CN111666719B - Gamma radiation multilayer shielding accumulation factor calculation method, device, equipment and medium - Google Patents

Abstract

The application discloses a method, a device, equipment and a medium for calculating a gamma radiation multilayer shielding accumulation factor, which comprise the following steps: determining various parameters influencing the accumulation factors, generating a plurality of groups of different shielding samples, and calculating corresponding accumulation factor values by combining an MCNP program; taking various determined parameters influencing the accumulation factors as input, taking the calculated corresponding accumulation factor values as output, and constructing a deep neural network; training the deep neural network, and finishing training until the set requirement is met by continuously debugging the learning parameters; inputting various actual parameters influencing the accumulation factors into the trained deep neural network, and directly predicting the corresponding gamma radiation multilayer accumulation factors. According to the method, the deep neural network is constructed, the pre-calculated data samples are adopted for deep neural network learning, the gamma radiation multilayer accumulation factors can be quickly and accurately calculated, the calculation time is short, a large number of accumulation factors can be calculated at one time, and the calculation precision is relatively high.

Description

Method, Apparatus, Equipment and Medium for Calculating Accumulation Factor of Gamma Radiation Multilayer Shielding

technical field

The invention relates to the field of radiation safety and protection, in particular to a method, device, equipment and medium for calculating the accumulation factor of gamma radiation multilayer shielding.

Background technique

At present, in radiation safety and protection, the gamma radiation accumulation factor (Built-up Factor) is a physical quantity that describes the influence of scattered photons. In general, it refers to the ratio of the actual value of a certain radiation dose at the inspection point to the radiation dose caused by rays emitted by the radioactive source that have not interacted with the shield in any way. The accumulation factor is usually different for different radiation doses. The physical quantities commonly used in radiation protection include fluence, exposure, and absorbed dose, and the corresponding accumulation factors are fluence accumulation factor, exposure accumulation factor, and absorbed dose accumulation factor. In radiation protection, in order to protect the safety of workers, it is often necessary to understand the radiation situation of the working environment, and it is necessary to calculate the three-dimensional radiation field. The calculation of 3D radiation field generally adopts point kernel integration, and the calculation of accumulation factor is the key of point kernel integration. The main task of the accumulation factor calculation is to correct and optimize the point kernel integral used in the calculation of the three-dimensional radiation field.

In the current practical engineering applications, the cumulative factor correction and optimization adopted by the three-dimensional radiation field calculation method adopts empirical formula or database interpolation method, but the empirical formula or database interpolation method is mainly applicable to the cumulative factor calculation in single-layer shielding. In a specific radiation field, the shield is often multi-layered, and each layer is composed of different materials. The empirical formula or database interpolation method cannot accurately calculate the multi-layer accumulation factor value, and the error is usually large, so it is impossible to ensure a three-dimensional radiation field. accuracy.

Therefore, how to accurately and quickly calculate the multi-layer radiation shielding accumulation factor calculated for the three radiation fields is a technical problem to be solved urgently by those skilled in the art.

SUMMARY OF THE INVENTION

In view of this, the purpose of the present invention is to provide a method, device, equipment and medium for calculating the accumulation factor of gamma radiation multilayer shielding, which not only takes less time to calculate, but also can calculate a large number of accumulation factors at one time, and its calculation accuracy is relatively high. Its specific plan is as follows:

A method for calculating accumulation factor of gamma radiation multilayer shielding, comprising:

Determine various parameters that affect the accumulation factor, generate multiple groups of different shielded samples, and calculate the corresponding accumulation factor value in combination with the MCNP program;

The determined various parameters affecting the accumulation factor are used as input, and the calculated corresponding accumulation factor value is used as output to construct a deep neural network;

The deep neural network is trained, and the training is ended by continuously debugging the learning parameters until the set requirements are met;

Various parameters that actually affect the accumulation factor are input into the trained deep neural network, and the corresponding multi-layer accumulation factor of gamma radiation is directly predicted.

Preferably, in the above-mentioned calculation method for the accumulation factor of the multi-layer shielding of gamma radiation provided by the embodiment of the present invention, the various parameters affecting the accumulation factor include incident particle energy, the density of each layer of shielding material, the number of shielding free paths of each layer, the Each layer shields the scattering cross section, each layer shields the photoelectric effect cross section, and each layer shields the electron pair effect cross section.

Preferably, in the above-mentioned gamma radiation multi-layer shielding accumulation factor calculation method provided in the embodiment of the present invention, multiple groups of different shielding samples are generated, and the corresponding accumulation factor values are calculated in combination with the MCNP program, which specifically includes:

According to the determined characteristics of various parameters affecting the accumulation factor, multiple groups of different models are established;

Batch generate MCNP input files of different particle energies, different shielding materials, and different combinations of shielding free path numbers;

According to the generated MCNP input file, call the MCNP program for calculation, and extract the dose considering scattering after shielding and the dose not considering scattering from the calculation result in batches;

A corresponding accumulation factor value is calculated by the ratio of the dose considering scattering to the dose not considering scattering.

Preferably, in the above-mentioned calculation method for the accumulation factor of the multi-layer shielding of gamma radiation provided by the embodiment of the present invention, while constructing the deep neural network, the method further includes:

The topology of the deep neural network is determined according to the number of input parameters and the number of output parameters.

Preferably, in the above-mentioned gamma radiation shielding accumulation factor calculation method provided by the embodiment of the present invention, the hidden layer of the deep neural network adopts double layers of neurons;

The node transfer function of the deep neural network includes a relu function and a linear function;

The training function of the deep neural network includes SDG function and momentum function.

Preferably, in the above-mentioned gamma radiation multi-layer shielding accumulation factor calculation method provided by the embodiment of the present invention, the set requirement includes that the average relative error of the validation set is less than a set prediction accuracy or reaches a set number of iterations.

The embodiment of the present invention also provides a gamma radiation multi-layer shielding accumulation factor calculation device, including:

The cumulative factor calculation module is used to determine various parameters affecting the cumulative factor, generate multiple groups of different shielded samples, and calculate the corresponding cumulative factor value in combination with the MCNP program;

A deep neural network building module, used for determining the various parameters affecting the accumulation factor as input, and using the calculated corresponding accumulation factor value as an output to construct a deep neural network;

A deep neural network training module, used to train the deep neural network, by continuously debugging the learning parameters, until the set requirements are met and the training ends;

The accumulation factor prediction module is used to input various parameters that actually affect the accumulation factor into the trained deep neural network, and directly predict the corresponding multi-layer accumulation factor of gamma radiation.

Preferably, in the above-mentioned gamma radiation multi-layer shielding accumulation factor calculation device provided in the embodiment of the present invention, the accumulation factor calculation module specifically includes:

a model establishment unit, used for establishing multiple groups of different models according to the determined characteristics of various parameters affecting the accumulation factor;

The MCNP file generation unit is used to batch generate MCNP input files of different particle energies, different shielding materials, and different combinations of shielding free path numbers;

The MCNP program calculation unit is used for invoking the MCNP program for calculation according to the generated MCNP input file, and extracting the dose considering scattering after shielding and the dose not considering scattering in batches from the calculation result;

The accumulation factor calculation unit is configured to calculate a corresponding accumulation factor value according to the ratio of the dose considering scattering to the dose not considering scattering.

An embodiment of the present invention further provides a gamma radiation multi-layer shielding accumulation factor calculation device, including a processor and a memory, wherein, when the processor executes a computer program stored in the memory, the above-mentioned embodiments of the present invention are implemented Calculation method of accumulation factor of gamma radiation multilayer shielding.

An embodiment of the present invention further provides a computer-readable storage medium for storing a computer program, wherein, when the computer program is executed by a processor, the above-mentioned method for calculating the cumulative factor of gamma radiation multilayer shielding provided by the embodiment of the present invention is implemented .

It can be seen from the above technical solutions that the method, device, equipment and medium for calculating the accumulation factor of gamma radiation multilayer shielding provided by the present invention include: determining various parameters affecting the accumulation factor, generating multiple sets of different shielding samples, and Combined with the MCNP program, the corresponding accumulation factor value is calculated; the determined parameters that affect the accumulation factor are used as input, and the calculated corresponding accumulation factor value is used as the output to construct a deep neural network; the deep neural network is trained by continuous Debug the learning parameters until the set requirements are met and end the training; input various parameters that actually affect the accumulation factor into the trained deep neural network, and directly predict the corresponding multi-layer accumulation factor of gamma radiation.

The invention proposes a multi-layer accumulation factor calculation based on a deep neural network, and provides a brand-new method for the calculation of the multi-layer shielding accumulation factor of gamma radiation. In the case of coupling, using pre-computed data samples for deep neural network learning can realize fast and accurate calculation of multi-layer accumulation factors of gamma radiation. , which can meet the accuracy requirements of the three-dimensional radiation field calculation for the correction of the point kernel integral.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the following briefly introduces the accompanying drawings required for the description of the embodiments or related technologies. Obviously, the accompanying drawings in the following description are only the For the embodiments of the invention, for those of ordinary skill in the art, other drawings can also be obtained according to the provided drawings without any creative effort.

1 is a flowchart of a method for calculating a cumulative factor of a multi-layer gamma radiation shielding provided by an embodiment of the present invention;

2 is a schematic diagram of a double-layer flat plate geometric model provided by an embodiment of the present invention;

3 is a deep neural network training error diagram provided by an embodiment of the present invention;

Fig. 4 is the regression curve diagram of the predicted value of the data including the training set, the verification set, and the test set and the original data during the double-layer masking provided by the embodiment of the present invention;

5 is a statistical diagram of error statistics of a prediction result of a deep neural network provided by an embodiment of the present invention;

6 is a comparison diagram of the predicted accumulation factor and the actual value of a deep neural network with an AL+FE double-layer shielding with an incident photon energy of 0.6MeV provided by an embodiment of the present invention;

FIG. 7 is a comparison diagram of a predicted accumulation factor and a real value of a deep neural network with an AL+FE double-layer shielding with incident photon energy of 1.5MeV provided by an embodiment of the present invention;

FIG. 8 is a comparison diagram of the predicted accumulation factor and the real value of a deep neural network with an AL+FE double-layer shielding with incident photon energy of 8MeV provided by an embodiment of the present invention;

FIG. 9 is a schematic structural diagram of an apparatus for calculating an accumulation factor of a multi-layer shielding of gamma radiation provided by an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

The present invention provides a method for calculating the accumulation factor of gamma radiation multilayer shielding, as shown in FIG. 1, including the following steps:

S101. Determine various parameters affecting the accumulation factor, generate multiple groups of different shielded samples, and calculate the corresponding accumulation factor value in combination with the MCNP program;

In specific implementation, various parameters affecting the accumulation factor may include incident particle energy, shielding material density of each layer, shielding free path number (thickness) of each layer, scattering cross section of each layer, photoelectric effect cross section of each layer, and shielding of each layer. Electron pair effect cross section. It should be noted that the MCNP (Monte Carlo N Particle Transport Code) program refers to the Monte Carlo particle transport calculation program, which can be used to calculate neutrons, photons, electrons or coupled neutrons/photons/ A generic package for electronic transport problems;

S102, the determined various parameters affecting the accumulation factor are used as input, and the calculated corresponding accumulation factor value is used as the output to construct a deep neural network (Deep Neural Networks, DNN);

It should be noted that the determined various parameters affecting the cumulative factor and the corresponding cumulative factor value are used as learning samples, the learning samples are preprocessed, and the learning samples are divided into input items and corresponding output items, and the learning samples are used for The learning of deep neural network can realize relatively accurate calculation of multi-layer accumulation factors;

S103, train the deep neural network, and end the training by continuously debugging the learning parameters until the set requirements are met;

In a specific implementation, the above-mentioned set requirements may include that the average relative error of the validation set is less than a set prediction accuracy or reaches a set number of iterations. That is to say, when using the neural network to train and learn the above cumulative factor calculation data, the learning parameters of the neural network are adjusted so that the average relative error of the validation set is less than the set prediction accuracy (such as 0.5%) or reaches The set number of iterations ends the neural network training;

S104, input various parameters that actually affect the accumulation factor into the trained deep neural network, and directly predict the corresponding multi-layer accumulation factor of gamma radiation;

Specifically, after the neural network learning is completed, the corresponding gamma radiation accumulation factor value can be quickly and accurately obtained by inputting the incident photon energy, the shielding free path number of each layer, the shielding scattering cross section of each layer, the photoelectric effect cross section, and the electron pair effect cross section during prediction. .

In the above-mentioned gamma radiation multi-layer shielding accumulation factor calculation method provided by the embodiment of the present invention, by constructing a deep neural network, without decoupling the complex physical relationship between the input and output, the pre-calculated data samples are used for the depth calculation. Neural network learning can realize fast and accurate calculation of multi-layer accumulation factors of gamma radiation. Not only does the calculation take less time, but also a large number of accumulation factors can be calculated at one time, and its calculation accuracy is relatively high, which can meet the requirements of three-dimensional radiation field calculation for point kernel integral correction. precision requirements.

During specific implementation, in the above-mentioned method for calculating the accumulation factor of gamma radiation multilayer shielding provided by the embodiment of the present invention, step S101 generates multiple groups of different shielding samples, and calculates the corresponding accumulation factor value in combination with the MCNP program, which may specifically include: first , according to the determined characteristics of various parameters affecting the accumulation factor, establish multiple groups of different models; then, generate MCNP input files with different particle energies, different shielding materials, and different combinations of shielding free path numbers in batches; then, according to the generated MCNP input files Input the file, call the MCNP program for calculation, and extract the dose considering scattering after shielding and the dose not considering scattering from the calculation results in batches; finally, calculate the corresponding accumulation factor by the ratio of the dose considering scattering and the dose not considering scattering value.

Specifically, the formula for calculating the corresponding cumulative factor value is as follows:

为考虑散射的剂量，D₁为未考虑散射的剂量，B为对应的累积因子值。in,

For the dose considering scattering, D ₁ is the dose not considering scattering, and B is the corresponding accumulation factor value.

In specific implementation, in the above-mentioned calculation method for the accumulation factor of gamma radiation multilayer shielding provided by the embodiment of the present invention, while executing step S102 to construct a deep neural network, it may also include: according to the number of input parameters n ₁ and the number of output parameters number n ₂ to determine the topology of the deep neural network.

Further, considering the complexity of the actual masking problem, the deep neural network structure has the following guiding principles:

First, for complex engineering problems, the hidden layer of the deep neural network adopts two layers of neurons;

Second, in a single-layer hidden layer neural network, the number of neurons in the entire neural network structure is recommended as:

n ₁ →2n ₁ ±1→n ₂ ;

Third, in the two-layer hidden layer neural network, the number of neurons in the entire neural network structure is recommended as:

n ₁ →1.5n ₁ →2n ₁ ±1→n ₂ ;

For deep neural network training parameters, the following recommendations are made:

First, the node transfer function of the deep neural network includes a relu function and a linear function; preferably, the node transfer function is: relu+relu+relu+linear;

Second, the training function of the deep neural network includes the SDG function and the momentum function; the training function is: SDG+momentum.

The following takes the double-layer flat plate model as an example as the calculation model of the accumulation factor, and the above-mentioned calculation method of the accumulation factor of the multi-layer gamma radiation shielding provided by the embodiment of the present invention is described in detail:

As shown in Figure 2, the two-layer slab model is composed of two slabs of different materials. Firstly, the problem of solving the accumulation factor is transformed into the ratio of the dose value after considering the scattering after double shielding to the dose value without considering the scattering by modeling; The shielded dose of the double-layer material combination; further organize the neural network learning sample data (including incident photon energy, material density of each layer, mean free path number of each layer, scattering cross section of each layer, photoelectric effect cross section of each layer, electron pair of each layer) effect cross section); then use the neural network to train the above samples, and the relative average error of prediction is less than 0.5%; finally, the learned neural network is used to directly predict the cumulative factor value to be sought. Specific steps are as follows:

Step 1: Determine various parameters affecting the accumulation factor, incident particle energy, shielding material of each layer, thickness of shielding layer of each layer (number of free paths), cross-sectional data of shielding material of each layer (photoelectric effect cross section, scattering cross section, electron pair effect) section), etc.;

Step 2: Establish N groups of different flat panel models according to the parameter features to be learned determined in step 1. Then use the sample batch generation program to batch generate MCNP input files with different energies, different shielding thicknesses and different shielding materials;

Step 3: Use the MCNP program to calculate the sample, and then batch extract the dose calculated by MCNP considering the scattering after shielding and the dose without considering the scattering, calculate the corresponding accumulation factor value, and finally use each parameter and its corresponding accumulation factor value as the neural network. Learning samples for the network. Some MCNP calculation data and corresponding cumulative factor data are shown in Table 1 below:

Table I

Scatter dose (Gy) not considered Consider scattering dose (Gy) Cumulative factor value 1.47234E-13 3.90952E-13 2.655310594 1.95883E-14 8.06854E-14 4.119060868 6.77E-15 3.41936E-14 5.052380235 3.09427E-14 1.16454E-13 3.763537119 3.25E-16 2.62E-15 8.047588175 2.89E-15 1.67444E-14 5.787881826 7.61E-15 3.73717E-14 4.911648238

Step 4: Preprocess the learning sample data, and divide the learning sample into input items and corresponding output items. It can be seen that there are 11 parameters of the double-layer input item, including: incident photon energy, density of the first layer shielding material, first layer photoelectric effect cross section, first layer scattering cross section, first layer electron pair effect cross section, first layer shielding freedom The number of paths, the density of the second layer shielding material, the second layer photoelectric effect cross section, the second layer scattering cross section, the second layer electron pair effect cross section, the second layer shielding free path number. The output term is the cumulative factor value.

Step 5: Determine the topology of the deep neural network according to the number of input parameters and the number of output parameters of the sample; wherein, the ratio between the input layer, the hidden layer, and the output layer of the neural network parameters can be set to 11: [50 80 50]: 1; the node transfer function adopts relu+relu+relu+linear; the training function adopts SDG+momentum;

Step 6: Use the neural network to train and learn the cumulative factor calculation data; continue to debug the neural network learning parameters until the prediction accuracy with an average relative error of less than 0.5% is reached or the set number of iterations is reached, and the neural network learning is ended; training error The changes are shown in Figure 3, and the regression curves of the predicted values and true values of the training set, validation set, and test set are shown in Figure 4. From Figure 4, it can be seen that the neural network fits the cumulative factor calculation data well, and there is no Fitting and underfitting;

Step 7. After the neural network learning is completed, the corresponding cumulative factor value can be quickly and accurately predicted by inputting the incident particle energy, shielding thickness (number of free paths), photoelectric effect cross section, scattering cross section, and electron pair effect cross section; the prediction result The error is shown in Figure 5.

The following table 2 shows part of the data display of the cumulative factor predicted by the neural network:

Table II

The following three groups of incident photon energies, 0.6MeV, 1.5MeV, and 8MeV, are selected for verification:

Incident photon energy 0.6MeV: Shielding model: Al+Fe (the first shield is aluminum, the second shield is iron), the number of shielding free paths: within 10 free paths, based on the predicted value and the actual value of the cumulative factor of the deep neural network For comparison, as shown in Figure 6. The mean absolute error is 1.06% and the maximum absolute error is 11.71%.

Incident photon energy 1.5MeV: Shielding model: Al+Fe (the first shield is aluminum, the second shield is iron), the number of shielding free paths: within 10 free paths, based on the predicted value and the actual value of the cumulative factor of the deep neural network For comparison, as shown in Figure 7. The mean absolute error is 2.70% and the maximum absolute error is 6.46%.

Incident photon energy 8MeV: shielding model: Al+Fe (the first shield is aluminum, the second shield is iron), the number of shielding free paths: within 10 free paths, based on the comparison between the predicted value of the deep neural network accumulation factor and the actual value , as shown in Figure 8. The mean absolute error is 3.08% and the maximum absolute error is 7.34%.

Under the conditions of this example, it can be seen from Table 2, Figure 6, Figure 7, and Figure 8 that the absolute average error of prediction is all within 5%, and the maximum error is also within an acceptable range. It shows that the use of neural network to calculate the accumulation factor can not only quickly calculate a large number of accumulation factor data, but also ensure the accuracy of the accumulation factor value to a large extent. Therefore, it is feasible to use the neural network to calculate the cumulative factor value.

Based on the same inventive concept, the embodiment of the present invention also provides a multi-layer gamma radiation shielding accumulation factor calculation device, because the principle of solving the problem of the gamma radiation multi-layer shielding accumulation factor calculation device is the same as that of the aforementioned gamma radiation multi-layer shielding accumulation factor The calculation methods are similar, so the implementation of the device for calculating the cumulative factor of the multi-layer gamma radiation shielding can refer to the implementation of the method for calculating the cumulative factor of the multi-layer gamma radiation shielding, and the repetition will not be repeated.

During specific implementation, the gamma radiation multi-layer shielding accumulation factor calculation device provided by the embodiment of the present invention, as shown in FIG. 9 , specifically includes:

The accumulation factor calculation module 11 is used to determine various parameters affecting the accumulation factor, generate multiple groups of different shielded samples, and calculate the corresponding accumulation factor value in combination with the MCNP program;

The deep neural network building module 12 is used to determine various parameters that affect the accumulation factor as input, and use the calculated corresponding accumulation factor value as an output to construct a deep neural network;

The deep neural network training module 13 is used to train the deep neural network, and the training is ended by continuously debugging the learning parameters until the set requirements are met;

The accumulation factor prediction module 14 is used to input various parameters that actually affect the accumulation factor into the trained deep neural network, and directly predict the corresponding multi-layer accumulation factor of gamma radiation.

In the above-mentioned gamma radiation multi-layer shielding accumulation factor calculation device provided by the embodiment of the present invention, through the interaction of the above-mentioned four modules, without decoupling the complex physical relationship between the input and output, a pre-calculated Deep neural network learning is performed on data samples to achieve fast and accurate calculation of multi-layer accumulation factors of gamma radiation. Not only does the calculation take less time, but a large number of accumulation factors can be calculated at one time, and its calculation accuracy is relatively high, which meets the point-to-point kernel integration of three-dimensional radiation field calculation. Corrected accuracy requirements.

Further, in the specific implementation, in the above-mentioned gamma radiation multi-layer shielding accumulation factor calculation device provided in the embodiment of the present invention, the accumulation factor calculation module 11 may specifically include:

The model establishment unit is used to establish multiple groups of different models according to the determined characteristics of various parameters affecting the accumulation factor;

The MCNP file generation unit is used to batch generate MCNP input files of different particle energies, different shielding materials, and different combinations of shielding free path numbers;

The MCNP program calculation unit is used to call the MCNP program for calculation according to the generated MCNP input file, and batch extract the dose considering the scattering after shielding and the dose not considering the scattering from the calculation result;

The accumulation factor calculation unit is configured to calculate the corresponding accumulation factor value by the ratio of the dose considering the scattering to the dose not considering the scattering.

For more specific working processes of the above-mentioned modules, reference may be made to the corresponding contents disclosed in the foregoing embodiments, which will not be repeated here.

Correspondingly, the embodiment of the present invention also discloses a multi-layer gamma radiation shielding accumulation factor calculation device, including a processor and a memory; wherein, when the processor executes the computer program stored in the memory, the multi-layer gamma radiation disclosed in the foregoing embodiments is implemented. Mask the cumulative factor calculation method.

For a more specific process of the above method, reference may be made to the corresponding content disclosed in the foregoing embodiments, which will not be repeated here.

Further, the present invention also discloses a computer-readable storage medium for storing a computer program; when the computer program is executed by a processor, the aforementioned method for calculating a multi-layer shielding accumulation factor of gamma radiation disclosed.

For a more specific process of the above method, reference may be made to the corresponding content disclosed in the foregoing embodiments, which will not be repeated here.

The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments may be referred to each other. For the apparatuses, devices, and storage media disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and reference may be made to the descriptions of the methods for related parts.

Professionals may further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two, in order to clearly illustrate the possibilities of hardware and software. Interchangeability, the above description has generally described the components and steps of each example in terms of functionality. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each particular application, but such implementations should not be considered beyond the scope of this application.

The steps of a method or algorithm described in conjunction with the embodiments disclosed herein may be directly implemented in hardware, a software module executed by a processor, or a combination of the two. A software module can be placed in random access memory (RAM), internal memory, read only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other in the technical field. in any other known form of storage medium.

To sum up, a method, device, equipment and medium for calculating the accumulation factor of gamma radiation multilayer shielding provided by the embodiments of the present invention include: determining various parameters affecting the accumulation factor, generating multiple sets of different shielding samples, and combining with the MCNP program to calculate The corresponding accumulation factor value is obtained; the determined parameters that affect the accumulation factor are used as input, and the corresponding calculated accumulation factor value is used as the output to construct a deep neural network; the deep neural network is trained, and the learning parameters are continuously debugged. End the training until the set requirements are met; input various parameters that actually affect the accumulation factor into the trained deep neural network, and directly predict the corresponding multi-layer accumulation factor of gamma radiation. In this way, by building a deep neural network, without decoupling the complex physical relationship between input and output, using pre-computed data samples for deep neural network learning, the multi-layer accumulation factor of gamma radiation can be quickly and accurately calculated, not only calculating It takes less time, and can calculate a large number of accumulation factors at one time, and its calculation accuracy is relatively high, which can meet the accuracy requirements of three-dimensional radiation field calculation for point kernel integral correction.

Finally, it should also be noted that in this document, relational terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply these entities or that there is any such actual relationship or sequence between operations. Moreover, the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device that includes a list of elements includes not only those elements, but also includes not explicitly listed or other elements inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in a process, method, article or apparatus that includes the element.

The method, device, equipment and medium for calculating the cumulative factor of gamma radiation multilayer shielding provided by the present invention have been described in detail above. The principles and implementations of the present invention are described with specific examples in this paper. The descriptions of the above examples are only It is used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific embodiments and application scope. The contents of the description should not be construed as limiting the present invention.

Claims (7)

1. A gamma radiation multilayer shielding accumulation factor calculation method is characterized by comprising the following steps:

determining various parameters influencing the accumulation factor, and establishing a calculation model of the accumulation factor according to the characteristics of the determined various parameters influencing the accumulation factor; the various parameters influencing the accumulation factor comprise incident particle energy, shielding material density of each layer, shielding free path number of each layer, shielding scattering cross section of each layer, shielding photoelectric effect cross section of each layer, and shielding electron pair effect cross section of each layer;

generating MCNP input files with different particle energies, different shielding materials and different shielding free path number combinations in batches;

calling an MCNP program to calculate according to the generated MCNP input file, and extracting the dose considering scattering and the dose not considering scattering after shielding in batch from a calculation result;

calculating a corresponding cumulative factor value by the ratio of the scattering considered dose to the non-scattering considered dose;

taking various determined parameters of the influence accumulation factors as input, and taking the corresponding calculated accumulation factor value as output to construct a deep neural network;

training the deep neural network, and finishing training until a set requirement is met by continuously debugging learning parameters;

inputting various actual parameters influencing the accumulation factors into the trained deep neural network, and directly predicting the corresponding gamma radiation multilayer accumulation factors.

2. The method for calculating the gamma radiation multilayer shielding accumulation factor according to claim 1, further comprising the following steps of, while constructing the deep neural network:

and determining the topological structure of the deep neural network according to the number of the input parameters and the number of the output parameters.

3. The gamma-radiation multilayer shielded accumulation factor calculation method as claimed in claim 2, wherein the hidden layer of the deep neural network employs double-layer neurons;

the node transfer functions of the deep neural network comprise a relu function and a linear function;

the training function of the deep neural network comprises an SDG function and a momentum function.

4. The gamma radiation multi-layer shielding accumulation factor calculation method as claimed in claim 1, wherein the set requirement includes that the average relative error of the validation set is less than a set prediction accuracy or reaches a set number of iterations.

5. A gamma radiation multilayer shield accumulation factor calculation apparatus, comprising:

an accumulation factor calculation module comprising: the model establishing unit is used for establishing a calculation model of the accumulation factor according to the determined characteristics of various parameters influencing the accumulation factor; the various parameters influencing the accumulation factor comprise incident particle energy, shielding material density of each layer, shielding free path number of each layer, shielding scattering cross section of each layer, shielding photoelectric effect cross section of each layer, and shielding electron pair effect cross section of each layer; the MCNP file generating unit is used for generating MCNP input files with different particle energies, different shielding materials and different shielding free path number combinations in batches; the MCNP program calculating unit is used for calling an MCNP program to calculate according to the generated MCNP input file, and extracting the dose considering scattering and the dose not considering scattering after shielding in batch from a calculation result; an accumulation factor calculation unit for calculating a corresponding accumulation factor value by a ratio of the scattering-considered dose to the non-scattering-considered dose;

the deep neural network construction module is used for taking various determined parameters of the influence accumulation factors as input and taking the corresponding calculated accumulation factor value as output to construct a deep neural network;

the deep neural network training module is used for training the deep neural network, and finishing training until a set requirement is met by continuously debugging learning parameters;

and the accumulation factor prediction module is used for inputting various actual parameters influencing the accumulation factors into the trained deep neural network and directly predicting the corresponding gamma radiation multilayer accumulation factors.

6. A gamma radiation multi-layer shielding accumulation factor calculation device comprising a processor and a memory, wherein the processor implements the gamma radiation multi-layer shielding accumulation factor calculation method according to any one of claims 1 to 4 when executing a computer program stored in the memory.

7. A computer-readable storage medium for storing a computer program, wherein the computer program when executed by a processor implements the gamma radiation multi-layered shielding accumulation factor calculation method of any one of claims 1 to 4.

Images