Method for digitally and automatically designing high-frequency cavity of compact cyclotron

Abstract

The invention discloses a method for digitally and automatically designing a high-frequency cavity of a compact cyclotron. The method comprises the following steps: inputting external parameters; setting an invariant parameter and a variable parameter of a cavity model required for calculating the quality factor Q0 of the resonant cavity; for the key point parameters, AI image recognition is utilized, and coordinates of the key points are extracted; key point derivative parameters are automatically calculated by using key point parameters identified by the AI image; forming N groups of parameters for calculating the cavity model according to the parameters in the variable range of each variable parameter; and screening a high-frequency cavity optimal model from the N groups of parameters for calculating the cavity model. According to the method, the key points which can be recognized by the AI image are utilized to automatically calculate the key points which can be recognized by the AI image, the number of cycles of each variable parameter value interval is multiplied, the variable parameters and the immutable parameters are input into simulation software together, and finally the optimal model related to the cavity is found out. The problem that in the prior art, it is difficult to pursue an optimal model by manually designing a cavity model is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of high-frequency cavities of cyclotrons, and particularly relates to a method for digitally and automatically designing a compact cyclotron high-frequency cavity. Background Art

[0002] Today, digital transformation has become a development trend in all industries. In the industrial development of cyclotrons, how to use the rapidly developing artificial intelligence algorithms to upgrade and optimize the design and construction of cyclotrons, save manpower, and improve efficiency is a problem we are exploring.

[0003] The existing design process of the high-frequency cavity of a cyclotron is as follows: The high-frequency cavity designer manually establishes the shell model, Dee plate model, inner rod model, tuning, and coupling model of the high-frequency cavity according to the external parameters: the main magnet model, resonance frequency, and the maximum radius of the estimated beam trajectory; then uses CST to simulate and calculate the frequency of the cavity, and then repeatedly adjusts the gap X between the Dee plate and the shell, the length Y of the inner rod, and the radius Z of the inner rod to make the frequency meet specific requirements;

[0004] The difficulty in manually designing the optimal model lies in that: the gap X between the Dee plate and the shell, the length Y of the inner rod, and the radius Z of the inner rod are all variable parameters, and each variable parameter has a parameter range. There are tens of thousands of combination relationships among the three parameter ranges. When manually selecting each parameter, at most a dozen or so combination relationships are selected, and the values are all selected according to experience. Finally, the one with the highest quality factor is selected as the optimal model among the dozen or so models that meet the frequency requirements. The problem with manually designing the optimal model is that the optimal frequency selected is not the optimal among tens of thousands of combination relationships, but the optimal among a dozen or so combination relationships. Therefore, it is difficult to pursue the optimal model when manually designing the cavity model. Summary of the Invention

[0005] In view of the problems existing in the prior art, the present invention proposes a method for digitally and automatically designing a compact cyclotron high-frequency cavity, aiming to solve the problem that it is difficult to pursue the optimal model when manually designing the cavity model.

[0006] The present invention adopts the following technical solutions to solve its technical problems:

[0007] A method for digitally and automatically designing a compact cyclotron high-frequency cavity is characterized by including the following steps:

[0008] Step 1: Input external parameters, which include the main magnet model file, the high-frequency cavity frequency, and the maximum radius of the estimated beam trajectory;

[0009] Step 2: Set the invariant parameters and variable parameters of the cavity model required to calculate the quality factor Q0 of the resonant cavity. The invariant parameters include key point parameters and key point derivative parameters;

[0010] Step 3: For the key point parameters, use AI image recognition to extract the coordinates of the key points in the main magnet model file;

[0011] Step 4: Automatically calculate the key point derivative parameters using the key point parameters of AI image recognition;

[0012] Step 5: Dynamically set the variable parameters with current relative changes and the variable parameters with relative invariance;

[0013] Step 6: Obtain the parameter values of the variable parameters with current relative changes and the variable parameters with current relative invariance;

[0014] Step 7: For the current variable parameter values, form a set of parameters for calculating the cavity model. Each cavity model can obtain an eigenfrequency and a quality factor through calculation. This set of parameters consists of the key point parameters of AI image recognition, the key point derivative parameters, the parameter values of the variable parameters with current relative changes, and the parameter values of the variable parameters with current relative invariance;

[0015] Step 8: Repeat Steps 5 to 7, and form N sets of parameters for calculating the cavity model according to the parameters within the variable range of each variable parameter;

[0016] Step 9: From the N sets of parameters for calculating the cavity model, screen out several sets of model parameters whose cavity frequencies meet the requirements, and compare to obtain the set of parameters with the highest quality factor, thereby determining the model of the high-frequency cavity.

[0017] Further, for the key point parameters in Step 3, using AI image recognition to extract the coordinates of the key points specifically means: using the AI image recognition algorithm to analyze the three-dimensional model of the main magnet, and extracting the coordinates and relevant dimensions of the points related to the high-frequency cavity in the main magnet:

[0018] 1) Two endpoints A and B of the hypotenuse of the valley region;

[0019] 2) Valley depth C;

[0020] 3) Main magnet radius D;

[0021] 4) Radius E of the opening in the magnet cover plate;

[0022] 5) Position F of the center of the opening in the magnet cover plate;

[0023] 6) Size G of the magnetic gap;

[0024] 7) Maximum radius H of the beam trajectory.

[0025] Further, the specific process of automatically calculating the key point derivative parameters in Step 4 using the key point parameters of AI image recognition is as follows:

[0026] A. Automatically establish the cavity housing model:

[0027] B. Automatically establish the cavity Dee plate model;

[0028] C. Automatically establish the inner rod model in the cavity;

[0029] D. Automatically establish the tuning and coupling device model.

[0030] Furthermore, for the process A of automatically establishing the cavity housing model in step four, the specific process is as follows:

[0031] 1) Taking the center point of the accelerator as the center of the circle, draw an arc with a radius of the main magnet radius D minus 5 mm as the boundary of the cavity tail;

[0032] 2) Parallel to the two endpoints A and B of the hypotenuse of the valley region, take the position 10 mm away from the line segment AB as the hypotenuse boundary of the cavity housing;

[0033] 3) The height of the cavity housing is the valley region depth C minus 5 mm;

[0034] 4) The radius of the cylinder extending from the cavity housing is the radius E of the magnet cover opening minus 10 mm,

[0035] 5) The center of the cylinder is the position F point of the center of the magnet cover opening;

[0036] 6) The height of the cylinder extending from the cavity housing is a parameter related to the variable parameter inner rod length Y,

[0037] Cylinder height = variable parameter inner rod length Y + the size G of the magnetic gap + 25 - the valley region depth C.

[0038] Furthermore, for the process B of automatically establishing the cavity Dee plate model in step four, the specific process is as follows:

[0039] 1) Let the distance between the hypotenuse of the Dee plate and the hypotenuse of the cavity housing be the variable parameter X, and the value range of this variable parameter X is 10 mm to 30 mm;

[0040] 2) The single-sided thickness of the Dee plate is 20 mm;

[0041] 3) The height of the single-sided Dee plate from the accelerator center plane is the size G of the magnetic gap;

[0042] 4) The distance between the tail of the Dee plate and the accelerator center point is the maximum radius H of the beam trajectory + 10 mm.

[0043] Furthermore, for the process C of automatically establishing the inner rod model in the cavity in step four, the specific process is as follows:

[0044] 1) The position F where the center of the inner rod is the center of the opening of the magnet cover plate;

[0045] 2) The length Y of the inner rod is a variable parameter, and the parameter range is 400 mm - 800 mm;

[0046] 3) The radius Z of the inner rod is a variable parameter, and the parameter range is 30 mm - 40 mm.

[0047] Furthermore, for the automatic establishment of the tuning and coupling device model in process D of step four, the specific process is as follows:

[0048] 1) The position of the tuning capacitor plate in the x-axis direction is the maximum radius H + 30 mm of the beam trajectory;

[0049] 2) The size of the tuning capacitor plate is 30 mm × 100 mm × 10 mm;

[0050] 3) The position of the coupling device in the x-axis direction is the maximum radius H + 40 mm of the beam trajectory;

[0051] 4) The radius of the head of the coupling capacitor is 20 mm.

[0052] Furthermore, the specific links from step five to step eight are as follows:

[0053] 1) Set the priority arrangement order of the variable parameters as: the inner rod radius Z, the inner rod length Y, the distance X between the hypotenuse of the Dee plate and the hypotenuse of the cavity housing;

[0054] 2) The interval of X is [10, 30], and let X = 10;

[0055] 3) The interval of Y is [400, 800], and let Y = 400;

[0056] 4) The interval of Z is [30, 40], and let Z = 30;

[0057] 5) According to the key point parameters and key point derivative parameters of AI image recognition, as well as the current values of the variable parameters, simulate and calculate the cavity model to obtain the frequency and quality factor;

[0058] 6) Z = Z + 1

[0059] 7) Whether Z is greater than 40, if not, return to process 5), if yes, continue to process 8);

[0060] 8) Y = Y + 1;

[0061] 9) Whether Y is greater than 800, if not, return to process 4), if yes, continue to process 10);

[0062] 10) Whether X is greater than 30, if not, return to process 3), if yes, end.

[0063] Advantages and effects of the present invention

[0064] 1. The present invention utilizes the key point parameters recognizable by AI images to automatically calculate the parameters required for establishing the cavity housing model, automatically calculate the parameters required for establishing the cavity Dee plate model, automatically calculate the parameters required for establishing the inner rod model of the cavity, and automatically calculate the parameters required for establishing the tuning and coupling device model. Only by importing the external parameter main magnet model file, high-frequency cavity frequency, and the maximum radius of the estimated beam trajectory into the AI image recognition software can the housing model, Dee plate model, inner rod model, and tuning and coupling model of the high-frequency cavity be automatically obtained, laying a foundation for pursuing the optimal model.

[0065] 2. The present invention divides the cavity model parameters into invariant parameters and variable parameters. For the variable parameters, the value ranges are set one by one, and then the values of each variable parameter are set as nested loops. Then, the models established with the current variable parameters and invariant parameters (the key point parameters recognized by AI and the derived parameters together constitute the invariant parameters) in each loop are input into the simulation software, and finally, the optimal model for the cavity design is found, solving the problem that it is difficult to pursue the optimal model in the prior art for manually designing the cavity model. Brief description of the drawings

[0066] Figure 1 is the flowchart of the method for automatically designing the high-frequency cavity of the compact cyclotron according to the present invention;

[0067] Figure 2 is the flowchart of obtaining N groups of cavity models with variable parameters X, Y, and Z according to the present invention;

[0068] Figure 3 is the schematic diagram of the main magnet model according to the present invention;

[0069] Figure 4 is the cavity housing model according to the present invention;

[0070] Figure 5 is the schematic diagram of the overall cavity structure according to the present invention;

[0071] Figure 6 is the schematic diagram of the Dee plate structure according to the present invention;

[0072] Figure 7 is the schematic diagram of the inner rod structure according to the present invention;

[0073] Figure 8 is the schematic diagram of the tuning and coupling device according to the present invention. Detailed implementation manners

[0074] Design principle of the present invention

[0075] 1. Innovation points of the present invention: The innovation lies in finding key point parameters for AI image recognition and extraction, and establishing the cavity shell model as shown in Figure 4 , establishing the cavity Dee plate model as shown in Figure 6 , establishing the inner rod model in the cavity as shown in Figure 7 , and establishing the correlation relationship between the parameters required for the tuning and coupling device model as shown in Figure 8 . By finding this correlation relationship, the parameters required for establishing the cavity shell model, the cavity Dee plate model, the inner rod model in the cavity, and the tuning and coupling device model can be automatically calculated through the key point parameters extracted by AI image recognition, so as to fully automate the shell model, Dee plate model, inner rod model, tuning, and coupling model of the high-frequency cavity that were originally manually established;

[0076] 2. Design principle of dynamically setting variable parameters with current relative changes and relatively invariant variable parameters

[0077] ① Sort the three types of variable parameters X, Y, and Z according to the priority Z, Y, X, and set their initial values respectively: 30, 400, 10; the number of loops within the value range of variable parameter X is 30 - 10 + 1 = 21; the number of loops within the value range of variable parameter Y is 800 - 400 + 1 = 401; the number of loops within the value range of variable parameter Z is 40 - 30 + 1 = 11; the total number of loops is 21×401×11 = 926310000 times. The specific method is as shown in Figure 2 . When the variable parameter Z finishes one loop, the number of loops is 11. At this time, Y = Y + 1. The total number of loops of variable parameter Y is 401. Each time Y + 1, a new round of Z loop is carried out, and each round of Z loop is 11 times. Therefore, when Y loops from 400 to 800, 401 times of Z loops are carried out; when the variable parameter Y finishes one loop, X = X + 1; the total number of loops of X is 21. Each time X + 1, a new round of Y loop and Z loop are carried out. A total of 21 times of Y loops and 21 times of Z loops are carried out. The Y loop is 401 times, and the Z loop is 11 times. Therefore, the total number of loops = 21×401×11 = 92631 times.

[0078] ② The reason for carrying out 92631 loops is that each change in value of 926310000 times will cause changes in the cavity frequency and the quality factor Q0 of the resonant cavity. Only by inputting the cavity frequencies and the quality factors Q0 of the resonant cavity for 92631 times into the simulation software can the simulation software find the optimal cavity model.

[0079] Based on the above principle, the present invention designs a method for digitally automatically designing a compact cyclotron high-frequency cavity as shown in Figure 1 ,Figure 2 As shown, it is characterized by including the following steps:

[0080] Step 1: Input external parameters, which include the main magnet model file (the main magnet model is as Figure 3 shown), the high-frequency cavity frequency, and the maximum radius of the estimated beam trajectory;

[0081] Step 2: Set the invariant parameters and variable parameters of the cavity model required to calculate the cavity quality factor Q0. The invariant parameters include key point parameters and key point derivative parameters;

[0082] Step 3: For the key point parameters, use AI image recognition to extract the coordinates of the key points in the main magnet model file;

[0083] Step 4: Automatically calculate the key point derivative parameters using the key point parameters of AI image recognition;

[0084] Step 5: Dynamically set the currently relatively changing variable parameters and relatively invariant variable parameters;

[0085] Step 6: Obtain the parameter values of the currently relatively changing variable parameters and the currently relatively invariant variable parameters;

[0086] Step 7: For the current variable parameter values, form a set of parameters for calculating the cavity model. Each cavity model can obtain an eigenfrequency and a quality factor through calculation. This set of parameters consists of the key point parameters of AI image recognition, the key point derivative parameters, the parameter values of the currently relatively changing variable parameters, and the parameter values of the currently relatively invariant variable parameters;

[0087] Step 8: Repeat steps 4 to 7, and form N sets of parameters for calculating the cavity model according to the parameters within the variable range of each variable parameter;

[0088] Step 9: From the N sets of parameters for calculating the cavity model, screen out several sets of model parameters whose cavity frequencies meet the requirements, and compare to obtain the set of parameters with the highest quality factor, thereby determining the model of the high-frequency cavity.

[0089] Furthermore, for the key point parameters in step 3, using AI image recognition to extract the coordinates of the key points specifically means: using the AI image recognition algorithm to analyze the three-dimensional model of the main magnet, and extracting the coordinates and relevant dimensions of the points related to the high-frequency cavity in the main magnet:

[0090] ① Two endpoints A and B of the hypotenuse of the valley region;

[0091] ② Valley depth C;

[0092] ③ Main magnet radius D;

[0093] ④Radius E of the opening in the magnet cover plate;

[0094] ⑤Position F of the center of the opening in the magnet cover plate;

[0095] ⑥Magnitude G of the magnetic gap;

[0096] ⑦Maximum radius H of the beam trajectory.

[0097] Furthermore, the specific process of automatically calculating the key point derivative parameters using the key point parameters of AI image recognition in Step 4 is as follows:

[0098] A. Automatically establish a cavity housing model, as Figure 4 shown:

[0099] B. Automatically establish a cavity Dee plate model, as Figure 6 shown;

[0100] C. Automatically establish an inner rod model in the cavity, as Figure 7 shown;

[0101] D. Automatically establish a tuning and coupling device model, as Figure 8 shown.

[0102] Furthermore, the automatic establishment of the cavity housing model in Process A of Step 4, as Figure 4 shown, the specific process is as follows:

[0103] 1) With the center point of the accelerator as the center, draw an arc with a radius equal to the main magnet radius D minus 5 mm as the boundary of the cavity tail;

[0104] 2) Parallel to the two endpoints A and B of the hypotenuse of the valley region, use the position 10 mm away from the line segment AB as the hypotenuse boundary of the cavity housing;

[0105] 3) The height of the cavity housing is the valley region depth C minus 5 mm;

[0106] 4) The radius of the cylinder extending from the cavity housing is the radius E of the opening in the magnet cover plate minus 10 mm,

[0107] 5) The center of the cylinder is the position F point of the center of the opening in the magnet cover plate;

[0108] 6) The height of the cylinder extending from the cavity housing is a parameter related to the length Y of the variable parameter inner rod. Cylinder height = variable parameter inner rod length Y + magnitude G of the magnetic gap + 25 - valley region depth C.

[0109] Furthermore, the automatic establishment of the cavity Dee plate model in Process B of Step 4, as Figure 6 shown, the specific process is as follows:

[0110] ⑴ Let the distance between the hypotenuse of the Dee plate and the hypotenuse of the cavity housing be the variable parameter X, and the value range of this variable parameter X is from 10 mm to 30 mm;

[0111] ⑵ The single-sided thickness of the Dee plate is 20 mm;

[0112] ⑶ The height of the single-sided Dee plate from the accelerator center plane is the size G of the magnetic gap;

[0113] ⑷ The distance between the tail of the Dee plate and the accelerator center point is the maximum radius H + 10 mm of the beam trajectory.

[0114] Further, for the automatic establishment of the inner rod model in process C of step four, as Figure 7 shown, the specific process is as follows:

[0115] 1) The center of the inner rod is at the position F, which is the center of the hole in the magnet cover plate;

[0116] 2) The length Y of the inner rod is a variable parameter, and the parameter range is 400 mm - 800 mm;

[0117] 3) The radius Z of the inner rod is a variable parameter, and the parameter range is 30 mm - 40 mm.

[0118] Further, for the automatic establishment of the tuning and coupling device model in process D of step four, as Figure 8 shown, the specific process is as follows:

[0119] 1) The position of the tuning capacitor plate in the x-axis direction is the maximum radius H + 30 mm of the beam trajectory;

[0120] 2) The size of the tuning capacitor plate is 30 mm × 100 mm × 10 mm;

[0121] 3) The position of the coupling device in the x-axis direction is the maximum radius H + 40 mm of the beam trajectory;

[0122] 4) The radius of the head of the coupling capacitor is 20 mm.

[0123] Further, the specific links from step five to step eight are as follows:

[0124] ① Set the priority arrangement order of the variable parameters as: the radius Z of the inner rod, the length Y of the inner rod, the distance X between the hypotenuse of the Dee plate and the hypotenuse of the cavity housing;

[0125] ② The interval of X is [10, 30], and let X = 10;

[0126] ③ The interval of Y is [400, 800], and let Y = 400;

[0127] ④ The interval of Z is [30, 40], and let Z = 30;

[0128] ⑤Simulate and calculate the cavity model based on the key point parameters and key point derivative parameters of AI image recognition, as well as the current values of the variable parameters, to obtain the frequency and quality factor;

[0129] ⑥Z = Z + 1

[0130] ⑦Is Z greater than 40? If no, return to step 5); if yes, continue;

[0131] ⑧Y = Y + 1;

[0132] ⑨Is Y greater than 800? If no, return to step 4); if yes, continue;

[0133] ⑩Is X greater than 30? If no, return to step 3); if yes, end.

[0134] It should be emphasized that the above specific embodiments are only explanations of the present invention and not limitations thereof. Those skilled in the art can make modifications to the above embodiments without creative contributions according to needs after reading this specification, but as long as they are within the scope of the claims of the present invention, they are protected by the patent law.

Claims

1. A method for digitally automatically designing a compact cyclotron high-frequency cavity, characterized in that: The following steps are involved: Step 1: input external parameters, including the main magnet model file, high-frequency cavity frequency, and the maximum radius of the estimated beam trajectory; Step 2: setting the constant parameters and variable parameters of the cavity model required for calculating the resonant cavity quality factor Q0, wherein the constant parameters include key point parameters and key point derived parameters; Step 3: For the key point parameters, use AI image recognition to extract the coordinates of the key points in the main magnet model file; Step 4: Use the key point parameters of AI image recognition to automatically calculate the key point derivative parameters; Step 5: Dynamically set the current relatively changing variable parameters and relatively unchanged variable parameters; Step 6: Obtain parameter values ​​of the current relatively changing variable parameters and the current relatively unchanged variable parameters; Step 7: For the current variable parameter value, a set of parameters for calculating the cavity model is formed. Each cavity model can obtain an eigenfrequency and a quality factor after calculation. The set of parameters is composed of the key point parameters of AI image recognition, key point derived parameters, current relatively changing variable parameters and current relatively unchanged variable parameters. Step 8, repeating step 54 to step 7, forming N groups of parameters for calculating the cavity model according to the parameters within the variable range of each variable parameter; Step nine: From the N groups of parameters for calculating the cavity model, select several groups of model parameters whose cavity frequencies meet the requirements, compare and obtain a group of parameters with the highest quality factor, and thereby determine the model of the high-frequency cavity.

2. The method for digitally automatically designing a compact cyclotron high-frequency cavity according to claim 1, characterized in that: The step 3 uses AI image recognition to extract the coordinates of the key point parameters, specifically: using the AI ​​image recognition algorithm to analyze the three-dimensional model of the main magnet, extracting the coordinates and relevant dimensions of the points in the main magnet related to the high-frequency cavity: 1) The two endpoints A and B of the hypotenuse of the valley area; 2) Valley depth C; 3) Main magnet radius D; 4) The radius E of the opening of the magnet cover; 5) The position F of the center of the hole in the magnet cover; 6) The size of the magnetic breath G; 7) The maximum radius H of the beam trajectory.

3. The method for digitally automatically designing a compact cyclotron high-frequency cavity according to claim 2, characterized in that: The step 4 uses the key point parameters of AI image recognition to automatically calculate the key point derived parameters. The specific process is as follows: A. Automatically build the cavity shell model: B. Automatically establish the cavity Dee plate model; C. Automatically build the rod model in the cavity; D. Automatically build tuning and coupling device models.

4. The method for digitally automatically designing a compact cyclotron high-frequency cavity according to claim 3, characterized in that: The process A of step 4 is to automatically establish the cavity shell model, and the specific process is as follows: 1) With the center point of the accelerator as the center of the circle, draw an arc with the radius D of the main magnet minus 5 mm as the radius as the boundary of the tail of the cavity; 2) The two endpoints A and B parallel to the hypotenuse of the valley area, with the position 10 mm away from the line segment AB as the hypotenuse boundary of the cavity shell; 3) The height of the cavity shell is the depth C of the valley area minus 5 mm; 4) The radius of the cylinder extending from the cavity shell is the radius E of the opening of the magnet cover minus 10 mm. 5) The center of the cylinder is point F, which is the center of the opening of the magnet cover; 6) The height of the cylinder extending from the cavity shell is a parameter related to the variable parameter inner rod length Y, and the cylinder height = variable parameter inner rod length Y + the size of the magnetic breath G + 25 - the valley depth C.

5. The method for digitally automatically designing a compact cyclotron high-frequency cavity according to claim 3, characterized in that: The specific process of automatically establishing the cavity Dee plate model in process B of step 4 is as follows: 1) Assume that the distance between the hypotenuse of the Dee plate and the hypotenuse of the cavity shell is a variable parameter X, and the value range of the variable parameter X is 10 mm to 30 mm; 2) The thickness of one side of Dee board is 20mm; 3) The height of the single-sided Dee plate from the central plane of the accelerator is the size G of the magnetic breath; 4) The distance between the tail of the Dee plate and the center point of the accelerator is the maximum radius H+10mm of the beam trajectory.

6. The method for digitally automatically designing a compact cyclotron high-frequency cavity according to claim 2, characterized in that: The process C of step 4 is to automatically establish the rod model in the cavity, and the specific process is as follows: 1) The center of the inner rod is the position F of the center of the opening of the magnet cover; 2) The inner rod length Y is a variable parameter with a parameter range of 400mm-800mm; 3) The inner rod radius Z is a variable parameter with a parameter range of 30mm-40mm.

7. The method for digitally automatically designing a compact cyclotron high-frequency cavity according to claim 3, characterized in that: The process D of step 4 is to automatically establish the tuning and coupling device model, and the specific process is as follows: 1) The position of the tuning capacitor plate in the x-axis direction is the maximum radius H+30 mm of the beam trajectory; 2) The size of the tuning capacitor plate is 30mm×100mm×10mm; 3) The position of the coupling device in the x-axis direction is the maximum radius H+40 mm of the beam trajectory; 4) The radius of the coupling capacitor head is 20mm.

8. The method for digitally automatically designing a compact cyclotron high-frequency cavity according to claim 3, characterized in that: The specific steps from step 5 to step 8 are as follows: 1) The priority order of setting the variable parameters is: inner rod radius Z, inner rod length Y, and distance X between the bevel of the Dee plate and the bevel of the cavity shell; 2) The interval of X is [10,30], let X = 10; 3) The interval of Y is [400,800], let Y = 400; 4) The interval of Z is [30,40], let Z = 30; 5) According to the key point parameters and key point derived parameters of AI image recognition and the current values ​​of variable parameters, the cavity model is simulated and calculated to obtain the frequency and quality factor; 6) Z = Z + 1 7) Is Z greater than 40? If not, return to step 5); if yes, continue with step 8); 8) Y = Y + 1; 9) Is Y greater than 800? If not, return to process 4); if yes, continue to process 10); 10) Is X greater than 30? If not, return to step 3; if yes, end.

Images