CN117742431B - High-voltage stabilization method and control equipment thereof - Google Patents

Abstract

The invention relates to the technical field of high-voltage output systems, in particular to a high-voltage stabilizing method and control equipment thereof, which comprises the steps of constructing a first algorithm model, a second algorithm model and a third algorithm model according to the high-voltage output system, and initializing parameters of the first algorithm model, the second algorithm model and the third algorithm model; obtaining a prediction covariance through a state equation of the first algorithm model, and updating the state and the covariance according to an observation equation of the first algorithm model to obtain a process noise variance; obtaining an optimal value of the process noise variance through a second algorithm model, substituting the optimal value of the process noise variance into the first algorithm model for iterative calculation until a relative optimal value of the output high voltage of the high voltage output system is obtained; and inputting the relative optimal value and the voltage set value into a third algorithm model to obtain a voltage output value. And the high voltage can be stably output based on software calculation, so that high voltage fluctuation is reduced, and the whole system is safer and more reliable.

Description

High-voltage stabilization method and control equipment thereof

Technical Field

The invention relates to the technical field of high-voltage output systems, in particular to a high-voltage stabilizing method and control equipment thereof.

Background

The application of the ion pump is an important research field in the technical field of super vacuum, and in the application process of the ion pump, how to output stable high voltage also becomes a great difficulty, and a traditional voltage output circuit can generate a great error when outputting high voltage.

The existing voltage output circuit generally utilizes pulse width modulation (Pulse Width Modulation, abbreviated as PWM) to output low voltage, then adjusts the low voltage to high voltage through a transformer, then continuously collects output voltage, and utilizes a Proportional-Integral-Derivative (abbreviated as PID) controller to adjust voltage fluctuation. The method can be suitable for the situations requiring no high voltage output, but the situations requiring strict high voltage fluctuation are slightly insufficient.

In view of this, overcoming the drawbacks of the prior art is a problem to be solved in the art.

Disclosure of Invention

The invention aims to solve the technical problems that: how to output stable high voltage in a higher voltage output scene.

The invention adopts the following technical scheme:

in a first aspect, a high pressure stabilization method is provided, comprising:

Constructing a first algorithm model, a second algorithm model and a third algorithm model according to a high-voltage output system, and initializing parameters of the first algorithm model, the second algorithm model and the third algorithm model respectively;

Obtaining a prediction covariance through a state equation of the first algorithm model, and updating the state and the covariance according to an observation equation of the first algorithm model to obtain a process noise variance;

obtaining an optimal value of the process noise variance through the second algorithm model, substituting the optimal value of the process noise variance into the first algorithm model for iterative calculation until a relative optimal value of the output high voltage of the high voltage output system is obtained;

and inputting the relative optimal value and the voltage set value into the third algorithm model to obtain a voltage output value.

Preferably, the building the first algorithm model, the second algorithm model and the third algorithm model according to the high-voltage output system, and initializing parameters of the first algorithm model, the second algorithm model and the third algorithm model respectively specifically includes:

the parameters for initializing the first algorithm model are as follows:

，，；

Wherein, Representing an estimate of the system state at time 0,Representing the process noise covariance matrix at time 0,The covariance at time 0 is indicated,Representing an L x 1-dimensional zero vector, which is the system response length; for the L x L-dimensional identity matrix, AndAre all given constants; l represents a length;

initializing parameters of the second algorithm model, wherein the initialized parameters comprise F, cr, np and G, F is a variation rate, cr is a cross probability, np is a population size, and G is the iteration number;

Initializing parameters of the third algorithm model, wherein the initialized parameters comprise Kp, ki and Kd, kp is a proportional gain, ki is an integral gain, and Kd is a differential gain.

Preferably, the obtaining the prediction covariance through the state equation of the first algorithm model, and updating the state and covariance according to the observation equation of the first algorithm model to obtain the process noise variance specifically includes:

acquiring a state equation of the first algorithm model, and obtaining a state predicted value and the prediction covariance according to the state equation;

Acquiring an observation equation of the first algorithm model, updating a state through the observation equation, obtaining a filter gain through the observation equation, and updating a covariance through the filter gain;

And obtaining the process noise variance according to the state equation and the observation equation.

Preferably, the obtaining a state equation of the first algorithm model, and obtaining a state predicted value and the prediction covariance according to the state equation specifically includes:

The state equation is obtained as follows:

;

Wherein, A state transfer function at time k-1; is the input matrix at time k-1; is the process noise at time k-1;

And (3) making:

;

The state prediction value is obtained as follows:

;

The predicted covariance is obtained as:

;

Wherein, An observation matrix at time k-1; is the prediction covariance matrix at time k; is the state transition matrix at time k; is a prediction covariance matrix at k-1; Is that Is a transpose of (2); the process noise covariance matrix at time k-1 is used to represent the uncertainty in the state equation.

Preferably, the obtaining the observation equation of the first algorithm model, updating a state through the observation equation, obtaining a filtering gain through the observation equation, and updating a covariance through the filtering gain, specifically includes:

the observation equation is obtained as follows:

;

Wherein, Representing the observed value at time k; h is an observation matrix for converting the state vectorMapping to an observation space; A predicted value indicating the time k; Represents observed noise at time k;

updating the state through the observation equation is as follows:

;

Wherein, Is an estimate of the state at time k,Is a state estimate of the prediction of time k-1,The filtering gain at the moment k;

the filter gain is obtained through the observation equation as follows:

;

Wherein, A filter gain representing time k for combining the observed data with the state estimate; is a prediction covariance matrix at time k-1, representing uncertainty of one-step prediction of the state at time k; Is the observation matrix at the moment k; Is that Is a transpose of (2); Is the observed noise covariance matrix at the moment k;

updating covariance by the filter gain to:

;

Wherein the method comprises the steps of Is the updated state covariance matrix; Is the state covariance matrix at time k-1; i is a known identity matrix; The filter gain at time k.

Preferably, said deriving said process noise variance from said state equation and said observation equation comprises:

The process noise obtained according to the state equation and the observation equation is as follows:

;

The process noise variance is:

;

Wherein, Is process noise; Process noise variance at time k; Is an operation to solve the process noise variance.

Preferably, the obtaining, by the second algorithm model, the optimal value of the process noise variance specifically includes:

Randomly initializing a population of solutions of the process noise variance, randomly selecting a preset number of solutions for each solution of the process noise variance, obtaining differences between each solution and other preset number of solutions, and generating a number solution;

And performing cross operation on each solution and the number of solutions to generate a test solution, and if the adaptability of the current test solution is better than that of the current solution, replacing the current solution by the current test solution to obtain the optimal value of the process noise variance.

Preferably, the randomly initializing the population of the solutions of the process noise variance, randomly selecting a preset number of solutions for each solution of the process noise variance, obtaining the difference between each solution and other preset number of solutions, and generating a number of solutions, specifically:

the step of population initialization of the solution of the process noise variance comprises the following steps:

;

Wherein, Is a random number uniformly distributed between 0 and 1,AndUpper and lower bounds for the search;

generating a quantitative solution as:

;

Wherein the method comprises the steps of Is a variant individual; The value range is 0-1 for the compression scale factor; 、 And Is the parent.

Preferably, the cross operation is performed on each solution and the number of solutions to generate a test solution, and if the fitness of the current test solution is better than that of the current solution, the current solution is replaced by the current test solution to obtain the optimal value of the process noise variance, which specifically includes:

The specific process of generating a test solution includes:

;

wherein r is a uniformly distributed random number of 0-1 generated by each variable; Cross probability as variable; Is an integer evenly distributed among 1~d;

the selection is made to obtain a relative optimum value of the process noise variance according to the following manner:

;

Wherein, For the current solution to be the one,For the current solution of the test to be performed,For the current solutionIs used for the degree of adaptation of the system,For the current test solutionIs used for the adaptation degree of the device.

In a second aspect, there is provided a control apparatus including: the device comprises a first control module, a second control module and a transformer module;

The first control module is used for receiving the instruction of the upper computer and executing the high-voltage stabilizing method to obtain an output signal;

The second control module is used for receiving the output signal, generating a corresponding voltage signal and collecting the voltage signal and transmitting the voltage signal back to the first control module to form a closed loop;

The transformer module is used for transforming the voltage signal generated by the second control module to obtain a required voltage value.

Compared with the prior art, the invention has the beneficial effects that:

The high-voltage stabilizing method based on the first algorithm model, the second algorithm model and the third algorithm model can stably output high voltage, reduce high-voltage fluctuation, enable the whole high-voltage system to be safer and more reliable, simultaneously, calculate by adopting a software algorithm, reduce errors, filter inherent error interference on hardware, and have reliability and accuracy compared with filtering by only relying on a hardware circuit. On the other hand, compared with hardware filtering, the method adopting software filtering saves more cost, and is convenient for a developer to carry out improvement and development according to actual conditions.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of a high-voltage stabilization method according to an embodiment of the present invention;

FIG. 2 is a schematic calculation flow diagram of a first algorithm model of a high-voltage stabilization method according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a calculation flow of a second algorithm model of a high-voltage stabilization method according to an embodiment of the present invention;

fig. 4 is a schematic structural diagram of a control device according to an embodiment of the present invention;

fig. 5 is a schematic structural diagram of a control device according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of a high-voltage stabilizing device according to an embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

The terms "first," "second," and the like herein are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first", "a second", etc. may explicitly or implicitly include one or more such feature. In the description of the present invention, unless otherwise indicated, the meaning of "a plurality" is two or more.

In the present invention, unless explicitly specified and limited otherwise, the term "connected" is to be construed broadly, and for example, "connected" may be either fixedly connected, detachably connected, or integrally formed; can be directly connected or indirectly connected through an intermediate medium.

In addition, the technical features of the embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.

Example 1:

In classical kalman filter algorithm applications, only for filtering in linear systems, both state and observation equations are required to be linear. In order to improve the applicable scene of the Kalman filtering algorithm, a plurality of improvements are sequentially proposed, wherein the extended Kalman filtering algorithm is widely applied to a scheme for processing a nonlinear system. The extended Kalman filtering algorithm is one of the most effective methods for solving the nonlinear state estimation problem, the basic idea is to perform Taylor expansion on a nonlinear function, omit Gao Jiexiang, reserve expanded low-order terms, realize linearization of the nonlinear function, and finally approximate the state estimation value and the variance estimation value of a system through the Kalman filtering algorithm. The embodiment provides a voltage output method based on combination of an improved extended Kalman filtering algorithm and PID control, which is used for improving the unstable condition of voltage output.

Specifically, a high-pressure stabilization method is proposed in the present embodiment. The method comprises the steps of constructing a first algorithm model, a second algorithm model and a third algorithm model according to a high-voltage output system, and initializing parameters of the first algorithm model, the second algorithm model and the third algorithm model respectively; obtaining a prediction covariance through a state equation of the first algorithm model, and updating the state and the covariance according to an observation equation of the first algorithm model to obtain a process noise variance; obtaining an optimal value of the process noise variance through the second algorithm model, substituting the optimal value of the process noise variance into the first algorithm model for iterative calculation until a relative optimal value of the output high voltage of the high voltage output system is obtained; and inputting the relative optimal value and the voltage set value into the third algorithm model to obtain a voltage output value. The high voltage output system may be a device, an apparatus or a circuit for outputting high voltage.

The high-voltage stabilizing method based on the first algorithm model, the second algorithm model and the third algorithm model can stably output high voltage, reduce high-voltage fluctuation, enable the whole high-voltage system to be safer and more reliable, simultaneously, calculate by adopting a software algorithm, reduce errors, filter inherent error interference on hardware, and have reliability and accuracy compared with filtering by only relying on a hardware circuit. On the other hand, compared with hardware filtering, the method adopting software filtering saves more cost, and is convenient for a developer to carry out improvement and development according to actual conditions. In this embodiment, the first algorithm model may be a kalman filter algorithm, the second algorithm model may be a differential evolution algorithm, and the third algorithm model may be a PID algorithm.

Next, specific steps of the high-pressure stabilization method will be described, as shown in fig. 1, including:

Step 101: and constructing a first algorithm model, a second algorithm model and a third algorithm model according to the high-voltage output system, and initializing parameters of the first algorithm model, the second algorithm model and the third algorithm model respectively.

The step 101 specifically includes: the parameters for initializing the first algorithm model are as follows:，， ; wherein x (0/0) represents a system state estimate at time 0, Q (0/0) represents a process noise covariance matrix at time 0, P (0/0) represents a covariance at time 0, Representing an L x 1-dimensional zero vector, which is the system response length; is L x L dimension identity matrix, delta and Are all given constants; l represents a length; initializing parameters of the second algorithm model, wherein the initialized parameters comprise F, cr, np and G, F is a variation rate, cr is a cross probability, np is a population size, and G is the iteration number; initializing parameters of the third algorithm model, wherein the initialized parameters comprise Kp, ki and Kd, kp is a proportional gain, ki is an integral gain, and Kd is a differential gain.

Taking the second algorithm model as an example of the differential evolution algorithm, initializing parameters f=0.5, cr=0.9, np=10 and g=100 of the differential evolution algorithm. Taking the third algorithm model as an example of the PID algorithm, parameters kp=150, ki=2, kd=50 of the PID algorithm can be initialized.

Step 102: and obtaining a prediction covariance through a state equation of the first algorithm model, and updating the state and the covariance according to an observation equation of the first algorithm model to obtain a process noise variance.

The processing procedure of the first algorithm model mainly comprises two parts: a prediction step and an updating step.

In the predicting step, the Kalman filter predicts the next state using the previous state of the system. This step is mainly divided into two parts:

State prediction: this is in part based on previous states and known behavior of the system (e.g., speed, direction, etc.) to predict what state the system will be in next. For example, tracking a car will estimate where it will next be based on its speed and direction of travel.

Increase in estimation error: since the prediction may never be completely accurate, the prediction step also includes an increase in estimation error. This error reflects the uncertainty of the prediction, meaning that even if the car is considered to be in a certain position, it may in fact deviate slightly from this position.

In the updating step, the Kalman filter uses the new measurement data to correct or update the prediction of the state. The update step can be divided into the following parts:

Calculating Kalman gain: this is a factor that determines the relative importance between the new measured data and the current prediction. If the measured data is very reliable, the Kalman gain will be large, meaning that the new data will have a large impact on the final estimate.

And (5) updating the state: the filter combines the predicted state and the new measured data to generate a more accurate state estimate. If the new measured data shows that the car position is different from the predicted position, the status update takes this difference into account, resulting in a more accurate position estimate.

Updating the estimation error: the kalman filter also updates information about the accuracy of the estimation. If the new measurement data is very reliable, confidence in the predicted state will increase and estimation errors will decrease.

By cycling the prediction step and the update step, the Kalman filter is able to continually integrate new information, correcting and improving the estimation of the system state.

First, a state equation is used to predict, and the current state is predicted to obtain the state at the next moment. Meanwhile, a predicted covariance is calculated by using a predicted value obtained by a state equation, uncertainty of state prediction is represented, an actual observed value is compared with the predicted value through an observation equation, and then the state and the covariance are updated according to the observation equation. This update process takes into account the differences between the actual observations and predictions to correct for the state and covariance. Finally, through this procedure, the latest state after the state update and covariance update can be obtained, and the Cheng Zaosheng variances are calculated therefrom.

Based on the present embodiment, the voltage value output by the high voltage output system needs to be estimated and updated, and in a preferred embodiment, as shown in fig. 2, the step 102 specifically includes:

step 1021: and acquiring a state equation of the first algorithm model, and obtaining a state predicted value and the prediction covariance according to the state equation.

Wherein, the acquiring the state equation is:

;

Wherein, A state transfer function at time k-1; is the input matrix at time k-1; is the process noise at time k-1;

And (3) making:

;

The state prediction value is obtained as follows:

;

The predicted covariance is obtained as:

;

Wherein, An observation matrix at time k-1; is the prediction covariance matrix at time k; is the state transition matrix at time k; is a prediction covariance matrix at k-1; Is that Is a transpose of (2); the process noise covariance matrix at time k-1 is used to represent the uncertainty in the state equation.

Step 1022: and obtaining an observation equation of the first algorithm model, updating a state through the observation equation, obtaining a filter gain through the observation equation, and updating a covariance through the filter gain.

Wherein, the acquisition of the observation equation is as follows:

;

Wherein, Representing the observed value at time k; h is an observation matrix for converting the state vectorMapping to an observation space; A predicted value indicating the time k; indicating observed noise at time k.

Updating the state through the observation equation is as follows:

;

Wherein, Is an estimate of the state at time k,Is a state estimate of the prediction of time k-1,The filter gain at time k.

The filter gain is obtained through the observation equation as follows:

;

Wherein, A filter gain representing time k for combining the observed data with the state estimate; is a prediction covariance matrix at time k-1, representing uncertainty of one-step prediction of the state at time k; Is the observation matrix at the moment k; Is that Is a transpose of (2); Is the observed noise covariance matrix at time k.

Updating covariance by the filter gain to:

;

Wherein the method comprises the steps of Is the updated state covariance matrix; Is the state covariance matrix at time k-1; i is a known identity matrix; The filter gain at time k.

Step 1023: and obtaining the process noise variance according to the state equation and the observation equation.

The process noise obtained according to the state equation and the observation equation is as follows:

;

The process noise variance is:

;

Wherein, Is process noise; Process noise variance at time k; Is an operation to solve the process noise variance.

Step 103: and obtaining an optimal value of the process noise variance through the second algorithm model, substituting the optimal value of the process noise variance into the first algorithm model for iterative calculation until the first algorithm model is ended, and obtaining a relative optimal value of the output high voltage of the high voltage output system.

In a preferred embodiment, as shown in fig. 3, the step 103 mainly includes: initialization, mutation, crossover, selection, and termination.

Initializing: an initial population is randomly generated, each individual in the population being a potential solution to the problem. The size of the population, the structure of the individual, and the range of initial values all depend on the particular problem. The population size should be large enough to ensure sufficient diversity and avoid premature convergence.

Variation: in this step, for each individual (target vector) in the population, three further individuals (base vector, difference vector 1 and difference vector 2) are randomly selected from the population in some way, and then a new individual (variance vector) is generated by performing a difference operation and a weighting operation on the three individuals.

The specific procedure of the initialization and mutation operation refers to step 1031.

Step 1031: randomly initializing a population of solutions of the process noise variance, randomly selecting a preset number of solutions for each solution of the process noise variance, obtaining differences between each solution and other preset number of solutions, and generating a number of solutions.

In this embodiment, the preset number is 3, that is, 3 solutions are randomly selected, so as to obtain the difference between each solution and the 3 solutions, and a number solution is generated, where the number solution is a vector.

The process noise variance solution population initialization method comprises the following steps:

;

Wherein, Is a random number uniformly distributed between 0 and 1,AndUpper and lower bounds for the search;

generating a quantitative solution as:

;

Wherein, Is a variant individual; The value range is 0-1 for the compression scale factor; 、 And Is the parent.

Crossing: the cross operation mixes the target vector and the variance vector to generate a test vector. This is typically done by some sort of interleaving strategy, such as a binary interleaving strategy, which decides with a certain probability for each dimension feature whether to choose the value of the target vector or the value of the variance vector.

Selecting: in each iteration, each test vector is evaluated (its fitness value is calculated). If the fitness value of the test vector is better than that of the target vector, the test vector can replace the target vector to become a new individual in the next iteration; otherwise, the target vector remains unchanged.

The specific procedure for the interleaving and selection operation refers to step 1032.

Step 1032: and performing cross operation on each solution and the number of solutions to generate a test solution, and if the adaptability of the current test solution is better than that of the current solution, replacing the current solution by the current test solution to obtain the optimal value of the process noise variance.

The specific process of generating a test solution includes:

;

wherein r is a uniformly distributed random number of 0-1 generated by each variable; Cross probability as variable; Is an integer evenly distributed between 1~d.

The selection is made to obtain a relative optimum value of the process noise variance according to the following manner:

;

Wherein, For the current solution to be the one,For the current solution of the test to be performed,For the current solutionIs used for the degree of adaptation of the system,For the current test solutionIs used for the adaptation degree of the device.

And (3) terminating: the above steps are repeated until the termination condition is satisfied. The termination condition may be that a maximum number of iterations is reached, or that the fitness value of the optimal individual in the population reaches a certain preset threshold.

Through the steps, the differential evolution algorithm can effectively search the space of the solution and gradually find the global optimal solution of the problem, namely the optimal value of the process noise variance.

The optimum value of the process noise variance is assigned to Qi for the next filtering. Until the kalman filter algorithm ends to obtain a relative optimum value of the voltage. The end condition of the kalman filter algorithm is typically based on a preset stopping criterion or that the desired filtering effect is achieved. Including the possibility to set the running time of the algorithm, and stop the algorithm when the set time limit is reached. Or by comparing the difference between the state estimation errors of two successive iterations, when the error variation is less than a certain threshold, the filtering can be considered to have converged and the algorithm stopped. It should be noted that the selection of the end condition may vary according to the specific application scenario and requirements. In practical application, flexible selection is required according to specific situations.

Step 104: and inputting the relative optimal value and the voltage set value into the third algorithm model to obtain a voltage output value.

And taking the relative optimal value and the voltage set value of the voltage obtained through Kalman filtering calculation as inputs of a PID algorithm, generating a voltage control signal by the PID algorithm, adjusting the voltage output of a high-voltage output system, and maintaining the voltage to be approximate to an expected value. Repeating the steps of calculating and updating until the PID algorithm is finished to obtain a final prediction result, and calculating the final voltage output value by the PID algorithm. Wherein the formula of the incremental PID is:

Wherein, For the output voltage output value of the output,Is a gain factor of a proportion of the gain,In order to integrate the gain factor(s),As a result of the differential gain coefficient,As a result of the error in the error,In order for the error to be a single time,Is a secondary error.

In the present embodiment, the purpose of the PID control using the PID algorithm is to minimize the difference between the set value (voltage set value) and the actual measured value (relative optimum value of the voltage estimated by the kalman filter). First, two input values need to be determined: a set point, which is the target voltage value that the system is expected to achieve, and a measured value. The measured value is the optimal value of the voltage calculated by the kalman filter. An error between the set point and the measured value is calculated. Error = set point-measured value. The goal of PID control is to reduce this error.

Regarding the PID algorithm:

Proportion (P) part: the ratio of the errors is calculated, typically the error multiplied by a scaling factor (P). The larger the error, the stronger the proportional control effect.

Integration (I) part: the integral of the error over time is calculated, typically by multiplying the error integral over time by an integral factor (I). The integral control eliminates steady state errors, i.e., small errors that exist over time.

The derivative (D) part: the differential of the error rate of change is calculated, typically by multiplying the error rate of change by a differential factor (D). Differential control predicts future trends in error, reducing overshoot.

And adding P, I and the D to obtain the output of the PID controller. This output value is a control signal used to adjust the system to reduce errors, such as adjusting a voltage regulator to change the output voltage. The output of the PID controller is used to adjust the system, for example by changing the output voltage of the power supply, and then measuring the actual voltage again and repeating this process. In this way, the PID controller can continuously adjust the voltage to ensure that the output voltage is as close as possible to the set target voltage. The system combining the Kalman filter and the PID controller can effectively cope with system noise and external interference, and improves the accuracy and stability of the whole system.

Example 2:

In embodiment 1, a high-voltage stabilization method is proposed, and in this embodiment, a control apparatus is proposed, as shown in fig. 4, comprising: the device comprises a first control module, a second control module and a transformer module; the first control module is used for receiving the instruction of the upper computer and executing the high-voltage stabilizing method to obtain an output signal; the second control module is used for receiving the output signal, generating a corresponding voltage signal and collecting the voltage signal and transmitting the voltage signal back to the first control module to form a closed loop; the transformer module is used for transforming the voltage signal generated by the second control module to obtain a required voltage value.

The first control module may be a microcontroller, the second control module may be a field programmable gate array (Field Programmable GATE ARRAY, abbreviated as FPGA) chip, as shown in fig. 5, where the control device mainly includes a microcontroller, an FPGA, a transformer, an output module, and a power supply module.

The upper computer is used for carrying out RS485 serial port communication with the microcontroller, and issuing serial port instructions to enable the microcontroller to make corresponding calculation. The microcontroller can adopt an STM32H750 chip, is used for receiving an instruction of an upper computer to generate feedback and performing algorithm calculation in the high-voltage stabilization method, and finally transmits an output result to the FPGA by utilizing fixed mobile fusion (Fixed Mobile Convergence, abbreviated as FMC) communication.

The FPGA is mainly used for receiving signals of the microcontroller, generating corresponding voltage and collecting voltage signals and transmitting the voltage signals back to the microcontroller to form a closed loop. The transformer module is mainly used for transforming the voltage signals generated by the FPGA and finally generating the required voltage values. The output module is mainly used for outputting high voltage. The power supply module is mainly used for supplying power to each module of the control equipment. The power supply module is mainly used for continuously generating stable voltage for the power supply module.

Specific steps of the high-voltage stabilization method are described in embodiment 1, and are not described in detail in this embodiment.

Example 3:

In embodiment 1, a high-voltage stabilization method is proposed, and in this embodiment, a high-voltage stabilization apparatus is proposed, as shown in fig. 6, including a construction module, a first obtaining module, a second obtaining module, and a third obtaining module;

The construction module is used for constructing a first algorithm model, a second algorithm model and a third algorithm model according to the high-voltage output system and initializing parameters of the first algorithm model, the second algorithm model and the third algorithm model respectively; the first obtaining module is used for obtaining a prediction covariance through a state equation of the first algorithm model, and updating the state and the covariance according to an observation equation of the first algorithm model so as to obtain a process noise variance; the second obtaining module is used for obtaining an optimal value of the process noise variance through the second algorithm model, substituting the optimal value of the process noise variance into the first algorithm model for iterative computation until the first algorithm model is finished, and obtaining a relative optimal value of the output high voltage of the high voltage output system; the third obtaining module is used for inputting the relative optimal value and the voltage set value into the third algorithm model to obtain a voltage output value.

Specific steps of the high-voltage stabilization method are described in embodiment 1, and are not described in detail in this embodiment.

The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.

Claims (4)

1. A method of high pressure stabilization comprising:

Establishing a first algorithm model, a second algorithm model and a third algorithm model according to a high-voltage output system, and initializing parameters of the first algorithm model, the second algorithm model and the third algorithm model respectively, wherein the first algorithm model is an improved Kalman filtering algorithm, the second algorithm model is a differential evolution algorithm, and the third algorithm model is a PID algorithm;

obtaining a predicted covariance through a state equation of the first algorithm model, and updating the state and the covariance according to an observation equation of the first algorithm model to obtain a process noise variance, wherein the method specifically comprises the following steps of:

acquiring a state equation of the first algorithm model, and obtaining a state predicted value and the prediction covariance according to the state equation, wherein the method specifically comprises the following steps:

The state equation is obtained as follows:

;

Wherein, A state transfer function at time k-1; is the input matrix at time k-1; is the process noise at time k-1;

And (3) making:

;

The state prediction value is obtained as follows:

;

The predicted covariance is obtained as:

;

Wherein, An observation matrix at time k-1; is the prediction covariance matrix at time k; is the state transition matrix at time k; is a prediction covariance matrix at k-1; Is that Is a transpose of (2); a process noise covariance matrix at time k-1, which is used for representing uncertainty in a state equation;

Obtaining an observation equation of the first algorithm model, updating a state through the observation equation, obtaining a filtering gain through the observation equation, and updating a covariance through the filtering gain, wherein the method specifically comprises the following steps:

the observation equation is obtained as follows:

;

Wherein, Representing the observed value at time k; h is an observation matrix for converting the state vectorMapping to an observation space; A predicted value indicating the time k; Represents observed noise at time k;

updating the state through the observation equation is as follows:

;

Wherein, Is an estimate of the state at time k,Is a state estimate of the prediction of time k-1,The filtering gain at the moment k;

the filter gain is obtained through the observation equation as follows:

;

Wherein, A filter gain representing time k for combining the observed data with the state estimate; is a prediction covariance matrix at time k-1, representing uncertainty of one-step prediction of the state at time k; Is the observation matrix at the moment k; Is that Is a transpose of (2); Is the observed noise covariance matrix at the moment k;

updating covariance by the filter gain to:

;

Wherein the method comprises the steps of Is the updated state covariance matrix; Is the state covariance matrix at time k-1; i is a known identity matrix; The filtering gain at the moment k; obtaining the process noise variance according to the state equation and the observation equation; obtaining an optimal value of the process noise variance through the second algorithm model, wherein the optimal value comprises the following specific steps:

Randomly initializing a population of solutions of the process noise variance, randomly selecting a preset number of solutions for each solution of the process noise variance, obtaining differences between each solution and other preset number of solutions, and generating a number solution; performing cross operation on each solution and the number of solutions to generate a test solution, and if the adaptability of the current test solution is better than that of the current solution, replacing the current solution by the current test solution to obtain an optimal value of the process noise variance;

substituting the optimal value of the process noise variance into the first algorithm model for iterative calculation until the relative optimal value of the output high voltage of the high voltage output system is obtained;

and inputting the relative optimal value and the voltage set value into the third algorithm model to obtain a voltage output value.

2. The high-voltage stabilization method according to claim 1, wherein the constructing a first algorithm model, a second algorithm model and a third algorithm model according to the high-voltage output system, and initializing parameters of the first algorithm model, the second algorithm model and the third algorithm model respectively, specifically comprises:

the parameters for initializing the first algorithm model are as follows:

，，；

Wherein, Representing an estimate of the system state at time 0,Representing the process noise covariance matrix at time 0,The covariance at time 0 is indicated,Representing an L x 1-dimensional zero vector, which is the system response length; for the L x L-dimensional identity matrix, AndAre all given constants; l represents a length;

initializing parameters of the second algorithm model, wherein the initialized parameters comprise F, cr, np and G, F is a variation rate, cr is a cross probability, np is a population size, and G is the iteration number;

Initializing parameters of the third algorithm model, wherein the initialized parameters comprise Kp, ki and Kd, kp is a proportional gain, ki is an integral gain, and Kd is a differential gain.

3. The high voltage stabilization method of claim 1 wherein the deriving the process noise variance from the state equation and the observation equation comprises:

The process noise obtained according to the state equation and the observation equation is as follows:

;

The process noise variance is:

;

Wherein, Is process noise; Process noise variance at time k; Is an operation to solve the process noise variance.

4. A control device for performing the high pressure stabilization method according to any one of claims 1-3, comprising: the device comprises a first control module, a second control module and a transformer module; the first control module is used for receiving the instruction of the upper computer and executing the high-voltage stabilizing method to obtain an output signal;

The second control module is used for receiving the output signal, generating a corresponding voltage signal and collecting the voltage signal and transmitting the voltage signal back to the first control module to form a closed loop;

The transformer module is used for transforming the voltage signal generated by the second control module to obtain a required voltage value.