Intelligent control method for realizing operation of high-confinement mode of magnetic confinement nuclear fusion plasma

Abstract

The invention discloses an intelligent control method for realizing operation of a magnetic confinement nuclear fusion plasma high confinement mode, and the method comprises the steps: obtaining a historical test data set of a magnetic confinement nuclear fusion device, and constructing a secondary heating power threshold calculation model; acquiring input data in real time from a diagnosis system of the magnetic confinement nuclear fusion device, performing calculation based on a secondary heating power threshold calculation model, acquiring a secondary heating power threshold, and feeding back the secondary heating power threshold to the auxiliary heating system; and analyzing the dependency relationship between the secondary heating power threshold value and each plasma state parameter, obtaining a control scheme for reducing the secondary heating power threshold value, providing the control scheme for a plasma control system for adjustment, and entering a next control period. According to the invention, the required secondary heating power can be dynamically estimated according to the real-time states of the plasma and the device, so that the estimation of the secondary heating power threshold is more accurate; and the plasma is adjusted in real time, the required secondary heating power is reduced, and conversion to a high confinement mode is realized with lower engineering cost.

Description

Technical Field

[0001] The present invention relates to the technical field of operation control of magnetic confinement nuclear fusion devices, and in particular to an intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma. Background Art

[0002] High confinement mode (H mode) is an advanced operating mode in magnetic confinement nuclear fusion devices represented by tokamaks. Its main feature is the formation of a steep pressure gradient at the edge of the plasma, which significantly improves the core plasma parameters and supports the realization of higher fusion reaction efficiency.

[0003] The most common method to obtain high confinement mode in experiments is to inject high enough secondary heating power to drive the transformation of low confinement mode to high confinement mode (Plasma Physics and Controlled Fusion.65(1):014001). Internationally, several scaling law formulas for the power threshold required to achieve high confinement mode have been proposed (Journal of Physics:Conference series 123(2008):012033, Nuclear Fusion 54(2014):083003). As long as the secondary heating power significantly greater than the scaling law threshold is injected, the expected confinement mode transformation can be achieved. However, this approach has several problems:

[0004] 1. The calibration law only takes the device operating parameters and plasma macro parameters as inputs, and the estimated conversion power threshold has a large uncertainty. Depending on the level of impurity ions in the plasma during device operation, the heating and feeding methods, etc., the actual required power may deviate greatly from the calibration law, resulting in a strong uncertainty in the design of the control scheme;

[0005] 2. Simply relying on increasing the secondary heating power to achieve H-mode will be greatly restricted by factors such as engineering feasibility and construction capital investment, and may not be feasible in practice;

[0006] 3. The power threshold required to obtain and maintain the H-mode will fluctuate with the change of plasma state. If the heating power is designed only based on the calibration law, the power may not be sufficient to maintain the H-mode at some time during the experiment, resulting in unexpected confinement mode degradation and affecting the stability of plasma control.

[0007] In view of this, this application is hereby filed. Summary of the invention

[0008] The purpose of the present invention is to provide an intelligent control method for realizing the operation of high confinement mode of magnetic confinement nuclear fusion plasma. On the basis of calibration of magnetic confinement nuclear fusion device and plasma macroscopic parameters, the present invention considers more detailed real-time state information of plasma and performs an artificial intelligence algorithm for dynamic high confinement mode conversion threshold estimation. The diagnostic system, secondary heating system (i.e. auxiliary heating system) and plasma control system of the magnetic confinement nuclear fusion device are connected in series with the algorithm to form an intelligent control method for realizing the high confinement mode of the magnetic confinement fusion device. On the one hand, the method of the present invention reduces the secondary heating power required for the high confinement mode by adjusting the plasma state, and on the other hand, it provides a more accurate power threshold estimation than the traditional calibration law, and invests in appropriate secondary heating power, so as to realize the operation of high confinement mode stably and controllably at a lower engineering cost.

[0009] The present invention is achieved through the following technical solutions:

[0010] In a first aspect, the present invention provides an intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma, the method comprising:

[0011] Acquire a historical test data set of a magnetic confinement nuclear fusion device, and construct a secondary heating power threshold calculation model based on the historical test data set;

[0012] Acquire input data from the diagnostic system of the magnetic confinement nuclear fusion device in real time, perform calculations based on the secondary heating power threshold calculation model to obtain the secondary heating power threshold required to achieve a high confinement mode, and feed the secondary heating power threshold back to the auxiliary heating system;

[0013] The dependence between the secondary heating power threshold and each plasma state parameter is analyzed to obtain a control scheme for reducing the secondary heating power threshold, and the control scheme is provided to the plasma control system for adjustment and enters the next control cycle.

[0014] The present invention firstly constructs a secondary heating power threshold calculation model based on the historical test data set of the magnetic confinement nuclear fusion device. Secondly, the model obtains real-time input data from the diagnostic system of the magnetic confinement nuclear fusion device, calculates the secondary heating power threshold required to achieve a high confinement mode, and guides the secondary heating system to input power. Then, the secondary heating power threshold calculation model is subjected to parameter dependency analysis to obtain the maximum influencing factor of the power threshold, thereby providing adjustment variables and adjustment directions for reducing the power threshold, which are controlled by the plasma control system.

[0015] Furthermore, a historical test data set of a magnetic confinement nuclear fusion device is obtained, and a secondary heating power threshold calculation model is constructed according to the historical test data set, including:

[0016] Obtain historical test data sets of magnetic confinement fusion devices;

[0017] From the historical test data set, the magnetic confinement nuclear fusion device and plasma state data are obtained to form input data X; the secondary heating total power data P is collected to analyze whether the plasma can achieve high confinement mode conversion, and the data that has undergone high confinement mode conversion is assigned a label Y=0, and the data that has not undergone high confinement mode conversion is assigned a label Y=1;

[0018] Based on the recurrent neural network and the self-attention neural network, the relationship between the input data X, the secondary heating total power data P and the first output label Y in the historical test data is fitted to obtain a function F satisfying the relationship Y=F(X, P), which is recorded as an intermediate model;

[0019] The intermediate model is inversely solved, the first output label Y is eliminated, and a function G that satisfies the relationship P=G(X) is obtained, that is, a secondary heating power threshold calculation model.

[0020] Furthermore, the intermediate model is inversely solved, the first output label Y is eliminated, and a function G satisfying the relationship P=G(X) is obtained, namely, a secondary heating power threshold calculation model, including:

[0021] For each input data X, scan P and obtain the turning point where Y changes from 0 to 1, and obtain the secondary heating power threshold corresponding to the plasma state of the magnetic confinement nuclear fusion device;

[0022] Based on the recurrent neural network and the self-attention neural network, the relationship between the input data X and the second output label P in the historical test data is fitted to obtain a function G satisfying the relationship P=G(X), that is, a secondary heating power threshold calculation model.

[0023] Furthermore, the input data is data formed by the magnetic confinement nuclear fusion device and plasma state parameters.

[0024] Furthermore, feeding back the secondary heating power threshold to the auxiliary heating system includes:

[0025] Set the heating margin k, and input heating power to the auxiliary heating system according to the value of P×(1+k), where P is the secondary heating power threshold;

[0026] The setting of the heating margin k includes: when a high constrained mode needs to be stably achieved, k=1; when a heating power needs to be saved, k=0.2.

[0027] Furthermore, the control scheme for lowering the secondary heating power threshold includes an adjustment variable and an adjustment direction for lowering the secondary heating power threshold.

[0028] Further, the dependence between the secondary heating power threshold and each plasma state parameter is analyzed to obtain a control scheme for reducing the secondary heating power threshold, and the control scheme is provided to the plasma control system for adjustment, including:

[0029] From the data list contained in the input data X, select the variables that can be adjusted by the plasma control system to form a variable list {x1, x2, ..., xn};

[0030] For each variable xi in the variable list {x1,x2,…,xn}, find the partial derivative To estimate its influence on the required secondary heating power, obtain the variable with the greatest influence on the secondary heating power; where G is the calculation model of the secondary heating power threshold;

[0031] The variable with the greatest influence on the secondary heating power is sent to the plasma control system, which adjusts the direction of the variable xi. Then reduce the variable xi; if the partial derivative Then the variable xi is increased; thereby achieving the effect of reducing the secondary heating power threshold Y.

[0032] Furthermore, the method further comprises:

[0033] After entering the next control cycle, the updated secondary heating power threshold is calculated and fed back to the auxiliary heating system to adjust the plasma in real time, reduce the required secondary heating power, and achieve the transition to a high confinement mode at a lower engineering cost.

[0034] In a second aspect, the present invention provides an intelligent control system for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma, the system comprising:

[0035] An acquisition unit, used for acquiring a historical test data set of a magnetic confinement nuclear fusion device;

[0036] A model building unit, used to build a secondary heating power threshold calculation model according to the historical test data set;

[0037] a calculation unit, configured to obtain input data from a diagnostic system of the magnetic confinement nuclear fusion device in real time, perform calculations based on the secondary heating power threshold calculation model to obtain the secondary heating power threshold required to achieve a high confinement mode, and feed the secondary heating power threshold back to the auxiliary heating system;

[0038] An analysis unit, used for analyzing the dependency between the secondary heating power threshold and each plasma state parameter, and obtaining a control scheme for reducing the secondary heating power threshold;

[0039] The control adjustment unit is used to provide the control scheme to the plasma control system for adjustment and enter the next control cycle.

[0040] Furthermore, the control scheme for lowering the secondary heating power threshold includes an adjustment variable and an adjustment direction for lowering the secondary heating power threshold.

[0041] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0042] Compared with the existing method of analyzing the secondary heating power required for confinement mode conversion by a calibration law and inputting a power significantly higher than a threshold value to realize the high confinement mode, the intelligent control method of the present invention realizes the high confinement mode operation of magnetic confinement nuclear fusion plasma:

[0043] 1. The present invention can dynamically estimate the required secondary heating power according to the real-time status of the plasma and the device, thereby estimating the secondary heating power threshold more accurately;

[0044] 2. The present invention can provide a control scheme for reducing the power threshold for the control system, thereby adjusting the plasma in real time, reducing the required secondary heating power, and achieving the conversion to a high confinement mode at a lower engineering cost;

[0045] 3. The present invention can configure the injection power of the heating system according to the implemented secondary heating power threshold, so as to more stably realize and maintain the high confinement mode and avoid unexpected confinement mode degradation of the plasma. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0047] Figure 1 It is a schematic diagram of the overall technical solution of the present invention;

[0048] Figure 2 A flow chart of the intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to the present invention;

[0049] Figure 3 This is a schematic diagram of the discharge condition of Example 1 of the present invention;

[0050] Figure 4 This is a schematic diagram of using a new P=G(X) function to estimate the secondary heating power threshold in real time in Example 1 of the present invention;

[0051] Figure 5 The present invention is a structural block diagram of an intelligent control system for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma. DETAILED DESCRIPTION

[0052] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0053] The existing method of analyzing the secondary heating power required for the constraint mode conversion through the scaling law and investing a power significantly higher than the threshold to achieve a high constraint mode has the following problems:

[0054] 1. The calibration law only takes the device operating parameters and plasma macro parameters as inputs, and the estimated conversion power threshold has a large uncertainty. Depending on the level of impurity ions in the plasma during device operation, the heating and feeding methods, etc., the actual required power may deviate greatly from the calibration law, resulting in a strong uncertainty in the design of the control scheme;

[0055] 2. Simply relying on increasing the secondary heating power to achieve H-mode will be greatly restricted by factors such as engineering feasibility and construction capital investment, and may not be feasible in practice;

[0056] 3. The power threshold required to obtain and maintain the H-mode will fluctuate with the change of plasma state. If the heating power is designed only based on the calibration law, the power may be insufficient to maintain the H-mode at some time during the test, resulting in unexpected confinement mode degradation and affecting the stability of plasma control.

[0057] To solve the above problems, it is necessary to consider more detailed real-time plasma state information based on the calibration of magnetic confinement nuclear fusion devices and plasma macroscopic parameters, and dynamically estimate the high confinement mode conversion power threshold as the basis for the investment of the secondary heating system. The present invention proposes an effective method for achieving this goal.

[0058] The main points of the present invention, that is, the new technical means adopted are summarized as follows:

[0059] (1) Real-time plasma state data is used in combination with artificial intelligence models with time series analysis capabilities such as recurrent neural networks and self-attention neural networks to perform dynamic calculations of the secondary heating power threshold.

[0060] (2) According to the secondary heating power threshold model P = G(X), by calculating the partial derivative Analyze the variables that have the greatest impact on the secondary heating power threshold and implement real-time control to reduce the power threshold.

[0061] (3) The auxiliary heating system input power is configured according to the dynamic secondary heating power threshold estimated by Y=F(X).

[0062] Specifically, Figure 1 As shown, Figure 1 This is a schematic diagram of the overall technical solution of the present invention. The present invention first constructs a secondary heating power threshold calculation model based on the historical test data set of the magnetic confinement nuclear fusion device. Secondly, the model obtains real-time input data from the diagnostic system of the magnetic confinement nuclear fusion device, calculates the secondary heating power threshold required to achieve a high confinement mode, and guides the secondary heating system to input power; then, a parameter dependency analysis is performed on the secondary heating power threshold calculation model to obtain the maximum influencing factor of the power threshold, thereby providing adjustment variables and adjustment directions for reducing the power threshold, which are controlled by the plasma control system.

[0063] Example 1

[0064] like Figure 2 As shown, the present invention realizes an intelligent control method for high confinement mode operation of magnetic confinement nuclear fusion plasma, the method comprising:

[0065] Step 1, obtaining a historical test data set of a magnetic confinement nuclear fusion device, and constructing a secondary heating power threshold calculation model based on the historical test data set;

[0066] In this embodiment, step 1 specifically includes:

[0067] Step 11, obtaining a historical test data set of a magnetic confinement nuclear fusion device;

[0068] Step 12, obtaining the magnetic confinement nuclear fusion device and plasma state data from the historical test data set to form input data X; the input data X is data formed by the magnetic confinement nuclear fusion device and plasma state parameters;

[0069] Collect the secondary heating total power data P, analyze whether the plasma can achieve high confinement mode conversion, assign a label Y=0 to the data that has high confinement mode conversion, and assign a label Y=1 to the data that has not high confinement mode conversion;

[0070] The source of input data X includes but is not limited to the following Table 1:

[0071] Table 1

[0072] Input physical quantity Measurement methods Device toroidal magnetic field Electromagnetic measurement Plasma current Electromagnetic measurement Plasma density Optical polarimetry measurements Plasma large radius Electromagnetic measurement + data analysis Small plasma radius Electromagnetic measurement + data analysis Plasma elongation ratio Electromagnetic measurement + data analysis Plasma three angles Electromagnetic measurement + data analysis Plasma boundary safety factor Electromagnetic measurement + data analysis Plasma X-point position Electromagnetic measurement + data analysis Plasma temperature Spectroscopic or microwave measurement Plasma speed Microwave measurement Plasma effective atomic number Spectroscopy measurement + data analysis Plasma recirculation level Measure the intensity of the deuterium alpha spectral line and indirectly estimate

[0073] Step 13, based on the recurrent neural network and the self-attention neural network, fit the relationship between the input data X, the secondary heating total power data P and the first output label Y in the historical test data to obtain a function F satisfying the relationship Y=F(X, P), which is recorded as an intermediate model;

[0074] Step 14, inversely solve the intermediate model, eliminate the first output label Y and obtain a function G that satisfies the relationship P=G(X), that is, a secondary heating power threshold calculation model. Step 14 specifically includes:

[0075] A, for each input data X, scan P and obtain the turning point where Y changes from 0 to 1, and obtain the secondary heating power threshold corresponding to the plasma state of the magnetic confinement nuclear fusion device;

[0076] B. Based on the recurrent neural network and the self-attention neural network, the relationship between the input data X and the second output label P in the historical test data set is fitted to obtain a function G that satisfies the relationship P=G(X), i.e., a secondary heating power threshold calculation model.

[0077] Step 2, acquiring input data X in real time from the diagnostic system of the magnetic confinement nuclear fusion device, performing calculation based on the secondary heating power threshold calculation model P=G(X), obtaining the secondary heating power threshold required to achieve a high confinement mode, and feeding back the secondary heating power threshold to the auxiliary heating system;

[0078] In this embodiment, the secondary heating power threshold P is fed back to the auxiliary heating system, and a certain heating margin k is set, and the heating power is input to the auxiliary heating system according to the value of P×(1+k); when it is hoped to stably achieve a high constraint mode, k should be set to a larger value, for example, a value of 1; when it is hoped to save heating power, the k value can be appropriately reduced, for example, a value of 0.2.

[0079] Step 3, analyzing the dependency between the secondary heating power threshold and each plasma state parameter, obtaining a control scheme for reducing the secondary heating power threshold, and providing the control scheme to the plasma control system for adjustment, and entering the next control cycle.

[0080] In this embodiment, the control scheme for lowering the secondary heating power threshold includes an adjustment variable and an adjustment direction for lowering the secondary heating power threshold.

[0081] In this embodiment, step 3 specifically includes:

[0082] (1) Selecting variables that can be adjusted by the plasma control system from the data list contained in the input data X to form a variable list {x1, x2, …, xn};

[0083] (2) For each variable xi in the variable list {x1,x2,…,xn}, find the partial derivative To estimate its influence on the required secondary heating power, obtain the variable with the greatest influence on the secondary heating power; where G is the calculation model of the secondary heating power threshold;

[0084] (3) The variable with the greatest influence on the secondary heating power is sent to the plasma control system, and the plasma control system adjusts the direction of the variable xi. Then reduce the variable xi; if the partial derivative Then the variable xi is increased; thereby achieving the effect of reducing the secondary heating power threshold Y.

[0085] Step 4, after entering the next control cycle, calculate and obtain the updated secondary heating power threshold, and feed the updated secondary heating power threshold back to the auxiliary heating system, adjust the plasma in real time, reduce the required secondary heating power, and achieve the conversion to the high confinement mode at a lower engineering cost.

[0086] In specific implementation, the present invention takes the high confinement mode control of the China Flood Control Experimental Reactor No. 2 A (HL-2A) device as an example, and the specific steps are as follows:

[0087] (1) 275 discharges with H mode and 30 discharges without H mode of HL-2A were collected to form a historical test database, and the input diagnostic data X and secondary heating power P were collected.

[0088] (2) Manually determine the H-mode conversion time point of each experimental gun, and assign a label Y to the data before and after the conversion time point according to the conversion time point. Figure 3 Taking the discharge condition shown as an example, the H-mode conversion occurs at about 670ms and lasts until about 820ms. The typical time delay of the H-mode conversion of the HL-2A device is 10 to 80ms. Therefore, the data before 590ms and after 820ms will be marked as Y=0, and the data during 590ms to 820ms will be marked as Y=1.

[0089] (3) Use a time-domain convolutional neural network (a type of recurrent neural network) to fit the function Y = F (X, P), and then for each X in the data set, scan the threshold P that makes Y convert from 0 to 1, and then use the scan result to fit a new function P = G (X). Figure 4 A schematic diagram of using this function to estimate the secondary heating power threshold in real time is given. In this step, the accuracy of the Y=F(X, P) function can reach 93%, thus ensuring the reliability of the P=G(X) function obtained by scanning. Figure 4 As shown in the figure, if the P=G(X) curve is plotted together with the actual injected power curve, it can be seen that in the early stage of discharge, the actual injected power is lower than the power required for the H mode, so it is not immediately converted to the H mode after the secondary heating injection begins. The subsequent plasma state changes cause the required power to decrease, which is lower than the actual injected power, so the mode conversion occurs around 670ms. After a period of time, the plasma state changes again cause the power required for the H mode to increase, causing the plasma to exit the H mode again.

[0090] (4) Using the automatic differentiation operator provided by the deep learning framework, the controllable parameters on HL-2A, such as the plasma large radius, small radius, density, and current, are calculated. The variables with the greatest impact are obtained and handed over to the device control system for adjustment.

[0091] (5) According to the secondary heating power threshold P calculated in real time, secondary heating is performed at a power of 1.5×P, thereby obtaining a stable high confinement mode plasma.

[0092] Example 2

[0093] like Figure 5 As shown, the difference between this embodiment and embodiment 1 is that this embodiment provides an intelligent control system for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma, and the functions of the system correspond one-to-one with the intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma in embodiment 1; the system includes:

[0094] An acquisition unit, used for acquiring a historical test data set of a magnetic confinement nuclear fusion device;

[0095] A model building unit, used to build a secondary heating power threshold calculation model according to the historical test data set;

[0096] a calculation unit, configured to obtain input data from a diagnostic system of the magnetic confinement nuclear fusion device in real time, perform calculations based on the secondary heating power threshold calculation model to obtain the secondary heating power threshold required to achieve a high confinement mode, and feed the secondary heating power threshold back to the auxiliary heating system;

[0097] An analysis unit, used for analyzing the dependency between the secondary heating power threshold and each plasma state parameter, and obtaining a control scheme for reducing the secondary heating power threshold; wherein the control scheme for reducing the secondary heating power threshold includes a regulating variable and a regulating direction for reducing the secondary heating power threshold;

[0098] The control adjustment unit is used to provide the control scheme to the plasma control system for adjustment and enter the next control cycle.

[0099] Among them, the execution process of each unit can be executed according to the process steps of the intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma in Example 1, and will not be repeated one by one in this embodiment.

[0100] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0101] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0102] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0103] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0104] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma, characterized in that: The method includes: Acquire a historical test data set of a magnetic confinement nuclear fusion device, and construct a secondary heating power threshold calculation model based on the historical test data set; Acquire input data from the diagnostic system of the magnetic confinement nuclear fusion device in real time, perform calculation based on the secondary heating power threshold calculation model to obtain the secondary heating power threshold, and feed the secondary heating power threshold back to the auxiliary heating system; The dependence between the secondary heating power threshold and each plasma state parameter is analyzed to obtain a control scheme for reducing the secondary heating power threshold, and the control scheme is provided to the plasma control system for adjustment and enters the next control cycle.

2. The intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to claim 1, characterized in that: Acquiring a historical test data set of a magnetic confinement nuclear fusion device, and constructing a secondary heating power threshold calculation model based on the historical test data set, including: Obtain historical test data sets of magnetic confinement fusion devices; From the historical test data set, the magnetic confinement nuclear fusion device and plasma state data are obtained to form input data X; the secondary heating total power data P is collected to analyze whether the plasma can achieve high confinement mode conversion, and the data that has undergone high confinement mode conversion is assigned a label Y=0, and the data that has not undergone high confinement mode conversion is assigned a label Y=1; Based on the recurrent neural network and the self-attention neural network, the relationship between the input data X, the secondary heating total power data P and the first output label Y in the historical test data is fitted to obtain a function F satisfying the relationship Y=F(X, P), which is recorded as an intermediate model; The intermediate model is inversely solved, the first output label Y is eliminated, and a function G that satisfies the relationship P=G(X) is obtained, that is, a secondary heating power threshold calculation model.

3. The intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to claim 2, characterized in that: The intermediate model is inversely solved, the first output label Y is eliminated, and a function G satisfying the relationship P=G(X) is obtained, namely, a secondary heating power threshold calculation model, including: For each input data X, scan P and obtain the turning point where Y changes from 0 to 1, and obtain the secondary heating power threshold corresponding to the plasma state of the magnetic confinement nuclear fusion device; Based on the recurrent neural network and the self-attention neural network, the relationship between the input data X and the second output label P in the historical test data set is fitted to obtain a function G satisfying the relationship P=G(X), i.e., a secondary heating power threshold calculation model.

4. The intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to claim 1, characterized in that: The input data are data formed by the magnetic confinement nuclear fusion device and plasma state parameters.

5. The intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to claim 1, characterized in that: Feeding back the secondary heating power threshold to the auxiliary heating system includes: Set the heating margin k, and input heating power to the auxiliary heating system according to the value of P×(1+k), where P is the secondary heating power threshold; The setting of the heating margin k includes: when a high constrained mode needs to be stably achieved, k=1; when a heating power needs to be saved, k=0.

2.

6. The intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to claim 1, characterized in that: The control scheme for reducing the secondary heating power threshold includes a regulating variable and a regulating direction for reducing the secondary heating power threshold.

7. The intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to claim 6, characterized in that: Analyzing the dependency between the secondary heating power threshold and each plasma state parameter, obtaining a control scheme for reducing the secondary heating power threshold, and providing the control scheme to a plasma control system for adjustment, including: From the data list contained in the input data X, select the variables that can be adjusted by the plasma control system to form a variable list {x1, x2, ..., xn}; For each variable xi in the variable list {x1,x2,…,xn}, find the partial derivative To estimate its influence on the required secondary heating power, obtain the variable with the greatest influence on the secondary heating power; where G is the calculation model of the secondary heating power threshold; The variable with the greatest influence on the secondary heating power is sent to the plasma control system, which adjusts the direction of the variable xi. Then reduce the variable xi; if the partial derivative Then increase the variable xi.

8. The intelligent control method for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to claim 1, characterized in that: The method further includes: After entering the next control cycle, the updated secondary heating power threshold is calculated and fed back to the auxiliary heating system to adjust the plasma in real time and reduce the required secondary heating power.

9. An intelligent control system for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma, characterized in that: The system includes: An acquisition unit, used for acquiring a historical test data set of a magnetic confinement nuclear fusion device; A model building unit, used to build a secondary heating power threshold calculation model according to the historical test data set; a calculation unit, configured to obtain input data from a diagnostic system of the magnetic confinement nuclear fusion device in real time, perform calculations based on the secondary heating power threshold calculation model to obtain the secondary heating power threshold, and feed the secondary heating power threshold back to the auxiliary heating system; An analysis unit, used for analyzing the dependency between the secondary heating power threshold and each plasma state parameter, and obtaining a control scheme for reducing the secondary heating power threshold; The control adjustment unit is used to provide the control scheme to the plasma control system for adjustment and enter the next control cycle.

10. The intelligent control system for realizing high confinement mode operation of magnetic confinement nuclear fusion plasma according to claim 9, characterized in that: The control scheme for reducing the secondary heating power threshold includes a regulating variable and a regulating direction for reducing the secondary heating power threshold.

Images