A high-voltage circuit breaker mechanical fault diagnosis method and system based on MIDBO-SVM model

Abstract

The present invention discloses a method and system for diagnosing mechanical faults of high-voltage circuit breakers based on a MIDBO-SVM model. The method first obtains a vibration signal of the circuit breaker, and then uses the obtained vibration signal in combination with a pre-built MIDBO-SVM model to diagnose the mechanical faults of the circuit breaker. During the construction of the MIDBO-SVM model, a good point set strategy and a circle chaos map are used to jointly generate an initial population method to improve the uniform ergodicity of the population. A sine-cosine search strategy and an adaptive weight factor are introduced during the individual position update stage of a rolling ball dung beetle to cause the solution to attenuate and oscillate until it approaches the global optimal solution. A Cauchy-Gauss mutation strategy is used to mutate the individual with the best current fitness in a manner away from the optimal solution, thereby expanding the spatial search range. The method also solves the problems of slow search speed and susceptibility to falling into local optimality in the later stages of traditional algorithms, thereby improving the accuracy of circuit breaker fault classification.

Description

Technical Field

[0001] The present invention belongs to the technical field of high-voltage circuit breaker mechanical fault diagnosis, and relates to a high-voltage circuit breaker mechanical fault diagnosis method and system based on a MIDBO-SVM model. Background Art

[0002] As a core component of power transmission and distribution systems, the stable and reliable operation of high-voltage circuit breakers is directly related to the safety and stability of the power grid. In complex and volatile power environments, circuit breakers must not only quickly and accurately perform opening and closing operations to isolate fault zones, but also maintain long-term stable operation. However, over time, the operating mechanisms and control circuits of circuit breakers often experience frequent failures such as refusal to open or close due to issues such as spring fatigue, mechanical component wear, and aging electrical components (such as core sticking and failed opening and closing coils). If these failures are not promptly identified and addressed, they pose a serious threat to the safety and stability of the power grid. To address this issue, vibration signal analysis, as a non-invasive, real-time monitoring method, is demonstrating significant potential for high-voltage circuit breaker mechanical condition monitoring and fault diagnosis. Vibration signals contain a wealth of information about the circuit breaker's operation, such as operating time, signal amplitude variations, and frequency components. These characteristics can directly reflect the circuit breaker's internal mechanical state and potential fault modes. Therefore, by accurately extracting and analyzing the characteristic parameters in the vibration signal, timely early warning and accurate diagnosis of circuit breaker mechanical failures can be achieved, providing strong support for the safe operation and maintenance of the power grid.

[0003] The rapid development of artificial intelligence (AI), particularly the mature application of machine learning algorithms, has revolutionized high-voltage circuit breaker fault diagnosis. Neural networks and support vector machines (SVMs), two major mainstream technologies, each exhibit unique advantages and limitations. Neural networks, with their powerful self-learning and nonlinear mapping capabilities, offer significant advantages in handling complex and ever-changing system models. However, their complex training process, susceptibility to local optima, and large sample data requirements limit their widespread practical application. In contrast, SVMs, with their excellent adaptability to small sample sizes, efficient processing of high-dimensional feature spaces, and strong global generalization capabilities, have become another important tool in high-voltage circuit breaker fault diagnosis. However, the performance of SVMs is highly dependent on their parameter settings, particularly the choice of the penalty factor C and kernel parameter g, which directly impact the model's classification accuracy and generalization ability. Therefore, efficiently and accurately optimizing SVM parameters is crucial for improving diagnostic effectiveness.

[0004] To address the challenge of SVM parameter optimization, the Dung Beetle Optimization (DBO) algorithm, a novel heuristic optimization algorithm, has been introduced into the field of circuit breaker fault diagnosis due to its strong search capabilities and rapid convergence. It is used to automatically search for and optimize SVM parameters. The DBO algorithm simulates the natural behavior of dung beetles rolling dung balls in search of food sources, gradually approaching the global optimal solution by continuously iteratively updating the population's position. However, in practical applications, the DBO algorithm also faces challenges such as decreased population diversity and a tendency to fall into local optimality. This is particularly true when dealing with the complex and ever-changing problem of circuit breaker fault diagnosis. Targeted algorithmic improvements and optimizations, such as the introduction of hybrid strategies and dynamic adjustment of search parameters, are necessary to balance global exploration with local exploitation capabilities, further improving the accuracy and efficiency of fault diagnosis. Summary of the Invention

[0005] In response to the problems existing in the prior art, the present invention provides a high-voltage circuit breaker mechanical fault diagnosis method and system based on the MIDBO-SVM model, thereby solving the technical problem that the dung beetle optimization algorithm in the prior art has an imbalance between global exploration and local development during the iteration process, and still needs to be specifically optimized according to the circuit breaker fault diagnosis problem.

[0006] The present invention is achieved through the following technical solutions:

[0007] A method for diagnosing mechanical faults of high-voltage circuit breakers based on a MIDBO-SVM model comprises the following steps:

[0008] S1: Obtain vibration signal of circuit breaker;

[0009] S2: The mechanical fault of the circuit breaker is diagnosed using the acquired vibration signal in combination with a pre-built MIDBO-SVM model. During the construction of the MIDBO-SVM model, the DBO algorithm adopts a good point set strategy and a circle chaos map to jointly generate an initial population. In the individual position update stage, a sine-cosine search strategy and an adaptive weight factor are introduced. The Cauchy-Gauss mutation strategy is implemented on the individual with the best current fitness to obtain the MIDBO. The MIDBO is then used to optimize the penalty factor C and the kernel parameter g of the SVM to complete the model construction of the MIDBO-SVM.

[0010] Preferably, the good point set strategy and circle chaotic mapping are jointly used to generate the initial population, specifically: first, the basic framework of the initial population is generated by using the good point set strategy, and then the circle chaotic mapping is applied to fine-tune each individual position.

[0011] Preferably, the sine-cosine search strategy and adaptive weight factor are introduced in the individual position update stage, specifically: in each iteration of DBO, the search direction and step size are adjusted in combination with the sine-cosine function according to the relative relationship between the current individual position and the target position, and the weight factor is dynamically adjusted to adapt to different stages of the search process.

[0012] Preferably, the Cauchy-Gaussian mutation strategy is implemented on the individual with the best current fitness, specifically: for the individual with the highest current fitness, the mutation direction and distance are generated according to the Cauchy distribution, and fine-tuned in combination with the randomness of the Gaussian distribution, so that the mutated individual can move away from the current optimal solution.

[0013] Preferably, the vibration signal of the circuit breaker is obtained by an acceleration sensor.

[0014] Preferably, the vibration signal includes amplitude, frequency and phase.

[0015] A high-voltage circuit breaker mechanical fault diagnosis system based on the MIDBO-SVM model includes:

[0016] Data acquisition unit: The data acquisition unit is used to acquire the vibration signal of the circuit breaker;

[0017] Fault diagnosis unit: The fault diagnosis unit is used to diagnose the mechanical fault of the circuit breaker using the acquired vibration signal in combination with a pre-built MIDBO-SVM model. During the construction of the MIDBO-SVM model, the DBO algorithm adopts a good point set strategy and a circle chaos map to jointly generate an initial population. In the individual position update stage, a sine-cosine search strategy and an adaptive weight factor are introduced. The Cauchy-Gauss mutation strategy is implemented on the individual with the best current fitness to obtain the MIDBO. The MIDBO is then used to optimize the penalty factor and kernel parameters of the SVM to complete the model construction of the MIDBO-SVM.

[0018] A computer device / apparatus / system comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above method.

[0019] A computer-readable storage medium stores a computer program, wherein the computer program / instructions are executed by a processor to implement the steps of the above method.

[0020] A computer program product comprises a computer program / instruction, which implements the steps of the above method when executed by a processor.

[0021] Compared with the prior art, the present invention has the following beneficial technical effects:

[0022] The present invention discloses a mechanical fault diagnosis method for high-voltage circuit breakers based on a MIDBO-SVM model. The method adopts the MIDBO-SVM model, that is, a multi-strategy fusion-improved adaptive dung beetle optimization algorithm to optimize the support vector machine parameter model. During the construction process of the model, first, a good point set strategy and a circle chaos map are adopted for the DBO algorithm to jointly generate an initial population. This method helps to increase the diversity and distribution uniformity of the population, thereby improving the global search capability of the algorithm; secondly, a sine-cosine search strategy and an adaptive weight factor are introduced in the individual position update stage. The sine-cosine search strategy simulates the optimization behavior of certain organisms in nature and can avoid falling into local optimality to a certain extent, while the adaptive weight factor is dynamically adjusted according to feedback information in the search process to balance the capabilities of global search and local search; in addition, during the construction process, a Cauchy-Gaussian mutation strategy is implemented for the individual with the best current fitness, and the Cauchy distribution has a heavy-tail characteristic. , so that the mutated individuals can move away from the current optimal solution and explore a broader solution space. At the same time, fine-tuning combined with the randomness of the Gaussian distribution can ensure that the mutation process is both directional and random, which helps to find a better solution; and through the optimized DBO, that is, the MIDBO algorithm, the optimized SVM parameters, namely the penalty factor C and the kernel parameter g, are obtained. These two parameters have a crucial impact on the performance of the SVM classifier. The penalty factor C controls the degree of penalty for misclassified samples, while the kernel parameter g determines the width of the radial basis function (RBF) in the SVM, which in turn affects the complexity and generalization ability of the model. Therefore, the high-voltage circuit breaker mechanical fault diagnosis method based on the MIDBO-SVM model effectively improves the performance of the SVM classifier by introducing a series of optimization measures such as the good point set strategy, circle chaos mapping, sine and cosine search strategy, adaptive weight factor and Cauchy-Gauss mutation strategy, making fault diagnosis more accurate and efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 Schematic diagram of a flow chart of a high-voltage circuit breaker mechanical fault diagnosis method based on the MIDBO-SVM model in the present invention;

[0025] Figure 2 This is a flow chart of high-voltage circuit breaker fault diagnosis based on the MIDBO-SVM model in Example 3 of the present invention;

[0026] Figure 3 Actual photos of the vibration signal acquisition platform for high-voltage circuit breakers built by the present invention, wherein: (a) is a photo of the vibration signal acquisition platform; (b) is a photo of the high-voltage circuit breaker;

[0027] Figure 4 This is a schematic diagram of the circuit breaker mechanical fault data acquisition system in Example 3 of the present invention;

[0028] Figure 5 This is a structural diagram of a high-voltage circuit breaker mechanical fault diagnosis system based on the MIDBO-SVM model in the present invention. DETAILED DESCRIPTION

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0030] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0031] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0032] In the description of the embodiments of the present invention, it should be noted that if the terms "upper," "lower," "horizontal," "inner," etc. appear, the orientation or positional relationship indicated is based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the inventive product is typically placed when in use. These terms are merely for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. In addition, the terms "first," "second," etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0033] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0034] In the description of the embodiments of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0035] The present invention is described in further detail below with reference to the accompanying drawings:

[0036] Example 1

[0037] like Figure 1 As shown, the present invention discloses a high-voltage circuit breaker mechanical fault diagnosis method based on the MIDBO-SVM model, comprising the following steps:

[0038] S1: Acquire a vibration signal of a circuit breaker; wherein the vibration signal includes amplitude, frequency, and phase, and an acceleration sensor may be used to acquire the vibration signal of the circuit breaker.

[0039] S2: The mechanical fault of the circuit breaker is diagnosed using the acquired vibration signal in combination with a pre-built MIDBO-SVM model. During the construction of the MIDBO-SVM model, the DBO algorithm adopts a good point set strategy and a circle chaos map to jointly generate an initial population. In the individual position update stage, a sine-cosine search strategy and an adaptive weight factor are introduced. The Cauchy-Gauss mutation strategy is implemented on the individual with the best current fitness to obtain the MIDBO. The MIDBO is then used to optimize the penalty factor C and the kernel parameter g of the SVM to complete the model construction of the MIDBO-SVM.

[0040] Among them, the good point set strategy and circle chaotic mapping are jointly used to generate the initial population, specifically: first, the basic framework of the initial population is generated by using the good point set strategy, and then the circle chaotic mapping is applied to each individual position for fine-tuning; the sine-cosine search strategy and adaptive weight factor are introduced in the individual position update stage, specifically: in each iteration of DBO, the search direction and step size are adjusted in combination with the sine-cosine function according to the relative relationship between the current individual position and the target position, and the weight factor is dynamically adjusted to adapt to the different stages of the search process; the Cauchy-Gaussian mutation strategy is implemented for the individual with the best current fitness, specifically: for the individual with the highest current fitness, the mutation direction and distance are generated according to the Cauchy distribution, and fine-tuned in combination with the randomness of the Gaussian distribution, so that the mutated individual can be away from the current optimal solution.

[0041] The present invention proposes an adaptive dung beetle optimization algorithm based on multi-strategy fusion to optimize support vector machine parameters, thereby solving the problem of identifying different types of high-voltage circuit breaker faults. First, a method of jointly generating the initial population using the good point set strategy and circle chaos mapping is proposed to improve the uniform ergodicity of the population; secondly, a sine-cosine search strategy and an adaptive weight factor are introduced in the individual position update stage of the rolling ball dung beetle, so that the solution attenuates and oscillates until it approaches the global optimal solution, avoiding falling into the local optimum; finally, a Cauchy-Gauss mutation strategy is adopted to mutate the individual with the best current fitness in a way that moves away from the optimal solution, thereby expanding the spatial search range. By combining the improved DBO algorithm with the support vector machine, the MIDBO-SVM method not only maintains a fast convergence speed, but also solves the problems of slow search speed and easy falling into local optimum in the later stage of the traditional algorithm, thereby improving the accuracy of circuit breaker fault classification and providing a more practical solution for circuit breaker fault diagnosis technology.

[0042] Example 2

[0043] In optimization algorithms, the uniformity and diversity of the initial population distribution significantly impacts the algorithm's overall performance, including convergence speed and accuracy. Traditional logistic chaotic mapping, due to its uneven distribution, often causes the algorithm to fall into local optima during the search process, limiting the algorithm's search capabilities and the quality of the final solution. To address this issue, researchers are continuously exploring more stable and uniformly distributed chaotic mapping methods.

[0044] As an improvement on the Logistic mapping, the Tent mapping improves ergodic uniformity to a certain extent and maintains a relatively consistent distribution density for different parameters. However, its inherent unstable periodicity and susceptibility to falling into fixed points still limit its application in complex optimization problems. This has prompted us to seek more stable and reliable chaotic mapping mechanisms.

[0045] This application innovatively proposes a population initialization method that combines the Circle chaotic map with the good point set strategy. The Circle chaotic map is known for its stable dynamics and uniform distribution in phase space, making it an ideal choice for population initialization. The initial population generated by the Circle chaotic map is evenly distributed in the solution space, effectively avoiding the risk of falling into a local optimum early in the algorithm.

[0046] To further enhance the diversity of the population and the robustness of initialization, this application introduces a good point set strategy. Good point set theory is a mathematical method for evenly distributing points in multidimensional space. It ensures that, given a limited number of points, the point set is distributed as evenly as possible in space. Applying the good point set strategy to population initialization ensures that the initial population is not only evenly distributed in a single dimension, but also maintains good distribution characteristics in multidimensional space, thereby improving the algorithm's global search capabilities.

[0047] Therefore, this application proposes to use a relatively stable and evenly distributed Circle chaos map and a good point set to initialize the population, while making up for the singleness of the group initialization method. The population mapping method is as follows:

[0048]

[0049] Among them, mod is the remainder function, x i Indicates the group position, a is 0.2, b is 0.5; o is a random number uniformly distributed between [0,1]; r k represents the optimal point; ub and lb represent the upper and lower limits of the group position respectively.

[0050] In order to overcome the problems of premature convergence and local optimal trap caused by the fast search speed of the dung beetle search algorithm, this paper proposes an innovative improvement strategy, which combines the sine and cosine search strategies with the dynamic inertia weight mechanism to optimize the producer's position update process.

[0051] Among them, the sine-cosine search strategy is a mechanism that guides the search direction through the periodic changes of the sine and cosine functions. These functions can produce oscillatory behavior within their period, allowing the algorithm to achieve a balance between exploration and exploitation. In the present invention, the sine-cosine search strategy is integrated into the producer's position update formula, which allows the algorithm to gradually approach the potential optimal solution in an oscillatory manner rather than a linear approach during the search process. This method helps the algorithm escape the local optimal area and explore a wider solution space.

[0052] Furthermore, the inertia weight ω plays a crucial role in the optimization algorithm, determining the extent to which the algorithm maintains or changes its current motion state during the search process. In this paper, by introducing random inertia weights into the individual position update equations, the algorithm can dynamically adjust its search step size and direction according to the different stages of the search process. In particular, this randomness ensures that the algorithm has varying exploration capabilities at different iterations, helping to avoid being trapped in local optima.

[0053] However, while the traditional linearly decreasing weight strategy (i.e., the weight gradually decreases with the number of iterations) helps the algorithm converge to a certain extent in the later stages, if a good solution is not found in the early stages of the algorithm, it may cause the weight to decay prematurely, limiting the algorithm's global search capabilities. Therefore, the present invention proposes further optimization strategies for the inertia weight, such as nonlinear decreasing and dynamic weighting based on fitness adjustment, to ensure that the algorithm maintains high search efficiency and flexibility throughout the entire search process.

[0054] By combining the sine and cosine search strategy with the dynamic inertia weight mechanism, the improved algorithm proposed in the present invention uses the oscillation characteristics of the sine and cosine functions to enable the algorithm to more effectively escape from local optimality, while the dynamic inertia weight ensures that the algorithm maintains appropriate exploration capabilities at different stages, thereby enhancing the global search capability. The randomness and dynamically adjusted weight strategy reduce the risk of the algorithm falling into local optimality in the early stages of the search, thereby avoiding premature convergence. By fine-tuning the search direction and step size, the algorithm can approach the global optimal solution more quickly and maintain high precision after reaching stability, thereby improving the convergence speed and accuracy. This strategy has good adaptability to different types of optimization problems and can adapt to different search needs and scenarios by adjusting parameters.

[0055] Therefore, in order to prevent the dung beetle from searching too fast, causing the algorithm to converge prematurely and fall into a local optimum, this patent proposes to introduce a sine-cosine search strategy to update the producer's position update formula, using the sine-cosine function to make the solution oscillate towards the global optimal solution, and at the same time introduce a random inertia weight into the individual position update equation. The size of the inertia weight ω clarifies the level of the algorithm's search ability and can reflect the ability of the individual to maintain its historical state. When using a linearly decreasing weight optimization process, if the algorithm cannot find the optimal point in the early stage, the weight will gradually decay as the iteration increases, and the population optimization will be more likely to fall into local convergence. The improved formula is as follows:

[0056]

[0057] ω=ω max +(ω max -ω min )rand()+δ·randn(); (3)

[0058] Where t represents the current number of iterations, X i (t) represents the position information of the dung beetle during the t-th iteration; k∈(0,0.2] represents the deflection coefficient; b represents a constant value in the range of (0,1); α is the natural coefficient; ω max is the maximum value of random inertia weight, which is 0.9; r1 and r2 are proportional coefficients; ω minis the minimum value of the random inertia weight, which is 0.4; rand() is a random number uniformly distributed in [0,1]; randn() is a random number with a standard normal distribution; σ is used to measure the deviation between the random inertia weight and the mathematical expectation, which is 0.3.

[0059] During the individual position update phase of a rolling dung beetle, a sine-cosine search strategy and adaptive weighting factors are introduced to dampen the oscillation of the solution until it approaches the global optimal solution. In this description, the rolling dung beetle refers to an algorithmic role that simulates the dung beetle's dung ball rolling behavior. The Dung Beetle Optimizer (DBO) is a swarm intelligence optimization algorithm inspired by the biological behavior of dung beetles. In the algorithm, rolling dung beetles represent individuals or solutions, which move through the search space by mimicking the dung beetle's dung ball rolling behavior to find the optimal solution.

[0060] In the later stages of DBO iterations, as the dung beetle population gradually becomes more homogeneous—meaning that the differences between individuals decrease—the algorithm tends to become stuck in a local optimum, leading to update stagnation and a significant decrease in search efficiency. To effectively address this issue, the present invention innovatively introduces a Cauchy-Gaussian mutation strategy, which aims to enhance the algorithm's global search capabilities by increasing population diversity, thereby avoiding local optimum stagnation.

[0061] First, the Cauchy-Gaussian mutation strategy combines the characteristics of the Cauchy and Gaussian distributions, providing a powerful mutation mechanism for the algorithm. The Cauchy distribution is known for its long tail, which can produce large mutations far away from the current position, helping the algorithm escape the local optimum. The Gaussian distribution, on the other hand, is characterized by its symmetry and concentration, which can produce smaller mutations near the current position, facilitating the algorithm's refined search.

[0062] In this invention, the Cauchy-Gaussian mutation strategy is applied to the individuals with the best current fitness. This is because, while these individuals perform well in the current search space, they may only be local optima, not global optima. By mutating them, we can force them to move away from their current optimal regions and explore a wider solution space, thereby increasing the likelihood of finding the global optimal solution. During the mutation process, individuals are first selected for mutation. In each iteration, the best-performing individual is selected as the mutation target based on its fitness. Then, a variable step length is generated using a hybrid strategy of Cauchy and Gaussian distributions. Specifically, either Cauchy or Gaussian distributions can be used to generate the step length based on a certain probability, or the results of both can be combined in some way. Next, the position is updated. This variable step length is applied to the selected individual, updating its position in the solution space. This step is critical, as it directly determines the search direction and range of the mutated individual. Finally, the mutated individual is evaluated, and its fitness value determines whether to accept the mutation. If the mutated individual performs better, the original individual is replaced; otherwise, a specific rule (such as acceptance probability) can be used to determine whether to retain the mutated result.

[0063] This method uses the Cauchy-Gauss mutation strategy, so that the algorithm can still maintain a high global search capability in the later stages of iteration, effectively avoiding the phenomenon of local optimal stagnation and enhancing the global search capability; the mutation operation causes some individuals in the population to update their positions in a way that is far away from the optimal solution, thereby increasing the diversity of the population, helping the algorithm to search in a wider solution space and improving the population diversity; the parameters in the Cauchy-Gauss mutation strategy (such as the ratio of Cauchy distribution and Gaussian distribution, mutation probability, etc.) can be adjusted according to the characteristics of the specific problem, making the algorithm more flexible and adaptable, thereby improving the flexibility and adaptability of the algorithm. Therefore, the Cauchy-Gauss mutation strategy proposed in the present invention provides an effective solution to the problem of local optimal stagnation that may occur in the DBO algorithm in the later stages of iteration. By increasing the diversity of the population and the global search capability, this strategy can significantly improve the performance and stability of the algorithm.

[0064] In the later stages of DBO iterations, the rapid assimilation of the dung beetle population can easily lead to local optimal stagnation. To address this issue, this patent proposes an optimization strategy using a Cauchy-Gaussian mutation strategy. This strategy selects individuals with the best current fitness and mutates them, causing some individuals in the population to update their positions away from the optimal solution, thereby increasing the diversity of the global space explored.

[0065] Gaussian mutation is to add a random vector that obeys Gaussian distribution to the state of the original individual. The one-dimensional Gaussian density function with mean μ and variance σ is:

[0066]

[0067] In the formula, x is a random variable, which means the value for which the probability density needs to be calculated; f G (x) represents the probability density of the random variable x at a specific value; μ is the mean of the distribution, which determines the symmetric center of the distribution; σ is the standard deviation of the distribution. The larger the standard deviation, the wider the data distribution.

[0068] The probability density function of the one-dimensional Cauchy distribution is:

[0069]

[0070] Where x is a random variable; f(x) represents the probability density of the random variable x at a specific value; t represents the scale parameter, which indicates the width of the distribution. The larger the value, the wider the distribution, and the smaller the value, the narrower the distribution.

[0071] During the iteration process, the Cauchy-Gauss mutation strategy is used to compare the population positions before and after the mutation, and a competition mechanism is used to select the solution with better fitness to enter the next iteration. The details are as follows:

[0072] P gbest =X gbest [1+λ1Cauchy(0,σ 2 )+λ2Gauss(0,σ 2 )] (6)

[0073]

[0074]

[0075] Where, P gbest represents the position of the optimal individual after mutation; X gbest represents the optimal individual; X newbest is the optimal individual after individual mutation; σ 2 represents the standard deviation of the Cauchy-Gaussian mutation strategy; Cauchy(0,σ 2 ) is a random variable that satisfies the Cauchy distribution; Gauss(0,σ 2 ) is a random variable with a Gaussian distribution; t represents the current iteration number; T is the total number of iterations; λ1 and λ2 are dynamic parameters that are adaptively adjusted with the number of iterations. In the early stages of the iteration, the Cauchy distribution has a strong perturbation capability, which can improve the algorithm's global search capabilities. As the number of iterations increases, λ1 gradually decreases, reducing the impact of the Cauchy mutation perturbation. In the later stages of the iteration, λ1 decreases and λ2 increases, primarily using Gaussian perturbations to enhance the algorithm's local exploration capabilities and improve convergence accuracy. The use of adaptive parameters demonstrates the advantages of both the Cauchy and Gaussian distributions, balancing the algorithm's global exploration and local search capabilities.

[0076] Example 3

[0077] In addition, if Figure 2 To fully verify the effectiveness of the proposed method (i.e., the MIDBO-SVM model, which combines MIDBO (an improved dung beetle behavior optimization algorithm) with SVM (support vector machine)) in high-voltage circuit breaker fault diagnosis, this example designed an experimental scheme, selected the VS1-ZN63A high-voltage vacuum circuit breaker as the experimental object, and constructed a fault simulation system. The following describes the experimental process, data acquisition, signal processing, model training and optimization, and the final verification steps.

[0078] (1) Experimental design and fault simulation

[0079] Selection of experimental objects: VS1-ZN63A high-voltage vacuum circuit breaker was selected as the experimental object due to its wide application and typical failure mode.

[0080] Fault simulation system construction: In order to simulate mechanical failures in actual operation, a fault simulation system was designed and built, which can simulate various mechanical failure conditions such as low operating voltage, high operating voltage, loose base bolts, as well as operating conditions under normal working conditions.

[0081] (2) Data collection and preprocessing

[0082] Signal acquisition platform: Figure 3 As shown in the figure, the signal acquisition platform integrates an accelerometer, a high-speed acquisition card, and an industrial computer. The accelerometer accurately captures the vibration signals of the circuit breaker during the closing process. These signals are then transmitted in real time to the industrial computer via the high-speed acquisition card for storage and display.

[0083] Data Collection: With the circuit breaker in the no-load state, we collected closing vibration signals under four operating conditions: normal operating conditions, low operating voltage, high operating voltage, and loose base bolts. Under each operating condition, we ensured that a sufficient number of samples were collected to ensure representative and reliable data.

[0084] Preprocessing: Perform denoising, filtering and other preprocessing operations on the collected raw vibration signals to improve signal quality and lay a good foundation for subsequent feature extraction and model training.

[0085] (3) Feature extraction

[0086] FFT feature extraction: Utilizing the Fast Fourier Transform (FFT) method, the preprocessed vibration signal is analyzed in the frequency domain, extracting key features such as frequency, amplitude, and phase, which serve as input features for the support vector machine. The FFT method effectively reveals the signal's characteristics in the frequency domain and helps distinguish vibration modes under different operating conditions.

[0087] (4) Model training and optimization

[0088] MIDBO-SVM model construction: Figure 4 As shown in Figure 2, the core of the MIDBO-SVM model is to optimize the SVM parameters using the MIDBO algorithm. The MIDBO algorithm uses its unique search mechanism to iteratively optimize the SVM penalty factor C and kernel parameter g to find the parameter combination that optimizes the model performance.

[0089] Parameter optimization: The training set data after feature extraction is input into the MIDBO-SVM model. Through the iterative optimization process of the MIDBO algorithm, the parameters of the SVM are continuously adjusted until the preset stopping conditions (such as the maximum number of iterations, convergence of the objective function, etc.) are reached.

[0090] Model testing: Use test set data to perform performance testing on the optimized MIDBO-SVM model to evaluate key indicators such as diagnostic accuracy and robustness under different working conditions.

[0091] (5) Result verification and discussion

[0092] Result analysis: The performance of the MIDBO-SVM model is compared with that of the traditional SVM model or other fault diagnosis methods on the same data set to verify the effectiveness of the MIDBO algorithm in optimizing SVM parameters and the superior performance of the MIDBO-SVM model in high-voltage circuit breaker fault diagnosis.

[0093] Example 4

[0094] In addition, if Figure 5 As shown, the present invention also discloses a high-voltage circuit breaker mechanical fault diagnosis system based on the MIDBO-SVM model, comprising:

[0095] Data acquisition unit: The data acquisition unit is used to acquire the vibration signal of the circuit breaker;

[0096] Fault diagnosis unit: The fault diagnosis unit is used to diagnose the mechanical fault of the circuit breaker using the acquired vibration signal in combination with a pre-built MIDBO-SVM model. During the construction of the MIDBO-SVM model, the DBO algorithm adopts a good point set strategy and a circle chaos map to jointly generate an initial population. In the individual position update stage, a sine-cosine search strategy and an adaptive weight factor are introduced. The Cauchy-Gauss mutation strategy is implemented on the individual with the best current fitness to obtain the MIDBO. The MIDBO is then used to optimize the penalty factor C and kernel parameter g of the SVM to complete the model construction of the MIDBO-SVM.

[0097] In addition, a schematic diagram of a terminal device is provided in one embodiment of the present invention. The terminal device in this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of each of the aforementioned method embodiments are implemented. Alternatively, when the processor executes the computer program, the functions of each module / unit in each of the aforementioned device embodiments are implemented.

[0098] The computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to accomplish the present invention.

[0099] The terminal device may be a computing device such as a desktop computer, a notebook computer, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0100] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0101] The memory may be used to store the computer programs and / or modules, and the processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory.

[0102] If the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electric carrier signals and telecommunication signals. The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A high-voltage circuit breaker mechanical fault diagnosis method based on the MIDBO-SVM model, characterized in that: The following steps are involved: S1: Obtain vibration signal of circuit breaker; S2: Using the acquired vibration signal and combining it with a pre-built MIDBO-SVM model, the mechanical fault of the circuit breaker is diagnosed. During the construction of the MIDBO-SVM model, the DBO algorithm adopts a good point set strategy and a circle chaos map to jointly generate an initial population. During the individual position update phase, a sine-cosine search strategy and an adaptive weight factor are introduced. A Cauchy-Gauss mutation strategy is implemented on the individual with the best current fitness to obtain a MIDBO. The MIDBO is then used to optimize the penalty factor C and kernel parameter g of the SVM to complete the construction of the MIDBO-SVM model. The initial population is generated by combining the good point set strategy and the circle chaotic map. Specifically, the basic framework of the initial population is first generated by using the good point set strategy, and then the circle chaotic map is applied to each individual position for fine-tuning. The population mapping method is as follows: Among them, mod is the remainder function, x i Indicates the group position, a is 0.2, b is 0.5; o is a random number uniformly distributed between [0,1]; r k represents the optimal point; ub and lb represent the upper and lower limits of the group position respectively.

2. A high-voltage circuit breaker mechanical fault diagnosis method based on the MIDBO-SVM model according to claim 1, characterized in that: The sine-cosine search strategy and adaptive weight factor are introduced in the individual position update stage. Specifically, in each iteration of DBO, the search direction and step size are adjusted in combination with the sine-cosine function according to the relative relationship between the current individual position and the target position, and the weight factor is dynamically adjusted to adapt to different stages of the search process.

3. The method for diagnosing mechanical faults of high-voltage circuit breakers based on the MIDBO-SVM model according to claim 1, characterized in that: The Cauchy-Gaussian mutation strategy is implemented on the individual with the best current fitness. Specifically, for the individual with the highest current fitness, the mutation direction and distance are generated according to the Cauchy distribution, and fine-tuned in combination with the randomness of the Gaussian distribution so that the mutated individual can move away from the current optimal solution.

4. The method for diagnosing mechanical faults of high-voltage circuit breakers based on the MIDBO-SVM model according to claim 1, characterized in that: A vibration signal of the circuit breaker is obtained through an acceleration sensor.

5. The method for diagnosing mechanical faults of high-voltage circuit breakers based on the MIDBO-SVM model according to claim 1, characterized in that: The vibration signal includes amplitude, frequency and phase.

6. A high-voltage circuit breaker mechanical fault diagnosis system based on the MIDBO-SVM model, used to implement the method according to any one of claims 1 to 5, characterized in that: include: Data acquisition unit: The data acquisition unit is used to acquire the vibration signal of the circuit breaker; Fault diagnosis unit: The fault diagnosis unit is used to diagnose the mechanical fault of the circuit breaker using the acquired vibration signal in combination with a pre-built MIDBO-SVM model. During the construction of the MIDBO-SVM model, the DBO algorithm adopts a good point set strategy and a circle chaos map to jointly generate an initial population. In the individual position update stage, a sine-cosine search strategy and an adaptive weight factor are introduced. The Cauchy-Gauss mutation strategy is implemented on the individual with the best current fitness to obtain the MIDBO. The MIDBO is then used to optimize the penalty factor and kernel parameters of the SVM to complete the model construction of the MIDBO-SVM.

7. A computer device / apparatus / system comprising a memory, a processor, and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

9. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images