阅读说明：本技术 一种针对离子电推进系统的连续变推力最优控制系统和方法 (Continuous variable thrust optimal control system and method for ionic electric propulsion system ) 是由 刘晨星 王成飞 方正 杨富全 刘贵忠 耿海 于 2021-08-27 设计创作，主要内容包括：本发明公开了一种针对离子电推进系统的连续变推力最优控制系统和方法,该方法能够在完成推力需求值的同时达到工作效率最优,实现离子电推进系统的最优控制。在具体操作时,控制器在每个控制周期接收上位机下达的推力需求值,并在每个调节周期接收推力检测单元返回的当前推力值,仅依据推力需求值与当前推力值,通过阳极流率开环控制,阳极电流、励磁电流闭环控制以实现推进器的实时最优推力控制。(The invention discloses a continuous variable thrust optimal control system and method for an ionic electric propulsion system. During specific operation, the controller receives a thrust requirement value issued by the upper computer in each control period, receives a current thrust value returned by the thrust detection unit in each regulation period, and controls the anode current and the exciting current in a closed loop mode through anode flow rate open loop control according to the thrust requirement value and the current thrust value so as to realize real-time optimal thrust control of the propeller.)

1. A continuous variable thrust optimal control method for an ionic electric propulsion system is characterized by comprising the following steps:

step 1, establishing a control quantity, a propeller thrust network, a working efficiency network, a thrust network loss function and a working efficiency loss function, training parameters in the thrust network through a loss function of output thrust, training a loss function of the working efficiency network through a loss function of working efficiency, and obtaining the trained thrust network and working efficiency network, wherein the control quantity comprises an anode flow rate, an anode current and an exciting current;

step 2, establishing an optimization model of output thrust through the trained thrust network and working efficiency network, wherein the optimization model aims at minimizing an error value between the output thrust and a thrust requirement and maximizing the working efficiency of a propeller;

step 3, establishing a controller, wherein the controller comprises an optimizer and a regulator, the optimizer is a control network of a controlled variable, the control network of the controlled variable is obtained after parameters are optimized through a loss function of the controlled variable, the loss function of the controlled variable is obtained through an optimization model, and the regulator is an anode flow rate open-loop control and anode current and exciting current closed-loop incremental PID control regulator;

step 4, dividing a thrust control stage into a control period and an adjustment period, and during the control period, issuing a thrust demand instruction by the upper computer; during the regulation period, the controller outputs the real-time optimal control quantity, the whole propeller obtains the optimal thrust, and the working efficiency of the whole propeller is optimal.

2. The optimal control method for the continuous variable thrust of the ionic electric propulsion system as claimed in claim 1, wherein in step 1, the thrust network and the work efficiency network are established through ground test data.

3. The optimal control method for the continuous variable thrust of the ionic electric propulsion system according to claim 1, wherein in the step 1, the thrust network is as follows:

f₁(u；θ₁)＝f₁(u_x,u_y,u_z；θ₁) (2)

the loss function of the thrust network is:

L₁(θ₁)＝(F-f₁(u；θ₁))² (1)；

u is a control quantity, wherein u_xIs the anode flow rate u_yAnodic current u_zF is the excitation current and is the true thrust value corresponding to u.

4. The optimal control method for the continuous variable thrust of the ionic electric propulsion system as claimed in claim 1, wherein in step 1, the work efficiency network is:

f₂(u；θ₂)＝f₂(u_x,u_y,u_z；θ₂) (4)

said work efficiency network loss function is

L₂(θ₂)＝(E-f₂(u；θ₂))² (3)

u is a control quantity, wherein u_xIs the anode flow rate u_yAnodic current u_zAnd E is the excitation current, and the value of the real work efficiency corresponding to u is the E.

5. The optimal control method for the continuous variable thrust of the ionic electric propulsion system according to claim 1, wherein in the step 2, the optimization model is as follows:

wherein N is the number of regulation periods in the control period, F_targetA value is required for thrust; f. of₁(u (k)) and f₂(u (k)) respectively representing the output thrust and the working efficiency of the ionic electric propulsion system in the k regulation period; χ is a terminal constraint set;a system control quantity constraint set.

6. The optimal control method for the continuous variable thrust of the ionic electric propulsion system according to claim 1, wherein in the step 3, the loss function of the control quantity is as follows:

L(θ₃)＝(F_target-f₁(u(F_target；θ₃)))²-λ·f₂(u(F_target；θ₃)) (8)

the above formula is an optimal control network u (f; theta)₃) λ is a hyper-parameter that balances thrust error and work efficiency.

7. The method for optimally controlling the continuously variable thrust of the ionic electric propulsion system according to claim 6, wherein in the step 3, the optimization process of the optimization model comprises the following steps: firstly, modeling input is thrust demand, and a control network u (f; theta) of a control quantity is output₃) According to the loss function, updating control network parameters by using a gradient back propagation algorithm to obtain an optimal control network, discretizing and quantizing the network to obtain an optimization table, and adjusting control quantity along the thrust gradient direction by using a gradient rising method through each single point data in the optimization table to obtain finally determined optimal control quantity corresponding to different thrusts.

8. The optimal control method for the continuous variable thrust of the ionic electric propulsion system as claimed in claim 1, wherein in the step 3, the anode flow rate in the regulator is controlled in an open loop manner by setting the optimal anode flow rate u to be within the corresponding optimal control quantity according to the thrust requirement in the optimizer_xFeeding the anode to a flow rate control unit to gradually adjust the anode to an optimal anode flow rate;

the control process of the anode current and the exciting current is that the anode current and the exciting current are controlled in a closed loop mode by using an incremental PID control method in the process that the anode flow rate is changed slowly.

9. A method for optimal control of continuously variable thrust for an ionic electric propulsion system as set forth in claim 1, characterized in that the function in the regulator is:

e(k)＝F(k)-F(k-1) (10)

wherein e (k) is a thrust error between the k moment and the k-1 moment; Δ u_y(k +1) is the anode current increment, Δ u_z(k +1) is the excitation current increment,is the coefficient of the anodic current proportional term,is the integral term coefficient of the anode current,is the coefficient of the differential term of the anode current,is the coefficient of the proportional term of the exciting current,is the coefficient of the integral term of the exciting current,is the excitation current differential term coefficient.

10. A control system for implementing the continuously variable thrust optimum control method for an ionic electric propulsion system according to claim 1, characterized by comprising:

the function relation establishing module is used for establishing a control quantity, a thrust network and a working efficiency network of the thruster, a thrust network loss function and a working efficiency loss function, training parameters in the thrust network through the loss function of the output thrust, training the loss function of the working efficiency network through the loss function of the working efficiency, and obtaining the trained thrust network and working efficiency network;

the optimization model establishing module is used for establishing an optimization model for outputting thrust through the trained thrust network and the trained work efficiency network;

the controller establishing module is used for establishing a controller, and the controller comprises an optimizer and a regulator;

the control module is used for dividing thrust control into a control period and an adjusting period, and during the control period, the upper computer sends a thrust demand instruction; during the regulation period, the controller outputs the optimal control quantity in real time.

Technical Field

The invention belongs to the field of space plasma electricity, and particularly relates to a continuous variable thrust optimal control system and method for an ionic electric propulsion system.

Background

The main working characteristics of the ionic electric propulsion system are wide-range high-precision continuous variable thrust, and the regulation of the thrust needs to be realized by controlling the combination regulation of the anode flow rate, the anode current and the exciting current. In order to realize the wide-range high-precision continuous adjustment of the thrust, the change of the control quantity of the ion thruster needs to be controlled in real time by means of a specific control algorithm to realize the required thrust requirement value. The change relation between the output thrust and the control quantity of the ion thruster cannot be expressed by the related physical law of plasma discharge at present, and the change relation must be expressed by fitting laws through test data obtained by a large number of detailed tests. The thrust requirement value issued by the upper computer is a function which is unknown in advance and changes randomly along with time, so the system is a typical follow-up system, and the design of the system is mainly used for researching the rapidity and the accuracy of the follow-up reference quantity (the thrust requirement value) of the controlled quantity (the thrust output value).

In recent years, with the continuous improvement of computing power, the deep neural network technology is continuously developed, the deep neural network can learn and adapt to the dynamic characteristics of an uncertain system, and the deep neural network has strong robustness and fault tolerance. Meanwhile, according to the universal approximation theorem, the deep neural network can approximate any continuous function and non-continuous function by connecting a plurality of characteristic values and combining linearity and nonlinearity. Thus, deep neural networks provide a powerful tool for modeling and control of nonlinear systems.

In engineering practice, the most widely used regulator control law is proportional integral derivative control, PID control for short. The PID control is one of the main technologies of industrial control due to its simple structure, good stability, reliable operation and convenient adjustment. However, in the field of ionic electric propulsion, the traditional PID control method only depends on thrust error to perform decision control, does not optimize the working efficiency of the system, cannot ensure that the system is in an optimal working state under any condition, and cannot meet the working requirements of low error and high efficiency (optimal control) of the ionic propeller.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provide a continuous variable thrust optimal control system and method for an ionic electric propulsion system, so as to solve the technical problems that the traditional PID control method in the prior art is not optimized for the working efficiency of the system and is difficult to ensure that the system is in an optimal working state.

In order to achieve the purpose, the invention adopts the following technical scheme to realize the purpose:

a method for optimal control of continuously variable thrust for an ionic electric propulsion system, comprising the steps of:

step 1, establishing a control quantity, a propeller thrust network, a working efficiency network, a thrust network loss function and a working efficiency loss function, training parameters in the thrust network through a loss function of output thrust, training a loss function of the working efficiency network through a loss function of working efficiency, and obtaining the trained thrust network and working efficiency network, wherein the control quantity comprises an anode flow rate, an anode current and an exciting current;

step 2, establishing an optimization model of output thrust through the trained thrust network and working efficiency network, wherein the optimization model aims at minimizing an error value between the output thrust and a thrust requirement and maximizing the working efficiency of a propeller;

step 3, establishing a controller, wherein the controller comprises an optimizer and a regulator, the optimizer is a control network of a controlled variable, the control network of the controlled variable is obtained after parameters are optimized through a loss function of the controlled variable, the loss function of the controlled variable is obtained through an optimization model, and the regulator is an anode flow rate open-loop control and anode current and exciting current closed-loop incremental PID control regulator;

step 4, dividing a thrust control stage into a control period and an adjustment period, and during the control period, issuing a thrust demand instruction by the upper computer; during the regulation period, the controller outputs the real-time optimal control quantity, the whole propeller obtains the optimal thrust, and the working efficiency of the whole propeller is optimal.

The invention is further improved in that:

preferably, in step 1, a thrust network and a work efficiency network are established through ground test data.

Preferably, in step 1, the thrust network is:

f₁(u；θ₁)＝f₁(u_x,u_y,u_z；θ₁) (2)

the loss function of the thrust network is:

L₁(θ₁)＝(F-f₁(u；θ₁))² (1)；

u is a control quantity, wherein u_xIs the anode flow rate u_yAnodic current u_zF is the excitation current and is the true thrust value corresponding to u.

Preferably, in step 1, the work efficiency network is:

f₂(u；θ₂)＝f₂(u_x,u_y,u_z；θ₂) (4)

said work efficiency network loss function is

L₂(θ₂)＝(E-f₂(u；θ₂))² (3)

u is a control quantity, wherein u_xIs the anode flow rate u_yAnodic current u_zAnd E is the excitation current, and the value of the real work efficiency corresponding to u is the E.

Preferably, in step 2, the optimization model is:

wherein N is the number of regulation periods in the control period, F_targetA value is required for thrust; f. of₁(u (k)) and f₂(u (k)) respectively representing the output thrust and the working efficiency of the ionic electric propulsion system in the k regulation period; χ is a terminal constraint set;for controlling the systemAnd (5) constraint collection.

Preferably, in step 3, the loss function of the control amount is:

L(θ₃)＝(F_target-f₁(u(F_target；θ₃)))²-λ·f₂(u(F_target；θ₃)) (8)

the above formula is an optimal control network u (f; theta)₃) λ is a hyper-parameter that balances thrust error and work efficiency.

Preferably, in step 3, the optimization process of the optimization model is as follows: firstly, modeling input is thrust demand, and a control network u (f; theta) of a control quantity is output₃) According to the loss function, updating control network parameters by using a gradient back propagation algorithm to obtain an optimal control network, discretizing and quantizing the network to obtain an optimization table, and adjusting control quantity along the thrust gradient direction by using a gradient rising method through each single point data in the optimization table to obtain finally determined optimal control quantity corresponding to different thrusts.

Preferably, in step 3, the open-loop control process of the anode flow rate in the regulator is to adjust the optimal anode flow rate in the corresponding optimal control quantity according to the thrust requirement in the optimizerThe anode flow rate is transmitted to a flow rate control unit and is gradually adjusted to the optimal anode flow rate;

the control process of the anode current and the exciting current is that the anode current and the exciting current are controlled in a closed loop mode by using an incremental PID control method in the process that the anode flow rate is changed slowly.

Preferably, the function in the regulator is:

e(k)＝F(k)-F(k-1) (10)

wherein e (k) is a thrust error between the k moment and the k-1 moment; Δ u_y(k +1) is the anode current increment, Δ u_z(k +1) is the excitation current increment,is the coefficient of the anodic current proportional term,is the integral term coefficient of the anode current,is the coefficient of the differential term of the anode current,is the coefficient of the proportional term of the exciting current,is the coefficient of the integral term of the exciting current,is the excitation current differential term coefficient.

A control system for implementing the above continuous variable thrust optimal control method for an ionic electric propulsion system, comprising:

the function relation establishing module is used for establishing a control quantity, a thrust network and a working efficiency network of the thruster, a thrust network loss function and a working efficiency loss function, training parameters in the thrust network through the loss function of the output thrust, training the loss function of the working efficiency network through the loss function of the working efficiency, and obtaining the trained thrust network and working efficiency network;

the optimization model establishing module is used for establishing an optimization model for outputting thrust through the trained thrust network and the trained work efficiency network;

the controller establishing module is used for establishing a controller, and the controller comprises an optimizer and a regulator;

the control module is used for dividing thrust control into a control period and an adjusting period, and during the control period, the upper computer sends a thrust demand instruction; during the regulation period, the controller outputs the optimal control quantity in real time.

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

the invention discloses a continuous variable thrust optimal control method for an ionic electric propulsion system, which can achieve optimal working efficiency while completing a thrust requirement value and realize optimal control of the ionic electric propulsion system. The optimal control method of PID continuous variable thrust based on the optimization-regulation structure in the ionic electric propulsion system meets the thrust control requirement of the ionic electric propulsion system by combining the deep neural network technology and the PID control method. During specific operation, the controller receives a thrust requirement value issued by the upper computer in each control period, receives a current thrust value returned by the thrust detection unit in each regulation period, and controls the anode current and the exciting current in a closed loop mode through anode flow rate open loop control according to the thrust requirement value and the current thrust value so as to realize real-time optimal thrust control of the propeller. The invention can adapt to the variable thrust requirement in a certain range aiming at the physical characteristics and the working requirements of the ionic electric propulsion system, jointly optimizes the thrust error and the working efficiency, and can realize the low-error high-efficiency continuous variable thrust optimal control of the ionic electric propulsion system under the condition of meeting the physical constraint of the control quantity.

The invention also discloses a continuous variable thrust optimal control system for the ionic electric propulsion system, which comprises a functional relation establishing module, an optimization model establishing module, a controller establishing module and a control module, wherein when the system is operated specifically, the controller receives a thrust required value issued by an upper computer in each control period, receives a current thrust value returned by the thrust detection unit in each regulation period, and realizes the real-time optimal thrust control of the propeller through anode flow rate open-loop control and anode current and excitation current closed-loop control only according to the thrust required value and the current thrust value.

Drawings

FIG. 1 is a schematic flow diagram of the present invention;

FIG. 2 is a controller layout;

FIG. 3 is a regulator layout;

FIG. 4 is a diagram of the effect of large target thrust control;

FIG. 5 is a graph of the variation of the operating efficiency of the large target thrust;

FIG. 6 is a diagram of the effect of small target thrust control;

FIG. 7 is a graph of small target thrust work efficiency variation;

FIG. 8 is a graph showing the effect of varying the target thrust control;

FIG. 9 is a graph illustrating a variation in the thrust operating efficiency for a varying target;

Detailed Description

The invention is described in further detail below with reference to the accompanying drawings:

in the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention; the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance; furthermore, unless expressly stated or limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly and encompass, for example, both fixed and removable connections; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.

Referring to fig. 1, the optimal control method for PID continuous variable thrust based on optimization-regulation structure in the ionic electric propulsion system according to the present invention comprises the following steps:

1) modeling the ionic electric propeller by utilizing a large amount of ground test data, and determining a control quantity u: anode flow rate u_xAnode current u_yExciting current u_zA functional relationship with the propeller output thrust F and the propeller working efficiency E;

abstracting the ion electric thruster as an input to a control quantity u: anode flow rate u_xAnode current u_yExciting current u_zAnd the output thrust F of the propeller and the working efficiency E of the propeller. And constructing a functional relation by utilizing a multi-layer perceptron MLP model according to ground test data. Because the value range difference of the thrust and the efficiency is large, the thrust and the efficiency are modeled separately, a loss function is set, and a gradient back propagation algorithm is utilized to update model parameters. Wherein the thrust network f₁(u；θ₁) The loss function of (d) is:

L₁(θ₁)＝(F-f₁(u；θ₁))² (1)

f₁(u；θ₁)＝f₁(u_x,u_y,u_z；θ₁) (2)

work efficiency network f₂(u；θ₂) The loss function of (d) is:

L₂(θ₂)＝(E-f₂(u；θ₂))² (3)

f₂(u；θ₂)＝f₂(u_x,u_y,u_z；θ₂) (4)

wherein u is a controlled quantity, wherein u_xIs the anode flow rate u_yAnodic current u_zAnd F is the excitation current, the actual thrust value corresponding to u is F, and the actual working efficiency value corresponding to u is E.

In the network training process, randomly extracting a mini-batch sample from a ground experiment data set every time, and updating a thrust network f according to a loss function (1) and the following formula₁(u；θ₁) Parameters are as follows:

updating the efficiency network f according to the loss function (4) as follows₂(u；θ₂) Parameters are as follows:

where α represents the learning rate and X represents the mini-batch size.

2) Dividing thrust control time sequence of ionic electric propulsion system into control periods T_cAnd adjusting the period T_aOne control cycle comprises a plurality of adjustment cycles. Control period upper computer lower thrust requirement value F_target(ii) a One adjustment of the controlled variable u is completed within the adjustment period. A multi-objective optimization model with constraints is set, and on the premise of meeting the physical limit of the controlled variable, optimal control of continuous time, namely, the thrust error is minimized as much as possible and the propeller efficiency is maximized, is realized;

specifically, the system control period is set to T₁The system regulation period is set to T₂. Setting a controller optimization model in a control period, and establishing the following optimization model by taking the minimum error between output thrust and a thrust requirement value and the maximum working efficiency of a propeller as the targets as possible on the premise that a control quantity meets a physical limit U:

wherein N is the number of regulation cycles in the control cycle, F_targetA value is required for thrust; f. of₁(u (k)) and f₂(u (k)) respectively representing the output thrust and the working efficiency of the ionic electric propulsion system in the k regulation period; χ is a terminal constraint set;a system control quantity constraint set.

3) Aiming at the characteristics of the ionic electric propulsion system,the thrust controller is modeled, and a controller with a unique optimization-regulation structure is designed, namely the controller comprises two parts: an optimizer and a regulator; the optimizer executes each control period according to the thrust requirement value F issued by the upper computer_targetCalculating to obtain the optimal control quantity u corresponding to the thrust demand value^*. In each adjusting period, the adjuster calculates real-time anode current and exciting current according to the thrust requirement value and the current thrust and anode flow rate, so that the thrust reaches the thrust requirement value and the working efficiency is optimal;

as shown in fig. 2, the controller is designed as an optimizer and a regulator for the physical characteristics of the ion electric propulsion system. Wherein the optimizer part: firstly, a neural network model u (f; theta) with target thrust as input and optimal control quantity as output is modeled₃) Updating model parameters by using a gradient back propagation algorithm according to a set loss function formula (8) to obtain an optimal control network, discretizing and quantizing the network to obtain an optimization table, and adjusting control quantity along the thrust gradient direction by using a gradient rising method for each single point data in the optimization table to obtain finally determined optimal control quantity corresponding to different thrustsAnd obtaining the final optimizer.

L(θ₃)＝(F_target-f₁(u(F_target；θ₃)))²-λ·f₂(u(F_target；θ₃)) (8)

The above formula is a loss function of the optimal control network u (f; theta), and lambda is a hyperparameter for balancing thrust error and working efficiency.

In the network training process, the thrust requirement value of the mini-batch size is randomly sampled, and the optimal control network u (f; theta) is updated according to the following formula according to the loss function (8)₃) Parameters are as follows:

where α represents the learning rate and X represents the mini-batch size.

A regulator portion: as shown in FIG. 3, aiming at the characteristic of slow regulation of the anode flow rate in the ionic electric propulsion system, a regulator for anode flow rate open-loop control and anode current and exciting current closed-loop incremental PID control is designed. Wherein, the anode flow rate control: according to the thrust requirement in the optimizer at the corresponding optimal control quantity, the optimal anode flow rate is obtainedAnd the anode flow rate is sent to an anode flow rate control unit so as to be gradually adjusted to the optimal anode flow rate. Controlling the anode current and the exciting current: in the process of slowly changing the anode flow rate, the anode current and the exciting current are controlled in a closed loop mode by an incremental PID control method, and the output thrust of the system is adjusted to a thrust requirement value quickly and stably. The specific functions in the regulator are:

e(k)＝F(k)-F(k-1) (10)

wherein e (k) is a thrust error between the k moment and the k-1 moment; Δ u_y(k +1) is the anode current increment, Δ u_z(k +1) is the excitation current increment,is the coefficient of the anodic current proportional term,is the integral term coefficient of the anode current,is the coefficient of the differential term of the anode current,is the coefficient of the proportional term of the exciting current,is the coefficient of the integral term of the exciting current,is the excitation current differential term coefficient.

The closed-loop PID control of the anode current and the exciting current is to input the current thrust value and the thrust requirement value in each adjusting period and control the anode current and the exciting current according to the difference between the current thrust value and the thrust requirement value. These two quantities can respond quickly. The open-loop control of the anode flow rate only needs to be started in a control period, and the optimizer obtains the corresponding optimal anode flow rate according to the thrust requirement value issued by the upper computer. Acting on the flow rate control unit to slowly change itself. The effect of the optimizer is to solve the optimal anode flow rate and open-loop adjust the anode flow rate. The regulator is used for closed-loop regulation according to the current thrust and the target thrust.

Through the setting, after the parameters of the optimizer and the regulator in the controller are respectively determined, in the specific implementation process, the optimal control quantity of the ionic electric propeller can be obtained only by inputting the thrust requirement value and the current thrust into the controller.

4) After the controller with the optimized-adjusted structure is obtained, in a specific application process, in each control period, the upper computer issues a thrust demand instruction, and in each adjustment period, the controller outputs a corresponding control quantity instruction, so that the optimal control of the continuous variable thrust of the ion electric thruster is realized.

The following further analysis is made in conjunction with the specific examples:

examples

Suppose an ion electric propulsion system includes a propeller having an input of an anode flow rate u and a controller_xAnode current u_yExciting current u_zThe output thrust range is 1 mN-20 mN; in each control period T_cLower thrust requirement value F of upper computer_targetSetting T_c0.1 s; each control cycle is divided into 50 regulation cycles T_aSetting T_aThe control quantity is adjusted once in each adjusting period at 0.002s, and the optimizer in the controller obtains the optimal anode flow rate according to the thrust demand valueActing on the flow rate control unit; and a regulator in the controller dynamically controls the anode current and the exciting current according to the thrust demand value and the current thrust and anode flow rate.

Under the condition of meeting the physical limitation of the control quantity, the invention sets the following optimization model by taking the minimum output thrust and the required thrust value error in the control period and the maximization and the system working efficiency as a combined target:

the optimal control method for PID continuous variable thrust based on an optimization-regulation structure in the ionic electric propulsion system comprises the following steps:

1) modeling the ionic electric propeller by utilizing a large amount of ground test data, and determining a control quantity u: anode flow rate u_xAnode current u_yExciting current u_zAn evolution relation between the output thrust F of the propeller and the working efficiency E of the propeller;

step 11): abstracting the ion electric thruster as an input to a control quantity u: anode flow rate u_xAnode current u_yExciting current u_zAnd the output thrust F of the propeller and the working efficiency E of the propeller. And constructing a functional relation by utilizing a multi-layer perceptron MLP model according to ground test data. Because the value range difference of the thrust and the efficiency is large, the thrust and the efficiency are modeled separately, a loss function is set, and a gradient back propagation algorithm is utilized to update model parameters. Wherein the thrust network f₁(u；θ₁) The loss function of (d) is:

L₁(θ₁)＝(F-f₁(u；θ₁))² (1)

f₁(u；θ₁)＝f₁(u_x,u_y,u_z；θ₁) (2)

work efficiency network f₂(u；θ₂) The loss function of (d) is:

L₂(θ₂)＝(E-f₂(u；θ₂))² (3)

f₂(u；θ₂)＝f₂(u_x,u_y,u_z；θ₂) (4)

wherein u is a controlled quantity, wherein u_xIs the anode flow rate u_yAnodic current u_zFor the exciting current, F is the real thrust value corresponding to u, E is the real work efficiency value corresponding to u, theta₁As parameters of the thrust network model, theta₂Are the operating efficiency model parameters.

In the network training process, randomly extracting a mini-batch sample from a ground experiment data set every time, and updating a thrust network f according to a loss function (1) and the following formula₁(u；θ₁) Parameters are as follows:

updating the efficiency network f according to the loss function (4) as follows₂(u；θ₂) Parameters are as follows:

where α represents the learning rate and X represents the mini-batch size.

2) Dividing thrust control time sequence of ionic electric propulsion system into control periods T_cAnd adjusting the period T_aOne control cycle comprises a plurality of adjustment cycles. Control period upper computer lower thrust requirement value F_target(ii) a One adjustment of the controlled variable u is completed within the adjustment period. Multi-objective optimization of set band constraintsThe model realizes the optimal control of continuous time, namely minimizing the thrust error and maximizing the propeller efficiency as far as possible on the premise of meeting the physical limit of the control quantity;

step 12): the system control period was set to 0.1s and the system regulation period to 0.002 s. Setting a controller optimization model in a control period, and on the premise that the control quantity meets a physical limit U, taking the goals of minimizing the error between the output thrust and the thrust requirement value as much as possible and maximizing the working efficiency of the propeller as follows:

wherein N is the number of regulation cycles in the control cycle, namely N is 50, F_targetA value is required for thrust; f. of₁(u (k)) and f₂(u (k)) respectively representing the output thrust and the working efficiency of the ionic electric propulsion system in the k regulation period; χ is a terminal constraint set;a system control quantity constraint set.

3) Aiming at the characteristics of the ionic electric propulsion system, a thrust controller is modeled, and a controller with a unique optimization-regulation structure is designed, namely the controller comprises two parts: an optimizer and a regulator; the optimizer executes each control period according to the thrust requirement value F issued by the upper computer_targetCalculating to obtain the optimal control quantity u corresponding to the thrust demand value^*. In each adjusting period, the adjuster calculates real-time anode current and exciting current according to the thrust requirement value and the current thrust and anode flow rate, so that the thrust reaches the thrust requirement value and the working efficiency is optimal;

step 13): aiming at the physical characteristics of the ionic electric propulsion system, the controller is designed into an optimizer and a regulator. Wherein the optimizer part: firstly, a neural network model u (f; theta) with target thrust as input and optimal control quantity as output is modeled₃) Updating model parameters by using a gradient back propagation algorithm according to a set loss function formula (8),obtaining an optimal control network, carrying out discrete quantization on the network to obtain an optimization table, and adjusting the control quantity along the thrust gradient direction by using a gradient rising method for each single-point data in the optimization table to obtain the finally determined optimal control quantity corresponding to different thrustsAnd obtaining the final optimizer.

L(θ₃)＝(F_target-f₁(u(F_target；θ₃)))²-λ·f₂(u(F_target；θ₃)) (8)

The above formula is an optimal control network u (f; theta)₃) λ is a hyper-parameter that balances thrust error and work efficiency, θ₃And controlling the parameters of the network model for the optimal.

In the network training process, the thrust requirement value of the mini-batch size is randomly sampled, and the optimal control network u (f; theta) is updated according to the following formula according to the loss function (8)₃) Parameters are as follows:

where α represents the learning rate and X represents the mini-batch size.

A regulator portion: aiming at the characteristic of slow regulation of anode flow rate in an ionic electric propulsion system, a regulator for anode flow rate open-loop control and anode current and exciting current closed-loop incremental PID control is designed. Wherein, the anode flow rate control: according to the thrust requirement in the optimizer at the corresponding optimal control quantity, the optimal anode flow rate u_x ^*And fed to a flow rate control unit to gradually adjust to the optimal anode flow rate. Controlling the anode current and the exciting current: in the process of slowly changing the anode flow rate, the anode current and the exciting current are controlled in a closed loop mode by an incremental PID control method, and the output thrust of the system is adjusted to a thrust requirement value quickly and stably.

e(k)＝F(k)-F(k-1) (10)

Wherein e (k) is a thrust error between the k moment and the k-1 moment; Δ u_y(k +1) is the anode current increment, Δ u_z(k +1) is the excitation current increment,is the coefficient of the anodic current proportional term,is the integral term coefficient of the anode current,is the coefficient of the differential term of the anode current,is the coefficient of the proportional term of the exciting current,is the coefficient of the integral term of the exciting current,is the excitation current differential term coefficient.

4) After the controller with the optimized-adjusted structure is obtained, in a specific application process, in each control period, the upper computer issues a thrust demand instruction, and in each adjustment period, the controller outputs a corresponding control quantity instruction, so that the optimal control of the continuous variable thrust of the ion electric thruster is realized.

Step 14): after the parameters of the optimizer and the regulator in the controller are respectively determined, in the specific implementation process, the optimal control quantity of the ionic electric propeller can be obtained only by inputting the thrust required value and the current thrust into the controller.

Simulation setup and analysis of experimental results are given below.

Simulation parameter setting

The specific simulation parameters of the simulation platform using DELL powerEdge (DELL-R940XA,4 GOLD-5117, RTX2080Ti) are shown in tables 1, 2 and 3. .

TABLE 1 thrust model and work efficiency training parameter settings

TABLE 2 optimizer training parameter settings

Number of training cycles	10000
		Batch size	1900
Learning rate	0.001
		Learning rate cosine anneal decay period	1000
Optimizer	Adam

TABLE 3 regulator parameter settings

Results and analysis of the experiments

Fig. 4, 5, 6, 7, 8 and 9 show the control effect of the invention on the thrust and work efficiency of the ionic electric propulsion system. From fig. 4 and 5, it can be seen that the algorithm of the present invention can realize rapid and accurate adjustment of a large target thrust, and is finally stabilized in a good working state, and no overshoot occurs, and it takes about 1-2 s to achieve a thrust requirement value. As can be seen from fig. 6 and 7, the algorithm provided by the present invention can realize fast and accurate adjustment of small target thrust, and is finally stabilized in a relatively good working state. In the experiment, the overshoot phenomenon is generated, the overshoot is small and is about 1mN, and the thrust requirement value is about 10 ms. As can be seen from fig. 8 and 9, the algorithm of the present invention can quickly and accurately follow the change of the target thrust when the change of the target thrust is changed, and a certain response time is required in the thrust up-regulation process, but the thrust down-regulation process can be almost instantaneously completed, and the whole control process is performed under the state of high work efficiency.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.