, the above embodiment has the following beneficial effects: the use of evenly distributed multi-channel audio acquisition devices can more comprehensively capture the sound signals of the circuit breaker, improve the spatial resolution of the sound data, ensure the integrity and accuracy of the collected data, enhance the reliability of fault diagnosis, and do not need to change the device structure of the circuit breaker itself, thereby reducing the impact of the data acquisition equipment on the operation of the circuit breaker.

[0031] In some embodiments of the first aspect of the present application, after obtaining the SOE data, the method further includes:

[0032] Performing time synchronization correction on the SOE data collected from different sensors;

[0033] According to the logic of occurrence before and after the event, the abnormal events in the SOE data are eliminated.

[0034] Compared with the prior art, the above embodiment has the following beneficial effects: through time synchronization correction and logical elimination of abnormal events, the accuracy and consistency of SOE data are ensured, misjudgment due to data errors is avoided, and the reliability and accuracy of fault identification results are further improved.

[0035] In a second aspect, an embodiment of the present application further provides a circuit breaker mechanical fault identification device, comprising: a data acquisition module, a first feature extraction module, a second feature extraction module and a fault identification module;

[0036] The data acquisition module is used to acquire SOE data and sound signal data of the circuit breaker within a preset time period;

[0037] The first feature extraction module is used to extract time features and action features from the SOE data, and identify several stages of the circuit breaker within the preset time period according to the action features;

[0038] The second feature extraction module is used to extract domain features corresponding to each of the stages from the sound signal data according to the time feature;

[0039] The fault identification module is used to input the domain features corresponding to each stage into the corresponding pre-trained prediction model to obtain a fault identification result.

[0040] In some embodiments of the second aspect of the present application, the first feature extraction module is used to extract time features and action features from the SOE data, and identify several stages of the circuit breaker in the preset time period according to the action features, including:

[0041] Extracting the time characteristics and action characteristics of the events occurring in the circuit breaker within the preset time period from the SOE data;

[0042] The plurality of stages of the circuit breaker within the preset time period are determined through the action characteristics.

[0043] In some embodiments of the second aspect of the present application, the method further includes: a first data preprocessing module; wherein the first data preprocessing module is used to preprocess the sound signal data after acquiring the sound signal data, including:

[0044] Performing noise reduction processing on the sound signal data by spectrum subtraction to obtain first power spectrum data;

[0045] Soft threshold processing is performed on the first power spectrum data to obtain the pre-processed sound signal data.

[0046] In some embodiments of the second aspect of the present application, the second feature extraction module is used to extract the domain features corresponding to each of the stages from the sound signal data according to the time features, including:

[0047] The domain features include: time domain features, frequency domain features and cepstrum domain features;

[0048] Determine the time interval corresponding to each of the stages through the time feature, divide the sound signal data according to the time interval, and obtain the sound signal data segments corresponding to each stage;

[0049] Extracting the time domain features by analyzing the time waveform of the sound signal data segment that has not been preprocessed and that appears over time;

[0050] Processing the preprocessed sound signal data segment by inverse short-time Fourier transform to obtain the frequency domain feature;

[0051] The cepstrum domain features are obtained by performing short-time Fourier transform on the frequency domain features, processing them with a Mel filter, and then performing logarithmic operation and discrete cosine transform.

[0052] In some embodiments of the second aspect of the present application, the prediction model is obtained by setting the circuit breaker in different mechanical states for actual operation, collecting analog sound signal data of corresponding stages, and then training a support vector machine based on the analog sound signal data.

[0053] In some embodiments of the second aspect of the present application, the analog sound signal data and the sound signal data are both collected and acquired through evenly distributed multi-channel audio collection devices.

[0054] In some embodiments of the second aspect of the present application, the method further includes: a second data preprocessing module; wherein the second data preprocessing module is used to preprocess the SOE data after acquiring the SOE data, including:

[0055] Performing time synchronization correction on the SOE data collected from different sensors;

[0056] According to the logic of occurrence before and after the event, the abnormal events in the SOE data are eliminated. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A schematic flow chart of a circuit breaker mechanical fault identification method provided in some embodiments of the present application;

[0058] Figure 2 This is a schematic diagram of the circuit breaker operation state when collecting simulated sound signal data in some embodiments of the present application;

[0059] Figure 3 This is a schematic structural diagram of a circuit breaker mechanical fault identification device provided in some embodiments of the present application. DETAILED DESCRIPTION

[0060] The existing technology uses sensors to measure the contact stroke, opening and closing speed, opening and closing coil current, current and voltage of the energy storage motor, and vibration signals during the operation process to identify mechanical faults of circuit breakers. However, during the actual operation of the circuit breaker, the above data collection will not only be affected by electromagnetic interference, but also affect the structure of the circuit breaker, making it difficult to achieve engineering application. Therefore, how to accurately identify the mechanical fault of the circuit breaker without changing the structure of the circuit breaker is a technical problem that needs to be solved at present.

[0061] In order to solve the above technical problems, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0062] Embodiment 1

[0063] Please refer to Figure 1 , a circuit breaker mechanical fault identification method provided by an embodiment of the present application, including S10 to S40, specifically:

[0064] S10: Acquire SOE data and sound signal data of the circuit breaker within a preset time period.

[0065] Furthermore, in some embodiments of the present application, both analog sound signal data and sound signal data are collected and acquired through evenly distributed multi-channel audio collection devices.

[0066] The use of evenly distributed multi-channel audio acquisition devices can more comprehensively capture the sound signals of the circuit breaker, improve the spatial resolution of the sound data, ensure the integrity and accuracy of the collected data, enhance the reliability of fault diagnosis, and reduce the impact of data acquisition equipment on the operation of the circuit breaker without changing the device structure of the circuit breaker itself.

[0067] Preferably, in some embodiments of the present application, the sound signal data or simulated sound signal data can be collected by the following method:

[0068] A 16-channel digital microphone array is preferably used for sound information collection. The digital microphones of the 16 channels are evenly distributed on a black circular panel with a radius of preferably 15 cm. The 16 channels are connected to the same data processing unit. The sampling time of each channel is synchronized. The sound sensor is placed at a distance of preferably about 1.5 m from the circuit breaker. The 16-channel digital microphone array can preferably be a device with the parameters described in the following Table 1.

[0069] Table 1

[0070]

[0071] Furthermore, in some embodiments of the present application, after acquiring the sound signal data, the method further includes:

[0072] Performing noise reduction processing on the sound signal data by spectrum subtraction to obtain first power spectrum data;

[0073] Soft threshold processing is performed on the first power spectrum data to obtain the pre-processed sound signal data.

[0074] Preferably, in some embodiments of the present application, the sound signal data may be subjected to noise reduction processing to obtain the first power spectrum data by using the following formula:

[0075] P V (k,m)=|V(k,m)| 2

[0076] P D (k,m)=max(0,|X(k,m)| 2 -P V (k,m))

[0077] Among them, V(k,m) is the frequency domain signal of the noise; P V The power spectrum of (k,m) noise; P D (k, m) is the first power spectrum data; X(k, m) is the original sound signal data in the frequency domain.

[0078] Preferably, in some embodiments of the present application, the first power spectrum data may be subjected to soft threshold processing by the following formula:

[0079] P D (k,m)=max(0,|X(k,m)| 2 -P V (k,m)+T)

[0080] Where T is the soft threshold, which is generally taken as twice the standard deviation of the noise estimate.

[0081] Preferably, in some embodiments of the present application, the collected sound signals may be arithmetic averaged to reduce random errors in the sampling process.

[0082] Since the circuit breaker will cause a lot of noise in the sound signal data during operation, the noise is mainly composed of the noise of other operating equipment. Therefore, in the case of known noise, the noise interference of other operating equipment can be effectively reduced through spectrum subtraction. In order to further reduce the small amount of residual noise caused by electromagnetic interference, the sound signal data is refined and denoised through the soft threshold processing method, retaining the key features in the sound signal data and improving the accuracy of subsequent fault identification. The above method has low computational requirements and can be used for real-time fault detection.

[0083] Furthermore, in some embodiments of the present application, after obtaining the SOE data, the method further includes:

[0084] Performing time synchronization correction on the SOE data collected from different sensors;

[0085] According to the logic of occurrence before and after the event, the abnormal events in the SOE data are eliminated.

[0086] Through time synchronization correction and logical elimination of abnormal events, the accuracy and consistency of SOE data are ensured, misjudgment due to data errors is avoided, and the reliability and accuracy of fault identification results are further improved.

[0087] S20: extracting time features and action features from the SOE data, and identifying the stages of the circuit breaker within the preset time period according to the action features.

[0088] Further, in some embodiments of the present application, extracting time features and action features from the SOE data, and identifying several stages of the circuit breaker in the preset time period according to the action features, includes:

[0089] Extracting the time characteristics and action characteristics of the events occurring in the circuit breaker within the preset time period from the SOE data;

[0090] The plurality of stages of the circuit breaker within the preset time period are determined through the action characteristics.

[0091] Since the SOE data generated by the circuit breaker at each stage includes the unique action characteristics of each stage, it is possible to first confirm which stages the circuit breaker has passed through within a preset time period based on the action characteristics, and perform independent fault identification for different stages, thereby improving the accuracy of subsequent fault identification based on sound signal data.

[0092] Preferably, in some embodiments of the present application, the steps of determining the several stages of the circuit breaker in the preset time period by the action characteristics mainly include:

[0093] The stages are arranged in the order of occurrence, mainly including: tripping stage, acceleration stage, deceleration stage and braking stage;

[0094] Among them, the tripping stage refers to the stage when the circuit breaker begins to respond to the overcurrent condition and prepares to trip. It is identified and determined by identifying the event of the tripping action in the SOE data. That is, if a tripping event occurs at the current time point and the current tripping moment is not the starting time point of the preset time period, it is determined that the circuit breaker has experienced the tripping stage within the preset time period.

[0095] The acceleration phase refers to the time period from tripping to the full opening of the circuit breaker. If after the tripping moment, the action of the main contacts of the circuit breaker starting to separate is queried from the SOE data, and the start time of the separation action is not the starting time point of the preset time period, it is determined that the circuit breaker has experienced the acceleration phase within the preset time period.

[0096] The deceleration stage occurs when the main contacts of the circuit breaker approach their final position. If, after an event after the end time of the acceleration stage (equivalent to the moment when the operating speed of the circuit breaker reaches its peak value), an event at the end time of the above-mentioned acceleration stage is queried from the SOE data, and the start time of the event is not the starting time point of the preset time period, it is determined that the circuit breaker has experienced the deceleration stage within the preset time period.

[0097] The braking stage refers to the stage before the circuit breaker reaches its final position. If after the end of the deceleration stage, it is found from the SOE data that the operating speed of the circuit breaker reaches 10% of the peak speed, and this moment is not the starting time point of the preset time period, then it is determined that the circuit breaker has experienced the braking stage within the preset time period, and the end time of the braking stage is the moment when the operating speed of the circuit breaker reaches 0.

[0098] S30: extracting domain features corresponding to each of the stages from the sound signal data according to the time features.

[0099] Furthermore, in some embodiments of the present application, extracting domain features corresponding to each of the stages from the sound signal data according to the time features includes:

[0100] The domain features include: time domain features, frequency domain features and cepstrum domain features;

[0101] Determine the time interval corresponding to each of the stages through the time feature, divide the sound signal data according to the time interval, and obtain the sound signal data segments corresponding to each stage;

[0102] Extracting the time domain features by analyzing the time waveform of the sound signal data segment that has not been preprocessed and that appears over time;

[0103] Processing the preprocessed sound signal data segment by inverse short-time Fourier transform to obtain the frequency domain feature;

[0104] The cepstrum domain features are obtained by performing short-time Fourier transform on the frequency domain features, processing them with a Mel filter, and then performing logarithmic operation and discrete cosine transform.

[0105] Through the time domain features, the waveform changes of the sound signal in time can be directly reflected, so as to capture information such as short-time energy changes and amplitude characteristics; the frequency domain features are obtained through the inverse short-time Fourier transform, and the denoised spectrum signal is converted into a time domain signal, so as to more accurately identify the vibration and noise patterns in the circuit breaker operation; and the cepstrum domain features can more effectively distinguish the periodic components and capture the nonlinear characteristics in the complex mechanical vibration signal by further transforming the frequency domain features. Through the multi-dimensional and staged sound signal data processing method, it is ensured that no matter what operating state or external environmental conditions the circuit breaker is in, the key features can be effectively extracted and fault diagnosis can be performed, so as to more accurately reflect the current operating status of the circuit breaker.

[0106] Preferably, in some embodiments of the present application, after the start and end times corresponding to each stage are extracted through the above preferred embodiments, the sound signal data segments corresponding to each stage can be determined based on the time features.

[0107] Preferably, in some embodiments of the present application, domain features can be extracted by the following preferred implementations:

[0108] The short-time energy mean in the time domain feature is extracted by the following formula (in addition to the short-time energy mean, the time domain feature also includes: maximum value, peak value, mean value, root mean square value, minimum value, and the calculation of the above statistics is no longer described in this embodiment of the application):

[0109] L i (n)=w(n)x[(i-1)d+n],1≤n≤L,1≤i≤f n

[0110]

[0111] Among them, L i (n) is the short-time energy of the i-th frame of sound signal data; E n is the short-time energy mean; f n is the total number of frames of sound signal data; L is the total frame length; n is the time domain index; d is the frame shift length; x[*] is the sound signal data that has not been preprocessed; w(*) is the window function.

[0112] The frequency domain features are extracted by the following formula:

[0113]

[0114] Among them, X(k,m) is the original sound signal data in the frequency domain, that is, the sound signal data without preprocessing; d(n) is the frequency domain feature, which is also the clean sound signal data in the time domain; N is the total number of frequency components; M is the number of frames; w(*) is the window function; is the basic frequency term in the Fourier series, where j is an imaginary number; ∠X(k,m) is the phase of the signal X(k,m); P D (k,m) is the first power spectrum data.

[0115] The cepstrum domain features are extracted by the following formula:

[0116]

[0117] Among them, MFCC[m] is the cepstrum domain feature; DCT(*) is the discrete cosine transform function; log(*) is the logarithmic function; d(n) is the frequency domain feature; F n (f) is the frequency response of the nth Mel filter; N is the total number of frequency components.

[0118] S40: Inputting the domain features corresponding to each stage into the corresponding pre-trained prediction model to obtain a fault identification result.

[0119] Furthermore, in some embodiments of the present application, the prediction model is obtained by setting the circuit breaker in different mechanical states for actual operation, collecting analog sound signal data of corresponding stages, and then training a support vector machine based on the analog sound signal data.

[0120] Preferably, in some embodiments of the present application, reference Figure 2 , after setting the mechanical state, the analog sound signal data can be collected by the following preferred implementation:

[0121] Four mechanical states are selected for simulation: normal, transmission guide rod stuck, too large clearance between the iron core and the closing spring release (refusal to move), and the opening spring release is loose (causing the opening after closing):

[0122] Normal state: Set the mechanical parts of the circuit breaker operating mechanism to normal, use the high-voltage switch tester to close the experimental circuit breaker, and the overall transmission mechanism of the circuit breaker is shown in the figure. Figure 2 (a) shown.

[0123] Drive guide rod stuck: The stuck fault of the circuit breaker operating mechanism includes mechanism stuck and iron core stuck. Mechanism stuck is mainly caused by wear and poor lubrication of parts. The energy stored in the closing spring is lost in the friction process of the stuck part. For the simulation of this fault, this application reduces the pre-compression amount of the closing spring to reduce the elastic potential energy. The closing spring is Figure 2 As shown in (b), the specific implementation method is to reduce the pre-compression amount of the closing spring during normal closing by 5 cm.

[0124] The clearance between the iron core and the closing spring is too large: Under normal mechanical conditions, the iron core ( Figure 2 (c)) and the positional relationship between the closing spring release and the closing spring release is as follows: Figure 2 As shown in (d), in order to simulate the fault, the distance between the iron core and the closing spring trip is manually increased until the iron core still cannot touch the trip when the electromagnet is attracted.

[0125] The tripping spring is loose: Use a wrench to loosen the tripping bolt. The loosening degree must reach the point where when the closing action occurs, the tripping spring can also act after storing energy, that is, the circuit breaker opens after closing.

[0126] Preferably, in some embodiments of the present application, the support vector machine can be trained in the following manner to obtain a prediction model:

[0127] 60% of the simulated sound signal data are selected for circuit breaker mechanical fault training in four stages: tripping, acceleration, deceleration and braking. The support vector machine (SVM) model is used for classification training, and the optimization objective function is:

[0128]

[0129] sty i (ω T x i +b)≥1-ε i ,ε i ≥0,i=1,2,…,n

[0130] where ω and b are the weight vector and bias term; ε i is the slack variable; C is the penalty parameter; x i is the input feature vector; y i is the category label; m is the total number of category labels.

[0131] By setting up a circuit breaker in a real environment, the circuit breaker can simulate sound signal data under different mechanical states, thereby improving the authenticity and diversity of the training data and the accuracy of the subsequent prediction model.

[0132] From the above, it can be seen that the circuit breaker mechanical fault identification method provided by the embodiment of the present application has the following beneficial effects: since when a mechanical fault occurs, abnormal sound is often mixed into the circuit breaker operation sound, the mechanical fault of the circuit breaker can be accurately identified by comparing the difference between the currently collected sound signal data and the sound signal data under normal operation through the pre-trained prediction model, and since a large number of sensors do not need to be set in each component of the circuit breaker when collecting sound signals, a large number of changes to the circuit breaker structure are avoided. However, the sound characteristics of the normal operation of the circuit breaker are different at different stages, so the combination of SOE data can accurately reflect the characteristics of the operation events of the circuit breaker at different operation stages, and the sound signal data at different stages are independently predicted, which effectively improves the accuracy of circuit breaker fault identification.

[0133] Embodiment 2

[0134] refer to Figure 3 , a circuit breaker mechanical fault identification device provided in an embodiment of the present application, includes: a data acquisition module 11, a first feature extraction module 12, a second feature extraction module 13, a fault identification module 14, a first data preprocessing module 15 and a second data preprocessing module 16.

[0135] Furthermore, in some embodiments of the present application, the data acquisition module 11 is used to acquire the SOE data and sound signal data of the circuit breaker within a preset time period; the first feature extraction module 12 is used to extract time features and action features from the SOE data, and identify the several stages of the circuit breaker within the preset time period according to the action features; the second feature extraction module 13 is used to extract domain features corresponding to each of the stages from the sound signal data according to the time features; the fault identification module 14 is used to input the domain features corresponding to each stage into the corresponding prediction model that has been pre-trained to obtain a fault identification result.

[0136] Furthermore, in some embodiments of the present application, the first feature extraction module 12 is used to extract time features and action features from the SOE data, and identify several stages of the circuit breaker within the preset time period based on the action features, including: extracting time features and action features of events occurring in the circuit breaker within the preset time period from the SOE data; and determining several stages of the circuit breaker within the preset time period through the action features.

[0137] Furthermore, in some embodiments of the present application, the first data preprocessing module 15 is used to preprocess the sound signal data after obtaining the sound signal data, including: performing noise reduction processing on the sound signal data through spectral subtraction to obtain first power spectrum data; performing soft threshold processing on the first power spectrum data to obtain the preprocessed sound signal data.

[0138] Furthermore, in some embodiments of the present application, the second feature extraction module 13 is used to extract domain features corresponding to each of the stages from the sound signal data according to the time features, including: the domain features include: time domain features, frequency domain features and cepstrum domain features; through the time features, the time interval corresponding to each of the stages is determined, the sound signal data is divided according to the time interval, and the sound signal data segments corresponding to each stage are obtained; the time domain features are extracted by analyzing the time waveforms of the sound signal data segments that have not been preprocessed as they migrate over time; the frequency domain features are obtained by inverse short-time Fourier transforming the preprocessed sound signal data segments; the cepstrum domain features are obtained by performing short-time Fourier transform on the frequency domain features, processing them with a Mel filter, and then performing logarithmic operations and discrete cosine transforms.

[0139] Furthermore, in some embodiments of the present application, the prediction model is obtained by setting the circuit breaker in different mechanical states for actual operation, collecting analog sound signal data of corresponding stages, and then training a support vector machine based on the analog sound signal data.

[0140] Furthermore, in some embodiments of the present application, the analog sound signal data and the sound signal data are both collected and acquired through evenly distributed multi-channel audio collection devices.

[0141] Furthermore, in some embodiments of the present application, the second data preprocessing module 16 is used to preprocess the SOE data after acquiring the SOE data, including: performing time synchronization correction on the SOE data collected from different sensors; and eliminating abnormal events in the SOE data according to the logic of occurrence before and after the event.

[0142] It can be understood that the above-mentioned device embodiment corresponds to the method embodiment of the present invention. A circuit breaker mechanical fault identification device provided by the embodiment of the present invention can implement any method embodiment of the present invention, that is, the circuit breaker mechanical fault identification method provided in embodiment one.

[0143] From the above, it can be seen that the circuit breaker mechanical fault identification device provided by the embodiment of the present application has the following beneficial effects: since when a mechanical fault occurs, abnormal sound is often mixed into the circuit breaker operation sound, the mechanical fault of the circuit breaker can be accurately identified by comparing the difference between the currently collected sound signal data and the sound signal data under normal operation through the pre-trained prediction model, and since a large number of sensors do not need to be set in each component of the circuit breaker when collecting sound signals, a large number of changes to the circuit breaker structure are avoided. However, the sound characteristics of the normal operation of the circuit breaker are different at different stages, so the combination of SOE data can accurately reflect the characteristics of the operation events of the circuit breaker at different operation stages, and the sound signal data at different stages are independently predicted, which effectively improves the accuracy of circuit breaker fault identification.

[0144] Embodiment 3

[0145] Based on the above-mentioned embodiment of the circuit breaker mechanical fault identification method, another embodiment of the present application provides a circuit breaker mechanical fault identification terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the circuit breaker mechanical fault identification method of any embodiment of the present application is implemented.

[0146] Exemplarily, in this embodiment, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present application. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program in the circuit breaker mechanical fault identification device.

[0147] The circuit breaker mechanical fault identification device may be a computing device such as a desktop computer, a notebook, a palm computer, a cloud server, etc. The circuit breaker mechanical fault identification terminal device may include, but is not limited to, a processor and a memory.

[0148] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the circuit breaker mechanical fault identification device, and uses various interfaces and lines to connect the various parts of the entire circuit breaker mechanical fault identification device. The memory may be used to store the computer program and / or module, and the processor implements various functions of the circuit breaker mechanical fault identification device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the mobile phone, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMed i aCard, SMC), a secure digital (SecureDi g ital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0149] Embodiment 4

[0150] Based on the above-mentioned embodiment of the circuit breaker mechanical fault identification method, another embodiment of the present application provides a storage medium, wherein the storage medium includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the circuit breaker mechanical fault identification method of any embodiment of the present application.

[0151] In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0152] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A circuit breaker mechanical fault identification method, characterized in that: include: Obtain SOE data and sound signal data of the circuit breaker within a preset time period; Extracting time features and action features from the SOE data, and identifying the stages of the circuit breaker within the preset time period according to the action features; Extracting domain features corresponding to each of the stages from the sound signal data according to the time features; The domain features corresponding to each stage are input into the corresponding pre-trained prediction model to obtain the fault identification result.

2. A circuit breaker mechanical fault identification method according to claim 1, characterized in that: The extracting time features and action features from the SOE data, and identifying the several stages of the circuit breaker in the preset time period according to the action features, includes: Extracting the time characteristics and action characteristics of the events occurring in the circuit breaker within the preset time period from the SOE data; The plurality of stages of the circuit breaker within the preset time period are determined through the action characteristics.

3. A circuit breaker mechanical fault identification method according to claim 1, characterized in that: After acquiring the sound signal data, the method further includes: Performing noise reduction processing on the sound signal data by spectrum subtraction to obtain first power spectrum data; Soft threshold processing is performed on the first power spectrum data to obtain the pre-processed sound signal data.

4. A circuit breaker mechanical fault identification method according to claim 3, characterized in that: The extracting, from the sound signal data, domain features corresponding to each of the stages according to the time features, comprises: The domain features include: time domain features, frequency domain features and cepstrum domain features; Determine the time interval corresponding to each of the stages through the time feature, divide the sound signal data according to the time interval, and obtain the sound signal data segments corresponding to each stage; Extracting the time domain features by analyzing the time waveform of the sound signal data segment that has not been preprocessed and that appears over time; Processing the preprocessed sound signal data segment by inverse short-time Fourier transform to obtain the frequency domain feature; The cepstrum domain features are obtained by performing short-time Fourier transform on the frequency domain features, processing them with a Mel filter, and then performing logarithmic operation and discrete cosine transform.

5. A circuit breaker mechanical fault identification method according to claim 1, characterized in that: The prediction model is obtained by setting the circuit breaker in different mechanical states for actual operation, collecting analog sound signal data of corresponding stages, and then training a support vector machine based on the analog sound signal data.

6. A circuit breaker mechanical fault identification method according to claim 5, characterized in that: include: The analog sound signal data and the sound signal data are both collected and acquired through evenly distributed multi-channel audio collection devices.

7. A circuit breaker mechanical fault identification method according to claim 1, characterized in that: After obtaining the SOE data, the method further includes: Performing time synchronization correction on the SOE data collected from different sensors; According to the logic of occurrence before and after the event, the abnormal events in the SOE data are eliminated.

8. A circuit breaker mechanical fault identification device, characterized in that: include: A data acquisition module, a first feature extraction module, a second feature extraction module and a fault identification module; The data acquisition module is used to acquire SOE data and sound signal data of the circuit breaker within a preset time period; The first feature extraction module is used to extract time features and action features from the SOE data, and identify several stages of the circuit breaker within the preset time period according to the action features; The second feature extraction module is used to extract domain features corresponding to each of the stages from the sound signal data according to the time feature; The fault identification module is used to input the domain features corresponding to each stage into the corresponding pre-trained prediction model to obtain a fault identification result.

9. A circuit breaker mechanical fault identification device as claimed in claim 8, characterized in that: The first feature extraction module is used to extract time features and action features from the SOE data, and identify several stages of the circuit breaker in the preset time period according to the action features, including: Extracting the time characteristics and action characteristics of the events occurring in the circuit breaker within the preset time period from the SOE data; The plurality of stages of the circuit breaker within the preset time period are determined through the action characteristics.

10. A circuit breaker mechanical fault identification device according to claim 8, characterized in that: Also includes: A first data preprocessing module; wherein the first data preprocessing module is used to preprocess the sound signal data after obtaining the sound signal data, including: Performing noise reduction processing on the sound signal data by spectrum subtraction to obtain first power spectrum data; Soft threshold processing is performed on the first power spectrum data to obtain the pre-processed sound signal data.

11. A circuit breaker mechanical fault identification device according to claim 10, characterized in that: The second feature extraction module is used to extract domain features corresponding to each of the stages from the sound signal data according to the time feature, including: The domain features include: time domain features, frequency domain features and cepstrum domain features; Determine the time interval corresponding to each of the stages through the time feature, divide the sound signal data according to the time interval, and obtain the sound signal data segments corresponding to each stage; Extracting the time domain features by analyzing the time waveform of the sound signal data segment that has not been preprocessed and that appears over time; Processing the preprocessed sound signal data segment by inverse short-time Fourier transform to obtain the frequency domain feature; The cepstrum domain features are obtained by performing short-time Fourier transform on the frequency domain features, processing them with a Mel filter, and then performing logarithmic operation and discrete cosine transform.

12. A circuit breaker mechanical fault identification device according to claim 8, characterized in that: The prediction model is obtained by setting the circuit breaker in different mechanical states for actual operation, collecting analog sound signal data of corresponding stages, and then training a support vector machine based on the analog sound signal data.

13. A circuit breaker mechanical fault identification device according to claim 12, characterized in that: The analog sound signal data and the sound signal data are both collected and acquired through evenly distributed multi-channel audio collection devices.

14. A circuit breaker mechanical fault identification device according to claim 8, characterized in that: Also includes: A second data preprocessing module; wherein the second data preprocessing module is used to preprocess the SOE data after acquiring the SOE data, including: Performing time synchronization correction on the SOE data collected from different sensors; According to the logic of occurrence before and after the event, the abnormal events in the SOE data are eliminated.

Images