A method for structural design of a multi-objective collaborative optimization PGNAA device

Abstract

The application discloses a structure design method of a multi-target cooperative optimization PGNAA device, and the method comprehensively considers the influence of multiple performance indexes, combines a Monte Carlo method, introduces a multi-target cooperative optimization mechanism with constraints, and thus obtains an optimal structure module combination of the PGNAA device meeting engineering application requirements, and can effectively improve the measurement and analysis level of the PGNAA device on samples.

Description

Technical Field

[0001] This invention belongs to the field of nuclear analysis equipment design and optimization technology, specifically a structural design method for a multi-objective collaborative optimization PGNAA device. Background Technology

[0002] Prompt Gamma Neutron Activation Analysis (PGNAA) is a non-destructive testing method that utilizes the capture or inelastic scattering reaction of neutrons with atomic nuclei in the analyte to detect the generated transient gamma rays, thereby enabling qualitative and quantitative analysis of elements. This method has been widely used in various fields such as industry, environment, and geological exploration.

[0003] Traditional PGNAA device optimization design methods typically employ a single metric for evaluation and optimization, such as maximizing effective signal strength or signal-to-noise ratio (SNR). However, multiple performance targets may conflict with each other. For instance, when the optimization objective is to increase the detected signal strength, the device's SNR is usually at a low level. Therefore, optimization methods that evaluate based on a single metric are insufficient to comprehensively and objectively characterize the overall measurement performance of a PGNAA device, exhibiting significant limitations. Summary of the Invention

[0004] The purpose of this invention is to provide a structural design method for a multi-objective collaborative optimization PGNAA device to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a structural design method for a multi-objective collaboratively optimized PGNAA device, wherein the PGNAA device comprises a neutron source, a detector, a neutron control module, and a protection module for device protection, the neutron control module comprising one or more of a neutron moderator module, a reflector module, a multiplier module, and a collimator module, characterized by comprising the following steps:

[0006] S1: Initial selection of structural materials for the PGNAA device and determination of optimization conditions and boundaries; Based on the Monte Carlo simulation method, the neutron flux distribution characteristics at the sample under different material conditions are calculated, and materials with superior performance are selected as candidate materials for the structural optimization design of the PGNAA device; At the same time, the neutron flux distribution in the region surrounding the neutron source of the PGNAA device is calculated, and the region with low neutron flux distribution is used as the boundary range for the structural size design and parameter optimization of the PGNAA device, and the size range of the component modules to be optimized is determined.

[0007] S2: Construct a parametric model of the PGNAA device; Based on the candidate materials obtained in step S1, establish a parametric model of the PGNAA device to be optimized in Monte Carlo simulation software, and clarify the structural form, geometric dimensions and material composition of the PGNAA device to be optimized.

[0008] S3: Simulation calculation of PGNAA device response under different conditions; Based on the Monte Carlo simulation method, the PGNAA device response under different parameter combinations is simulated for the parameters to be optimized in step S2, and the multi-objective evaluation function dataset corresponding to each parameter combination is obtained.

[0009] S4: Correlation analysis of multi-objective evaluation functions; perform linear correlation analysis on the multi-objective evaluation function dataset of the PGNAA device to clarify the interrelationship between each evaluation function, eliminate highly redundant evaluation functions, and thus select key evaluation functions for multi-objective collaborative optimization;

[0010] S5: Multi-objective collaborative optimization and constraint screening; Based on the multi-objective evaluation function dataset obtained in step S4, the multi-objective evaluation function dataset is processed and analyzed through multi-objective optimization theory; Under the premise of satisfying the engineering constraints of mass and volume, the device parameter space is systematically searched and screened to obtain the Pareto optimal parameter combination that satisfies the constraints;

[0011] S6: Determine the optimal parameter solution; For the Pareto optimal parameter combination obtained in step S5, further optimize each parameter combination through an objective weighted comprehensive evaluation method to determine the structural module combination under optimal conditions, thereby obtaining the optimal PGNAA device model parameters for the final design of the PGNAA device.

[0012] Furthermore, in step S1, the different materials include LiH, PE, W, Pb, C, and Al; the neutron source in the structure of the PGNAA device is one of the DT neutron generator, DD neutron generator, and Am-Be source; the detector in the structure of the PGNAA device is one of the BGO, sodium iodide, lanthanum bromide, and high-purity germanium detector; and the sample detected by the PGNAA device includes multiple elements selected from Mg, H, C, Ti, Fe, Mn, Cu, Al, and Cl.

[0013] Furthermore, in step S3, when performing simulation calculations of the PGNAA device response, it is necessary to consider the complete neutron and gamma transport processes of the PGNAA device. Specifically, this involves using various tallies to statistically analyze key physical quantities, including the following steps:

[0014] S31: Records the energy spectrum information deposited in the detector;

[0015] S32: Record the effective gamma signal generated by the sample;

[0016] S33: Record the neutron flux inside the detector;

[0017] S34: Record the thermal neutron flux on the sample surface.

[0018] Furthermore, in step S3, the thermal neutron flux at the sample, the neutron flux inside the detector, the effective characteristic gamma ray count, and the signal-to-noise ratio evaluation function parameters in the multi-objective evaluation function dataset are obtained through Monte Carlo simulation, and the signal-to-noise ratio of the PGNAA device is calculated using the following formula: .

[0019] Furthermore, in step S4, the linear correlation analysis process is as follows: Let there be m evaluation indicators Y:

[0020] ; Any two different evaluation indicators Y j and Y k The Pearson correlation coefficient between (j, k=1, 2, ..., m) is defined as:

[0021] ;

[0022] in, Let be the values ​​of the j-th and k-th evaluation indicators under the i-th sample condition, respectively. The evaluation index Y is respectively j Y k The sample mean; This indicates the degree of linear correlation between two evaluation indicators; when The closer the correlation is to 1, the higher the correlation between the evaluation indicators.

[0023] Furthermore, in step S5, when optimizing the neutron control module, the optimization objective formula is as follows: In the formula, F EF (X ) F is the effective signal target evaluation function. TNF (X i Let F be the target evaluation function for the thermal neutron flux at the sample. SNR (X i Mass(X) is the signal-to-noise ratio target evaluation function. i ) and Volume(X i ) represent the structural mass and volume constraint functions of the PGNAA device, with constraint values ​​determined based on specific engineering requirements; the optimization objective formula for the protection module is as follows: Among them, F DNF (X i) represents the neutron flux inside the detector.

[0024] Furthermore, in step S5, the systematic search and filtering of the device parameter space includes the following steps:

[0025] S51: Input multi-objective evaluation function dataset;

[0026] S52: Normalize the data;

[0027] S53: Perform non-dominated sorting of candidate solutions based on Pareto dominance;

[0028] S54: Extract candidate solutions that are not dominated by any other solutions to form a Pareto optimal solution set;

[0029] S55: Output the Pareto optimal solution set and the corresponding multi-objective function values.

[0030] Furthermore, the objective weighting comprehensive evaluation method is a weighted evaluation method that combines the entropy weight method with TOPSIS.

[0031] Compared with the prior art, the beneficial effects of the present invention are:

[0032] 1. Implement multi-parameter collaborative optimization design

[0033] This invention breaks through the limitation of using a single parameter as the optimization evaluation index in the structural design of traditional PGNAA devices. It comprehensively considers the changes in multiple performance indicators of the device when various structural parameters such as material type and geometric dimensions change, and realizes the multi-objective collaborative optimization design of the device's structural parameters.

[0034] 2. Effectively avoid redundancy and conflicting interference in optimization indicators.

[0035] By conducting correlation analysis on the multi-objective evaluation function dataset, the objective evaluation functions that play a dominant role in the device performance are screened out, reducing the interference of highly correlated or redundant indicators on the optimization results and improving the excellence of the multi-objective optimization process.

[0036] 3. No need for subjective bias in setting weights.

[0037] This invention first introduces a bias-free multi-objective optimization mechanism to search and filter the parameter space without pre-setting the weights of each optimization objective, thereby obtaining the Pareto optimal parameter combination that satisfies the engineering constraints. Then, based on the data distribution characteristics of the evaluation functions of each objective in the Pareto solution set, a comprehensive evaluation method based on objective weighting is adopted to further determine the optimal device parameter scheme from the Pareto optimal parameter combination, thus avoiding the uncertainty caused by relying on human experience to set weights in traditional methods.

[0038] 4. It enables efficient parameter selection under engineering constraints.

[0039] By introducing engineering constraints such as mass and volume into the multi-objective optimization process, the method can efficiently search and filter the parameter space while meeting the requirements of actual engineering applications, thereby improving the engineering feasibility of the device design results.

[0040] 5. Significantly improves the overall performance and analytical accuracy of the device.

[0041] By comprehensively evaluating the Pareto optimal parameter combination, the optimal device structure parameter scheme is determined, thereby effectively improving the accuracy and precision of the PGNAA device for sample analysis.

[0042] 6. The method is highly versatile and has a wide range of applications.

[0043] The method of this invention is based on Monte Carlo simulation and multi-objective optimization framework, and has good versatility and scalability, and can be applied to the optimization design requirements of PGNAA devices in different application scenarios. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the structure of the PGNAA device provided in an embodiment of the present invention;

[0045] Figure 2 This is a diagram showing the neutron energy spectrum distribution of different materials provided in the embodiments of the present invention;

[0046] Figure 3 This is a diagram showing the neutron flux distribution in the region surrounding the neutron source provided in an embodiment of the present invention. Figure 3 The upper figure shows the neutron flux distribution from the neutron source to the detector region, and the lower figure shows the neutron flux distribution from the neutron source to the sample region.

[0047] Figure 4 This is a graph showing the correlation analysis results of the slowing body module indicators provided in this embodiment of the invention;

[0048] Figure 5 This is a diagram showing the optimized structural parameters of the moderating body module provided in this embodiment of the invention.

[0049] Figure 6 This is a correlation analysis diagram of the shielding module indicators provided in the embodiments of the present invention;

[0050] Figure 7 This is a diagram showing the optimized structural parameters of the shielding module provided in an embodiment of the present invention.

[0051] Figure 8 This is a graph showing the fitting results of the characteristic peak count of Ti element under different device conditions as a function of element content, provided in the embodiments of the present invention.

[0052] Figure 9 This is a graph showing the fitting results of the Fe element characteristic peak count as a function of element content under different device conditions provided in the embodiments of the present invention;

[0053] In the figure, 1-neutron source, 2-detector, 3-neutron control module, 4-protection module, 5-sample. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Please see Figures 1-9 This invention provides a technical solution: a structural design method for a multi-objective collaborative optimization PGNAA device. The PGNAA device consists of a neutron source 1, a detector 2, a neutron control module 3, and a protection module 4 (shielding body) for device protection. The neutron source 1 is one of a DT neutron generator, a DD neutron generator, or an Am-Be source. The detector 2 of the PGNAA device is one of a BGO, sodium iodide, lanthanum bromide, or high-purity germanium detector. The neutron control module 3 includes one or more of a neutron moderator module, a reflector module, a multiplier module, and a collimator module.

[0056] The PGNAA device structural design method includes the following steps:

[0057] S1: Initial selection of structural materials and determination of optimization conditions for the PGNAA device; Based on the Monte Carlo simulation method, the neutron flux distribution characteristics at the sample under different material conditions are calculated, and materials with superior performance are selected as candidate materials for the structural optimization design of the PGNAA device; At the same time, the neutron flux distribution in the region surrounding neutron source 1 of the PGNAA device is calculated, and the region with low neutron flux distribution is used as the boundary range for the structural size design and parameter optimization of the PGNAA device, and the size range of the component modules to be optimized is determined; Since the distribution of neutron flux in space usually shows a gradual decay from near to far with neutron source 1 as the center, but its specific decay law is affected by the type of neutron source 1, energy spectrum characteristics, etc., it is difficult to obtain directly through analytical methods. Therefore, to quantitatively obtain the actual distribution of neutrons in the space surrounding neutron source 1, this invention employs the Monte Carlo particle transport simulation method to numerically calculate the neutron flux distribution. Specifically, the space surrounding neutron source 1 is gridded using the FMESH card in the MCNP program, and the neutron flux distribution at different spatial locations is calculated, thereby obtaining the distribution characteristics of neutron flux as a function of spatial location. Based on this, regions where neutron flux decays to a low level are selected according to the neutron flux distribution results, serving as the boundary range for related structural size design and parameter optimization. This method reasonably limits the spatial size range of simulation and optimization, thereby avoiding unnecessary simulation calculations in regions with minimal neutron flux contribution, reducing computational complexity, and improving the efficiency and relevance of the device's structural design and parameter optimization.

[0058] The different materials include LiH, PE, W, Pb, C, and Al; the samples detected by the PGNAA device include multiple elements from Mg, H, C, Ti, Fe, Mn, Cu, Al, and Cl; the neutron distribution of different materials in different energy regions is as follows: Figure 2 As shown in the figure, we can see that in the mid-energy region (0.1 MeV~1 MeV), W and Pb materials have high neutron fluxes under moderation. However, PGNAA devices usually require more thermal neutrons to increase the yield of captured gamma rays. Therefore, these two materials can be selected as materials for initial moderation of high-energy neutrons. In the thermal neutron energy region (0~1e-6 MeV), LiH and PE materials have high thermal neutron fluxes under moderation. Since LiH can both moderate and absorb neutrons, it is selected as the shielding material, and PE is selected as the moderator module material.

[0059] Without adding any structural materials, the neutron flux distribution from neutron source 1 to detector 2 and the sample region is as follows: Figure 3 As shown, according to Figure 3Analysis shows that the simulation calculations used neutron source 1 as the center and employed the FMESH grid statistical method to divide the spatial region, with a grid size of 2 cm × 2 cm, to obtain the variation of neutron flux at different radial distances; from Figure 3 It is known that the neutron flux is relatively high in the region near neutron source 1 and gradually decreases with increasing distance. When the distance to neutron source 1 exceeds about 14 cm, the neutron flux decreases significantly and tends to stabilize. Continuing to add structural materials outside this range has limited effect on improving the effective neutron utilization rate of the system, but instead increases the size and structural complexity of the device. At the same time, as the distance between neutron source 1 and the sample and detector 2 increases, the probability of effective characteristic gamma ray generation and geometric detection efficiency will decrease, which is not conducive to improving the overall measurement performance of the device. Therefore, the area of ​​about 14 cm around neutron source 1 is determined as the effective neutron interaction region. In this embodiment, considering that the radius of the neutron tube used is about 3.8 cm, after deducting the size of the neutron source 1, the maximum size of the final optimized structural design of the device is determined to be 10 cm.

[0060] S2: Construct a parametric model of the PGNAA device; Based on the candidate materials obtained in step S1, establish a parametric model of the PGNAA device to be optimized in Monte Carlo simulation software, clarifying the structural form, geometric dimensions, and material composition of the PGNAA device to be optimized. A schematic diagram of the PGNAA device is shown below. Figure 1 As shown in the figure, the arrows indicate the path of the neutrons generated by neutron source 1 that undergo inelastic scattering and capture reactions with sample 5, producing transient gamma rays that are recorded by detector 2.

[0061] S3: Simulation calculation of PGNAA device response under different conditions; Based on the Monte Carlo simulation method, the PGNAA device response under different parameter combinations is simulated for the parameters to be optimized in step S2, and the multi-objective evaluation function dataset corresponding to each parameter combination is obtained; In the multi-objective evaluation function dataset, the thermal neutron flux at the sample, the neutron flux in the detector, the effective characteristic gamma ray count, and the signal-to-noise ratio evaluation function parameters are obtained through Monte Carlo simulation, and the signal-to-noise ratio of the PGNAA device is calculated by the following formula: .

[0062] When performing simulation calculations of the PGNAA device response, it is necessary to consider the complete neutron and gamma transport processes of the PGNAA device. Specifically, this involves using various tallies to statistically analyze key physical quantities, including the following steps:

[0063] S31: Records the energy spectrum information deposited in the detector;

[0064] S32: Record the effective gamma signal generated by the sample;

[0065] S33: Record the neutron flux inside the detector;

[0066] S34: Record the thermal neutron flux on the sample surface.

[0067] S4: Correlation analysis of multi-objective evaluation functions; perform linear correlation analysis on the multi-objective evaluation function dataset of the PGNAA device to clarify the interrelationships between the evaluation functions, eliminate highly redundant evaluation functions, and thus select key evaluation functions for multi-objective collaborative optimization; the linear correlation analysis process is as follows: Suppose there are m evaluation indicators Y:

[0068] ;

[0069] Any two different evaluation indicators Y j and Y k The Pearson correlation coefficient between (j, k=1, 2, ..., m) is defined as:

[0070] ;

[0071] in, Let be the values ​​of the j-th and k-th evaluation indicators under the i-th sample condition, respectively. The evaluation index Y is respectively j Y k The sample mean; This indicates the degree of linear correlation between two indicators; when The closer the value is to 1, the higher the correlation between the evaluation indicators. It should be noted that in the simulation calculation stage in steps S2-S3, the present invention first obtains a multi-objective evaluation result set composed of multiple performance indicators. These evaluation indicators include multiple parameters characterizing the performance of the device, such as effective signal strength, signal-to-noise ratio, and thermal neutron flux at the sample location. There may be strong correlations or redundancies between different indicators.

[0072] If all evaluation indicators are directly introduced into the multi-objective evaluation or optimization process, it may lead to duplicate weighting of some indicators with similar physical meanings, thus affecting the objectivity and stability of the evaluation results. Furthermore, it will significantly increase the computational complexity and cost of multi-objective optimization. Therefore, this invention filters the initially obtained multi-index evaluation data before multi-objective evaluation, retaining evaluation function data that can represent the key performance characteristics of the system and have low correlation with each other, for subsequent multi-objective evaluation and parameter optimization.

[0073] Through the above screening process, the redundancy of evaluation indicators can be reduced while ensuring the integrity of the physical meaning of the evaluation results. This improves the computational efficiency and convergence stability of the multi-objective evaluation and optimization process, thereby enhancing the overall effectiveness of the device parameter optimization design.

[0074] S5: Multi-objective collaborative optimization and constraint screening; Based on the multi-objective evaluation function dataset obtained in step S4, the dataset is processed and analyzed using multi-objective optimization theory; Under the premise of satisfying the mass and volume engineering constraints, the device parameter space is systematically searched and screened to obtain the Pareto optimal parameter combination that satisfies the constraints; Among them, when optimizing the neutron control module 3, the optimization objective formula is as follows: In the formula, F EF (X i F is the effective signal target evaluation function. TNF (X i Let F be the target evaluation function for the thermal neutron flux at the sample. SNR (X i Mass(X) is the signal-to-noise ratio target evaluation function. i ) and Volume(X i ) represent the structural mass and volume constraint functions of the PGNAA device, with constraint values ​​determined based on specific engineering requirements; the optimization objective formula for the protection module is as follows: Among them, F DNF (X i ) represents the neutron flux inside the detector.

[0075] The systematic search and filtering of the device parameter space includes the following steps:

[0076] S51: Input multi-objective evaluation function dataset;

[0077] S52: Normalize the data;

[0078] S53: Perform non-dominated sorting of candidate solutions based on Pareto dominance;

[0079] S54: Extract candidate solutions that are not dominated by any other solutions to form a Pareto optimal solution set;

[0080] S55: Output the Pareto optimal solution set and the corresponding multi-objective function values.

[0081] By introducing the aforementioned multi-objective optimization mechanism, the system performance imbalance caused by single-index optimization can be avoided, ensuring that the selected parameter combination exhibits superior performance across multiple key performance indicators. This improves the rationality and engineering applicability of the overall structural design of the device. In this invention, the search and selection of the device parameter space is a multi-objective optimization process conducted under engineering constraints. Specifically, firstly, a device parameter space is constructed within the aforementioned defined parameter range, and the feasibility of parameter combinations is determined based on mass and volume engineering constraints, eliminating parameter combinations that do not meet the constraints.

[0082] Based on this, for parameter combinations that satisfy the constraints, multiple performance evaluation indices are compared according to the principle of multi-objective optimization. Parameter combinations that are not simultaneously outperformed by all other parameter combinations in each evaluation index are selected, thus forming a Pareto-optimal set of parameter combinations that satisfy the engineering constraints. The parameter combinations in this set achieve reasonable trade-offs among multiple performance indices and are suitable for the optimized design of the device's structure and parameters.

[0083] The above search and filtering process can be implemented through various multi-objective optimization methods. Its core lies in identifying and extracting Pareto optimal solutions that satisfy engineering constraints in the parameter space, without relying on any specific algorithm. The search for candidate solutions can employ methods including but not limited to population evolution, swarm intelligence, or random sampling. Subsequently, candidate solutions can be filtered through non-dominance relationship determination, crowding calculation, or other multi-objective optimization criteria, thereby retaining the set of Pareto optimal solutions that satisfy engineering constraints.

[0084] S6: Determine the optimal parameter solution; For the Pareto optimal parameter combination obtained in step S5, further optimize each parameter combination using an objectively weighted comprehensive evaluation method (a weighted evaluation method combining entropy weight method and TOPSIS) to determine the structural module combination under optimal conditions, thereby obtaining the optimal PGNAA device model parameters for the final structural design of the device; It should be noted that the objectively weighted comprehensive evaluation method of this invention is a technical solution for comprehensively ranking and optimizing multiple objective performance indicators for the obtained Pareto optimal solution set. There are multiple methods to achieve this objective. In this embodiment, the entropy weight method combined with the TOPSIS method is adopted.

[0085] Specifically, the entropy weight method is used to objectively determine the weight of each performance index based on the data distribution of each index in the Pareto solution set. This method does not rely on subjective human weighting, but determines the weight by analyzing the dispersion or information entropy of the index. The greater the dispersion and the higher the information content of the index, the greater its weight, thus reflecting its contribution to distinguishing different solutions.

[0086] After obtaining the weights of each indicator, the TOPSIS method is used for comprehensive evaluation. The principle of TOPSIS is: in the Pareto optimal solution set, the ideal solution (the best value of each indicator) and the negative ideal solution (the worst value of each indicator) are determined, and the weighted distance between each solution and the ideal solution and the negative ideal solution is calculated. The closer the solution is to the ideal solution and the farther it is from the negative ideal solution, the better its comprehensive performance evaluation result.

[0087] By combining the entropy weight method and TOPSIS, this invention can objectively select the device structure parameters with the best overall performance from the Pareto optimal solution set, achieve a reasonable trade-off between multiple objective performances, and does not rely on subjective weighting judgments, thereby improving the scientificity and repeatability of device optimization selection.

[0088] In this embodiment, for the moderating body module, correlation analysis of indicators is used, such as... Figure 4 As shown in the figure (where SNR is the signal-to-noise ratio, EF is the effective signal, and TNF is the thermal neutron flux at the sample), the three indicators were not highly correlated (correlation coefficient r < 0.9). Therefore, the three-indicator dataset was selected for multi-objective collaborative optimization, as shown in the figure. Figure 5 As shown, during the optimization process of the moderator module, based on the aforementioned correlation analysis results, three representative indicators were selected from multiple performance evaluation metrics as multi-objective collaborative optimization objectives: signal-to-noise ratio (SNR), effective gamma signal intensity (EF), and thermal neutron flux (TNF) at the sample. Figure 5 The three coordinate axes in the diagram; Figure 5 The blue scatter dots represent the Pareto optimal solution set obtained through a multi-objective optimization mechanism in step S5, under the constraints of mass and volume engineering. None of the parameter combinations in this solution set are simultaneously superior to all other parameter combinations among the three selected evaluation indicators, reflecting a reasonable trade-off between different performance indicators. Figure 5 The red pentagram in the middle represents the optimal combination of parameters for overall performance obtained by further selecting from the above Pareto optimal solution set using a comprehensive evaluation method based on objective weighting in step S6; it should be noted that... Figure 5 The numerical range of each coordinate axis is 0 to 1. This is because, in the process of multi-objective optimization and comprehensive evaluation, the evaluation indicators with different dimensions and numerical ranges are normalized to eliminate the influence of dimensional differences on the evaluation results and facilitate the comprehensive comparison between different indicators.

[0089] For shielding modules, correlation analysis of indicators, such as... Figure 6As shown (where DNF is the neutron flux inside the detector), a high negative correlation was found between SNR and DNF, and a high positive correlation between EF and DNF. The overall goal of device structure optimization is to improve the measurement performance of the PGNAA system. System measurement performance can be characterized by key indicators such as effective signal and signal-to-noise ratio. EF and SNR are used to characterize the device's ability to detect characteristic gamma rays of target elements and measurement stability. The higher the value, the better the system measurement performance. For the shielding module, the smaller the neutron flux inside the detector, the better, for the following two reasons: 1. Detector protection: Excessive neutrons entering the detector may cause radiation damage to the detector, affecting its long-term stability and service life; 2. Noise suppression: Neutrons are captured or scattered in the detector and its surrounding structure, generating non-target gamma background signals, thereby increasing the spectral background level and reducing the system signal-to-noise ratio. Therefore, in the optimization of the shielding module, DNF should be as small as possible, while EF and SNR should be as high as possible. This embodiment reduces the DNF target evaluation function dataset and selects the EF and SNR dual-index dataset for multi-objective collaborative optimization, such as... Figure 7 As shown, this is the optimal parameter model of the shielding module. During the optimization process of this module, based on the aforementioned correlation analysis, two indicators were selected from multiple performance evaluation indicators as multi-objective optimization objectives, namely the effective gamma signal strength (EF) and the signal-to-noise ratio (SNR), corresponding to the two coordinate axes in the figure. The blue scatter points in the figure represent the Pareto optimal solution set that satisfies the engineering constraints in step S5, and the red pentagrams represent the comprehensive performance optimal solution selected based on the objective weighted comprehensive evaluation method in step S6.

[0090] Figure 8 and Figure 9 The changes in the characteristic peak counts of Ti and Fe elements with their content are shown under different device conditions (device 1 is a device designed using the method of this invention, and device 2 is a device designed using the prior art through signal-to-noise ratio optimization). The slope represents the analytical sensitivity of the device for the element; higher sensitivity indicates better measurement performance. Figure 8 and Figure 9 It can be seen that the method proposed in this invention has higher sensitivity and statistical power, thus proving the effectiveness and superiority of the structural design method of the PGNAA device proposed in this invention.

[0091] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A structural design method for a multi-objective collaboratively optimized PGNAA device, wherein the PGNAA device comprises a neutron source, a detector, a neutron control module, and a protection module for device protection, wherein the neutron control module comprises one or more of a neutron moderator module, a reflector module, a multiplier module, and a collimator module, characterized in that, Includes the following steps: S1: Initial selection of structural materials for the PGNAA device and determination of optimization conditions and boundaries; Based on the Monte Carlo simulation method, the neutron flux distribution characteristics at the sample under different material conditions are calculated, and materials with superior performance are selected as candidate materials for the structural optimization design of the PGNAA device; At the same time, the neutron flux distribution in the region surrounding the neutron source of the PGNAA device is calculated, and the region with low neutron flux distribution is used as the boundary range for the structural size design and parameter optimization of the PGNAA device, and the size range of the component modules to be optimized is determined. S2: Construct a parametric model of the PGNAA device; Based on the candidate materials obtained in step S1, establish a parametric model of the PGNAA device to be optimized in Monte Carlo simulation software, and clarify the structural form, geometric dimensions and material composition of the PGNAA device to be optimized. S3: Simulation calculation of PGNAA device response under different conditions; Based on the Monte Carlo simulation method, the PGNAA device response under different parameter combinations is simulated for the parameters to be optimized in step S2, and the multi-objective evaluation function dataset corresponding to each parameter combination is obtained. S4: Correlation analysis of multi-objective evaluation functions; perform linear correlation analysis on the multi-objective evaluation function dataset of the PGNAA device to clarify the interrelationships between the evaluation functions, eliminate highly redundant evaluation functions, and thus select key evaluation functions for multi-objective collaborative optimization; the linear correlation analysis process is as follows: Suppose there are m evaluation indicators Y: ； Any two different evaluation indicators Y j and Y k The Pearson correlation coefficient between (j, k=1, 2, ..., m) is defined as: ； in, Let be the values ​​of the j-th and k-th evaluation indicators under the i-th sample condition, respectively. The evaluation index Y is respectively j Y k The sample mean; This indicates the degree of linear correlation between two evaluation indicators; when The closer the value is to 1, the higher the correlation between the evaluation indicators. S5: Multi-objective collaborative optimization and constraint selection; Based on the multi-objective evaluation function dataset obtained in step S4, the dataset is processed and analyzed using multi-objective optimization theory; When optimizing the neutron control module, the optimization objective formula is as follows: In the formula, F EF (X i F is the effective signal target evaluation function. TNF (X i Let F be the target evaluation function for the thermal neutron flux at the sample. SNR (X i Mass(X) is the signal-to-noise ratio target evaluation function. i ) and Volume(X i ) represent the structural mass and volume constraint functions of the PGNAA device, with constraint values ​​determined based on specific engineering requirements; the optimization objective formula for the protection module is as follows: Among them, F DNF (X i The neutron flux within the detector is denoted as ; and under the premise of satisfying mass and volume engineering constraints, a systematic search and screening of the device parameter space is performed, which includes the following steps: S51: Input multi-objective evaluation function dataset; S52: Normalize the data; S53: Perform non-dominated sorting of candidate solutions based on Pareto dominance; S54: Extract candidate solutions that are not dominated by any other solutions to form a Pareto optimal solution set; S55: Output the Pareto optimal solution set and the corresponding multi-objective function values; thereby obtaining the Pareto optimal parameter combination that satisfies the constraints; S6: Determine the optimal parameter solution; For the Pareto optimal parameter combination obtained in step S5, further optimize each parameter combination through an objective weighted comprehensive evaluation method to determine the structural module combination under optimal conditions, thereby obtaining the optimal PGNAA device model parameters for the final design of the PGNAA device.

2. The structural design method for the PGNAA device with multi-objective collaborative optimization according to claim 1, characterized in that: In step S1, the different materials include LiH, PE, W, Pb, C, and Al; the neutron source in the structure of the PGNAA device is one of the DT neutron generator, DD neutron generator, and Am-Be source; the detector in the structure of the PGNAA device is one of the BGO, sodium iodide, lanthanum bromide, and high-purity germanium detector; and the sample detected by the PGNAA device includes multiple elements selected from Mg, H, C, Ti, Fe, Mn, Cu, Al, and Cl.

3. The structural design method for the PGNAA device with multi-objective collaborative optimization according to claim 1, characterized in that: In step S3, when performing simulation calculations of the PGNAA device response, it is necessary to consider the complete neutron and gamma transport processes of the PGNAA device. Specifically, this involves using various tallies to statistically analyze key physical quantities, including the following steps: S31: Records the energy spectrum information deposited in the detector; S32: Record the effective gamma signal generated by the sample; S33: Record the neutron flux inside the detector; S34: Record the thermal neutron flux on the sample surface.

4. The structural design method for the PGNAA device with multi-objective collaborative optimization according to claim 1, characterized in that: In step S3, the evaluation function parameters for thermal neutron flux at the sample, neutron flux inside the detector, effective characteristic gamma ray count, and signal-to-noise ratio in the multi-objective evaluation function dataset are obtained through Monte Carlo simulation. The signal-to-noise ratio of the PGNAA device is calculated using the following formula: .

5. The structural design method for the PGNAA device with multi-objective collaborative optimization according to claim 1, characterized in that: The objective weighting comprehensive evaluation method is a weighted evaluation method that combines the entropy weight method with TOPSIS.

Images