An intelligent control method for an external circuit breaker based on a time-delay neural feedback network

Abstract

The present invention discloses an intelligent control method for an external circuit breaker based on a time-delay neural feedback network, comprising the following steps: Real-time collect the state signals and phase signals of the intelligent control system of the external circuit breaker, and after filtering processing, use the time-delay neural feedback network to perform non-linear conversion on the signals to generate a preliminary control signal; Use an optimization objective function to optimize the preliminary control signal to minimize system error and control signal fluctuation, and obtain an optimized control signal; Construct a non-linear oscillation model for simulating the dynamic response of the external circuit breaker and the intelligent control system, adjust the oscillation behavior through control input, generate a dynamic prediction value of the state based on the optimized control signal, and then update the control signal according to the difference between the dynamic prediction value and the actual state. The present invention has the advantages of strong dynamic adaptability, long service life, good stability, and timely response, etc.

Description

Technical Field

[0001] The present invention belongs to the technical field of control methods, and particularly relates to an external circuit breaker control method. Background Art

[0002] External circuit breakers are widely used in fields such as smart grids, industrial automation, power management, and smart energy scheduling, mainly for overload protection and short-circuit protection. In the case of dynamic changes in grid load, imbalance between power supply and demand, and frequent environmental disturbances, how to optimize the control strategy of external circuit breakers to improve the stability, response speed, and control accuracy of the system has become an important research direction in current intelligent control technology.

[0003] However, most traditional control methods adopt linear control or fixed-threshold strategies, thus having the following problems:

[0004] 1. It is impossible to accurately describe the non-linear dynamic characteristics of the circuit breaker system (such as load mutation, mechanical inertia delay), and it cannot adapt to the complex dynamic changes of the system state.

[0005] 2. Existing methods rely on simple feedback mechanisms, have a lag in response when facing sudden disturbances, and are prone to overcompensation or control failure.

[0006] 3. The fixed-threshold strategy may cause problems such as frequent switching of the circuit breaker under disturbances such as grid fluctuations and load changes, resulting in increased mechanical wear, reduced equipment life, and affecting the safe and stable operation of the system. Summary of the Invention

[0007] The present invention proposes an intelligent control method for an external circuit breaker based on a time-delay neural feedback network, and its purpose is to solve the problems of poor dynamic adaptability, lag in response when facing sudden disturbances, and frequent switching of existing methods.

[0008] The technical solution of the present invention is as follows:

[0009] An intelligent control method for an external circuit breaker based on a time-delay neural feedback network, comprising the following steps:

[0010] Step S1: Real-time collect the state signal and phase signal of the intelligent control system of the external circuit breaker. After filtering, use the time-delay neural feedback network to perform non-linear conversion on the signal to generate a preliminary control signal; use an optimization objective function to optimize the preliminary control signal to minimize system error and control signal fluctuations, and obtain an optimized control signal;

[0011] Step S2: Construct a non - linear oscillation model to simulate the dynamic response of the external circuit breaker and the intelligent control system. Adjust the oscillation behavior through the control input, generate a dynamic prediction value of the state based on the optimized control signal, and then update the control signal using a correction formula according to the difference between the dynamic prediction value and the actual state;

[0012] The corrected control signal is sent to the control system of the external circuit breaker of the electricity meter to perform the switching operation.

[0013] As a further improvement of the intelligent control method for the external circuit breaker based on the time - delay neural feedback network: In step S1, the input of the time - delay neural feedback network model is a combination of the system state signal and the phase signal. The connection weights and time - delay factors between neurons act together to determine the time - varying characteristics of the output signal.

[0014] As a further improvement of the intelligent control method for the external circuit breaker based on the time - delay neural feedback network: In step S1, the optimization objective function contains two weighted terms. One weighted term is based on the difference between the current state and the target state, and the other weighted term is based on the obtained preliminary control signal.

[0015] As a further improvement of the intelligent control method for the external circuit breaker based on the time - delay neural feedback network: In step S1, the optimization objective function is iteratively adjusted by the gradient descent method to solve the optimal control signal; the optimized control signal not only reflects the change in load demand but also plays a regulating role in the operating frequency of the circuit breaker and the stability of the intelligent control system.

[0016] As a further improvement of the intelligent control method for the external circuit breaker based on the time - delay neural feedback network: In step S1, the process of solving the optimal control signal is as follows: First, calculate the gradient of the optimization objective function, and dynamically adjust the control signal according to the direction and magnitude of the gradient; each iteration represents an update of the control signal.

[0017] As a further improvement of the intelligent control method for the external circuit breaker based on the time - delay neural feedback network: In step S2, construct a non - linear oscillation model to simulate the dynamic response of the external circuit breaker and the intelligent control system, and adjust it in combination with the intelligent control input; the non - linear oscillation model is based on a non - linear oscillation equation, which comprehensively considers the relationship between the acceleration, speed, and position changes of the external circuit breaker under different operating states, and introduces the influence of external disturbances, enabling the intelligent control system to more accurately predict the operating state and load response of the circuit breaker.

[0018] As a further improvement of the intelligent control method for the external circuit breaker based on the time-delay neural feedback network: In step S2, the method of correcting the control signal based on the deviation between the predicted state and the actual state is as follows: First, solve the non-linear oscillation model to obtain the predicted value of the future state, which is used to estimate the future evolution trend of the system; Then, compare the predicted value of the future state with the currently measured state value to quantify the error between the current state of the system and the predicted trend; Finally, superimpose the correction signal obtained based on the error onto the optimized control signal to generate a more smooth and stable corrected control signal, thereby significantly improving the control accuracy and response speed of the external circuit breaker.

[0019] Compared with the prior art, the present invention has the following positive effects:

[0020] 1. The present invention adopts a time-delay neural feedback network, which can analyze the system state signal and phase signal in real time, and perform non-linear conversion on the input data, enabling the control system to have the ability of adaptive adjustment.

[0021] 2. The present invention constructs an optimization objective function aiming at minimizing the state error and optimizing the smoothness of the control signal, effectively reducing the adjustment error of the system, avoiding excessive control signal fluctuations at the same time, reducing the frequent operation of the external circuit breaker, reducing mechanical losses, and improving the equipment life and operation stability.

[0022] 3. The present invention introduces a non-linear oscillation modeling method to simulate the dynamic response of the external circuit breaker and its controlled system. By establishing an oscillation model with inertia, damping and external disturbance terms, the system can accurately predict the future state and adjust the control signal, enhancing the system's adaptability to load fluctuations, sudden disturbances and environmental changes, avoiding response lags, and reducing situations of over-compensation and control failure.

[0023] 4. On the basis of optimizing the control signal, the present invention further designs an intelligent correction mechanism, which dynamically adjusts according to the error between the predicted state value and the current state value, can effectively reduce the prediction error, improve the intelligent response ability of the external circuit breaker, and ensure the accuracy of the control signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0025] The technical solution of the present invention will be described in detail below with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0026] As Figure 1 , an intelligent control method for an external circuit breaker based on a time-delay neural feedback network includes the following steps:

[0027] Step S1: Real-time collect the status signals and phase signals of the external circuit breaker intelligent control system. After filtering, use a time-delay neural feedback network to perform non-linear conversion on the signals to generate preliminary control signals. Optimize the preliminary control signals using an optimization objective function to minimize system error and control signal fluctuations, and obtain the optimized control signals.

[0028] Real-time collection is carried out through a variety of sensors installed in the intelligent control system to obtain system status signals and phase signals. The system status signals include the load status and grid status of the external circuit breaker, which can be provided by an intelligent electricity meter, monitoring equipment, or sensors, and reflect the operating parameters of each controlled object in the external circuit breaker intelligent control system. The phase signal is used to reflect the dynamic changes of the power grid and is obtained through a phase measurement device, reflecting the phase changes and angle information of each node in the power grid.

[0029] The collected signals are filtered to remove high-frequency noise and other interference components to ensure data accuracy. Specifically, a band-pass filter is used to filter the system status signals and phase signals, only receiving signals within a preset frequency band and removing low-frequency drift or high-frequency clutter. For power load signals, the filtered data can be used to calculate key information such as instantaneous power and power factor for use in subsequent control calculations.

[0030] Design a time-delay neural feedback network. Through the time-delay neural feedback network, perform non-linear conversion on the system status signals and phase signals to generate preliminary control signals. The input of the time-delay neural feedback network model is a combination of the system status signals and phase signals. The connection weights and time-delay factors between neurons act together to determine the time-varying characteristics of the output signal. The core calculation formula of the time-delay neural feedback network is:

[0031] ;

[0032] where, is the preliminary control signal output by the time-delay neural feedback network at time is the connection weight of the th neuron in the time-delay neural feedback network, representing the influence degree of the input signal; is the number of neurons; is the system status signal (such as the load status or grid status of the external circuit breaker) at time is the phase signal (used to reflect the dynamic changes of the power grid or system) at time is the The non - linear adjustment factor of a neuron, which controls the non - linear strength of the response of the input signal to the output signal, is obtained through experiments; is the bias term of the th neuron, used to simulate incomplete data or errors; is the time - delay attenuation factor of the th neuron, used to control the attenuation rate of the input signal over time; is the time - delay reference time point of the th neuron, determined by the communication time. The preliminary control signal is extracted from the system state signal and the phase signal through the time - delay neural feedback network

[0033] In the target optimization stage, in order to optimize the intelligent adjustment ability of the intelligent control system, an optimization objective function is designed to optimize the control signal, so as to ensure that the control system can adaptively adjust in a dynamic environment, minimize the error between the current state of the system and the target state, and at the same time keep the control signal stable to avoid the impact on the actuator (such as an external circuit breaker) caused by violent fluctuations. The expression of the optimization objective function is:

[0034] ;

[0035] where, is the optimization objective function, measuring the error of the current control state; is the target system state at time is the maximum value of the system state, used for normalization; is the maximum amplitude of the preset control signal; , are both weighting factors, used to adjust the weights of the system state error and the control signal amplitude.

[0036] The optimization objective function is iteratively adjusted by the gradient descent method to solve the optimal control signal. The specific process is as follows: First, calculate the gradient of the optimization objective function, and dynamically adjust the control signal according to the direction and magnitude of the gradient; each iteration represents an update of the control signal. The optimized control signal can not only accurately reflect the change of the load demand, but also effectively adjust the operating frequency of the circuit breaker, while improving the stability of the intelligent control system. The update formula of the control signal is:

[0037] ;

[0038] where, is the optimized control signal; is the learning rate, controlling the step size of each iteration, obtained through experiments; is the gradient of the optimization objective function with respect to the control signal, representing the relationship between the rate of change of the optimization objective function and the change in the control signal.

[0039] Step S2: Construct a non - linear oscillation model to simulate the dynamic response of the external circuit breaker and the intelligent control system. Adjust the oscillation behavior through the control input, generate a dynamic prediction value of the state based on the optimized control signal, and then update the control signal using a correction formula according to the difference between the dynamic prediction value and the actual state to ensure the accurate execution of the circuit breaker switching operation.

[0040] Specifically, to enhance the adaptive ability of the intelligent control system to dynamic behavior, a non - linear oscillation model is constructed to simulate the dynamic response of the external circuit breaker and the intelligent control system, and it is adjusted in combination with the intelligent control input. The non - linear oscillation model is based on a non - linear oscillation equation, which comprehensively considers the relationships among the acceleration, velocity, and position changes of the external circuit breaker under different operating states, and introduces the influence of external disturbances, enabling the intelligent control system to more accurately predict the operating state and load response of the circuit breaker. The equation of the non - linear oscillation model is as follows:

[0041] ;

[0042] where, represents the predicted state acceleration, that is, the second - order derivative of ; is the predicted state change rate, that is, the first - order derivative of ; is the predicted state value, that is, the system state at time ; is the predicted time step, which determines the time interval of the load prediction; is the damping coefficient, which controls the resistance of the oscillation and is obtained through experiments; is the stiffness coefficient, representing the restoring force of the intelligent control system to the offset; is the amplitude factor, which determines the amplitude of the oscillation; is the oscillation frequency, is the phase of the oscillation; is the predicted external disturbance term, which is obtained by predicting historical data through environmental monitoring and then through machine learning, such as factors like climate change and grid fluctuations; is the influence factor of the control input, representing the influence intensity of the optimized control signal on the state change and is obtained through experiments.

[0043] The core purpose of the above prediction is to describe how the system state responds to the control input Make a response. Although the obtained control signal is not actually implemented, mathematical modeling can still be used to simulate the system's response, so as to anticipate potential future load fluctuations in advance, and then make necessary adjustments before the control signal is implemented to avoid over-regulation or reaction lag.

[0044] To enhance the precise adjustment ability of the intelligent control system for the external circuit breaker, the present invention corrects the control signal based on the deviation between the predicted state and the actual state. The specific method is as follows: First, solve the nonlinear oscillation model to obtain the predicted value of the future state , which is used to estimate the future evolution trend of the system; then compare the predicted value of the future state with the currently measured state value to quantify the error between the current state of the system and the predicted trend; finally, superimpose the correction signal obtained based on the error onto the optimized control signal to generate a more smooth and stable corrected control signal, thereby significantly improving the control accuracy and response speed of the external circuit breaker. The corrected control signal is updated through the following formula:

[0045] ;

[0046] where, is the corrected control signal, which is used to adjust the operation of the circuit breaker; is the predicted state value obtained through the nonlinear oscillation model; is the amplitude adjustment coefficient of the correction signal, which controls the intensity of the correction signal and is obtained through experiments; is the time delay factor, which controls the attenuation speed of the time delay during the signal correction process and is obtained through experiments; is the weighting factor for the state change rate, indicating the contribution degree of the state change rate in the correction signal and is obtained through experiments; is the influence index of the state change rate, which controls the response sensitivity of the state change rate in the correction signal and is obtained through experiments; is the reference time. Through the corrected control signal, the error in the state control process is minimized to ensure the precise execution of the circuit breaker switching operation.

[0047] The corrected control signal is sent to the control system of the external circuit breaker of the electric energy meter to perform the switching operation and complete the intelligent control of the external circuit breaker.

[0048] It should be noted that for those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. The scope of the present invention is defined by the claims rather than the above description.

Claims

1. An intelligent control method for an external circuit breaker based on a time-delay neural feedback network, characterized in that: The following steps are involved: Step S1, real-time acquisition of state signals and phase signals of the external circuit breaker intelligent control system, after filtering, using a time-delay neural feedback network to perform nonlinear conversion on the signals to generate a preliminary control signal; using an optimization objective function to optimize the preliminary control signal to minimize system errors and control signal fluctuations, and obtain an optimized control signal; The input of the time-delay neural feedback network model is a combination of a system state signal and a phase signal, and the connection weights and the time-delay factors between neurons work together to determine the time-varying characteristics of the output signal; The calculation formula of the time-delay neural feedback network is: ； in, Delayed Neural Feedback Network The initial control signal output at all times; It is the first The connection weight of each neuron indicates the influence of the input signal; is the number of neurons; yes System status signal at each moment; yes The phase signal at the moment; It is The nonlinear adjustment factor of each neuron controls the nonlinear intensity of the response of the input signal to the output signal, which is obtained through experiments; It is The bias term of each neuron; It is The delay decay factor of each neuron is used to control the rate at which the input signal decays over time; It is The delay reference time point of each neuron is determined by the communication time; In step S1, the optimization objective function includes two weighted items, one of which is based on the difference between the current state and the target state, and the other is based on the obtained preliminary control signal; The expression of the optimization objective function is: ； in, It is the optimization objective function, which measures the error of the current control state; yes The target system state at the moment; is the maximum value of the system state, used for normalization; is the maximum amplitude of the preset control signal; , They are all weighting factors, used to adjust the weight of system state error and control signal amplitude; Step S2, constructing a nonlinear oscillation model to simulate the dynamic response of the external circuit breaker and the intelligent control system, adjusting the oscillation behavior through the control input, generating a dynamic prediction value of the state based on the optimized control signal, and then updating the control signal using a correction formula according to the difference between the dynamic prediction value and the actual state; The corrected control signal is sent to the control system of the external circuit breaker of the energy meter to perform the switching operation; In step S2, a nonlinear oscillation model is constructed to simulate the dynamic response of the external circuit breaker and the intelligent control system, and is adjusted in combination with the intelligent control input; the nonlinear oscillation model is based on a nonlinear oscillation equation, which comprehensively considers the relationship between the acceleration, speed and position changes of the external circuit breaker under different operating conditions, and introduces the influence of external disturbances, so that the intelligent control system can more accurately predict the operating state and load response of the circuit breaker; The equation of the nonlinear oscillation model is as follows: ； in, represents the predicted state acceleration, i.e. The second derivative of is the predicted rate of state change, i.e. The first derivative of ; is the predicted state value, that is, System status at the moment; is the forecast time step, which determines the time interval of load forecast; is the damping coefficient, which controls the resistance to oscillation and is obtained experimentally; is the stiffness coefficient, which represents the restoring force of the intelligent control system to the deviation; is the amplitude factor, which determines the amplitude of the oscillation; is the oscillation frequency, is the phase of the oscillation; It is the predicted external disturbance term, which is obtained by obtaining historical data through environmental monitoring and then predicted through machine learning; is the influence factor of the control input, which indicates the influence intensity of the optimized control signal on the state change; In step S2, the control signal is corrected based on the deviation between the predicted state and the actual state in the following manner: first, the nonlinear oscillation model is solved to obtain the predicted value of the future state, which is used to estimate the future evolution trend of the system; then, the predicted value of the future state is compared with the current measured state value to quantify the error between the current state of the system and the predicted trend; finally, the correction signal obtained based on the error is superimposed on the optimized control signal to generate a smoother and more stable corrected control signal, thereby significantly improving the control accuracy and response speed of the external circuit breaker; The corrected control signal is updated by the following formula: ； in, It is the corrected control signal, used to adjust the action of the circuit breaker; is the state prediction value obtained by the nonlinear oscillation model; is the amplitude adjustment coefficient of the correction signal, which controls the strength of the correction signal and is obtained through experiments; is the delay factor, which controls the decay rate of the delay during the signal correction process and is obtained through experiments; It is a weighting factor for the state change rate, indicating the contribution of the state change rate to the correction signal, obtained through experiments; It is the influence index of the state change rate, which controls the response sensitivity of the state change rate in the correction signal and is obtained through experiments; It is the base time.

2. The external circuit breaker intelligent control method based on the time-delay neural feedback network according to claim 1, characterized in that: In step S1, the optimization objective function is iteratively adjusted by the gradient descent method to solve the optimal control signal; the optimized control signal not only reflects the change in load demand, but also regulates the operating frequency of the circuit breaker and the stability of the intelligent control system.

3. The external circuit breaker intelligent control method based on the time-delay neural feedback network according to claim 2, characterized in that: In step S1, the process of solving the optimal control signal is: first calculate the gradient of the optimization objective function, and dynamically adjust the control signal according to the direction and amplitude of the gradient; each iteration represents an update of the control signal.

Images