Circuit breaker fault type identification method and device based on multi-dimensional signal characteristics, computer equipment, readable storage medium and program product

Abstract

The invention relates to the technical field of circuit breaker fault detection, and provides a circuit breaker fault type identification method and device based on multi-dimensional signal characteristics, computer equipment, a readable storage medium and a program product. The method comprises the steps of obtaining a stroke time sequence feature vector, a speed time sequence feature vector, an acceleration time sequence feature vector and a global feature vector according to a stroke signal, a speed signal and an acceleration signal in a circuit breaker fault state; performing time sequence windowing processing on the travel time sequence feature vector, the speed time sequence feature vector and the acceleration time sequence feature vector to obtain a first time sequence window feature vector and a second time sequence window feature vector; fusing the first time sequence window feature vector, the second time sequence window feature vector and the global feature vector to obtain a fused time sequence feature vector; and obtaining the fault type of the circuit breaker according to the fusion time sequence feature vector and a fault type identification model. By adopting the method, the circuit breaker fault type identification precision can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of circuit breaker fault detection, and in particular to a circuit breaker fault type identification method, apparatus, computer equipment, computer-readable storage medium and computer program product based on multi-dimensional signal characteristics. Background Art

[0002] Circuit breakers (such as high-voltage circuit breakers) are one of the important components in the power system. They need to work stably and reliably when the power system is normal or faulty to ensure the reliability and stability of the power system. In the case of circuit breaker failure, operating mechanism failure accounts for a high proportion, among which mechanical failure accounts for a high proportion of operating mechanism failures. Mechanical failures are mainly spring loosening, mechanism jamming and other failure types. Accurately identifying the type of circuit breaker failure is a necessary means to improve the operation and maintenance level of the power system and ensure the safe and stable operation of the power system.

[0003] Traditional circuit breaker fault type identification methods usually perform feature analysis based on single signals such as vibration signals or opening and closing coil current signals. However, the features of a single signal have a certain bias, ignoring the association of other signals such as vibration and travel, which reduces the accuracy of circuit breaker fault type identification. For example, during circuit breaker operation, the vibration signal may have a large error due to environmental noise or sensor position, or the electromagnet current signal may be affected by voltage fluctuations, resulting in misdiagnosis of circuit breaker fault type identification.

[0004] In addition, the operation process of the operating mechanism involves the coordinated action of more than 10 mechanical parts, and each action stage corresponds to different components, so it is necessary to pay attention to the timing signals in each action stage. The traditional method of identifying the fault type of the circuit breaker extracts the features of the vibration signal of the entire action sequence, but fails to fully consider the specific action characteristics of each time sequence interval, which also reduces the accuracy of the circuit breaker fault type identification. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for identifying circuit breaker fault types based on multi-dimensional signal characteristics in order to solve the above technical problems.

[0006] In a first aspect, the present application provides a method for identifying a circuit breaker fault type based on multi-dimensional signal features, comprising:

[0007] Acquire travel signal, speed signal and acceleration signal of the circuit breaker in fault state;

[0008] According to the travel signal, speed signal and acceleration signal of the circuit breaker in a fault state, a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector are obtained;

[0009] Performing time series window processing on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector;

[0010] Fusing the first time series window feature vector, the second time series window feature vector and the global feature vector to obtain a fused time series feature vector;

[0011] The fault type of the circuit breaker is obtained according to the fused time series feature vector and the pre-built fault type identification model.

[0012] In one embodiment, the step of obtaining a travel signal, a speed signal, and an acceleration signal of a circuit breaker in a fault state includes:

[0013] Collect the action process of the moving contact when the circuit breaker is in the opening and closing state under the fault state, and obtain the original action video;

[0014] Decomposing the original action video into single-frame action images, graying each frame of the action image, and performing edge monitoring to obtain a displacement signal of a moving contact of the circuit breaker;

[0015] Drawing a travel curve diagram of the moving contact according to the displacement signal and time series of the moving contact;

[0016] Mark the displacement node signal and time node signal of the moving contact on the travel curve diagram; the displacement node signal includes the opening position and the closing position; the time node signal includes the breaking time point and the closing time point;

[0017] Obtaining a travel signal of the circuit breaker according to the opening position and the closing position;

[0018] According to the opening position, closing position, breaking time point, closing time point and speed calculation formula, the speed signal of the circuit breaker is obtained;

[0019] According to the opening position, closing position, breaking time point, closing time point and acceleration calculation formula, the acceleration signal of the circuit breaker is obtained.

[0020] In one embodiment, obtaining a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector according to the travel signal, speed signal and acceleration signal in the circuit breaker fault state includes:

[0021] According to the opening and closing states of the circuit breaker under the fault state, the opening time period and the closing time period are obtained;

[0022] According to the travel signal and speed signal of the circuit breaker, the moving contact position change signal change time, the moving contact maximum change travel and the moving contact maximum speed travel are obtained;

[0023] According to the opening time period, closing time period, moving contact position change signal change time, moving contact maximum change stroke and moving contact maximum speed stroke, the stroke time sequence characteristic vector is obtained;

[0024] According to the speed signal of the circuit breaker, the average speed of the moving contact action, the speed when the moving contact displacement signal changes, the maximum speed of the moving contact and the time when the moving contact reaches the maximum speed are obtained to obtain a speed time series characteristic vector;

[0025] Performing variational mode decomposition processing on the acceleration signal to obtain an acceleration time series eigenvector;

[0026] The travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector are combined together to form a global feature vector.

[0027] In one embodiment, performing variational mode decomposition processing on the acceleration signal to obtain the acceleration time series feature vector includes:

[0028] Performing Hilbert transformation on the acceleration signal to decompose and obtain analytical signals and unilateral spectra of initial components of several intrinsic mode functions;

[0029] The initial component of the intrinsic mode function is iteratively updated, and when the change of the iterative result is less than a set threshold, the iteration is stopped to obtain the intrinsic mode function component to obtain the acceleration time series characteristic vector.

[0030] In one embodiment, performing time-series window processing on the travel time-series feature vector, the speed time-series feature vector, and the acceleration time-series feature vector to obtain a first time-series window feature vector and a second time-series window feature vector includes:

[0031] Obtaining a first time domain feature and a second time domain feature according to the travel time series feature vector and the speed time series feature vector;

[0032] According to the acceleration time series feature vector, a first acceleration time-frequency domain feature and a second acceleration time-frequency domain feature are obtained;

[0033] Obtaining a first time series window feature vector according to the first time domain feature and the first acceleration time-frequency domain feature;

[0034] A second time series window feature vector is obtained according to the second time domain feature and the second acceleration time-frequency domain feature.

[0035] In one embodiment, before obtaining the fault type of the circuit breaker according to the fused time series feature vector and the pre-built fault type identification model, the method further includes:

[0036] According to the nature-inspired rule optimization algorithm and the support vector machine model, the model to be trained is obtained;

[0037] According to the fused time series feature vector sample set, the model to be trained is trained and optimized until the prediction accuracy reaches convergence and the training is stopped to obtain the fault type recognition model.

[0038] In a second aspect, the present application also provides a circuit breaker fault type identification device based on multi-dimensional signal characteristics, comprising:

[0039] A signal acquisition module, used to acquire a travel signal, a speed signal and an acceleration signal of the circuit breaker in a fault state;

[0040] A feature vector acquisition module, used to obtain a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector according to a travel signal, a speed signal and an acceleration signal in a fault state of the circuit breaker;

[0041] A time series window processing module, used for performing time series window processing on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector;

[0042] A feature vector fusion module, used for fusing the first time series window feature vector, the second time series window feature vector and the global feature vector to obtain a fused time series feature vector;

[0043] The fault type identification module is used to obtain the fault type of the circuit breaker according to the fused time series feature vector and the pre-built fault type identification model.

[0044] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the above method.

[0045] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and the computer program is executed by a processor to execute the above method.

[0046] In a fifth aspect, the present application further provides a computer program product, wherein the computer program product comprises a computer program, and the computer program is executed by a processor to execute the above method.

[0047] The above-mentioned circuit breaker fault type identification method, device, computer equipment, computer-readable storage medium and computer program product based on multi-dimensional signal characteristics obtain the travel signal, speed signal and acceleration signal under the circuit breaker fault state; obtain the travel time series feature vector, speed time series feature vector, acceleration time series feature vector and global feature vector according to the travel signal, speed signal and acceleration signal under the circuit breaker fault state; perform time series window processing on the travel time series feature vector, speed time series feature vector and acceleration time series feature vector to obtain the first time series window feature vector and the second time series window feature vector; fuse the first time series window feature vector, the second time series window feature vector and the global feature vector to obtain the fused time series feature vector; obtain the fault type of the circuit breaker according to the fused time series feature vector and a pre-built fault type identification model. The present application extracts features from the travel signal, speed signal and acceleration signal under the circuit breaker fault state and performs feature fusion to obtain a fused time series feature vector to obtain the fault type of the circuit breaker, thereby avoiding the phenomenon of loss of other signal features caused by using a single signal for feature extraction, and can improve the accuracy of circuit breaker fault type identification; time series window processing is performed on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector, so that the signal features of different action intervals can be better obtained, and local information is taken into account, thereby further improving the accuracy of circuit breaker fault type identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 FIG. 1 is an application environment diagram of a circuit breaker fault type identification method based on multi-dimensional signal features in one embodiment;

[0050] Figure 2 It is a flowchart of a method for identifying a circuit breaker fault type based on multi-dimensional signal features in one embodiment;

[0051] Figure 3 is a schematic diagram of a travel signal, a speed signal and an acceleration signal in one embodiment;

[0052] Figure 4 is a flow chart of a method for identifying a circuit breaker fault type based on multi-dimensional signal features in another embodiment;

[0053] Figure 5is a structural block diagram of a circuit breaker fault type identification device based on multi-dimensional signal characteristics in one embodiment;

[0054] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] The circuit breaker fault type identification method based on multi-dimensional signal characteristics provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The terminal 102 can obtain the travel signal, speed signal and acceleration signal under the fault state of the circuit breaker, and then obtain the fault type of the circuit breaker. Among them, the terminal 102 can be but not limited to various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The server 104 can be an independent physical server, or it can be a server cluster or distributed system composed of multiple physical servers, or it can be a cloud server that provides cloud computing services.

[0057] In an exemplary embodiment, Figure 2 As shown in FIG, a circuit breaker fault type identification method based on multi-dimensional signal characteristics is provided. The method is applied to Figure 1 The terminal 102 in the example is used as an example to illustrate, including the following steps S201 to S205. Among them:

[0058] Step S201, obtaining a travel signal, a speed signal and an acceleration signal of a circuit breaker in a fault state.

[0059] The action process of the moving contact in the opening and closing state under the circuit breaker fault state can be collected by a high-speed camera, and the action process can be analyzed to obtain the travel signal, speed signal and acceleration signal under the circuit breaker fault state, such as Figure 3 shown.

[0060] Step S202, obtaining a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector according to the travel signal, the speed signal and the acceleration signal in the circuit breaker fault state.

[0061] The travel signal, speed signal and acceleration signal under the circuit breaker fault state can be respectively extracted to obtain the travel time series feature vector, speed time series feature vector and acceleration time series feature vector, which can be collectively referred to as a multi-dimensional signal time series feature vector.

[0062] The travel time series feature vector, the speed time series feature vector, and the acceleration time series feature vector can be combined to obtain a global feature vector.

[0063] Step S203, performing time series window processing on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector.

[0064] The opening time period and the closing time period can be obtained according to the opening and closing states of the circuit breaker to determine the first timing window and the second timing window.

[0065] According to the first timing window and the second timing window, the travel timing characteristic vector, the speed timing characteristic vector and the acceleration timing characteristic vector are subjected to timing window processing to obtain the first timing window characteristic vector and the second timing window characteristic vector.

[0066] Step S204, fusing the first time series window feature vector, the second time series window feature vector and the global feature vector to obtain a fused time series feature vector.

[0067] The first time window feature vector, the second time window feature vector and the global feature vector may be fused. For example, the first time window feature vector, the second time window feature vector and the global feature vector may be merged to obtain a higher-dimensional fused time window feature vector.

[0068] The feature vector set composed of the feature vectors of the first time series window can be recorded as A. Specifically, the feature vectors of the first time series window are The feature vector set composed of the feature vectors of the second time series window can be recorded as B. Specifically, the feature vectors of the second time series window are For time domain features, the global feature vector is V1, which is independent of the window cutting position and is recorded as .

[0069] According to the vector space merging strategy, the first time window feature vectors, the second time window feature vectors and the global feature vectors from multiple signal sources can be fused at the same level to obtain a fused time series feature vector as shown in formula (1).

[0070] (1)

[0071] Among them, R represents the fused time series feature vector set, which is 20 dimensions.

[0072] Step S205, obtaining the fault type of the circuit breaker according to the fused time series feature vector and the pre-built fault type identification model.

[0073] The fused time series feature vector can be input into a pre-built fault type identification model to obtain an output result of the fault type identification model; and the output result of the fault type identification model can be used as the fault type of the circuit breaker.

[0074] In the above-mentioned circuit breaker fault type identification method based on multi-dimensional signal features, feature extraction and feature fusion are performed on the travel signal, speed signal and acceleration signal under the circuit breaker fault state to obtain a fused time series feature vector to obtain the fault type of the circuit breaker, thereby avoiding the phenomenon of loss of other signal features caused by using a single signal for feature extraction, and can improve the accuracy of circuit breaker fault type identification; time series window processing is performed on the travel time series feature vector, speed time series feature vector and acceleration time series feature vector, so as to better obtain the signal characteristics of different action intervals, and take local information into consideration, thereby further improving the accuracy of circuit breaker fault type identification.

[0075] In one of the embodiments, a travel signal, a speed signal and an acceleration signal of a circuit breaker under a fault state are obtained, and the specific steps are as follows: the action process of the moving contact in the opening and closing states under the circuit breaker fault state is collected to obtain an original action video; the original action video is decomposed into a single-frame action image, each frame of the action image is grayed, and edge monitoring is performed to obtain a displacement signal of the moving contact of the circuit breaker; a travel curve diagram of the moving contact is drawn according to the displacement signal and time series of the moving contact; the displacement node signal and the time node signal of the moving contact are marked on the travel curve diagram; the displacement node signal includes an opening position and a closing position; the time node signal includes an opening time point and a closing time point; the travel signal of the circuit breaker is obtained according to the opening position and the closing position; the speed signal of the circuit breaker is obtained according to the opening position, the closing position, the opening time point, the closing time point and the speed calculation formula; the acceleration signal of the circuit breaker is obtained according to the opening position, the closing position, the opening time point, the closing time point and the acceleration calculation formula.

[0076] The action process of the moving contact of the circuit breaker in the opening and closing states under the fault state can be captured by a high-speed camera to obtain the original action video; the original action video can be decomposed into single-frame action images, each frame of the action image is grayed out, and edge monitoring is performed to obtain the displacement signal of the moving contact of the circuit breaker.

[0077] A travel curve diagram of the moving contact can be drawn based on the displacement signal and time series of the moving contact; the displacement node signal and time node signal of the moving contact are marked on the travel curve diagram; the displacement node signal includes the opening position and the closing position, and the time node signal includes the breaking time point and the closing time point.

[0078] The travel signal of the circuit breaker can be obtained according to the opening position and the closing position. The speed signal of the circuit breaker can be obtained according to the opening position, the closing position, the breaking time point, the closing time point and the speed calculation formula. The acceleration signal of the circuit breaker can be obtained according to the opening position, the closing position, the breaking time point, the closing time point and the acceleration calculation formula.

[0079] In this embodiment, according to the action process of the moving contact when the circuit breaker is in the opening and closing state, a stroke signal, a speed signal and an acceleration signal with high accuracy are analyzed and obtained.

[0080] In one of the embodiments, a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector are obtained according to a travel signal, a speed signal and an acceleration signal under a circuit breaker fault state, and the specific steps are as follows: according to the opening and closing state under the circuit breaker fault state, an opening time period and a closing time period are obtained; according to the travel signal and the speed signal of the circuit breaker, a moving contact displacement signal change time, a moving contact maximum change stroke and a moving contact maximum speed stroke are obtained; according to the opening time period, the closing time period, the moving contact displacement signal change time, the moving contact maximum change stroke and the moving contact maximum speed stroke, a travel time series feature vector is obtained; according to the speed signal of the circuit breaker, an average speed of the moving contact action, a speed when the moving contact displacement signal changes, a moving contact maximum speed and a moving contact maximum speed time are obtained to obtain a speed time series feature vector; a variational mode decomposition process is performed on the acceleration signal to obtain an acceleration time series feature vector; and the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector are combined together as a global feature vector.

[0081] The opening time period T can be obtained according to the opening and closing state of the circuit breaker under fault condition. 12 And closing period T 23 , where T 12 =T2-T1, T 23 =T3-T2, the opening movement is at time T1, the opening latch is unlocked, the opening and closing coil attracts the iron core to drive the tripping device to release the constraint, and then the spring begins to release energy to drive the crank arm to move; the closing movement is at time T2, the moving contact and the static contact are in contact, the contact position change signal changes, and the action ends at time T3.

[0082] According to the travel signal and speed signal of the circuit breaker, the moving contact position change signal change time T2, the moving contact maximum change travel S1 and the moving contact maximum speed travel S2 can be obtained; the opening time period T 12 , Closing period T 23 , the moving contact position change signal change time T2, the moving contact maximum change stroke S1 and the moving contact maximum speed stroke S2 ​​are used as the stroke timing feature vector to obtain the five-dimensional stroke timing feature vector v1=[T 12 , T 23 , T2, S1, S2].

[0083] According to the speed signal of the circuit breaker, the average speed V1 of the moving contact, the speed V2 when the moving contact position change signal changes, the maximum speed V3 of the moving contact and the time T4 when the moving contact reaches the maximum speed can be obtained; the average speed V1 of the moving contact, the speed V2 when the moving contact position change signal changes, the maximum speed V3 of the moving contact and the time T4 when the moving contact reaches the maximum speed can be used as the speed time series feature vector to obtain a four-dimensional speed time series feature vector v2=[V1, V2, V3, T4].

[0084] The acceleration signal can be subjected to variational mode decomposition (VMD) processing to obtain several intrinsic mode function components; based on the several intrinsic mode function components, the modal energy entropy corresponding to the several intrinsic mode function components can be obtained as the acceleration time series feature vector. Among them, the moving contact movement direction can be used as the direction of the acceleration signal, and the moving contact movement direction can include the opening direction and the closing direction.

[0085] The first time window feature vector, the second time window feature vector and the global feature vector may be fused. For example, the first time window feature vector, the second time window feature vector and the global feature vector may be merged to obtain a higher-dimensional fused time window feature vector.

[0086] In this embodiment, the opening period and the closing period are obtained according to the opening and closing states of the circuit breaker; the travel time series feature vector and the speed time series feature vector are obtained according to the travel signal and the speed signal of the circuit breaker; the acceleration signal is subjected to variational modal decomposition processing to obtain the acceleration time series feature vector; the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector are combined together as a global feature vector to identify the circuit breaker fault type.

[0087] In one of the embodiments, the acceleration signal is subjected to variational mode decomposition processing to obtain an acceleration time series characteristic vector, and the specific steps are as follows: the acceleration signal is subjected to Hilbert transform to decompose and obtain analytical signals and one-sided spectra of several initial components of the intrinsic mode function; the initial components of the intrinsic mode function are iteratively updated, and when the change in the iterative result is less than a set threshold, the iteration is stopped to obtain the intrinsic mode function components to obtain the acceleration time series characteristic vector.

[0088] The acceleration signal can be Hilbert transformed to decompose and obtain analytical signals and one-sided spectra of k initial components of the intrinsic mode function; the k initial components of the intrinsic mode function can be iteratively updated, and when the change of the iterative result is less than a set threshold, the iteration is stopped to obtain k intrinsic mode function (IMF) components; the above k intrinsic mode function components can be used as acceleration time series feature vectors.

[0089] Each natural mode function component at the center frequency The mathematical expression of the variational model during decomposition is shown in formula (2).

[0090] (2)

[0091] in, represents the analytical signal of the decomposed intrinsic mode function components, represents the unit pulse function, and t represents time.

[0092] The augmented Lagrangian method (ALM) can be introduced to find the optimal solution for the variational model. The mathematical expression is shown in formula (3).

[0093] (3)

[0094] in, represents the penalty factor, and uses the alternating multiplier direction algorithm to search for the saddle point of the problem and update , ,right Perform Fourier transform update to get , as shown in equations (4) to (6).

[0095] (4)

[0096] (5)

[0097] (6)

[0098] By performing Hilbert-Huang transform on each intrinsic mode function component, the Hilbert spectrum can be obtained, and the Hilbert transform marginal spectrum can be obtained by performing integral operation, and the energy value is normalized to obtain the normalized energy value of each intrinsic mode function component. .

[0099] The mathematical expression of the Hilbert marginal spectrum energy entropy value of the intrinsic mode function component is shown in formula (7).

[0100] (7)

[0101] The acceleration time series eigenvector can be obtained according to the Hilbert marginal spectrum energy entropy value of each intrinsic mode function component.

[0102] In this embodiment, the acceleration signal is subjected to Hilbert transform and decomposed to obtain analytical signals and one-sided spectra of several initial components of the intrinsic mode function; the initial components of the intrinsic mode function are iteratively updated, and when the change in the iterative result is less than a set threshold, the iteration is stopped to obtain the intrinsic mode function components, so as to obtain an acceleration time series feature vector with relatively high accuracy.

[0103] In one of the embodiments, a time series window processing is performed on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector. The specific steps are as follows: according to the travel time series feature vector and the speed time series feature vector, a first time domain feature and a second time domain feature are obtained; according to the acceleration time series feature vector, a first acceleration time-frequency domain feature and a second acceleration time-frequency domain feature are obtained; according to the first time domain feature and the first acceleration time-frequency domain feature, a first time series window feature vector is obtained; according to the second time domain feature and the second acceleration time-frequency domain feature, a second time series window feature vector is obtained.

[0104] When extracting the travel time series feature vector and the speed time series feature vector, the window feature extraction is performed according to the opening and closing state of the circuit breaker. Therefore, the opening period T in the travel time series feature vector can be 12 and the moving contact position change signal change time T2, and the speed V2 of the moving contact position change signal change in the speed time series feature vector as the first time domain feature; the closing period T in the travel time series feature vector can be 23 , the maximum change stroke of the moving contact S1 and the stroke when the moving contact speed is maximum S 2， And the time T4 when the moving contact has the maximum speed and the average speed V1 of the moving contact in the speed time series feature vector are used as the second time domain features.

[0105] When extracting the acceleration time series feature vector, the window feature extraction is performed according to the opening and closing states of the circuit breaker, in which 6 inherent mode function components are obtained for each action time series, and each acceleration time series feature vector is a 12-dimensional feature vector.

[0106] Therefore, the six inherent modal function components IMF1, IMF2, IMF3, IMF4, IMF5 and IMF6 corresponding to the opening motion action timing in the acceleration timing characteristic vector can be used as the first acceleration time-frequency domain features, and the six inherent modal function components IMF7, IMF8, IMF9, IMF10, IMF11 and IMF12 corresponding to the closing motion action timing in the acceleration timing characteristic vector can be used as the second acceleration time-frequency domain features.

[0107] The first time-series window feature vector [T 12 , T2, V2, IMF1, IMF2, IMF3, MF4, IMF5, IMF6]; According to the second time domain characteristics and the second acceleration time-frequency domain characteristics, the second time series window feature vector [T 23 , T4, S1, S2, V1, IMF7, IMF8, IMF9, IMF10, IMF11, IMF12].

[0108] In this embodiment, the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector are subjected to time series window processing, so that the signal characteristics of different action intervals can be better obtained, and the local information is taken into consideration, thereby further improving the accuracy of circuit breaker fault type identification.

[0109] In one of the embodiments, before obtaining the fault type of the circuit breaker based on the fused timing feature vector and a pre-built fault type identification model, the method provided by the present application also includes: obtaining a model to be trained based on a nature-inspired rule optimization algorithm and a support vector machine model; training and optimizing the model to be trained based on the fused timing feature vector sample set, stopping the training until the prediction accuracy reaches convergence, and obtaining a fault type identification model.

[0110] The model to be trained can be obtained according to the nature-inspired rule optimization algorithm and the support vector machine model.

[0111] Specifically, the optimal parameters of the kernel function of the support vector machine model can be found according to the NRBO (Nature-inspired Rule-based Optimization) algorithm. The specific steps are as follows:

[0112] The improved NRBO algorithm starts the search for the optimal solution by generating an initial random population within the boundary of candidate solutions. For a population of a given size Np, if the dimension of the optimization parameter is dim, the mathematical expression of the position of the nth individual in the jth dimension is shown in formula (8).

[0113] (8)

[0114] Among them, lb j andub j are the lower and upper bounds of the j-th dimension, respectively, and rand is a random number between 0 and 1;

[0115] According to the set fitness function, calculate the fitness value of each individual, and select the best fitness value and its corresponding position xb and the worst fitness value and its corresponding position xw;

[0116] For the nth individual in the tth iteration, the NRSR algorithm explores the individual position, and the mathematical expression of the new position is shown in formula (9).

[0117] (9)

[0118] Where t is the number of iterations; r1 and r2 represent random numbers between (0, 1), , ,

[0119] In order to improve the local search performance and global search performance, the three positions updated by the current position are respectively improved by the NRBO algorithm. The new position obtained by the NRSR algorithm is combined with the step size factor Rho to update the individual position. The mathematical expressions are shown in Equations (10) to (12).

[0120] (10)

[0121] (11)

[0122] (12)

[0123] Among them, Nr is the value calculated by the NRSR rule, It is an adaptive coefficient that changes with the number of iterations and is used to avoid local optimality and reduce the amount of calculation.

[0124] To avoid the local optimal trap, TAO is introduced to combine the optimal position xb and the vector position x obtained by NRSR. n t+1 To generate solutions with enhanced quality x t TAO, as shown in formula (13).

[0125] (13)

[0126] in, and are random numbers between [-1, 1] and [-0.5, 0.5], and It is a random number generated according to specific rules.

[0127] A fused time series feature vector sample set can be obtained based on the signal fault samples of the circuit breaker.

[0128] The signal fault samples include actual signal fault samples and simulated signal fault samples. Circuit breaker fault types may include spring relaxation, component wear, buffer oil leakage, lubrication failure, load increase, and dust accumulation in the operating mechanism. Bearing wear, lubrication failure, ejector pin emptying, spring relaxation, reduced number of turns, and increased number of turns can be simulated, and the simulated displacement signal of the moving contact can be obtained from the simulation software to obtain the travel signal sample, speed signal sample, and acceleration signal sample of the circuit breaker as simulated signal fault samples. The travel signal, speed signal, and acceleration signal of the circuit breaker in the fault state of bearing wear, lubrication failure, ejector pin emptying, spring relaxation, reduced number of turns, and increased number of turns can be collected as actual signal fault samples.

[0129] After obtaining the fused time series feature vector sample set, the fused time series feature vector sample set is randomly divided into a training set and a test set, where the number relationship between the training set and the test set is 2:1.

[0130] According to the training set, the model to be trained is optimized to obtain the optimal parameters of the model to be trained; according to the test set, the performance of the model to be trained is evaluated, and the training is stopped when the prediction accuracy of the model to be trained reaches convergence. The model to be trained at this time is used as a fault type identification model.

[0131] The model to be trained uses the predicted fault type as output, the actual fault type as the prediction target, and minimizing the sum of the prediction accuracies of all predicted fault types as the training target, and the training is stopped when the sum of the prediction accuracies reaches convergence.

[0132] In order to better understand the above method, an application example of the circuit breaker fault type identification method based on multi-dimensional signal characteristics of the present application is described in detail below. Figure 4 shown.

[0133] Circuit breakers, such as high-voltage circuit breakers, are one of the important components in the power system. They need to work stably and reliably when the power system is normal or faulty to ensure the reliability and stability of the power system. According to relevant surveys, among the failures of high-voltage circuit breakers, operating mechanism failures account for 61%, of which 83% are caused by mechanical failures. Mechanical failures are mainly spring relaxation, mechanism jamming and other failure types. Accurate diagnosis of high-voltage circuit breaker mechanical failures is a necessary means to improve the operation and maintenance level of the power system and ensure the safe and stable operation of the power system.

[0134] Traditional circuit breaker fault type identification methods (also known as circuit breaker fault diagnosis methods) are usually based on the feature analysis of single signals such as vibration signals or opening and closing coil current signals. However, the features of a single signal have a certain bias, ignoring the association of other signals such as vibration and travel, reducing the accuracy of fault type identification, and having a negative impact on operation and maintenance decisions. For example, during the operation of the circuit breaker, the vibration signal may have a large error due to environmental noise or sensor position, or the electromagnet current signal may be affected by voltage fluctuations, resulting in misdiagnosis. During the operation of high-voltage circuit breakers, signal detection methods such as travel, speed and acceleration are becoming more and more mature. The traditional monitoring method is to install travel sensors, acceleration sensors and grating encoders on high-voltage circuit breakers to provide a rich data basis for fault type identification. Through the coordinated diagnosis of multiple signals, the lack of features caused by the temporary absence of a certain signal during the operation process can be compensated. Collecting the opening and closing travel signal and current signal, and fusing and reducing the dimensionality of the two types of signal features, can achieve a better fault type identification effect in the case of small samples. The joint composite analysis of the sound signal and vibration features effectively improves the accuracy of fault type identification. Therefore, through collaborative diagnosis of multiple signals, not only can the comprehensiveness of fault extraction be improved, but also the risk of misdiagnosis caused by failure or interference of a single signal can be effectively reduced, thereby showing higher accuracy and stability in fault type identification.

[0135] On the other hand, the action process of the operating mechanism involves the coordinated action of more than 10 mechanical parts. Each action stage corresponds to different components, so it is necessary to pay attention to the timing signals in each action stage. The existing method extracts features from the vibration signal of the entire action sequence, which improves the feature extraction accuracy and optimizes the fault type identification results, but fails to fully consider the specific action characteristics of each timing interval.

[0136] In order to solve the above technical problems, this embodiment provides a circuit breaker fault type identification method based on multi-dimensional signal characteristics, the method comprising:

[0137] Step S1, obtaining signal fault samples of a travel signal, a speed signal and an acceleration signal in a circuit breaker fault state.

[0138] Step S2, according to the signal fault samples of the travel signal, speed signal and acceleration signal in the circuit breaker fault state, obtain the travel time series feature vector samples, speed time series feature vector samples, acceleration time series feature vector samples and global feature vector samples. Perform time series window processing on the travel time series feature vector samples, speed time series feature vector samples and acceleration time series feature vector samples to obtain the first time series window feature vector samples and the second time series window feature vector samples.

[0139] Step S4, fusing the first time series window feature vector samples, the second time series window feature vector samples and the global feature vector samples to obtain a fused time series feature vector sample set.

[0140] Step S5, obtaining a model to be trained based on the nature-inspired rule optimization algorithm and the support vector machine model; training and optimizing the model to be trained based on the fused time series feature vector sample set until the prediction accuracy reaches convergence and the training is stopped to obtain a fault type recognition model.

[0141] Step S6, obtaining a travel signal, a speed signal and an acceleration signal in a circuit breaker fault state.

[0142] Step S7, obtaining a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector according to the travel signal, the speed signal and the acceleration signal in the circuit breaker fault state.

[0143] Step S8, performing time series window processing on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector.

[0144] Step S9, fusing the first time series window feature vector, the second time series window feature vector and the global feature vector to obtain a fused time series feature vector.

[0145] Step S10, obtaining the fault type of the circuit breaker according to the fused time series feature vector and the pre-built fault type identification model.

[0146] The signal fault samples include actual signal fault samples and simulated signal fault samples. Circuit breaker fault types may include spring relaxation, component wear, buffer oil leakage, lubrication failure, load increase, and dust accumulation in the operating mechanism. Bearing wear, lubrication failure, ejector pin emptying, spring relaxation, reduced number of turns, and increased number of turns can be simulated, and the simulated displacement signal of the moving contact can be obtained from the simulation software to obtain the travel signal sample, speed signal sample, and acceleration signal sample of the circuit breaker as simulated signal fault samples. The travel signal, speed signal, and acceleration signal of the circuit breaker in the fault state of bearing wear, lubrication failure, ejector pin emptying, spring relaxation, reduced number of turns, and increased number of turns can be collected as actual signal fault samples.

[0147] In step S1, the steps of obtaining signal fault samples are as follows:

[0148] The action process of the moving contact of the circuit breaker in the opening and closing states under the fault state is captured by a high-speed camera to obtain the original action video; the original action video is decomposed into single-frame action images, each frame of the action image is grayed and edge monitored to obtain the actual displacement signal of the moving contact of the circuit breaker; the simulated displacement signal of the moving contact is obtained from the simulation software.

[0149] According to the displacement signal and time series of the moving contact in the actual displacement signal and the simulated displacement signal, the travel curve of the moving contact is drawn; the displacement node signal and time node signal of the moving contact are marked on the travel curve; the displacement node signal includes the opening position and the closing position; the time node signal includes the breaking time point and the closing time point; according to the opening position and the closing position, the travel signal sample of the circuit breaker is obtained; according to the opening position, the closing position, the breaking time point, the closing time point and the speed calculation formula, the speed signal sample of the circuit breaker is obtained; according to the opening position, the closing position, the breaking time point, the closing time point and the acceleration calculation formula, the acceleration signal sample of the circuit breaker is obtained. According to the travel signal sample, speed signal sample and acceleration signal sample of the circuit breaker, the signal fault sample is obtained.

[0150] During the opening movement at time T1, the opening latch is unlocked, the opening and closing coil attracts the iron core to drive the tripping device to release the constraint, and then the spring begins to release energy to drive the crank arm to move; during the closing movement at time T2, the moving contact and the static contact are in contact, the contact position change signal changes, and the action ends at time T3.

[0151] The steps to obtain the travel time series feature vector are as follows:

[0152] The opening time period (first time period) T can be obtained according to the opening and closing state (action sequence) of the circuit breaker under fault condition. 12 And closing period (second period) T 23, where T 12 =T2-T1, T 23 =T3-T2, extract the moving contact position change signal change time T2, the moving contact maximum change stroke S1 and the moving contact maximum speed stroke S2; the opening period T 12 , Closing period T 23 , the moving contact position change signal change time T2, the moving contact maximum change stroke S1 and the moving contact maximum speed stroke S2 ​​are used as the stroke timing feature vector to obtain the five-dimensional stroke timing feature vector v1=[T 12 , T 23 , T2, S1, S2].

[0153] In step S7, the steps of obtaining the speed time series feature vector are as follows:

[0154] According to the speed signal of the circuit breaker, the average speed V1 of the moving contact, the speed V2 when the moving contact position change signal changes, the maximum speed V3 of the moving contact and the time T4 when the moving contact reaches the maximum speed can be obtained; the average speed V1 of the moving contact, the speed V2 when the moving contact position change signal changes, the maximum speed V3 of the moving contact and the time T4 when the moving contact reaches the maximum speed can be used as the speed time series feature vector to obtain a four-dimensional speed time series feature vector v2=[V1, V2, V3, T4].

[0155] The steps to obtain the acceleration time series feature vector are as follows:

[0156] The moving contact movement direction is taken as the direction of the acceleration signal, and the moving contact movement direction includes the opening direction and the closing direction;

[0157] The acceleration signal collected in the opening direction or closing direction is processed by variational mode decomposition (VMD) and the energy entropy of each mode is obtained to obtain the acceleration time series feature vector.

[0158] The action time of the opening movement can be expressed as (T1-T2), and the action time of the closing movement can be expressed as (T2-T3) to achieve timing window processing.

[0159] The travel time series feature vector, velocity time series feature vector and acceleration time series feature vector of (T1-T2) and (T2-T3) can be combined together as a complete feature vector of the high-voltage circuit breaker, which is called a global feature vector.

[0160] The travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector may be subjected to time series window processing to obtain a first time series window feature vector and a second time series window feature vector. The specific steps are as follows:

[0161] The circuit breaker action is divided into two first timing windows and second timing windows according to the time sequence.

[0162] The first time series window includes a first time domain feature and a first acceleration time-frequency domain feature. The first time series window feature vector [T 12 , T2, V2, IMF1, IMF2, IMF3, MF4, IMF5, IMF6];

[0163] The second time series window includes a second time domain feature and a second acceleration time-frequency domain feature. The second time series window feature vector [T 23 , T4, S1, S2, V1, IMF7, IMF8, IMF9, IMF10, IMF11, IMF12].

[0164] The specific steps of performing variational mode decomposition (VMD) processing on the acceleration signal collected in the opening direction or closing direction and obtaining the energy entropy of each mode to obtain the acceleration time series feature vector are as follows:

[0165] The acceleration signal can be Hilbert transformed to decompose and obtain analytical signals and one-sided spectra of k initial components of the intrinsic mode function; the k initial components of the intrinsic mode function can be iteratively updated, and when the change of the iterative result is less than a set threshold, the iteration is stopped to obtain k intrinsic mode function (IMF) components; the above k intrinsic mode function components can be used as acceleration time series feature vectors.

[0166] Each natural mode function component at the center frequency The mathematical expression of the variational model during decomposition is shown in formula (2). (2)

[0168] in, represents the analytical signal of the decomposed intrinsic mode function components, represents the unit pulse function, and t represents time.

[0169] The augmented Lagrangian method (ALM) can be introduced to find the optimal solution for the variational model. The mathematical expression is shown in formula (3).

[0170] (3)

[0171] in, represents the penalty factor, and uses the alternating multiplier direction algorithm to search for the saddle point of the problem and update , ,right Perform Fourier transform update to get , as shown in equations (4) to (6).

[0172] (4)

[0173] (5)

[0174] (6)

[0175] By performing Hilbert-Huang transform on each intrinsic mode function component, the Hilbert spectrum can be obtained, and the Hilbert transform marginal spectrum can be obtained by performing integral operation, and the energy value is normalized to obtain the normalized energy value of each intrinsic mode function component. .

[0176] The mathematical expression of the Hilbert marginal spectrum energy entropy value of the intrinsic mode function component is shown in formula (7).

[0177] (7)

[0178] The acceleration time series eigenvector can be obtained according to the Hilbert marginal spectrum energy entropy value of each intrinsic mode function component.

[0179] In step S9, the first time series window feature vector, the second time series window feature vector and the global feature vector are fused to obtain a fused time series feature vector. The specific steps are as follows:

[0180] The feature vector set composed of the feature vectors of the first time series window can be recorded as A. Specifically, the feature vectors of the first time series window are ; The feature vector set composed of the feature vectors of the second timing window can be recorded as B. Specifically, the feature vectors of the second timing window For time domain features, the global feature vector is V1, which is independent of the window cutting position and is recorded as .

[0181] According to the vector space merging strategy, the first time window feature vectors, the second time window feature vectors and the global feature vectors from multiple signal sources can be fused at the same level to obtain a fused time series feature vector as shown in formula (1).

[0182] (1)

[0183] Among them, R represents the fused time series feature vector set, which is 20 dimensions.

[0184] In step S5, a model to be trained is obtained according to the optimization algorithm based on the natural inspiration rule and the support vector machine model; the model to be trained is trained and optimized according to the fused time series feature vector sample set until the prediction accuracy reaches convergence and the training is stopped to obtain a fault type recognition model. The specific steps are as follows:

[0185] The model to be trained can be obtained according to the nature-inspired rule optimization algorithm and the support vector machine model.

[0186] Specifically, the optimal parameters of the kernel function of the support vector machine model can be found according to the NRBO (Nature-inspired Rule-based Optimization) algorithm. The specific steps are as follows:

[0187] The improved NRBO algorithm starts the search for the optimal solution by generating an initial random population within the boundary of candidate solutions. For a population of a given size Np, if the dimension of the optimization parameter is dim, the mathematical expression of the position of the nth individual in the jth dimension is shown in formula (8).

[0188] (8)

[0189] Among them, lb j andub j are the lower and upper bounds of the j-th dimension, respectively, and rand is a random number between 0 and 1;

[0190] According to the set fitness function, calculate the fitness value of each individual, and select the best fitness value and its corresponding position xb and the worst fitness value and its corresponding position xw;

[0191] For the nth individual in the tth iteration, the NRSR algorithm explores the individual position, and the mathematical expression of the new position is shown in formula (9).

[0192] (9)

[0193] Where t is the number of iterations; r1 and r2 represent random numbers between (0, 1), ,

[0194] , In order to improve the local search performance and global search performance, the three positions updated by the current position are respectively improved by the NRBO algorithm. The new position obtained by the NRSR algorithm is combined with the step size factor Rho to update the individual position. The mathematical expressions are shown in Equations (10) to (12).

[0195] (10)

[0196] (11)

[0197] (12)

[0198] Among them, Nr is the value calculated by the NRSR rule, It is an adaptive coefficient that changes with the number of iterations and is used to avoid local optimality and reduce the amount of calculation.

[0199] To avoid the local optimal trap, TAO is introduced to combine the optimal position xb and the vector position x obtained by NRSR. n t+1 To generate solutions with enhanced quality x t TAO , as shown in formula (13).

[0200] (13)

[0201] in, and are random numbers between [-1, 1] and [-0.5, 0.5], and It is a random number generated according to specific rules.

[0202] After obtaining the fused time series feature vector sample set, the fused time series feature vector sample set is randomly divided into a training set and a test set, where the number relationship between the training set and the test set is 2:1.

[0203] According to the training set, the model to be trained is optimized to obtain the optimal parameters of the model to be trained; according to the test set, the performance of the model to be trained is evaluated, and the training is stopped when the prediction accuracy of the model to be trained reaches convergence. The model to be trained at this time is used as a fault type identification model.

[0204] The model to be trained uses the predicted fault type as output, the actual fault type as the prediction target, and minimizing the sum of the prediction accuracies of all predicted fault types as the training target, and the training is stopped when the sum of the prediction accuracies reaches convergence.

[0205] The beneficial effects of the circuit breaker fault type identification method based on multi-dimensional signal features provided in this embodiment are as follows:

[0206] 1) Compared with the traditional circuit breaker fault type identification method which uses a single signal, the circuit breaker fault type identification method provided in this embodiment uses the three signals of travel, speed and acceleration generated by the circuit breaker action to fuse the signal features, thereby avoiding the phenomenon of loss of other signal features caused by the use of a single signal, and can improve the accuracy of circuit breaker fault type identification, which has a good effect on comprehensively reflecting circuit breaker faults.

[0207] 2) Based on the circuit breaker action sequence, the signal is divided into windows for feature extraction, which can better obtain the signal characteristics of different action intervals and obtain local information, so as to better and more completely represent the fault signal.

[0208] 3) The support vector machine model is optimized using the NRBO optimization algorithm, which effectively improves the recognition accuracy of the fault type recognition model. At the same time, the fault type recognition model is not affected by the fault sample size and data dispersion, and has good application value in circuit breaker fault type recognition.

[0209] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0210] Based on the same inventive concept, the embodiment of the present application also provides a circuit breaker fault type identification device based on multi-dimensional signal features for implementing the circuit breaker fault type identification method based on multi-dimensional signal features involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the circuit breaker fault type identification device based on multi-dimensional signal features provided below can refer to the limitations of the circuit breaker fault type identification method based on multi-dimensional signal features above, and will not be repeated here.

[0211] In an exemplary embodiment, Figure 5 As shown, a circuit breaker fault type identification device based on multi-dimensional signal characteristics is provided, wherein:

[0212] The signal acquisition module 501 is used to acquire the travel signal, speed signal and acceleration signal of the circuit breaker in a fault state;

[0213] A feature vector acquisition module 502 is used to obtain a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector according to the travel signal, speed signal and acceleration signal in the circuit breaker fault state;

[0214] A time series window processing module 503 is used to perform time series window processing on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector;

[0215] A feature vector fusion module 504 is used to fuse the first time series window feature vector, the second time series window feature vector and the global feature vector to obtain a fused time series feature vector;

[0216] The fault type identification module 505 is used to obtain the fault type of the circuit breaker according to the fused time series feature vector and the pre-built fault type identification model.

[0217] In one embodiment, the signal acquisition module 501 is further used to: collect the action process of the moving contact when the circuit breaker is in the opening and closing state under the fault state to obtain the original action video; decompose the original action video into a single-frame action image, grayscale each frame of the action image, and perform edge monitoring to obtain the displacement signal of the moving contact of the circuit breaker; draw a travel curve diagram of the moving contact according to the displacement signal and time series of the moving contact; mark the displacement node signal and time node signal of the moving contact on the travel curve diagram; the displacement node signal includes the opening position and the closing position; the time node signal includes the breaking time point and the closing time point; according to the opening position and the closing position, obtain the travel signal of the circuit breaker; according to the opening position, the closing position, the breaking time point, the closing time point and the speed calculation formula, obtain the speed signal of the circuit breaker; according to the opening position, the closing position, the breaking time point, the closing time point and the acceleration calculation formula, obtain the acceleration signal of the circuit breaker.

[0218] In one embodiment, the feature vector acquisition module 502 is further used to: obtain an opening period and a closing period according to the opening and closing states of the circuit breaker under a fault state; obtain a moving contact position change signal change time, a moving contact maximum change stroke, and a moving contact maximum speed stroke according to the travel signal and speed signal of the circuit breaker; obtain a travel time series feature vector according to the opening period, the closing period, the moving contact position change signal change time, the moving contact maximum change stroke, and the moving contact maximum speed stroke; obtain an average speed of the moving contact action, a speed when the moving contact position change signal changes, a maximum speed of the moving contact, and a time when the moving contact has the maximum speed according to the speed signal of the circuit breaker to obtain a speed time series feature vector; perform variational mode decomposition processing on the acceleration signal to obtain an acceleration time series feature vector; and combine the travel time series feature vector, the speed time series feature vector, and the acceleration time series feature vector as a global feature vector.

[0219] In one of the embodiments, the feature vector acquisition module 502 is further used to: perform Hilbert transform on the acceleration signal to decompose and obtain analytical signals and one-sided spectra of several initial components of the intrinsic mode function; iteratively update the initial components of the intrinsic mode function, and when the change of the iterative result is less than a set threshold, stop the iteration to obtain the intrinsic mode function components to obtain the acceleration time series feature vector.

[0220] In one embodiment, the time series window processing module 503 is also used to: obtain a first time domain feature and a second time domain feature according to the travel time series feature vector and the speed time series feature vector; obtain a first acceleration time-frequency domain feature and a second acceleration time-frequency domain feature according to the acceleration time series feature vector; obtain a first time series window feature vector according to the first time domain feature and the first acceleration time-frequency domain feature; obtain a second time series window feature vector according to the second time domain feature and the second acceleration time-frequency domain feature.

[0221] In one of the embodiments, the device also includes a model training module, which is used to: obtain a model to be trained based on a nature-inspired rule optimization algorithm and a support vector machine model; train and optimize the model to be trained based on a fused time series feature vector sample set until the prediction accuracy reaches convergence, and stop the training to obtain a fault type identification model.

[0222] Each module in the above-mentioned circuit breaker fault type identification device based on multi-dimensional signal characteristics can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0223] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data of an embodiment of a circuit breaker fault type identification method based on multi-dimensional signal characteristics. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a circuit breaker fault type identification method based on multi-dimensional signal characteristics is implemented.

[0224] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0225] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0226] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0227] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0228] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0229] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0230] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0231] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A circuit breaker fault type identification method based on multi-dimensional signal characteristics, characterized in that: The method comprises: Acquire travel signal, speed signal and acceleration signal of the circuit breaker in fault state; According to the travel signal, speed signal and acceleration signal of the circuit breaker in a fault state, a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector are obtained; Performing time series window processing on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector; Fusing the first time series window feature vector, the second time series window feature vector and the global feature vector to obtain a fused time series feature vector; The fault type of the circuit breaker is obtained according to the fused time series feature vector and the pre-built fault type identification model.

2. The method according to claim 1, characterized in that The step of obtaining a travel signal, a speed signal and an acceleration signal of the circuit breaker in a fault state includes: Collect the action process of the moving contact when the circuit breaker is in the opening and closing state under the fault state, and obtain the original action video; Decomposing the original action video into single-frame action images, graying each frame of the action image, and performing edge monitoring to obtain a displacement signal of a moving contact of the circuit breaker; Drawing a travel curve diagram of the moving contact according to the displacement signal and time series of the moving contact; Mark the displacement node signal and time node signal of the moving contact on the travel curve diagram; the displacement node signal includes the opening position and the closing position; the time node signal includes the breaking time point and the closing time point; Obtaining a travel signal of the circuit breaker according to the opening position and the closing position; According to the opening position, closing position, breaking time point, closing time point and speed calculation formula, the speed signal of the circuit breaker is obtained; According to the opening position, closing position, breaking time point, closing time point and acceleration calculation formula, the acceleration signal of the circuit breaker is obtained.

3. The method according to claim 1, characterized in that The method of obtaining a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector according to the travel signal, speed signal and acceleration signal in the circuit breaker fault state comprises: According to the opening and closing states of the circuit breaker under the fault state, the opening time period and the closing time period are obtained; According to the travel signal and speed signal of the circuit breaker, the moving contact position change signal change time, the moving contact maximum change travel and the moving contact maximum speed travel are obtained; According to the opening time period, closing time period, moving contact position change signal change time, moving contact maximum change stroke and moving contact maximum speed stroke, the stroke time sequence characteristic vector is obtained; According to the speed signal of the circuit breaker, the average speed of the moving contact action, the speed when the moving contact displacement signal changes, the maximum speed of the moving contact and the time when the moving contact reaches the maximum speed are obtained to obtain a speed time series characteristic vector; Performing variational mode decomposition processing on the acceleration signal to obtain an acceleration time series eigenvector; The travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector are combined together to form a global feature vector.

4. The method according to claim 3, characterized in that The performing variational mode decomposition processing on the acceleration signal to obtain the acceleration time series feature vector includes: Performing Hilbert transformation on the acceleration signal to decompose and obtain analytical signals and unilateral spectra of initial components of several intrinsic mode functions; The initial component of the intrinsic mode function is iteratively updated, and when the change of the iterative result is less than a set threshold, the iteration is stopped to obtain the intrinsic mode function component to obtain the acceleration time series characteristic vector.

5. The method according to claim 1, characterized in that The performing time series window processing on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector comprises: Obtaining a first time domain feature and a second time domain feature according to the travel time series feature vector and the speed time series feature vector; According to the acceleration time series feature vector, a first acceleration time-frequency domain feature and a second acceleration time-frequency domain feature are obtained; Obtaining a first time series window feature vector according to the first time domain feature and the first acceleration time-frequency domain feature; A second time series window feature vector is obtained according to the second time domain feature and the second acceleration time-frequency domain feature.

6. The method according to claim 1, characterized in that Before obtaining the fault type of the circuit breaker according to the fused time series feature vector and the pre-built fault type identification model, the method further includes: According to the nature-inspired rule optimization algorithm and the support vector machine model, the model to be trained is obtained; According to the fused time series feature vector sample set, the model to be trained is trained and optimized until the prediction accuracy reaches convergence and the training is stopped to obtain the fault type recognition model.

7. A circuit breaker fault type identification device based on multi-dimensional signal characteristics, characterized in that: The device comprises: A signal acquisition module, used to acquire a travel signal, a speed signal and an acceleration signal of the circuit breaker in a fault state; A feature vector acquisition module, used to obtain a travel time series feature vector, a speed time series feature vector, an acceleration time series feature vector and a global feature vector according to a travel signal, a speed signal and an acceleration signal in a fault state of the circuit breaker; A time series window processing module, used for performing time series window processing on the travel time series feature vector, the speed time series feature vector and the acceleration time series feature vector to obtain a first time series window feature vector and a second time series window feature vector; A feature vector fusion module, used for fusing the first time series window feature vector, the second time series window feature vector and the global feature vector to obtain a fused time series feature vector; The fault type identification module is used to obtain the fault type of the circuit breaker according to the fused time series feature vector and the pre-built fault type identification model.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images