Power distribution control system of intelligent low-voltage power distribution cabinet

Abstract

The invention discloses a power distribution control system of an intelligent low-voltage power distribution cabinet. The power distribution control system comprises a data acquisition module, a preprocessing module, a low-voltage power distribution cabinet fault monitoring model establishment module, a parameter tuning module and a power distribution control module. The invention belongs to the field of power distribution control, and particularly relates to an intelligent low-voltage power distribution cabinet power distribution control system, which performs kernel smoothing processing on a power distribution cabinet feature vector by introducing a lag attenuation coefficient, reduces the interference of noise on subsequent analysis, shortens the lag of the system on a sudden change condition, and is beneficial to finding potential faults in time. Time memory factors are introduced to design a power distribution shield loss function, frequent control caused by small-amplitude jitter of voltage and current is avoided, and power supply reliability is improved; continuously occurring abnormal conditions are restrained, so that the control accuracy of the low-voltage power distribution cabinet is improved; by optimizing the artificial fish swarm algorithm, parameters are dynamically adjusted, and local convergence is accelerated; fish school behaviors and reverse learning are used together, so that the control efficiency of the low-voltage power distribution cabinet is improved.

Description

Technical Field

[0001] The present invention relates to the field of distribution control, and specifically refers to a distribution control system for an intelligent low-voltage distribution cabinet. Background Art

[0002] The distribution control system of a low-voltage distribution cabinet is an important part of the power system to ensure stable, safe and efficient low-voltage power supply. It is mainly used for distributing, controlling and protecting electric energy, and is widely used in industrial, commercial and civil buildings and other fields. However, the general low-voltage distribution cabinet control system is sensitive to sensor noise, transient spikes and subtle jitters caused by switch actions, resulting in false alarms or frequent triggering of protection actions, and being unable to respond quickly to mutations; there is a problem that the parameter settings of the general low-voltage distribution cabinet control system are improper, and the parameter adjustment lacks directionality, and it is easy to fall into a local optimum, thus resulting in low control efficiency of the low-voltage distribution cabinet. Summary of the Invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a distribution control system for an intelligent low-voltage distribution cabinet. Aiming at the problem that the general low-voltage distribution cabinet control system is sensitive to sensor noise, transient spikes and subtle jitters caused by switch actions, resulting in false alarms or frequent triggering of protection actions, and being unable to respond quickly to mutations, this solution performs kernel smoothing processing on the characteristic vectors of the distribution cabinet by introducing a hysteresis attenuation coefficient, which can not only erase small noises and reduce the interference of noises on subsequent analysis, but also not overly blunt normal fluctuations, retain important information on the operating state of the distribution cabinet, shorten the hysteresis of the system to mutations, be able to respond to these mutations more timely, and is conducive to timely discovery of potential faults; introduce a time memory factor to design a distribution shield loss function to avoid frequent control caused by small fluctuations in voltage and current, reduce the frequent tripping and closing of circuit breakers and switches in the cabinet, and improve power supply reliability; constrain continuously occurring abnormal situations, improve the sensitivity to faults, and thus improve the control accuracy of the low-voltage distribution cabinet; aiming at the problem that the general low-voltage distribution cabinet control system has improper parameter settings, the parameter adjustment lacks directionality, and it is easy to fall into a local optimum, thus resulting in low control efficiency of the low-voltage distribution cabinet, this solution optimizes the artificial fish swarm algorithm, initializes the fish swarm position based on the transformation sequence to improve the traversability of the search; dynamically adjusts the parameters based on the sine growth base and nonlinear amplification to balance global-local exploration and accelerate local convergence; combines the fish swarm behavior with reverse learning to enhance the ability to jump out of the local optimum; and thus improves the control efficiency of the low-voltage distribution cabinet.

[0004] The technical solution adopted by the present invention is as follows: A distribution control system for an intelligent low-voltage distribution cabinet provided by the present invention includes a data acquisition module, a preprocessing module, a low-voltage distribution cabinet fault monitoring model establishment module, a parameter tuning module and a distribution control module;

[0005] The data acquisition module collects historical power distribution cabinet sensor data and constructs a feature vector of the power distribution cabinet;

[0006] The preprocessing module introduces a lag decay coefficient to perform kernel smoothing on the feature vector of the power distribution cabinet to obtain a power distribution cabinet data set;

[0007] The low-voltage power distribution cabinet fault monitoring model establishment module introduces a time memory factor to design a power distribution shield loss function and designs a low-voltage power distribution cabinet fault monitoring model based on the power distribution cabinet data set;

[0008] The parameter tuning module tunes the parameters of the kernel smoothing process and the low-voltage power distribution cabinet fault monitoring model to establish a final low-voltage power distribution cabinet fault monitoring model;

[0009] The power distribution control module performs power distribution control on real-time power distribution cabinet sensor data based on the low-voltage power distribution cabinet fault monitoring model.

[0010] Further, the data acquisition module collects historical power distribution cabinet sensor data, labels the power distribution cabinet status as a data label, and normalizes the historical power distribution cabinet operation data to construct a feature vector of the power distribution cabinet.

[0011] Further, the preprocessing module performs kernel smoothing on the feature vector of the power distribution cabinet and gives a lag decay coefficient to the remote frame. The formula used is: ; ; where, is the kernel smoothing function, and u is the function variable; is the Gamma function; is the degree-of-freedom parameter; is the hysteresis coefficient; is the feature after kernel smoothing of the nth frame; M is the half-width of the sliding window, and k is the frame offset within the sliding window; is the normalized eigenvalue; is the kernel bandwidth; in a sliding window manner, the kernel-smoothed power distribution cabinet feature vectors of the nearest T frames are assembled into a matrix to obtain a power distribution cabinet data set.

[0012] Further, the low-voltage power distribution cabinet fault monitoring model establishment module selects a one-dimensional convolution + LSTM hybrid network based on the power distribution cabinet data set to establish a low-voltage power distribution cabinet fault monitoring model, expressed as: ; The input is a matrix from time to t for a total of T frames ; is the one-dimensional convolution operation; is the feature map after convolution; is the long short-term memory network, is the output of the long short-term memory network, It is a fully connected transformation; the final output is the predicted score of the power distribution cabinet status ; and embed the power distribution shield loss function, and introduce the time memory factor , expressed as: ; the final power distribution shield loss function is expressed as: ; where , and are the residuals at time t, time t-1, and time t-2 respectively, which is the difference between the true label and the predicted value; and are the memory coefficients; is the indicator function; is the tolerance threshold; is the overall loss scaling factor; a is the steepness parameter; e is the base of the natural logarithm.

[0013] Furthermore, the parameter tuning module optimizes the hyperparameters involved in the kernel smoothing process and the low-voltage power distribution cabinet fault monitoring model based on the optimized artificial fish swarm algorithm; ensures that the performance of the finally established model meets the standards; specifically includes the following:

[0014] Initialization; establish a search space based on the hyperparameters to be tuned, set k = 0, and randomly generate a sequence , expressed as: , and press Initialize the position of each fish; where and are the s-th and (s-1)-th chaotic values respectively; controls the demarcation point of the sequence transformation; is the initial position of the j-th dimension of the i-th individual search space of the fish swarm; and are the minimum and maximum values of the j-th dimension of the search space respectively;

[0015] Preset; divide the power distribution cabinet dataset into a test set and a training set, and select the prediction accuracy rate of the test set as the individual fitness value when the loss of the low-voltage power distribution cabinet fault monitoring model converges for the training set; the initial global optimum ; where is the fitness value of the initial position of the i-th individual of the fish swarm; initialize the convergence factor and the inertia weight ;

[0016] Iterative search; for k = 0 to execute: update the parameters based on the sine growth base and nonlinear amplification ; ; execute the fish swarm behavior, and the formula used is: ; Reverse learning; for each fish, generate a reverse individual and define the range limit and ; Generate reverse points, , ; And retain the best according to the individual fitness value and update the global optimum; where is the maximum number of iterations; , and control the convergence speed and curve shape; m adjusts the degree of non-linearity; and are the positions of the i-th individual in the fish swarm at the (k + 1)-th iteration and the k-th iteration respectively; and are the inertia weight and convergence factor at the k-th iteration respectively; Step is the maximum step size; is the optimal individual at the k-th iteration; ; N is the population size; is the position of the j-th dimension of the search space; is the generated initial reverse point; is the k-th chaotic value; is the final reverse point;

[0017] Search determination; set the determination threshold. During the iterative search process, when there is an individual fitness value higher than the determination threshold, the search ends, and a low-voltage power distribution cabinet fault monitoring model is established based on the individual positions; if , then go to initialization.

[0018] Furthermore, the power distribution control module collects real-time power distribution cabinet sensor data, preprocesses it and inputs it into the low-voltage power distribution cabinet fault monitoring model, and performs power distribution control based on the power distribution cabinet status output by the low-voltage power distribution cabinet fault monitoring model.

[0019] The beneficial effects achieved by the present invention using the above solution are as follows:

[0020] (1) Aiming at the problem that the general low-voltage power distribution cabinet control system is sensitive to sensor noise, transient spikes, and subtle jitters caused by switch actions, resulting in false alarms or frequent triggering of protection actions and being unable to respond quickly to mutations, this solution performs kernel smoothing processing on the characteristic vectors of the power distribution cabinet by introducing a hysteresis attenuation coefficient. It can not only erase small noises and reduce the interference of noises on subsequent analysis, but also not overly blunt normal fluctuations, retaining important information on the operating state of the power distribution cabinet, shortening the lag of the system to mutations, being able to respond to these mutations more timely, and being conducive to timely discovery of potential faults; introducing a time memory factor to design the power distribution shield loss function to avoid frequent control caused by small fluctuations in voltage and current, reducing the frequent tripping and closing of circuit breakers and switches in the cabinet, and improving power supply reliability; restricting continuously occurring abnormal situations, improving the sensitivity to faults, and thus improving the control accuracy of the low-voltage power distribution cabinet.

[0021] (2) Aiming at the problem that the general low-voltage power distribution cabinet control system has improper parameter settings, and the parameter tuning lacks directionality, easily falling into local optima, which in turn leads to low control efficiency of the low-voltage power distribution cabinet, this solution optimizes the artificial fish swarm algorithm, initializes the fish swarm position based on the transformation sequence to improve the traversability of the search; dynamically adjusts parameters based on the sine growth base and nonlinear amplification to balance global-local exploration and accelerate local convergence; combines fish swarm behavior with reverse learning to enhance the ability to jump out of local optima; and thus improves the control efficiency of the low-voltage power distribution cabinet. Brief Description of the Drawings

[0022] Figure 1 It is a schematic flow chart of an intelligent low-voltage power distribution cabinet power distribution control system provided by the present invention;

[0023] Figure 2 It is a schematic flow chart of the parameter tuning module.

[0024] The drawings are used to provide further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. Detailed Embodiments

[0025] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0026] In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc. indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0027] Embodiment 1, referring to Figure 1 , an intelligent low-voltage power distribution control system provided by the present invention includes a data acquisition module, a preprocessing module, a low-voltage power distribution cabinet fault monitoring model establishment module, a parameter optimization module, and a power distribution control module;

[0028] The data acquisition module collects historical power distribution cabinet sensor data, constructs a power distribution cabinet feature vector; and sends the data to the preprocessing module;

[0029] The preprocessing module introduces a hysteresis attenuation coefficient to perform kernel smoothing processing on the power distribution cabinet feature vector to obtain a power distribution cabinet data set; and sends the data to the low-voltage power distribution cabinet fault monitoring model establishment module;

[0030] The low-voltage power distribution cabinet fault monitoring model establishment module introduces a time memory factor to design a power distribution shield loss function, and designs a low-voltage power distribution cabinet fault monitoring model based on the power distribution cabinet data set; and sends the data to the parameter optimization module;

[0031] The parameter optimization module optimizes the parameters of the kernel smoothing processing and the low-voltage power distribution cabinet fault monitoring model to establish a final low-voltage power distribution cabinet fault monitoring model; and sends the data to the power distribution control module;

[0032] The power distribution control module performs power distribution control on the real-time power distribution cabinet sensor data based on the low-voltage power distribution cabinet fault monitoring model.

[0033] Embodiment 2, referring to Figure 1 , based on the above embodiment, the data acquisition module collects historical power distribution cabinet sensor data and labels the power distribution cabinet status as a data label; the power distribution cabinet status includes undervoltage, normal, and overload; the historical power distribution cabinet operation data is normalized to construct a power distribution cabinet feature vector; the historical power distribution cabinet operation data includes three-phase current, three-phase voltage, temperature, humidity, leakage current, and time period coding.

[0034] Embodiment 3, referring to Figure 1 , based on the above embodiment, the preprocessing module performs kernel smoothing processing on the power distribution cabinet feature vector, which not only erases small noises but also does not overly blunt normal fluctuations, and gives a remote frame a hysteresis attenuation coefficient to shorten the system's mutation hysteresis for night low load → early morning startup and arc short circuit. The formula used is: ; ;in, is the kernel smoothing function, u is the function variable; is the Gamma function; is the degree of freedom parameter; is the hysteresis coefficient; is the kernel-smoothed feature of the nth frame; M is the half-width of the sliding window, and k is the frame offset within the sliding window; is the normalized eigenvalue; is the kernel bandwidth; using a sliding window method, the kernel smoothed distribution cabinet feature vectors of the latest T frames are pieced together into a matrix to obtain the distribution cabinet dataset.

[0035] Example 4, see Figure 1 This embodiment is based on the above embodiment. The low-voltage distribution cabinet fault monitoring model establishment module is based on the distribution cabinet data set. A one-dimensional convolution + LSTM hybrid network is selected to establish a low-voltage distribution cabinet fault monitoring model, taking into account local time series characteristics and long-term and short-term dependencies. It is expressed as: ; Input from time To the matrix of t total T frame ; It is a one-dimensional convolution operation; It is the feature map after convolution, which is a multiple one-dimensional activation sequence extracted by one-dimensional convolution on the time axis of the input matrix of the distribution cabinet dataset; is a long short-term memory network, is the LSTM network output, It is a fully connected transformation; the final output is the predicted score of the distribution cabinet status ; and embed the power distribution shield loss function. In the low-voltage distribution cabinet, the voltage and current measurements will be subject to subtle fluctuations caused by sensor noise, transient spikes and switch actions. If each tiny fluctuation triggers a control action, the circuit breakers and switches in the cabinet will frequently trip and close, which will not only reduce the power supply reliability but also shorten the life of the equipment. Therefore, the power distribution shield loss function is designed to Small fluctuations of zero loss can avoid frequent control caused by small jitters of voltage / current; when the residual When the loss is close to the upper limit, the model gradient will not explode due to abnormal surge, lightning shock, and short circuit mutation; and the time memory factor is introduced. , increasing the penalty for the residuals of consecutive violations, expressed as: ; Final distribution shield loss function Expressed as: ;in, 、 and are the residuals at time t, time t-1, and time t-2, which are the differences between the true label and the predicted value; and is the memory coefficient; is the indicator function; is the tolerance threshold; is the overall loss scaling factor; a is the steepness parameter; e is the base of the natural logarithm.

[0036] By performing the above operations, for the general low-voltage power distribution cabinet control system, it is sensitive to sensor noise, transient spikes, and subtle jitters caused by switch actions, resulting in false alarms or frequent triggering of protection actions, and being unable to respond quickly to mutations. In this solution, by introducing a hysteresis attenuation coefficient, kernel smoothing processing is performed on the feature vectors of the power distribution cabinet. It can not only erase small noises and reduce the interference of noises on subsequent analysis, but also not overly blunt normal fluctuations, retain important information about the operating state of the power distribution cabinet, shorten the hysteresis of the system to mutations, be able to respond to these mutations more timely, and is conducive to detecting potential faults in a timely manner; introducing a time memory factor to design the power distribution shield loss function, avoiding frequent control caused by small fluctuations in voltage and current, reducing the frequent tripping and closing of circuit breakers and switches in the cabinet, and improving power supply reliability; restricting continuously occurring abnormal situations, improving the sensitivity to faults, and thus improving the control accuracy of low-voltage power distribution cabinets.

[0037] Example 5, refer to Figure 1 and Figure 2 Based on the above example, the parameter tuning module optimizes the hyperparameters involved in kernel smoothing processing and the low-voltage power distribution cabinet fault monitoring model based on the optimized artificial fish swarm algorithm, including the half-width of the sliding window, kernel bandwidth, hysteresis coefficient, number of LSTM hidden units, learning rate, weight decay, and maximum number of training epochs; ensuring that the performance of the finally established model meets the standards; specifically including the following content:

[0038] Initialization; establish a search space based on the hyperparameters to be tuned, set k = 0, and randomly generate a sequence which is expressed as: and initialize the position of each fish according to where, and are the s-th and (s - 1)-th chaotic values respectively; is the demarcation point for controlling sequence transformation; is the initialized position of the j-th dimension of the i-th individual search space of the fish swarm; and are the minimum and maximum values of the j-th dimension of the search space respectively;

[0039] Preset; divide the power distribution cabinet dataset into a test set and a training set, and select the prediction accuracy rate of the test set when the loss of the low-voltage power distribution cabinet fault monitoring model converges to the training set as the individual fitness value; the initial global optimum where, is the fitness value of the initial position of the i-th individual in the fish swarm; initialize the convergence factor and the inertia weight ;

[0040] Iterative search; for k = 0 to Execute: Update parameters based on the sine growth base and nonlinear amplification ; ; Execute the fish swarm behavior, and the formula used is: ; Reverse learning; for each fish, generate a reverse individual, define the range limit and ; Generate reverse points, , ; And retain the best according to the individual fitness value, and update the global optimum; where is the maximum number of iterations; , and control the convergence speed and curve shape; m adjusts the nonlinear degree; and are the positions of the i-th individual in the fish swarm at the (k + 1)-th iteration and the k-th iteration respectively; and are the inertia weight and the convergence factor at the k-th iteration respectively; Step is the maximum step size; is the optimal individual at the k-th iteration; ; N is the population size; is is the position of the j-th dimension in the search space; is the generated initial reverse point; is the k-th chaotic value; is the final reverse point;

[0041] Search determination; set the determination threshold. During the iterative search process, when there is an individual fitness value higher than the determination threshold, the search ends, and a fault monitoring model for the low-voltage power distribution cabinet is established based on the individual positions; if , then go to initialization.

[0042] By performing the above operations, aiming at the problems existing in the general low-voltage power distribution cabinet control system, such as improper parameter settings, lack of direction in parameter adjustment, easy to fall into local optimum, and thus resulting in low control efficiency of the low-voltage power distribution cabinet, this solution optimizes the artificial fish swarm algorithm, initializes the fish swarm position based on the transformation sequence to improve the traversability of the search; dynamically adjusts the parameters based on the sine growth base and nonlinear amplification to balance the global-local exploration and accelerate the local convergence; combines the fish swarm behavior with reverse learning to enhance the ability to jump out of the local optimum; and thus improves the control efficiency of the low-voltage power distribution cabinet.

[0043] Example Six, refer toFigure 1 , this embodiment is based on the above - mentioned embodiment. The power distribution control module collects real - time sensor data of the power distribution cabinet, and after pre - processing, inputs it into the low - voltage power distribution cabinet fault monitoring model. Power distribution control is performed based on the power distribution cabinet status output by the low - voltage power distribution cabinet fault monitoring model. If the power distribution cabinet status is normal, no equipment is switched, and the status quo is maintained. If the power distribution cabinet status is undervoltage, the actuators are the voltage tap - changer switch T and the dynamic reactive power compensation device Q, and the control law is expressed as: ; ; where is the nominal voltage; is the real - time voltage measurement value; and are the tap - changer switching steps and the reactive power compensation switching amount respectively; is the voltage regulation gain; According to and , tap - up and reactive power compensation switching are performed according to their magnitudes. If the power distribution cabinet status is overloaded, the actuators are the controllable circuit breaker B and the sectional load shedding device S, and the control law is expressed as: ; ; where is the maximum allowable current; is the current overload amount; is the actual bus current value; is the load shedding ratio coefficient; is the load level to be shed. First, perform sectional load shedding S. If it is still overloaded after shedding, then remotely trip B.

[0044] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "including", "comprising" or any other variant thereof is intended to cover non - exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0045] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention.

[0046] The above describes the present invention and its implementation manners. Such description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual structure is not limited thereto. In general, if those of ordinary skill in the art are inspired by it and, without departing from the gist of the present invention, design similar structural manners and embodiments to this technical solution without creative efforts, they shall fall within the protection scope of the present invention.

Claims

1. An intelligent low-voltage power distribution control system for power distribution cabinets, characterized in that: The system includes a data acquisition module, a preprocessing module, a low-voltage power distribution cabinet fault monitoring model establishment module, a parameter tuning module, and a power distribution control module; The data acquisition module collects historical power distribution cabinet sensor data and constructs a power distribution cabinet feature vector; The preprocessing module introduces a lag decay coefficient to perform kernel smoothing processing on the power distribution cabinet feature vector to obtain a power distribution cabinet data set; The low-voltage power distribution cabinet fault monitoring model establishment module introduces a time memory factor to design a power distribution shield loss function and designs a low-voltage power distribution cabinet fault monitoring model based on the power distribution cabinet data set; The parameter tuning module tunes the parameters of the kernel smoothing processing and the low-voltage power distribution cabinet fault monitoring model to establish a final low-voltage power distribution cabinet fault monitoring model; The power distribution control module performs power distribution control on real-time power distribution cabinet sensor data based on the low-voltage power distribution cabinet fault monitoring model.

2. The intelligent low-voltage power distribution control system for a power distribution cabinet according to claim 1, characterized in that: The preprocessing module performs kernel smoothing on the feature vectors of the power distribution cabinet and gives a lag decay coefficient to the remote frame. The formula used is: ; ; where is the kernel smoothing function and u is the function variable; is the Gamma function; is the degree-of-freedom parameter; is the hysteresis coefficient; is the feature after kernel smoothing of the nth frame; M is the half-width of the sliding window and k is the frame offset within the sliding window; is the normalized eigenvalue; is the kernel bandwidth; In a sliding window manner, the kernel-smoothed feature vectors of the power distribution cabinet for the most recent T frames are assembled into a matrix to obtain the power distribution cabinet dataset.

3. The intelligent low-voltage power distribution control system for a power distribution cabinet according to claim 2, characterized in that: The low-voltage power distribution cabinet fault monitoring model establishment module selects a one-dimensional convolution + LSTM hybrid network based on the power distribution cabinet data set to establish a low-voltage power distribution cabinet fault monitoring model, which is expressed as: ; The input is a matrix of T frames from time to t ; is a one-dimensional convolution operation; is the feature map after convolution; is a long short-term memory network, is the output of the long short-term memory network, is a fully connected transformation; the final output is the predicted score of the power distribution cabinet state ; And embed the power distribution shield loss function, and introduce the time memory factor , which is expressed as: ; The final power distribution shield loss function is expressed as: ; Among them, , and are the residuals at time t, t-1, and t-2 respectively, which is the difference between the true label and the predicted value; and are memory coefficients; is an indicator function; is a tolerance threshold; is the overall loss scaling factor; a is the steepness parameter; e is the base of the natural logarithm.

4. The intelligent low-voltage power distribution control system of the power distribution cabinet according to claim 3, characterized in that: The parameter tuning module optimizes the hyperparameters involved in the kernel smoothing processing and the low-voltage power distribution cabinet fault monitoring model based on the optimized artificial fish swarm algorithm; Ensure that the performance of the finally established model meets the standard; specifically Includes the following: Initialization; establish a search space based on the hyperparameters to be optimized, set k = 0, and randomly generate a sequence , expressed as: , and according to Initialize the position of each fish; where and are the s-th and (s - 1)-th chaotic values respectively; The demarcation point for controlling the sequence transformation; is the initialized position of the j-th dimension of the i-th individual in the fish swarm search space; and are the minimum and maximum values of the j-th dimension of the search space respectively; Preset; divide the power distribution cabinet dataset into a test set and a training set, and select the prediction accuracy rate of the test set as the individual fitness value when the loss of the low-voltage power distribution cabinet fault monitoring model converges for the training set; initial global optimum ; among them, is the fitness value of the initial position of the i-th individual in the fish swarm; initialize the convergence factor and the inertia weight ; Iterative search; Search determination; set a determination threshold. During the iterative search process, when there is an individual fitness value higher than the determination threshold, the search ends, and a fault monitoring model for the low-voltage power distribution cabinet is established based on the individual position; if , then go to initialization.

5. The intelligent low-voltage power distribution control system for a power distribution cabinet according to claim 4, wherein: The iterative search is for k = 0 to Execute: Update parameters based on sinusoidal growth basis and nonlinear amplification ; ; Execute fish school behavior, and the formula used is: ; Opposite learning; For each fish, generate reverse individuals and define range bounds and ; Generate reverse points, , ; And retain the best according to the individual fitness value and update the global optimum; where, is the maximum number of iterations; , and control the convergence speed and curve shape; m adjusts the degree of non-linearity; and are the positions of the i-th individual in the fish swarm at the (k + 1)-th iteration and the k-th iteration respectively; and are the inertia weight and convergence factor at the k-th iteration respectively; Step is the maximum step size; is the optimal individual at the k-th iteration; ; N is the population size; is is the position of the j-th dimension of the search space; is the generated initial reverse point; is the k-th chaotic value; is the final reverse point.

6. The intelligent low-voltage power distribution control system for a power distribution cabinet according to claim 5, wherein: The data acquisition module collects historical power distribution cabinet sensor data and labels the power distribution cabinet status as a data label; performs normalization processing on the historical power distribution cabinet operation data and constructs a power distribution cabinet feature vector.

7. An intelligent low-voltage power distribution control system for a power distribution cabinet according to claim 6, characterized in that: The power distribution control module collects real-time power distribution cabinet sensor data, inputs it into the low-voltage power distribution cabinet fault monitoring model after preprocessing, and performs power distribution control based on the power distribution cabinet status output by the low-voltage power distribution cabinet fault monitoring model.

Images