CN111709183B - An optimization method for neutron tube acceleration system based on genetic algorithm - Google Patents

Abstract

The invention discloses a neutron tube acceleration system optimization method based on a genetic algorithm, which comprises the steps of firstly taking geometric structural parameters of a neutron tube acceleration system as parameters to be optimized, randomly generating N binary numbers to represent the parameters to be optimized to form an individual, constructing an initial population, constructing a geometric model of the neutron tube acceleration system, calculating beam current performance indexes of the neutron tube acceleration system corresponding to the genetic individual, and establishing a target function; measuring fitness of the genetic individual by using the objective function value; carrying out iteration to carry out genetic operation, generating a new population, calculating the fitness of each individual of the new population until the iteration stopping condition is met, and outputting an optimal solution as a value of a parameter to be optimized; the invention calculates the beam performance index of the neutron tube accelerating system by utilizing a finite element method, then applies a genetic algorithm, and takes the beam performance of the neutron tube accelerating system as a target to globally optimize the geometric structure parameter of the neutron tube accelerating system, thereby obtaining the neutron tube accelerating system under the optimized geometric structure parameter.

Description

An optimization method for neutron tube acceleration system based on genetic algorithm

Technical field

The invention belongs to the technical field of neutron tube acceleration systems, and particularly relates to a neutron tube acceleration system optimization method based on genetic algorithms.

Background technique

The neutron tube is a small accelerator, which completely seals the ion source, accelerator, target and storage in a glass vacuum tube or ceramic vacuum tube; it has the advantages of compact structure, high safety and easy use; currently, neutron tubes are widely used It is used in fields such as neutron photography, neutron therapy and petroleum logging.

The neutron tube acceleration system draws out the beam from the ion source plasma, and focuses and accelerates the beam through the electric field force, which affects the beam performance that finally reaches the target surface; among them, the beam intensity and uniformity are the parameters for evaluating the beam performance. Important indicators; in the existing technology, the finite element simulation method is usually used to optimize the parameters of the acceleration system structure to improve the uniformity of the beam, thereby improving the performance of the neutron tube; in the existing technology, only finite element simulation is used When optimizing parameters using this method, the optimal solution is often not obtained, resulting in low beam uniformity, damaging the service life and performance of the target, and thus affecting the life and performance of the neutron tube.

Contents of the invention

In view of the technical problems existing in the prior art, the present invention provides a neutron tube acceleration system optimization design method and a neutron tube acceleration system to solve the problem that the prior art only uses the finite element simulation method to optimize the neutron tube acceleration system parameters. At this time, the optimal solution is often not obtained, resulting in low beam uniformity, which in turn affects the technical problems of the life and performance of the neutron tube.

In order to achieve the above objects, the technical solutions adopted by the present invention are:

The invention provides a neutron tube acceleration system optimization method based on genetic algorithm, which includes the following steps:

Step 1. Initialization

Taking the geometric structure parameters of the neutron tube acceleration system as the parameters to be optimized, N binary numbers are randomly generated to represent the parameters to be optimized and become an individual to form the initial population;

Step 2. Calculate individual fitness

Convert the obtained binary numbers to generate the geometric structure parameters of the neutron tube acceleration system expressed in decimal system; construct a geometric model of the neutron tube acceleration system based on the geometric structure parameters of the neutron tube acceleration system expressed in decimal system; use the finite element method to calculate the genetic The beam performance index of the individual corresponding neutron tube acceleration system;

According to the beam performance index of the neutron tube acceleration system, an objective function is established; then the fitness of the genetic individual is measured by the value of the objective function;

Step 3: Perform genetic operations iteratively to generate a new population, calculate the fitness of each individual in the new population until the iteration stop condition is met, and output the optimal solution when the iteration stop condition is met as the value of the parameter to be optimized, and follow the steps to be taken. Optimize the values of parameters to create a neutron tube acceleration system.

Further, in step 1, the parameters to be optimized include the accelerating gap of the neutron tube accelerating system, the accelerating electrode pole diameter, the accelerating electrode length, the lead pole aperture, the magnetic ring slope diameter, the magnetic cylinder slope aperture and the accelerating electrode aperture.

Further, in step 2, the beam performance indicators of the neutron tube acceleration system corresponding to the genetic individual include beam unevenness, beam miss number and beam radius;

The mathematical expression of the objective function is:

Target function ObjV=target/base+outtarget;

Among them, target is the beam unevenness; outtarget is the number of beam misses, and base is the beam radius; the smaller the objective function value, the better the fitness of the genetic individual;

Among them, the mathematical expression of beam unevenness target is:

Among them, o is the non-uniformity; Z _x is the number of particles in a certain ring x on the target surface under ideal conditions; A _x is the area of a certain ring x on the target surface; A _is the total area of the target surface; U _x is the actual number of particles in a certain ring x on the target surface; x=1,2,…,n.

Furthermore, the finite element method and Matlab software are used to solve the objective function.

Further, in step 3, genetic operations include selection, crossover and mutation; among them, the selection operation uses a mechanism combining elite selection and roulette based on the fitness value of the genetic individual; the crossover operation uses a uniform crossover method and randomly selects individuals. Implement row crossover or column crossover; mutation operation uses bit mutation mechanism.

Further, in step 3, the iteration stop condition is met to reach the maximum evolutionary generation, and the optimal solution at this time is the parameter to be optimized corresponding to the individual with the smallest fitness in the current population.

Further, in step 3, the iteration stop condition is satisfied when the fitness reaches the set requirements. At this time, the optimal solution is the parameters to be optimized corresponding to the individuals whose fitness reaches the set requirements.

Compared with the prior art, the beneficial effects of the present invention are:

The present invention provides a neutron tube acceleration system optimization method based on a genetic algorithm. The finite element method is used to calculate the beam performance index of the neutron tube acceleration system, and then the genetic algorithm is applied to calculate the beam performance of the neutron tube acceleration system as The goal is to globally optimize the geometric structure parameters of the neutron tube accelerating system; this invention automatically searches for the globally optimal structural size according to the performance index requirements of the neutron tube accelerating system, thereby obtaining the neutron under optimized geometric structure parameters. Tube acceleration system; solves the problem that the optimal solution is often not obtained when optimizing the parameters of the neutron tube acceleration system using only finite element simulation methods, significantly improves the beam performance of the neutron tube; and saves a lot of time in program selection. , improves work efficiency and effectively improves the service life and performance of the neutron tube.

Furthermore, the extraction pole side and the accelerating pole side of the accelerating area in the neutron tube accelerating system have a greater impact on the focusing of the particle beam. According to the results of the potential and particle trajectories, the accelerating gap and accelerating electrode in the neutron tube accelerating system are The extremely small diameter, the length of the accelerating electrode, the diameter of the extraction pole, the diameter of the magnetic ring slope, the diameter of the magnetic cylinder slope and the aperture of the accelerating electrode are the parameters to be optimized, which can effectively improve the beam performance of the neutron tube acceleration system, thereby improving the neutron tube's Service life and performance; the geometric structure parameters of the above seven neutron tube acceleration systems form an individual genetic algorithm. The genetic algorithm selects cross-mutation based on the fitness of the individual. The individual genes with good fitness have a greater probability of being inherited by future generations. Genetic Going down is the optimal solution.

Furthermore, the beam unevenness, beam miss number and beam radius are selected as the beam performance indicators of the neutron tube acceleration system, which can intuitively reflect the beam performance. The quality of the acceleration system can be measured through the beam unevenness, beam radius. It is evaluated by the flow radius and the number of beam misses. The smaller the value of the objective function, the better the individual.

Furthermore, the objective function is jointly calculated using the finite element method and Matlab software, fully combining the advantages of COMSOL simulation and MATLAB programming. It can be widely adapted to the optimization needs of magnetic nanoparticle simulation test platforms using magnetic nanoparticle devices of different sizes, and can be used through different Parameter transfer between software directly generates and saves magnetic nanoparticle measurement simulation models that meet the requirements.

In summary, the present invention uses the finite element method, the genetic algorithm and the uniformity evaluation method to combine each other, and uses COMSOL and MATLAB software to optimize the design of the neutron tube acceleration system, and finally obtains an acceleration system structure with better performance, improving The performance of neutron tubes is understood and used to guide the design and manufacture of neutron tubes, which is highly practical.

Description of drawings

Figure 1 is a schematic flow chart of the neutron tube acceleration system optimization method according to the present invention;

Figure 2 is a flow chart of the individual fitness evaluation method in the neutron tube acceleration system optimization method according to the present invention;

Figure 3 is a distribution diagram of beam particles on the target surface of the optimal solution neutron tube acceleration system described in the embodiment;

Figure 4 is a schematic diagram of the convergence of the genetic algorithm of the neutron tube acceleration system described in the embodiment.

Detailed ways

In order to make the technical problems, technical solutions and beneficial effects solved by the present invention clearer, the following specific examples will further describe the present invention in detail. It should be understood that the specific embodiments described here are only used to explain the present invention and are not intended to limit the present invention.

As shown in Figures 1 and 2, a neutron tube acceleration system optimization method based on genetic algorithm includes the following steps:

Step 1. Initialization

Taking the geometric structure parameters of the neutron tube acceleration system as the parameters to be optimized, N binary numbers are randomly generated to represent the parameters to be optimized and become an individual to form the initial population; among them, the parameters to be optimized include the acceleration gap and acceleration of the neutron tube acceleration system. The diameter of the electrode pole, the length of the accelerating electrode, the aperture of the lead pole, the diameter of the magnetic ring slope, the diameter of the magnetic cylinder slope and the aperture of the accelerating electrode.

Step 2. Calculate individual fitness

Convert the obtained binary numbers to generate the geometric structure parameters of the neutron tube acceleration system expressed in decimal system; construct a geometric model of the neutron tube acceleration system based on the geometric structure parameters of the neutron tube acceleration system expressed in decimal system; use the finite element method to calculate the genetic The beam performance indicators of the individual corresponding neutron tube accelerating system; among them, the beam performance indicators of the neutron tube accelerating system include beam unevenness, beam miss number and beam radius.

According to the beam performance evaluation index of the neutron tube acceleration system, an objective function is established; then the objective function value is used to measure the fitness of the genetic individual; among them, the mathematical expression of the objective function is:

Target function ObjV=target/base+outtarget;

Among them, target is the beam unevenness; outtarget is the number of beam misses, and base is the beam radius; the smaller the objective function value, the better the fitness of the genetic individual;

Among them, the mathematical expression of beam unevenness target is:

Among them, o is the non-uniformity; Z _x is the number of particles in a certain ring x on the target surface under ideal conditions; A _x is the area of a certain ring x on the target surface; A _is the total area of the target surface; U _x is the actual number of particles in a certain ring x on the target surface; x=1,2,…,n.

The number of beam misses is based on the finite element simulation calculation results, reading the three-dimensional coordinates of the particles, and counting whether the coordinates are obtained on the target.

The beam radius is obtained by calculating the distance from the particle coordinates on the target surface to the target center coordinates.

Step 3: Perform genetic operations iteratively to generate a new population, calculate the fitness of each individual in the new population until the iteration stop condition is met, and output the optimal solution when the iteration stop condition is met as the value of the parameter to be optimized, and follow the steps to be taken. Optimize the values of parameters to create a neutron tube acceleration system; among them, genetic operations include selection, crossover and mutation; among them, the selection operation uses a combination of elite selection and roulette based on the fitness value of the genetic individual; the crossover operation uses In the uniform crossover method, individuals are randomly selected to perform row crossover or column crossover; the mutation operation adopts a bit mutation mechanism; when the iteration stop condition is met to reach the maximum evolutionary generation or the fitness reaches the set requirements, when the iteration stop condition is met to reach the maximum evolutionary generation, The optimal solution is the parameter to be optimized corresponding to the individual with the smallest fitness in the current population; when the iteration stop condition is met and the fitness reaches the set requirement, the optimal solution is the parameter to be optimized corresponding to the individual whose fitness reaches the set requirement. parameter.

The present invention provides a neutron tube acceleration system optimization method based on a genetic algorithm. The finite element method is used to calculate the beam performance index of the neutron tube acceleration system, and then the genetic algorithm is applied to calculate the beam performance of the neutron tube acceleration system as The goal is to globally optimize the geometric structure parameters of the neutron tube accelerating system; this invention automatically searches for the globally optimal structural size according to the performance index requirements of the neutron tube accelerating system, thereby obtaining the neutron under optimized geometric structure parameters. Tube acceleration system; it solves the problem that the optimal solution is often not obtained when optimizing the parameters of the neutron tube acceleration system using only finite element simulation methods. It significantly improves the beam performance of the neutron tube acceleration system and saves a lot of solution selection. Allocation time improves work efficiency and effectively improves the service life and performance of the neutron tube.

Example

The structures related to particle acceleration in the neutron tube acceleration system mainly include the extraction electrode, the accelerating electrode and the target. The extraction electrode is grounded, and the accelerating electrode applies high voltage; a potential difference is formed between the extraction electrode and the accelerating electrode, and the particles at the outlet of the extraction electrode are affected by the electric field. Moves toward the acceleration pole and forms a particle beam; after the acceleration process, due to inertia, the particle beam continues to move toward the target. In the process of drifting toward the target, the particle beam radius becomes larger and larger. When the particle beam radius becomes larger, , the larger the effective area of the target, the longer the life of the target under the same conditions, and thus the longer the service life of the neutron tube; the finite element and genetic algorithm are used to parameterize the geometric structure of the neutron tube acceleration system, and by improving The uniformity of beam distribution on the target will improve the service life and performance of the neutron tube acceleration system and the neutron tube.

The method for optimizing the neutron tube acceleration system based on genetic algorithm described in this embodiment includes the following steps:

Step 1. Select the accelerating gap D in the neutron tube accelerating system, the accelerating electrode extremely small diameter D ₂ , the accelerating electrode length L, the lead pole aperture phi ₂ , the magnetic ring slope diameter fm ₂ , the magnetic cylinder slope diameter fm ₃ and the accelerating electrode Aperture D ₁ is used as the parameter to be optimized for the neutron tube acceleration system; the value range and accuracy of the above seven parameters to be optimized are set respectively; the value range is set according to the actual engineering experience and does not conflict with the structural size. The parameter range is set according to the genetic algorithm. The parameter accuracy is set by simulation experience. Generally, the parameter range is divided into 2 ^{to 10} parts as the parameter accuracy.

The value of each group of parameters to be optimized is regarded as an individual in the population, that is, the acceleration gap D, the accelerating electrode minimum diameter D ₂ , the accelerating electrode length L, and the extraction pole aperture phi ₂ in each group of neutron tube accelerating system are used , the magnetic ring slope diameter fm ₂ , the magnetic cylinder slope diameter fm ₃ and the accelerating electrode aperture D ₁ are used as an individual in the population to set the control parameters of the genetic algorithm; among them, the maximum evolutionary generation is 100, and the initial population size is 50, the crossover probability is 0.7, and the mutation probability is 0.1.

In the neutron tube acceleration system, the extraction pole side of the acceleration area and the extraction pole side and acceleration pole side of the acceleration area have a greater impact on the particle beam focusing process, which in turn affects the beam distribution on the target. Therefore, the selection of the neutron tube In the accelerating system, the accelerating gap D, the accelerating electrode pole diameter D ₂ , the accelerating electrode length L, the lead pole aperture phi ₂ , the magnetic ring slope aperture fm ₂ , the magnetic cylinder slope aperture fm ₃ and the accelerating electrode aperture D ₁ are used as neutrons The optimized variables of the geometric parameters of the tube acceleration system can improve the uniformity and beam radius of the beam and reduce the number of beam misses, thus improving the service life and performance of the neutron tube.

Step 2. According to the accuracy set in step 1, determine the individual length, combine with the random function, randomly generate N binary numbers to represent the parameters to be optimized, and become an individual to form the initial population;

Step 3. For the individuals in the initial population, convert the binary number to decimal, convert the seven parameters to be optimized in each individual into real numbers within the corresponding value range, and calculate the individual fitness;

The individual fitness calculation process is as follows:

For each individual, MATLAB software is used to cooperate with the COMSOL simulation software to transfer the seven parameters to be optimized of the corresponding individual to the COMSOL simulation software. Other model parameters, material properties, boundary conditions, etc. are set in the COMSOL simulation software through the program. , construct and calculate the geometric model of the neutron tube acceleration system; call COMSOL software to calculate the particle coordinates, and obtain three performance indicators of the beam performance of the neutron tube acceleration system, including beam unevenness target and beam miss number outtarget and beam radius base;

Specifically, for each individual, the seven parameters to be optimized in each individual are converted into real numbers within the corresponding value range, and a three-dimensional finite element model of the neutron tube acceleration system is established. In the model structure of the neutron tube acceleration system On the basis of , divide the mesh; set the grounding in the model structure, apply voltage to the accelerating electrode, and connect the target plane to potential.

After selecting materials, applying boundary conditions, and meshing operations, the electric field simulation calculation of the neutron tube acceleration system is performed. Under the action of the electromagnetic field, a large number of particles move in a certain direction to form a beam. In the simulation, due to the particle The order of magnitude is too large, it is advisable to use some particles to simulate the overall beam, and for the beam intensity of the neutron tube, there is no need to consider the role of space charge.

The simulation of particle trajectory is completed based on the previous electric potential calculation. In the particle tracking module of COMSOL simulation software, electric field force is added in the particle movement area, particle attributes such as mass, charge, etc. are set, and particle inlet boundary conditions are added at the exit of the extraction pole. Release 10,000 particles. The particle density obeys the Gaussian distribution. The particles move toward the target direction under the action of the electric field force. A wall boundary condition is added to the ion entrance surface of the target model to freeze the particles on the surface and stop moving forward.

When the particle beam is distributed on the target surface, R is the distance between the particles and the target center, and R will be used to calculate the uniformity of the particles on the target; the uniformity of the particle beam on the target is evaluated by non-uniformity. The lower the uniformity, the more uniform the proof is. The calculation principle of unevenness is as follows:

Assume that there are 10,000 particles in a circle with radius R _max , and R _max is divided into 10 parts;

Under ideal conditions, 10,000 particles are calculated according to the area ratio and the number of particles Z ₁ =10000×(1/10) ² =100 should be in the circle of R _1/10 = 0.665mm. Assume that in the actual situation Z ₁ =14, both The absolute value of the subtraction is the uneven number Z=100-14=86;

Under ideal conditions, according to the area calculation, there should be Z ₁ = 10000 × [(2/10)2-(1/10)2] = 300 in the ring of R _1/10 - R _2/10 . Assume the actual situation Z ₁ = 184, the absolute value of the subtraction between the two is Z = 300-184 = 126.

By analogy, until the last ring R _9/10 -R _max is calculated, the difference Z between the ideal value and the actual value of each ring is added and divided by the total number of particles 10,000 to obtain the unevenness within the circle. Assuming that the calculated total Z value is 4496, it can be obtained that the unevenness o=Z/10000=0.449. The smaller the unevenness of the geometric structure, the better the acceleration system effect.

The mathematical expression of beam unevenness target is:

Among them, o is the non-uniformity; Z _x is the number of particles in a certain ring x on the target surface under ideal conditions; A _x is the area of a certain ring x on the target surface; A _is the total area of the target surface; U _x is the actual number of particles in a certain ring x on the target surface; x is the number of equally divided rings on the target surface, x=1,2,…,10.

The number of beam misses is based on the finite element simulation calculation results, reading the three-dimensional coordinates of the particles, and counting whether the coordinates are obtained on the target.

The beam radius is obtained by calculating the distance from the particle coordinates on the target surface to the target center coordinates; among them, the beam radius is selected to be the distance from the target center that can cover 90% of the particles; that is, assuming that the distance from 10,000 particles to the target center is as follows: Select the 9000th particle in the large array and determine it as the beam radius.

Based on the obtained three performance indicators of the beam performance of the neutron tube acceleration system, the objective function ObjV is constructed to measure the fitness of each individual, and then measure the performance of the neutron tube acceleration system corresponding to each individual in the population.

Among them, the mathematical expression of the objective function ObjV is:

Target function ObjV=target/base+outtarget;

Among them, target is the beam unevenness; outtarget is the number of beam misses, and base is the beam radius; the smaller the objective function value, the better the fitness of the genetic individual.

In order to speed up the convergence of the genetic algorithm, the three performance indicators of the beam performance are amplified. Among them, once the number of beam outtargets is not equal to 0, it is equal to infinity. The purpose is to screen out the genetic individuals that will miss the target, so that Its genes will not be effectively passed on to the next generation; the beam uniformity target and beam radius base are exponentially amplified; the smaller the objective function value, the better the individual, the smaller the objective function value, the smaller the unevenness, and the beam radius. The larger it is, the number of beam misses is zero.

Step 4. After the fitness calculation of the population is completed, the genetic operations of selection, crossover and mutation are iteratively performed. The genetic operations include selection, crossover and mutation. Among them, the selection operation uses elite selection and mutation based on the fitness value of the genetic individual. Roulette combined mechanism; crossover operation adopts uniform crossover method, randomly selecting individuals to perform row crossover or column crossover; mutation operation adopts bit mutation mechanism.

After the genetic operation is completed, a new population is generated, the binary gene bit string of each individual in the new population is separated, decoded, re-divided into seven parameters to be optimized, and converted from binary to decimal, according to the fitness calculation in step 3 Method to calculate the fitness of each individual in the new population.

Step 5: Decode the new gene population into the geometric structure parameters of the neutron tube acceleration system through genetic operations, and carry out neutron tube beam performance analysis of the new population and fitness evaluation of the gene population, in accordance with the conditions that are conducive to improving the efficiency of the neutron tube acceleration system. Select the direction of sub-tube beam performance and evolve the gene population to determine the next generation's gene population and achieve continuous evolution of the population;

Step 6. The population continues to evolve and meets the iteration stop condition, that is, when the set maximum evolutionary generation ** is reached, the individual with the smallest fitness appears. After decoding, the acceleration space D and acceleration electrode minimum diameter D ₂ corresponding to the individual are obtained. The actual values of the seven parameters to be optimized _, namely the length of the accelerating electrode L, the diameter of the lead pole phi ₂ , the diameter of the magnetic ring slope fm ₂ , the diameter of the magnetic cylinder slope fm ₃ and the aperture D of the accelerating electrode, are used as the neutron tube acceleration system to be optimized. The optimal solution of the geometric structure parameters, and create a neutron tube acceleration system according to the values of the parameters to be optimized;

As shown in Figures 3 and 4, after 100 generations and 300 individual iterative calculations, the optimal solution of the population is stable, the fitness of the optimal solution is stable at 10 ^-4 , and the genetic algorithm converges; among them, in the initial stage, the population The fitness is basically stable around 894.4, with a slight decrease, but not obvious. Around 50 generations, some individuals in the population have better fitness due to crossover and mutation, and the optimal solution fitness of the population drops off a cliff, and then its genes Rapidly spread within the population, the fitness of the final optimal solution stabilizes at 10 ^-4 , that is, the optimal solution is obtained, and the particles cover the entire target surface at 100 generations; it is observed that when the fitness becomes 0.14 around 60 generations, the particles become the largest The radius is only 0.5mm from the target edge, and the optimal solution of around 100 generations requires too high a dimensional accuracy of 10 ^-5 mm. Based on the engineering reality, around 60 generations and a fitness of 0.14 are selected as the optimal solution for this problem. Solution; among them, compared with the neutron tube acceleration system in the unoptimized state, the beam unevenness in the optimal solution is smaller than that in the unoptimized state, and the beam radius in the optimal solution is greater than that in the unoptimized state. Under the beam radius, the maximum distance between the particles and the target center in the optimal solution is greater than the maximum distance between the particles and the target center in the unoptimized state. In addition, the number of beam misses in the optimal solution is zero; by comparing the optimal solution, that is, the individual with the smallest fitness , decode and obtain the optimized geometric structure parameters of the neutron tube accelerating system. Under the optimal solution structure, the neutron tube beam performance is greatly improved; the electric potential in the neutron tube accelerating system is improved before the focusing ability of the accelerating electrode intersects. It is larger, which improves the particle focusing ability and expands the radius of particle distribution on the target surface; the radius of particles on the target surface is significantly expanded, and the number of off-target particles is zero, which is beneficial to improving the pressure resistance performance of the neutron tube and improving the neutron tube. tube life and performance.

In this invention, the finite element method, the genetic algorithm and the uniformity evaluation method are combined with each other, and the acceleration system of the neutron tube is optimized and designed with the help of COMSOL simulation software and MATLAB software, and finally a better acceleration system structure is obtained, which improves the It improves the service life and performance of neutron tubes and is used to guide the design and manufacturing of neutron tubes, which is highly practical.

The above embodiment is only one of the ways to realize the technical solution of the present invention. The scope of protection claimed by the present invention is not only limited by this embodiment, but also includes any technical scope disclosed by the present invention. Changes, substitutions and other implementations may be easily imagined by those skilled in the art.

Claims (6)

1. A neutron tube acceleration system optimization method based on genetic algorithm, which is characterized by including the following steps: Step 1. Initialization Taking the geometric structure parameters of the neutron tube acceleration system as the parameters to be optimized, N binary numbers are randomly generated to represent the parameters to be optimized and become an individual to form the initial population; Step 2. Calculate individual fitness Convert the obtained binary numbers to generate the geometric structure parameters of the neutron tube acceleration system expressed in decimal system; construct a geometric model of the neutron tube acceleration system based on the geometric structure parameters of the neutron tube acceleration system expressed in decimal system; use the finite element method to calculate the genetic The beam performance index of the individual corresponding neutron tube acceleration system; According to the beam performance index of the neutron tube acceleration system, an objective function is established; then the fitness of the genetic individual is measured by the value of the objective function; Step 3: Perform genetic operations iteratively to generate a new population, calculate the fitness of each individual in the new population until the iteration stop condition is met, and output the optimal solution when the iteration stop condition is met as the value of the parameter to be optimized, and follow the steps to be taken. Optimize the values of parameters to create a neutron tube acceleration system; In step 2, the beam performance indicators of the neutron tube acceleration system corresponding to the genetic individual include beam unevenness, beam miss number and beam radius; The mathematical expression of the objective function is: Target function ObjV=target/base+outtarget; Among them, target is the beam unevenness; outtarget is the number of beam misses, and base is the beam radius; the smaller the objective function value, the better the fitness of the genetic individual; Among them, the mathematical expression of beam unevenness target is: Among them, o is the non-uniformity; Z _x is the number of particles in a certain ring x on the target surface under ideal conditions; A _x is the area of a certain ring x on the target surface; A _is the total area of the target surface; U _x is the actual number of particles in a certain ring x on the target surface; x=1,2,…,n. 2. A method for optimizing the neutron tube acceleration system based on genetic algorithm according to claim 1, characterized in that in step 1, the parameters to be optimized include the acceleration gap of the neutron tube acceleration system, the minimum diameter of the acceleration electrode, the acceleration The length of the electrode, the aperture of the lead pole, the diameter of the magnetic ring slope, the diameter of the magnetic cylinder slope and the aperture of the accelerating electrode. 3. A neutron tube acceleration system optimization method based on genetic algorithm according to claim 1, characterized in that the objective function is solved using the finite element method and Matlab software. 4. A kind of neutron tube acceleration system optimization method based on genetic algorithm according to claim 1, characterized in that, in step 3, genetic operations include selection, crossover and mutation; wherein, the selection operation is based on the genetic individual fitness value. The size of the algorithm uses a mechanism combining elite selection and roulette; the crossover operation uses a uniform crossover method, randomly selecting individuals to perform row crossover or column crossover; the mutation operation uses a bit mutation mechanism. 5. A neutron tube acceleration system optimization method based on genetic algorithm according to claim 1, characterized in that, in step 3, satisfying the iteration stop condition is to reach the maximum evolutionary generation, and the optimal solution at this time is in the current population. The parameters to be optimized corresponding to the individual with the smallest fitness. 6. A neutron tube acceleration system optimization method based on genetic algorithm according to claim 1, characterized in that in step 3, satisfying the iteration stop condition is that the fitness reaches the set requirement, and the optimal solution at this time is adaptation. Parameters to be optimized corresponding to individuals whose degree reaches the set requirements.