Switch cabinet circuit breaker mechanical fault state identification method based on voiceprint recognition

Abstract

The invention discloses a switch cabinet circuit breaker mechanical fault state identification method based on voiceprint identification. The method comprises three parts of switch cabinet circuit breaker opening and closing sound pressure signal on-line monitoring, switch cabinet circuit breaker mechanical fault state inversion model construction and switch cabinet circuit breaker mechanical fault state on-line identification. The switch cabinet circuit breaker opening and closing sound pressure signal on-line monitoring process is as follows: a voiceprint sensor is arranged at a specific position outside a switch cabinet and is used for monitoring sound pressure signals during opening and closing of the switch cabinet circuit breaker. A large number of sound pressure signals under the mechanical fault of the switch cabinet circuit breaker are obtained through the multi-physical field simulation technology to serve as training samples, non-contact and online identification of the mechanical fault state of the switch cabinet circuit breaker is achieved, compared with a traditional method, the method has the advantages of being low in development cost, short in development period, simple in development process and the like, and due to the fact that the number of samples is larger, the identification accuracy is higher. Therefore, the training precision of the support vector machine algorithm is higher, and the mechanical fault state identification effect of the method is better.

Description

Technical Field

[0001] The present invention relates to the technical field of online monitoring and fault diagnosis of power equipment, and in particular to a method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition. Background Art

[0002] As a core component of distribution networks, switchgear circuit breakers perform both control and protection functions. Their reliable operation is directly linked to the safe operation of the power grid. According to statistics from major power grid organizations, mechanical failures account for approximately 83% of all circuit breaker failures. However, current diagnostic methods for mechanical circuit breaker failures in my country primarily rely on regular inspections and post-fault repairs. These methods exhibit a certain lag, making timely detection and troubleshooting difficult. Therefore, research on online detection technologies for switchgear circuit breakers is of great significance.

[0003] Most existing online detection technologies for mechanical faults in circuit breakers use vibration signals and the coil current during the circuit breaker opening and closing process as characteristic quantities. However, both methods have certain limitations. In the former, due to the contact installation of the vibration sensor, the sensor is easily affected by the installation position and method during the monitoring of the vibration signal; in the latter, relying solely on a single current signal, it is difficult to fully identify various types of circuit breaker faults, especially faults occurring outside the electromagnet area. Sound signals and vibration signals have the same origin and contain relevant information about the circuit breaker vibration. Therefore, compared with current signals, using them as characteristic quantities makes it easier to achieve comprehensive detection of mechanical faults in circuit breakers. In addition, sound sensors are easy to install and are non-contact measurements, which can avoid problems caused by installation position and installation method.

[0004] Currently, most methods for identifying mechanical circuit breaker faults based on sound signals, both domestically and internationally, rely on experimental methods to obtain sound signal samples from the circuit breaker fault. This method is time-consuming and labor-intensive to develop, with a long development cycle and irreversible damage to the test circuit breaker. Furthermore, the number of sound signal samples obtained through experimental methods is limited, and some circuit breaker fault states are difficult to simulate experimentally. This results in poor identification results for mechanical circuit breaker fault identification methods based on these experiments.

[0005] Therefore, this patent uses multi-physics field simulation technology to obtain a large number of sound signal samples under different mechanical fault states of switch cabinet circuit breakers. On this basis, a method for identifying the mechanical fault state of switch cabinet circuit breakers based on voiceprint recognition technology is proposed, which can achieve non-contact, online identification of the mechanical fault state of switch cabinet circuit breakers. Compared with traditional methods, the method described in this patent has the advantages of low development cost, short development cycle, and simple development process. In addition, due to the larger number of samples, the training accuracy of the support vector machine algorithm is higher, and the mechanical fault state identification effect of the method is better. Summary of the Invention

[0006] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0007] In view of the above problems existing in the existing method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition, the present invention is proposed.

[0008] Therefore, the present invention aims to provide a method for identifying mechanical faults in switchgear circuit breakers based on voiceprint recognition. This method employs a voiceprint sensor placed at a specific location near the switchgear to monitor the sound pressure signals during the opening and closing of the switchgear circuit breaker. Based on the characteristics of the sound pressure signals, the switchgear circuit breaker is then identified online and contactlessly. This method provides a basis for operational and maintenance decisions regarding switchgear circuit breakers, helping operators promptly identify and troubleshoot mechanical faults, thereby preventing serious accidents.

[0009] To solve the above technical problems, the present invention provides the following technical solution: a method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition, comprising the following steps:

[0010] Step 1: Online monitoring of the sound pressure signal of the switch cabinet circuit breaker during opening and closing. Place the soundprint sensor on the front of the switch cabinet, facing the sound source, and keep a distance of 0.5m from the switch cabinet. The soundprint sensor monitors the sound pressure signal of the switch cabinet circuit breaker during opening, closing, and operation in real time.

[0011] Step 2: Use the gamma filter cepstral coefficients to extract the sound pressure signal of the switchgear circuit breaker, reflecting the spectrum envelope and details of the sound pressure signal under different mechanical fault states of the circuit breaker, and then reflecting the frequency domain distribution characteristics of the sound pressure signal;

[0012] Step 3: Constructing an inversion model for the mechanical fault state of the switchgear circuit breaker, including constructing a switchgear sound field simulation model, dividing the switchgear circuit breaker into opening and closing states, simulating the sound pressure signals at the locations where the soundprint sensors are located under different mechanical fault states of the switchgear circuit breaker, extracting the gamma filter cepstral coefficients of the sound pressure signal simulation values ​​as the identification features of the mechanical fault state of the circuit breaker, and using the support vector machine algorithm to train the mechanical fault state samples;

[0013] Step 4: Online identification of the mechanical fault state of the switchgear circuit breaker. The measured value of the switchgear circuit breaker sound pressure signal obtained by online monitoring of the switchgear circuit breaker opening and closing sound pressure signals in step 1 is input into the switchgear circuit breaker mechanical fault state inversion model constructed in step 3 to obtain the mechanical fault state of the switchgear circuit breaker.

[0014] As a preferred solution of the method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition according to the present invention, the gamma filter realizes auditory filtering by using the impulse response characteristics of the equivalent human ear basilar membrane, and the center frequency of each filter corresponds to the basilar membrane position. Formula (1) is the step response function B of the mth filter: m (t):

[0015]

[0016] Where V is the filter gain; β is the filter order; f m is the center frequency of the mth filter; is the initial phase; θ(t) is the step function; c m is the attenuation factor, which is related to f m The relationship is:

[0017]

[0018] To extract the sound signal The gamma filter cepstral coefficients are first processed by windowing and then fast Fourier transform to obtain the spectrum sequence X i (k), then make the spectrum sequence X i (k) After passing through the gamma filter bank, the output result is exponentially compressed, and the compression coefficient is 0.2. Formula (3) represents the effect of the mth filter on the signal Filtering and compression process:

[0019]

[0020] Where, is the compression result; n is X i (k) the number of FFT points; H m (k) is the mth filter in X i (k) The gain of the k-th point corresponding to the frequency;

[0021] The sound signal can be obtained by decorrelating the compressed output of each filter through discrete cosine transformation The gamma filter cepstral coefficients, formula (4) is the discrete cosine transform formula:

[0022]

[0023] Where, d i (j) is the j-th dimension cepstral coefficient; I is the number of filters.

[0024] As a preferred solution of the method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition according to the present invention, the cepstral coefficient of the gamma filter is basically 0 after exceeding 30 dimensions, and thus the first 31 dimensions are taken as the feature quantity for identifying the mechanical fault state of the switch cabinet circuit breaker and the principal component analysis method is used to perform dimensionality reduction processing on it.

[0025] As an optimal solution of the method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition described in the present invention, the switch cabinet sound field simulation model construction process in step three includes constructing a geometric model of the switch cabinet and its surrounding air domain, assigning material parameters to various parts of the switch cabinet, setting sound field excitation and boundary conditions, meshing the geometric model of the switch cabinet and its surrounding air domain, setting solver parameters and solving the problem.

[0026] As a preferred solution of the method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition according to the present invention, in step three, the fault states of the switch cabinet circuit breaker are divided into five types: normal closing, normal opening, closing delay, opening delay, and refusal to operate; wherein, normal opening and closing states are normal states, and the other states are fault states; a large number of sound pressure signals at the arrangement positions of the voiceprint sensors under the five mechanical fault states of the switch cabinet are obtained by simulation, and their gamma filter cepstral coefficients are extracted, which are used as feature quantities for identifying the mechanical fault state of the circuit breaker.

[0027] As a preferred solution of the method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition described in the present invention, wherein: in the step three, a support vector machine algorithm is used to train the gamma filter cepstral coefficient samples of the sound pressure signal under different mechanical fault states of the circuit breaker, so as to construct an inversion model of the mechanical fault state of the switch cabinet circuit breaker, wherein the input is the gamma filter cepstral coefficient of the sound pressure signal, and the output is the faulty mechanical state of the switch cabinet circuit breaker.

[0028] As a preferred solution of the method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition according to the present invention, the sound pressure satisfies the constraint of formula (5) in the time and space domain:

[0029]

[0030] Where ρ0 is the fluid density; q is the dipole source; Q is the monopole source; p is the sound pressure; and c is the flow velocity. In the switchgear circuit breaker acoustic field simulation model, the sound pressure changes rapidly, so the time step is set to 0.5 ms to capture the sound pressure changes in a very short time.

[0031] Beneficial effects of the present invention: The present invention provides a method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition technology, which can effectively realize non-contact and online identification of the mechanical fault state of the switch cabinet circuit breaker, and help on-site operation and maintenance personnel to promptly discover the mechanical fault state of the switch cabinet circuit breaker and eliminate the fault, thereby avoiding more severe consequences. Compared with the circuit breaker mechanical fault identification method using current signals as feature quantities, the identification range is more comprehensive; compared with the method using vibration signals as feature quantities, the sensor installation is more convenient, and it is a non-contact measurement, which can avoid problems caused by the installation position and installation method. In addition, compared with the traditional method for identifying the mechanical fault of the circuit breaker based on voiceprint recognition technology, the method of the present invention uses multi-physics field simulation technology to obtain a large number of sound pressure signal samples under circuit breaker faults, which has the advantages of low development cost, short development cycle, simple development process, etc., and because the number of samples is larger, the training accuracy of the support vector machine algorithm is higher, and the mechanical fault state identification effect of the method is better. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0033] Figure 1 It is the overall flow chart of the present invention;

[0034] Figure 2 Arrange a location for the voiceprint sensor;

[0035] Figure 3 Build a process for switchgear acoustic field simulation model;

[0036] Figure 4 is the actual geometry of the switchgear;

[0037] Figure 5 is the switchgear geometric model;

[0038] Figure 6 The sound pressure distribution of the switch cabinet at different times when the circuit breaker is normally closed (sound pressure: Pa);

[0039] Figure 7 The sound pressure signal measured by the soundprint sensor under different mechanical fault states of the circuit breaker;

[0040] Figure 8 This is the fault classification flow chart based on the support vector machine algorithm. DETAILED DESCRIPTION

[0041] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0042] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0043] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0044] Furthermore, the present invention is described in detail with reference to schematic diagrams. For ease of illustration, when describing the embodiments of the present invention, cross-sectional views illustrating device structures may be partially enlarged and not to scale. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of protection of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0045] Reference Figures 1-8 , provides a method for identifying the mechanical fault state of a switch cabinet circuit breaker based on voiceprint recognition, comprising the following steps:

[0046] Step 1: Online monitoring of the sound pressure signal of the switch cabinet circuit breaker during opening and closing. Place the soundprint sensor on the front of the switch cabinet, facing the sound source, and keep a distance of 0.5m from the switch cabinet. The soundprint sensor monitors the sound pressure signal of the switch cabinet circuit breaker during opening, closing, and operation in real time.

[0047] Step 2: Use the gamma filter cepstral coefficients to extract the sound pressure signal of the switchgear circuit breaker, reflecting the spectrum envelope and details of the sound pressure signal under different mechanical fault states of the circuit breaker, and then reflecting the frequency domain distribution characteristics of the sound pressure signal;

[0048] Step 3: Constructing an inversion model for the mechanical fault state of the switchgear circuit breaker, including constructing a switchgear sound field simulation model, dividing the switchgear circuit breaker into opening and closing states, simulating the sound pressure signals at the locations where the soundprint sensors are located under different mechanical fault states of the switchgear circuit breaker, extracting the gamma filter cepstral coefficients of the sound pressure signal simulation values ​​as the identification features of the mechanical fault state of the circuit breaker, and using the support vector machine algorithm to train the mechanical fault state samples;

[0049] Step 4: Online identification of the mechanical fault state of the switchgear circuit breaker. The measured value of the switchgear circuit breaker sound pressure signal obtained by online monitoring of the switchgear circuit breaker opening and closing sound pressure signals in step 1 is input into the switchgear circuit breaker mechanical fault state inversion model constructed in step 3 to obtain the mechanical fault state of the switchgear circuit breaker.

[0050] Among them, the gamma filter realizes auditory filtering by using the impulse response characteristics of the equivalent human ear basilar membrane. The center frequency of each filter corresponds to the same position of the basilar membrane. Formula (1) is the step response function B of the mth filter m (t):

[0051]

[0052] Where V is the filter gain; β is the filter order; f m is the center frequency of the mth filter; is the initial phase; θ(t) is the step function; c m is the attenuation factor, which is related to f m The relationship is:

[0053]

[0054] To extract the sound signal The gamma filter cepstral coefficients are first processed by windowing and then fast Fourier transform to obtain the spectrum sequence X i (k), then make the spectrum sequence X i (k) After passing through the gamma filter bank, the output result is exponentially compressed, and the compression coefficient is 0.2. Formula (3) represents the effect of the m-th filter on the signal Filtering and compression process:

[0055]

[0056] Where, is the compression result; n is X i (k) the number of FFT points; H m (k) is the mth filter in X i (k) The gain of the k-th point corresponding to the frequency;

[0057] The sound signal can be obtained by decorrelating the compressed output of each filter through discrete cosine transformation The gamma filter cepstral coefficients, formula (4) is the discrete cosine transform formula:

[0058]

[0059] Where, d i (j) is the j-th dimension cepstral coefficient; I is the number of filters.

[0060] Furthermore, the cepstral coefficients of the gamma filter are basically 0 after exceeding 30 dimensions, so the first 31 dimensions are taken as the feature quantities for identifying the mechanical fault state of the switch cabinet circuit breaker and the principal component analysis method is used to reduce the dimension.

[0061] Among them, the construction process of the switch cabinet sound field simulation model in step three includes the construction of the switch cabinet and its surrounding air domain geometric model, the assignment of material parameters of each part of the switch cabinet, the setting of sound field excitation and boundary conditions, the meshing of the switch cabinet and its surrounding air domain geometric model, the setting of solver parameters and the solution.

[0062] Furthermore, in step three, the fault states of the switch cabinet circuit breaker are divided into five types: normal closing, normal opening, closing delay, opening delay, and refusal to operate; among them, normal opening and closing states are normal states, and the other states are fault states; the simulation obtains the sound pressure signals at the arrangement position of the soundprint sensor under the five mechanical fault states of a large number of switch cabinets and extracts their gamma filter cepstral coefficients, which are used as the feature quantity for identifying the mechanical fault state of the circuit breaker.

[0063] Furthermore, in step three, a support vector machine algorithm is used to train the gamma filter cepstral coefficient samples of the sound pressure signal under different mechanical fault states of the circuit breaker to construct an inversion model of the mechanical fault state of the switchgear circuit breaker, where the input is the gamma filter cepstral coefficient of the sound pressure signal and the output is the fault mechanical state of the switchgear circuit breaker.

[0064] Specifically, the sound pressure satisfies the constraint of formula (5) in the time and space domain:

[0065]

[0066] Where ρ0 is the fluid density; q is the dipole source; Q is the monopole source; p is the sound pressure; and c is the flow velocity. In the switchgear circuit breaker acoustic field simulation model, the sound pressure changes rapidly, so the time step is set to 0.5 ms to capture the sound pressure changes in a very short time.

[0067] Specific embodiment:

[0068] Taking a 10kV switchgear as an example, the modeling process of the mechanical fault state inversion model of the switchgear circuit breaker is explained.

[0069] 1. Build a switch cabinet sound field simulation model

[0070] The process of constructing a switch cabinet sound field simulation model using multi-physics field simulation technology is as follows: Figure 3As shown, it includes five parts: building the geometric model of the switch cabinet and its surrounding air domain, assigning material parameters to each part of the switch cabinet, setting the acoustic field excitation and boundary conditions, meshing the geometric model of the switch cabinet and its surrounding air domain, setting the solver parameters and solving the problem. The specific process is as follows:

[0071] The switchgear geometry is as follows Figure 4 As shown in the figure, the circuit breaker is located in the orange area. When opening and closing, the circuit breaker contacts collide and vibrate, thereby generating sound pressure at the contact position and spreading to the surrounding area. The sound pressure satisfies the constraint of formula (5) in the time and space domain:

[0072]

[0073] Where ρ0 is the fluid density; q is the dipole source; Q is the monopole source; p is the sound pressure; c is the flow velocity; In order to improve the calculation speed while ensuring the accuracy, Figure 4 The actual structure of the switch cabinet shown in the figure is appropriately simplified, and the simplified geometric model is as follows Figure 5 As shown in the figure, the orange dots indicate the contact locations. The circuit breaker, located in the circuit breaker compartment at the center front of the switchgear, is one of the switchgear's core components. Table 1 shows the material properties of the switchgear's solid and air domains, where pA represents absolute pressure in Pa and T represents temperature in Kelvin. The physical properties of solid materials are relatively stable and largely unaffected by temperature, so they are assumed to be constants. However, the density and speed of sound in air are significantly affected by temperature and absolute pressure and are functions of both.

[0074] Table 1 Materials of various parts of switchgear

[0075]

[0076] Will Figure 5 The switch cabinet boundary shown is set as a "hard acoustic boundary" and the sound pressure and its first-order time derivative are set to 0. The sound source is located at Figure 5 A "monopole source" is set at the circuit breaker contact point shown. In the switchgear circuit breaker acoustic field simulation model, the sound pressure changes rapidly, so a time step of 0.5ms is set to capture very short-term changes. Running the program on a computer with a Core™ i7-12700 processor and 16GB of RAM, it took 31 hours to complete the simulation of the switchgear in one circuit breaker state. Figure 6 The sound pressure distribution inside the switch cabinet at 0s, 0.5s, 1s and 1.5s when the circuit breaker is normally closed is shown. As time goes by, the absolute value of the sound pressure inside the switch cabinet first increases and then decreases. At 0.5s, as the circuit breaker is closed, the absolute value of the sound pressure reaches a maximum value of 0.53Pa.

[0077] 2. Classify the opening and closing states of the switch cabinet circuit breaker

[0078] Switchgear circuit breakers can experience a variety of mechanical failures, including common ones like spring mechanism jamming, transmission guide rod jamming, and operating mechanism jamming. These failures affect the circuit breaker contacts through mechanical components like gears and springs, causing them to typically exhibit the following five states during opening and closing: normal closing, normal opening, closing delay, opening delay, and refusal to operate. Normal opening and closing are considered normal, while the remaining states are considered faulty. The sound pressure at the circuit breaker contacts varies in different states at the same time, and thus the sound pressure transmitted to any location in space also varies. This characteristic provides a theoretical basis for identifying mechanical fault states in switchgear circuit breakers.

[0079] 3. Simulate and obtain the sound pressure signal at the location of the soundprint sensor under different mechanical fault conditions of the switch cabinet circuit breaker

[0080] Based on the switchgear sound field simulation model established above, a large number of sound pressure signal samples at the arrangement positions of the soundprint sensors under different mechanical fault states of the switchgear circuit breakers are obtained through simulation. Figure 7 The sound pressure signals measured by the soundprint sensor under five typical mechanical fault states of the switch cabinet circuit breaker are shown.

[0081] 4. Extract the gamma filter cepstral coefficients of the sound pressure signal simulation value as the characteristic quantity for identifying the mechanical fault state of the circuit breaker

[0082] The method of extracting the gamma filter cepstral coefficients of the sound signal in process (1) is used to extract the characteristic value of the sound pressure signal simulation value at the layout position of the soundprint sensor under different mechanical fault conditions of the switch cabinet circuit breaker. Table 2 shows the characteristics of the sound pressure signal simulation value at the layout position of the soundprint sensor under different mechanical fault conditions of the switch cabinet circuit breaker. Figure 7 The characteristic vector of the sound pressure signal simulation value under the typical mechanical fault state of the circuit breaker after the gamma filter cepstral coefficient extraction of the sound pressure signal simulation value.

[0083] Table 2 Characteristic vectors of the simulated sound pressure signal under five typical mechanical fault conditions of the switchgear circuit breaker

[0084]

[0085]

[0086] 5. Use support vector machine algorithm to train mechanical fault state samples

[0087] Support vector machines (SVMs) are common algorithms for supervised learning and intelligent classification. They effectively address the problem of previous machine learning algorithms failing to achieve satisfactory classification results for practical problems involving high dimensions and nonlinearity. The basic principle of SVMs is to map problems that cannot be classified in low dimensions to higher dimensions by selecting an appropriate kernel function, and they exhibit good generalization capabilities.

[0088] The present invention uses this algorithm to train the gamma filter cepstral coefficient samples of the sound pressure signal simulation value under different mechanical fault states of the switch cabinet circuit breaker to realize the identification of the mechanical fault state of the switch cabinet circuit breaker. The 400 sets of gamma filter cepstral coefficient samples of the sound pressure signal simulation value under different mechanical fault states of the switch cabinet circuit breaker obtained by the above process are divided into a training set and a test set, wherein the number of training set samples is 300 and the number of test set samples is 100. Figure 8 The process trains on characteristic vector samples of sound pressure signals measured by the soundprint sensor under different mechanical fault conditions of the circuit breaker, thereby constructing an inversion model for the mechanical fault state of the switchgear circuit breaker. Testing has shown that the recognition accuracy of the entire model is 93.67%.

[0089] This invention provides a method for identifying mechanical faults in switchgear circuit breakers based on voiceprint recognition technology. This method uses a voiceprint sensor placed at a specific location near the switchgear to monitor the sound pressure signals during the opening and closing of the switchgear circuit breaker. Based on the characteristics of the sound pressure signals, the method performs non-contact, online identification of the switchgear circuit breaker. This method provides a basis for operational and maintenance decisions regarding switchgear circuit breakers, helping operators promptly identify and troubleshoot mechanical faults, thereby preventing serious accidents.

[0090] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for identifying mechanical fault status of a switch cabinet circuit breaker based on voiceprint recognition, characterized in that: The following steps are involved: Step 1: Online monitoring of the sound pressure signal of the switch cabinet circuit breaker during opening and closing. Place the soundprint sensor on the front of the switch cabinet, facing the sound source, and keep a distance of 0.5m from the switch cabinet. The soundprint sensor monitors the sound pressure signal of the switch cabinet circuit breaker during opening, closing, and operation in real time. Step 2: Use the gamma filter cepstral coefficients to extract the sound pressure signal of the switchgear circuit breaker, reflecting the spectrum envelope and details of the sound pressure signal under different mechanical fault states of the circuit breaker, and then reflecting the frequency domain distribution characteristics of the sound pressure signal; Step 3: Constructing an inversion model for the mechanical fault state of the switchgear circuit breaker, including constructing a switchgear sound field simulation model, dividing the switchgear circuit breaker into opening and closing states, simulating the sound pressure signals at the locations where the soundprint sensors are located under different mechanical fault states of the switchgear circuit breaker, extracting the gamma filter cepstral coefficients of the sound pressure signal simulation values ​​as the identification features of the mechanical fault state of the circuit breaker, and using the support vector machine algorithm to train the mechanical fault state samples; Step 4: Online identification of the mechanical fault state of the switchgear circuit breaker. The measured value of the switchgear circuit breaker sound pressure signal obtained by online monitoring of the switchgear circuit breaker opening and closing sound pressure signals in step 1 is input into the switchgear circuit breaker mechanical fault state inversion model constructed in step 3 to obtain the mechanical fault state of the switchgear circuit breaker.

2. The method for identifying mechanical fault status of a switch cabinet circuit breaker based on voiceprint recognition according to claim 1, characterized in that: The gamma filter realizes auditory filtering by using the impulse response characteristics of the equivalent human ear basilar membrane. The center frequency of each filter corresponds to the same position of the basilar membrane. Formula (1) is the step response function B of the mth filter: m (t): Where V is the filter gain; β is the filter order; f m is the center frequency of the mth filter; is the initial phase; θ(t) is the step function; c m is the attenuation factor, which is related to f m The relationship is: To extract the sound signal The gamma filter cepstral coefficients are first processed by windowing and then fast Fourier transform to obtain the spectrum sequence X i (k), then make the spectrum sequence X i (k) After passing through the gamma filter bank, the output result is exponentially compressed, and the compression coefficient is 0.

2. Formula (3) represents the effect of the mth filter on the signal Filtering and compression process: Where, is the compression result; n is X i (k) the number of FFT points; H m (k) is the mth filter in X i (k) The gain of the k-th point corresponding to the frequency; The sound signal can be obtained by decorrelating the compressed output of each filter through discrete cosine transformation The gamma filter cepstral coefficients, formula (4) is the discrete cosine transform formula: Where, d i (j) is the j-th dimension cepstral coefficient; I is the number of filters.

3. The method for identifying mechanical fault status of a switch cabinet circuit breaker based on voiceprint recognition according to claim 2, characterized in that: The cepstral coefficients of the gamma filter are substantially 0 after exceeding 30 dimensions, so the first 31 dimensions are taken as feature quantities for identifying the mechanical fault state of the switch cabinet circuit breaker and are subjected to dimensionality reduction processing using the principal component analysis method.

4. The method for identifying mechanical fault status of a switch cabinet circuit breaker based on voiceprint recognition according to claim 1, characterized in that: The construction process of the switch cabinet sound field simulation model in step three includes constructing a geometric model of the switch cabinet and its surrounding air domain, assigning material parameters to various parts of the switch cabinet, setting sound field excitation and boundary conditions, meshing the geometric model of the switch cabinet and its surrounding air domain, setting solver parameters and solving the problem.

5. The method for identifying mechanical fault status of a switch cabinet circuit breaker based on voiceprint recognition according to claim 4, characterized in that: In step three, the switchgear circuit breaker fault states are divided into five types: normal closing, normal opening, closing delay, opening delay, and refusal to operate; among them, normal opening and closing states are normal states, and the other states are fault states; a large number of sound pressure signals at the arrangement locations of the soundprint sensors under the five mechanical fault states of the switchgear are simulated and their gamma filter cepstral coefficients are extracted, which are used as feature quantities for identifying the mechanical fault states of the circuit breaker.

6. The method for identifying mechanical fault status of a switch cabinet circuit breaker based on voiceprint recognition according to claim 5, characterized in that: In step three, a support vector machine algorithm is used to train the gamma filter cepstral coefficient samples of the sound pressure signal under different mechanical fault states of the circuit breaker to construct an inversion model of the mechanical fault state of the switchgear circuit breaker, wherein the input is the gamma filter cepstral coefficient of the sound pressure signal, and the output is the fault mechanical state of the switchgear circuit breaker.

7. The method for identifying mechanical fault status of a switch cabinet circuit breaker based on voiceprint recognition according to claim 5, characterized in that: The sound pressure satisfies the constraint of formula (5) in the time-space domain: Where ρ0 is the fluid density; q is the dipole source; Q is the monopole source; p is the sound pressure; and c is the flow velocity. In the switchgear circuit breaker acoustic field simulation model, the sound pressure changes rapidly, so the time step is set to 0.5 ms to capture the sound pressure changes in a very short time.

Images