Irradiation prediction method for irradiated goods

Abstract

The invention provides an irradiation prediction method for irradiated goods, and the method comprises the steps: optimizing irradiation parameters according to an irradiation uniformity evaluation result, including adjusting irradiation time, irradiation intensity and an irradiation mode, and obtaining an optimal irradiation parameter combination through an iterative optimization method; according to the optimized irradiation parameters, an irradiation scheme including irradiation time, irradiation intensity and an irradiation mode is made, the irradiation scheme is input into an irradiation equipment control system, and the irradiation process is automatically executed; and carrying out sampling detection on the goods after actual irradiation, measuring irradiation doses at different positions in the goods, evaluating an actual irradiation effect, verifying the accuracy of a simulation model and optimization parameters, and correcting the simulation model and an optimization algorithm according to a verification result.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for predicting the irradiation of irradiated goods. Background Art

[0002] The irradiation dose control of irradiated goods is a complex technical problem. The unevenness of the material, density, and shape of the goods leads to extremely uneven distribution of radiation inside the goods, making it difficult to achieve consistency in dose in various parts. Especially for goods with large density differences, it is difficult for radiation to penetrate high-density areas, while over-irradiation may occur in low-density areas, with some areas under-irradiated and some areas over-irradiated, posing great challenges to irradiation quality control. At the same time, the irradiation process lacks effective real-time monitoring means, irradiation dose data cannot be obtained in real time, the irradiation process lacks controllability, abnormal situations are difficult to detect and handle in a timely manner, and there is a risk of irradiation out of control. In summary, the complexity and variability of cargo characteristics, the difficulty of optimizing irradiation parameters, and the blindness of irradiation process monitoring have jointly led to many challenges in the dose control technology of irradiated goods. It is urgent to carry out targeted research to overcome key technical difficulties, achieve uniform and controllable dose distribution of irradiated goods, and improve irradiation quality and efficiency. Summary of the invention

[0003] The present invention provides a method for predicting the irradiation of irradiated goods, which mainly includes:

[0004] According to the characteristic parameters of the irradiated goods, including the density, shape and size of the goods, a three-dimensional model of the irradiated goods is established to simulate the radiation dose distribution during the irradiation process, obtain the radiation dose values ​​at various positions inside the irradiated goods, and generate a radiation dose distribution map to identify the irradiation blind spots or over-irradiated areas and evaluate the irradiation uniformity;

[0005] According to the irradiation uniformity evaluation results, optimize the irradiation parameters, including adjusting the irradiation time, irradiation intensity and irradiation mode, and obtain the optimal irradiation parameter combination through iterative optimization method;

[0006] Input the optimized irradiation parameters into the computer simulation model, re-simulate the irradiation process, obtain the optimized irradiation dose distribution diagram, and evaluate the optimization effect. If the expected irradiation uniformity and efficiency targets are not achieved, continue to iterate and optimize to obtain the optimal irradiation parameter combination until the requirements are met;

[0007] According to the optimized irradiation parameters, formulate an irradiation plan, including irradiation time, irradiation intensity and irradiation method, and input the irradiation plan into the irradiation equipment control system to automatically execute the irradiation process;

[0008] During the irradiation process, sensors or detectors arranged inside the cargo are used to collect irradiation dose data in real time and monitor the dose distribution during the irradiation process. If an abnormal deviation from the expected dose distribution is found, an early warning is triggered, the irradiation process is interrupted, and the irradiation parameters are readjusted;

[0009] Compare the real-time collected irradiation dose data with the simulated predicted dose distribution map, calculate the deviation value, and if the deviation exceeds the preset threshold, update the simulation model parameters to improve the simulation accuracy. At the same time, associate the measured data, cargo information and irradiation parameters to form a traceable irradiation quality record;

[0010] Sampling inspection is carried out on the goods after actual irradiation, the radiation dose at different positions inside the goods is measured, the actual irradiation effect is evaluated, the accuracy of the simulation model and optimization parameters is verified, and the simulation model and optimization algorithm are corrected according to the verification results.

[0011] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0012] The present invention discloses an irradiation prediction method for irradiated goods. The method first establishes a three-dimensional model according to the characteristics of the goods, simulates the irradiation dose distribution through simulation technology, and identifies the irradiation dead angle or over-irradiation area. Then, a dynamic γ irradiation device with automatic surface change of a single-plate source is used to calculate the overall average absorbed dose, and the irradiation parameters are optimized according to the uniformity evaluation results. The optimized parameters are input into the simulation model for iterative verification until the requirements are met. During the actual irradiation process, the dose distribution is monitored in real time by sensors, and an early warning is triggered and the parameters are adjusted when an abnormality is found. At the same time, the measured data is compared with the simulation prediction, and the model is continuously updated to improve the accuracy. Finally, the irradiated goods are sampled and tested to verify the accuracy of the simulation and optimization, forming a closed-loop improvement mechanism. The present invention combines computer simulation with real-time monitoring to achieve precise control and continuous optimization of the irradiation process, significantly improve the irradiation uniformity and efficiency, and ensure the traceability of the irradiation quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 The present invention is a flowchart of a method for predicting irradiation of irradiated goods.

[0014] Figure 2 The figure is a schematic diagram of a method for predicting irradiation of irradiated goods according to the present invention.

[0015] Figure 3 It is another schematic diagram of the irradiation prediction method of irradiated goods according to the present invention. DETAILED DESCRIPTION

[0016] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0017] like Figure 1-3 In this embodiment, a method for predicting the irradiation of irradiated goods may specifically include:

[0018] S101. Based on the characteristic parameters of the irradiated cargo, including the density, shape and size of the cargo, a three-dimensional model of the irradiated cargo is established to simulate the radiation dose distribution during the irradiation process, obtain the radiation dose values ​​at various positions inside the irradiated cargo, and generate a radiation dose distribution map to identify irradiation blind spots or over-irradiated areas and evaluate irradiation uniformity.

[0019] Generate a corresponding three-dimensional grid model based on the density, shape and size parameters of the cargo, perform adaptive grid refinement on complex shape areas, evaluate grid quality and optimize grid structure, and obtain a high-precision three-dimensional model database of irradiated cargo; set the energy spectrum of the radiation source and the initial ray parameters of spatial distribution, and use the Monte Carlo method to simulate the transmission process of photons in the three-dimensional model, including random number generation, photon path tracing and energy deposition calculation; record the energy absorbed by the grid unit, and obtain the irradiation dose distribution data through cumulative statistics; use a three-dimensional interpolation algorithm to spatially reconstruct the dose data of discrete grid points, map the dose distribution data to a three-dimensional visualization model, and use the HSV color space to map dose values ​​to colors; based on the dose distribution map, use the DBSCAN density clustering algorithm to identify continuous areas that are higher or lower than the expected dose, obtain irradiation dead angles and over-irradiation areas, calculate the volume proportion and dose deviation degree of each area, and quantitatively evaluate the overall irradiation uniformity index by calculating the dose variation coefficient.

[0020] Specifically, a three-dimensional model database of irradiated goods is constructed, and the corresponding three-dimensional grid model is generated according to the density, shape and size parameters of the goods. The octree algorithm is used to adaptively refine the grid in complex shape areas. The grid quality is evaluated and the grid structure is optimized through the Jacobian matrix condition number to obtain a high-precision three-dimensional model of irradiated goods. The radiation source parameters and ray tracing algorithm are set, and the initial ray parameters such as the radiation source energy spectrum and spatial distribution are set. The Monte Carlo method is used to simulate the transmission process of photons in the three-dimensional model, including random number generation, photon path tracing and energy deposition calculation. The energy absorbed by each grid unit is recorded, and the irradiation dose distribution data is obtained by cumulative statistics. The irradiation dose distribution map is generated, and the dose data of discrete grid points are spatially reconstructed using a three-dimensional interpolation algorithm. The dose distribution data is mapped to a three-dimensional visualization model using volume rendering technology. The HSV color space is used to map the dose value to color. Cold colors are used in low-dose areas and warm colors are used in high-dose areas. The isodose surface is drawn to depict the uniformity of irradiation. Identify irradiation abnormal areas and evaluate uniformity. Based on the generated dose distribution map, use the DBSCAN density clustering algorithm to identify continuous areas with higher or lower doses than the expected dose, obtain irradiation dead angles and over-irradiation areas, calculate the volume proportion of each area and the degree of dose deviation, and quantitatively evaluate the overall irradiation uniformity index by calculating the dose variation coefficient (ratio of standard deviation to mean). When constructing the 3D model of irradiated cargo, assume that the cargo is cylindrical with a density of 1.2 g / cm 3, with a height of 50 cm and a diameter of 30 cm. The octree algorithm was used for adaptive mesh refinement, with the initial mesh size set to 5 cm and gradually refined to 0.5 cm on the cylindrical surface and edge area. The Jacobian matrix condition number was used to evaluate the mesh quality, with a threshold set to 10, and cells exceeding the threshold were locally reconstructed. The optimized three-dimensional model contained about 100,000 mesh cells. A cobalt-60 radiation source was set with energies of 1.17 MeV and 1.33 MeV, and the spatial distribution was a cylindrical envelope cargo. The Monte Carlo method was used to simulate the transmission of 100 million photons in the model, and the Mersenne Twi ster algorithm was used to generate random numbers, trace the path of each photon, and calculate the energy deposition. The dose distribution data of each grid cell was obtained by cumulative statistics. The dose data of discrete grid points were spatially reconstructed using the cubic spline interpolation algorithm to form a continuous three-dimensional dose distribution function. The dose distribution data was mapped to the visualization model using volume rendering technology, and the HSV color space was used for mapping. The dose range of 0-10kGy corresponds to hue 0°-360°, the saturation is fixed at 100%, and the brightness changes linearly between 80% and 100%. A 5kGy isodose surface was drawn to show the irradiation uniformity. Based on the generated dose distribution map, the DBSCAN density clustering algorithm was used to identify abnormal areas, setting the neighborhood radius to 2cm and the minimum number of points to 10. Two irradiation dead spots were identified, located at the bottom edge of the cylinder, accounting for 0.5% of the volume; one over-irradiated area was located at the central axis of the cylinder, accounting for 1.2% of the volume. The overall dose variation coefficient was calculated to be 0.15, indicating that the irradiation uniformity reached a good level.

[0021] The total energy of ionizing radiation absorbed by the material is analyzed and obtained by using the calculation method of the overall average absorbed dose of the dynamic γ irradiator with automatic surface replacement of a single plate source in a given irradiated product, combined with the characteristics of the dynamic γ irradiator with automatic surface replacement of a single plate source and the dose distribution law in the irradiation container.

[0022] A three-dimensional geometric model of the irradiation device was established according to the size of the irradiation container and the activity of the plate source. The Monte Carlo method was used to simulate the transmission process of gamma rays in the container to obtain the spatial dose rate distribution data in the container. The dose rate distribution data was processed by a three-dimensional spline interpolation algorithm to generate a continuous dose rate distribution function, and the cumulative absorbed dose at any point in the irradiation container was calculated in combination with the plate source replacement frequency and the total irradiation time. The dose distribution was corrected for the cumulative absorbed dose in combination with the attenuation coefficient of the irradiated product to obtain the actual dose distribution. Based on the actual dose distribution, a dosimeter layout plan was set up, and dosimeters were virtually placed at key positions in the irradiated product. The maximum dose point, minimum dose point and average dose point were determined through dose gradient analysis. The virtual dosimeter measurement values ​​were converted into actual measurement values ​​in combination with the dosimeter response characteristics and calibration factors, and the overall average absorbed dose was calculated using the volume weighted average method.

[0023] Specifically, according to the size of the irradiation container and the activity of the plate source, a three-dimensional geometric model of the irradiation device is established, and the Monte Carlo method is used to simulate the transmission process of gamma rays in the container, including the generation, interaction and energy deposition of gamma photons. In the simulation, the position of the source is periodically changed to reflect the automatic face-changing characteristics of the single-plate source, and the spatial dose rate distribution data in the container is calculated. According to the dose rate distribution data, a continuous dose rate distribution function is generated by a three-dimensional spline interpolation algorithm, and the cumulative absorbed dose at any point in the irradiation container is calculated by combining the plate source face-changing frequency and the total irradiation time. Considering the attenuation coefficient of the irradiated product, the dose distribution is corrected using the exponential decay law to obtain the actual dose distribution considering the product influence. Based on the corrected cumulative absorbed dose distribution, a dosimeter layout scheme is set, and dosimeters are virtually placed at key positions in the irradiated product. The maximum dose point, minimum dose point and average dose point are determined by dose gradient analysis, and virtual dosimeters are arranged at these positions and surrounding areas, and the absorbed dose values ​​of each measurement point are recorded. The recorded virtual dosimeter measurement values ​​are converted into actual measurement values ​​in combination with the dosimeter response characteristics and calibration factors. The overall average absorbed dose was calculated by the volume weighted average method, and the total energy of ionizing radiation absorbed by the material was calculated by energy deposition calculation in combination with the mass and volume of the irradiated product. A three-dimensional model of an irradiation container with a size of 2m×2m×3m was established for a dynamic γ irradiation device with a single plate source automatically changing its surface, and the plate source activity was set to 100,000 Curies. The Monte Carlo method was used to simulate 100 million γ photons, in which the photoelectric effect, Compton scattering and electron pair effect were considered, and the energy deposition was approximated by continuous slowing down. The plate source automatically changed its surface every 60 seconds, and the source position was updated every 1 million photons in the simulation. The 10×10×15 grid dose rate distribution data was obtained by this method. The discrete data was converted into a continuous function using the cubic spline interpolation algorithm, and the total irradiation time was set to 10 hours to calculate the cumulative absorbed dose distribution in the container. Assuming that the irradiated product has a density of 1.2g / cm 3 A cylindrical object with a height of 1.5m and a diameter of 1m, using a mass attenuation coefficient of 0.0495cm 2 / g corrected dose distribution. In the corrected dose distribution, the maximum dose point of 2000Gy, the minimum dose point of 800Gy and the average dose point of 1400Gy were determined by gradient analysis. A virtual dosimeter was placed every 10cm within 20cm around these positions, for a total of 27. Considering the energy response characteristic of the dosimeter of 0.95 and the calibration factor of 1.05, the virtual measurement value was converted into the actual value. The overall average absorbed dose was calculated to be 1380Gy by the volume weighted average method. Combined with the product mass of 1413kg, the total energy of ionizing radiation absorbed by the material was calculated to be 1.95×106J.

[0024] S102. According to the irradiation uniformity evaluation result, the irradiation parameters are optimized, including adjusting the irradiation time, irradiation intensity and irradiation mode, and the optimal irradiation parameter combination is obtained through an iterative optimization method.

[0025] An irradiation parameter optimization model is established, which includes irradiation time, irradiation intensity and irradiation mode as independent variables, and dose uniformity as the objective function; according to the irradiation parameter optimization model, a genetic algorithm is used to generate an initial irradiation parameter combination population, in which the irradiation time and irradiation intensity are coded with real numbers, and the irradiation mode is coded with integers; the dose distribution under each group of irradiation parameter combinations is calculated by Monte Carlo simulation to obtain dose uniformity; arithmetic crossover and Gaussian mutation operations are performed on the irradiation parameter combinations in the population to obtain a new generation of irradiation parameter combinations; it is determined whether the optimization results of the new generation of irradiation parameter combinations meet the preset convergence conditions, and if so, the optimal irradiation parameter combination is output; if not, the optimal irradiation parameter combination is returned to continue iterating until the optimal irradiation parameter combination is obtained or the preset maximum number of iterations is reached.

[0026] Specifically, according to the irradiation uniformity evaluation results, an irradiation parameter optimization model is established, with irradiation time 0-24 hours, irradiation intensity 0-1000Gy / h and irradiation mode (single-sided, double-sided, multi-sided) as independent variables, and dose uniformity (the inverse of the ratio of maximum dose to minimum dose) as the objective function, setting the value range and constraints of each parameter. Genetic algorithm is used to generate the initial irradiation parameter combination population, the population size is set to 100, the irradiation time and intensity are coded with real numbers, and the irradiation mode is coded with integers. The dose distribution under each set of parameters is calculated by Monte Carlo simulation, the number of simulated photons is set to 10^6, and the dose uniformity is evaluated. Arithmetic crossover and Gaussian mutation operations are performed on the parameter combinations in the population, the crossover probability is set to 0.8, and the mutation probability is set to 0.1, to generate a new generation of parameter combinations. Repeat the Monte Carlo simulation and uniformity evaluation, and select high-quality individuals by roulette selection. Determine whether the optimization result meets the preset convergence conditions, that is, the optimal solution changes less than 1% for 10 consecutive generations or reaches the maximum number of iterations of 1000 generations. If satisfied, the optimal irradiation parameter combination is output, including the optimal irradiation time, intensity and mode. If not satisfied, return to the cross-mutation step to continue iteration until the optimal solution is obtained or the maximum number of iterations is reached. In the process of irradiation uniformity optimization, a parameter optimization model is first established, and the irradiation time range is set to 0-24 hours, the irradiation intensity range is set to 0-1000Gy / h, and the irradiation mode is set to 1 (single-sided), 2 (double-sided) or 3 (multi-sided). The dose uniformity (the inverse of the ratio of the maximum dose to the minimum dose) is used as the objective function, and the genetic algorithm is used for optimization. The initial population size is set to 100, and the real number encoding is used to represent the irradiation time and intensity, and the integer encoding is used to represent the irradiation mode. For example, an individual may be encoded as [15.5,800,2], indicating 15.5 hours of irradiation time, 800Gy / h irradiation intensity, and double-sided irradiation. The Monte Carlo method is used to simulate 10^6 photons and calculate the dose distribution under each set of parameters. In the evaluation process, it is assumed that the maximum dose is 2000Gy and the minimum dose is 1600Gy, so the uniformity is 0.8. The population is subjected to arithmetic crossover (probability 0.8) and Gaussian mutation (probability 0.1, standard deviation is 5% of the parameter range) to generate a new generation. For example, two parents

[0027] The crossover between [15.5,800,2] and [18.0,750,1] may produce the offspring [16.75,775,2]. The roulette wheel selection method is used to select high-quality individuals, and individuals with high uniformity are more likely to be selected. During the iteration process, the optimal solution of each generation is recorded, and the optimal solution is stopped when the change of the optimal solution for 10 consecutive generations is less than 1% (such as from 0.85 to 0.8585) or when 1000 generations are reached. The final output is the optimal parameter combination, such as [17.2,820,2], which means that the best irradiation plan is 17.2 hours, 820Gy / h, and double-sided irradiation.

[0028] S103, input the optimized irradiation parameters into the computer simulation model, re-simulate the irradiation process, obtain the optimized irradiation dose distribution diagram, evaluate the optimization effect, and if the expected irradiation uniformity and efficiency targets are not achieved, continue to iterate the optimization to obtain the optimal irradiation parameter combination until the requirements are met.

[0029] According to the optimized parameters of irradiation time, irradiation intensity and irradiation mode, the parameters are input into the Monte Carlo simulation program, which is provided with the geometric structure of the irradiation container, the characteristics of the radiation source and the properties of the irradiated objects; the Mersenne rotation algorithm is used to generate random numbers, which are used to simulate the trajectory of photons to obtain energy deposition distribution data; a three-dimensional irradiation dose distribution map is generated according to the energy deposition distribution data, and the resolution of the voxelized data is improved by a trilinear interpolation algorithm; the irradiation uniformity index of the three-dimensional irradiation dose distribution map is obtained, and the irradiation uniformity index includes the ratio of the maximum dose to the minimum dose and the dose variation coefficient, and the irradiation efficiency is calculated; if the irradiation uniformity index and the irradiation efficiency do not reach the preset target, the optimization algorithm parameters are adjusted, and the optimization algorithm parameters include the search range and the number of iterations, and the simulated annealing algorithm is used for re-optimization until the requirements are met or the maximum number of iterations is reached.

[0030] Specifically, the optimized irradiation time, irradiation intensity and irradiation mode parameters are input into the Monte Carlo simulation program, the geometric structure of the irradiation container, the characteristics of the radiation source and the properties of the irradiated object (including density, shape and size) are set, and the Mersenne rotation algorithm is used to generate random numbers to simulate the photon trajectory and calculate the energy deposition distribution. A three-dimensional irradiation dose distribution map is generated based on the simulation results, and a 0.1 cm 3 The voxel size was voxelized, the continuous spatial dose data was discretized, the trilinear interpolation algorithm was applied to improve the resolution, and the dose distribution was intuitively presented through the heat map color mapping scheme. The irradiation uniformity index was calculated, including the ratio of the maximum dose to the minimum dose and the dose variation coefficient (standard deviation divided by the mean value). The irradiation time was considered to calculate the irradiation efficiency (total dose divided by time), and the uniformity weight was set to 0.7 and the efficiency weight was set to 0.3 to comprehensively evaluate the optimization effect. It was judged whether the optimization effect reached the preset goal. If it reached, the optimal parameter combination was output. If not, the optimization algorithm parameters were adjusted, such as expanding the search range by 10% or increasing the number of iterations by 50 times, returning to the first step for re-optimization, and using the simulated annealing algorithm to avoid falling into the local optimal solution until the requirements were met or the maximum number of iterations was 1000 times. In the process of irradiation parameter optimization, the optimized irradiation time of 18 hours, irradiation intensity of 800Gy / h and double-sided irradiation method were first input into the Monte Carlo simulation program. The irradiation container is set to a rectangular parallelepiped of 2m×2m×3m, the activity of the cobalt-60 radiation source is 100,000 Curies, and the irradiated object has a density of 1.2g / cm 3, a cylinder with a diameter of 1m and a height of 1.5m. The Mersenne rotation algorithm was used to generate 10^9 random numbers to simulate the photon trajectory and calculate the energy deposition distribution. A three-dimensional irradiation dose distribution map was then generated, and the spatial dose data was discretized into 200×200×300 0.1cm 3 The resolution of the voxels was increased to 400×400×600 by using the trilinear interpolation algorithm. The dose distribution was presented by the heat map color mapping scheme, where blue indicated the low-dose area 0-500Gy and red indicated the high-dose area 1500-2000Gy. The maximum dose was calculated to be 1950Gy, the minimum dose was 1560Gy, the dose ratio was 1.25, the dose variation coefficient was 0.08, and the irradiation efficiency was 88.9Gy / h. The uniformity weight was set to 0.7 and the efficiency weight was set to 0.3, and the comprehensive score was 0.85. Since the preset target of 0.9 was not reached, the optimization algorithm parameters were automatically adjusted, the irradiation time search range was expanded to 16-20 hours, the irradiation intensity range was expanded to 720-880Gy / h, and the number of iterations was increased by 50 times. The simulated annealing algorithm with an initial temperature of 100 and a cooling coefficient of 0.95 was used to avoid falling into the local optimal solution, and the optimization process was restarted until the score target was reached or 1000 iterations were completed.

[0031] S104. Formulate an irradiation plan based on the optimized irradiation parameters, including irradiation time, irradiation intensity and irradiation method, and input the irradiation plan into the irradiation equipment control system to automatically execute the irradiation process.

[0032] A standard irradiation plan file including irradiation time, irradiation intensity and irradiation mode is generated according to the optimized irradiation parameters. The standard irradiation plan file converts the parameters into an instruction set recognizable by the equipment through a lookup table and a linear interpolation algorithm; the equipment compatibility check program is called to verify the compatibility of the standard irradiation plan file with the irradiation equipment. The compatibility check program determines whether the irradiation parameters are within the allowable range of the equipment. If they are exceeded, they are adjusted to the nearest feasible value; the safety limit check program is started to conduct a safety assessment on the standard irradiation plan file. The safety assessment uses a threshold comparison method to ensure that the plan meets safety specifications, including the maximum dose rate, the upper limit of the cumulative dose and radiation protection requirements; the checked and adjusted irradiation plan is transmitted to the irradiation equipment control system through a data interface, and the real-time monitoring module is activated to monitor the fluctuation of irradiation intensity; the automatic execution program is started to continuously record the irradiation process data. When the dose rate deviates from the set value and exceeds the preset threshold, the automatic execution program triggers the PID control algorithm for real-time adjustment, and starts the slow power reduction program when the cumulative dose reaches the preset percentage of the target value. When an equipment failure occurs or the dose rate fluctuates abnormally and exceeds the preset threshold, the emergency handling subroutine is called to execute the safe shutdown process.

[0033] Specifically, according to the optimized irradiation parameters, a standard irradiation plan file in JSON format is generated, which includes irradiation time, irradiation intensity and irradiation mode. The parameters are converted into an instruction set recognizable by the device through a lookup table and a linear interpolation algorithm. The device compatibility check program is called to verify the compatibility of the irradiation plan with the irradiation equipment, and to check whether the irradiation parameters are within the allowable range of the equipment. If exceeded, the nearest neighbor algorithm is used to automatically adjust to the closest feasible value. The safety limit check program is started to conduct a safety assessment of the irradiation plan, including the maximum dose rate not exceeding 1000Gy / h, the upper limit of the cumulative dose of 20kGy and radiation protection requirements, and the threshold comparison method is used to ensure that the plan meets safety specifications. The checked and adjusted irradiation plan is transmitted to the irradiation equipment control system through the data interface, and the real-time monitoring module is activated to collect dose rate data once a second to monitor the fluctuation of irradiation intensity. The automatic execution program is started to continuously record the irradiation process data. If the dose rate deviates from the set value by more than 5%, the PID control algorithm is triggered for real-time adjustment. When the cumulative dose reaches 95% of the target value, the slow power reduction program is started. If a device failure occurs or the dose rate fluctuates abnormally by more than 10%, the emergency handling subroutine is automatically called to execute the safety shutdown process. In actual application, the optimized irradiation parameters are irradiation time of 18 hours, irradiation intensity of 800Gy / h, and double-sided irradiation. Generate an irradiation plan file in JSON format, which contains these parameter information. Through the preset lookup table, the irradiation intensity of 800Gy / h is converted into the device control instruction "IR800", and the irradiation time of 18 hours is converted into "T1080". The device compatibility checker finds that the maximum irradiation intensity of the device is 750Gy / h, and the nearest neighbor algorithm is used to automatically adjust the intensity to 750Gy / h. The safety limit checker verifies that the adjusted parameters meet the safety specifications: the maximum dose rate is 750Gy / h<1000Gy / h, and the cumulative dose is 13.5kGy<20kGy. The adjusted plan is transmitted to the control system, and the real-time monitoring module is started to collect dose rate data once a second. At the fifth hour, the dose rate was monitored to drop to 710 Gy / h, and the PID control algorithm automatically increased the source strength by 5%. When the cumulative dose reached 12.825 kGy (95% of the target value), the system started the slow power reduction program, gradually reducing the dose rate to 375 Gy / h in the last hour. The irradiation equipment automatically performs double-sided irradiation according to the preset program, and automatically flips the irradiated items every 60 minutes to ensure uniformity.

[0034] S105. During the irradiation process, the irradiation dose data is collected in real time through sensors or detectors arranged inside the cargo to monitor the dose distribution during the irradiation process. If an abnormal deviation from the expected dose distribution is found, an early warning is triggered, the irradiation process is interrupted, and the irradiation parameters are readjusted.

[0035] Determine key locations inside the cargo based on cargo volume and shape information, and install thermoluminescent dosimeters at key locations; collect real-time dose data through the thermoluminescent dosimeter, and transmit the real-time dose data to the central control unit; calculate the dose rate and cumulative dose at each measuring point based on the received real-time dose data to obtain the dose distribution result; determine whether the dose distribution result triggers an early warning based on a preset dose deviation threshold, and if the dose distribution result exceeds the preset dose deviation threshold, a warning signal of the corresponding level is triggered; after receiving the early warning signal, start the adaptive adjustment program, automatically adjust the irradiation parameters based on the dose distribution result, and generate an optimized irradiation plan.

[0036] Specifically, according to the volume and shape of the cargo, a sensor layout scheme is designed, 20 thermoluminescent dosimeters are installed at key locations inside the cargo, and high-dose gradient areas are determined as installation points through Monte Carlo simulation. The wireless transmission module is used to transmit real-time dose data to the central control unit. Dose distribution monitoring is carried out, sensor data is received in real time, the dose rate and cumulative dose of each measuring point are calculated, and a continuous dose distribution is generated using a three-dimensional spline interpolation method. It is compared with the expected dose distribution to generate a real-time dose distribution map. A three-level warning mechanism is established, and the dose deviation threshold is set. The first-level warning threshold is 5%, the second-level is 10%, and the third-level is 15%. When the dose deviation of any measuring point exceeds the preset threshold, the corresponding level of warning signal is triggered, and the third-level warning automatically interrupts the irradiation process. Start the adaptive adjustment program, use the decision tree algorithm to analyze the cause of dose deviation, automatically adjust the irradiation parameters, including irradiation time, intensity and mode, use the PID control algorithm to fine-tune the parameters, generate the optimized irradiation plan, and re-enter the irradiation equipment for execution. Without interrupting irradiation, the sensor data is continuously monitored. If a sensor failure or abnormal reading is detected, it automatically switches to the backup sensor or removes the abnormal data point to ensure the reliability of the monitoring system. In a 2m×2m×3m irradiation container, the Monte Carlo method was used to simulate 100 million photon trajectories and determine 20 high-dose gradient points as the installation locations of thermoluminescent dosimeters. Each dosimeter collects data once per second and sends it to the central control unit through a 2.4GHz wireless transmission module. The three-dimensional spline interpolation algorithm was used to expand the 20 discrete point data into a continuous distribution of 10,000 grid points. The system calculated the dose rate and cumulative dose in real time and compared them with the expected distribution. The three-level warning mechanism was set up, including triggering the first-level warning when any point deviation reached 5%, and the control system automatically fine-tuned the irradiation intensity by ±2%; the deviation reached 10% to trigger the second-level warning and adjust the irradiation direction; the deviation reached 15% to trigger the third-level warning and immediately interrupt the irradiation. The C4.5 decision tree algorithm was used to analyze the cause of the deviation, and combined with the PID control algorithm, the irradiation parameters were automatically adjusted with a proportional coefficient of 0.5, an integral coefficient of 0.1, and a differential coefficient of 0.05. If a sensor abnormality is detected, such as a sudden change in the reading exceeding 50%, the system automatically switches to an adjacent backup sensor or removes the abnormal point during data processing.

[0037] S106. Compare the irradiation dose data collected in real time with the dose distribution map predicted by simulation, calculate the deviation value, and if the deviation exceeds the preset threshold, update the simulation model parameters to improve the simulation accuracy; at the same time, associate the measured data, cargo information and irradiation parameters and store them to form a traceable irradiation quality record.

[0038] Acquire irradiation dose data and the generated predicted dose distribution map; align data points using a cubic spline interpolation algorithm based on the irradiation dose data and the predicted dose distribution map to obtain deviation values ​​of corresponding points; calculate the overall deviation distribution for the deviation values ​​of corresponding points, and set a three-level deviation threshold, which includes a first threshold, a second threshold, and a third threshold; if the overall deviation distribution exceeds any of the three-level deviation thresholds, trigger the automatic optimization program for simulation model parameters, which uses a gradient descent method to adjust model parameters; integrate irradiation dose data, basic cargo information, and irradiation parameters into a unified data structure, and assign a unique identification code using a UUID algorithm; receive the unified data structure and transmit it to a distributed database system, which is used to store data, wherein each data block contains a hash value, a timestamp, and an irradiation record of the previous block.

[0039] Specifically, the irradiation dose data is obtained from the real-time monitoring system, and the predicted dose distribution map is retrieved from the simulation module. The data points are aligned by the cubic spline interpolation algorithm, the deviation value of each corresponding point is calculated, and the root mean square error method is used to obtain the overall deviation distribution. Three-level deviation thresholds are set, which are 5%, 10%, and 15% respectively. When the calculated deviation exceeds the preset threshold, the simulation model parameter automatic optimization program is triggered, and the model parameters are adjusted by the gradient descent method assisted by automatic differentiation technology. The simulation calculation is re-run to generate an updated dose distribution prediction map. The root mean square error evaluation accuracy is improved by comparing the optimization before and after. A data association storage mechanism is established to integrate the measured dose data, basic cargo information (density, shape, size) and irradiation parameters (time, intensity, method) into a unified data structure in JSON format, and a unique identification code is assigned using the UUID algorithm. The associated data is transmitted to the distributed database system, and the blockchain technology of the proof-of-work consensus algorithm is used to ensure that the data cannot be tampered with. Each data block contains the hash value, timestamp and irradiation record of the previous block, and a complete irradiation quality record chain is constructed to achieve full process traceability. At the same time, the data integrity is automatically verified through smart contracts. In actual application, the irradiation device collects dose data from 20 thermoluminescent dosimeters per second, and obtains the predicted dose distribution of a 1000×1000×1000 grid from the Monte Carlo simulation module. The measured data is expanded to the same resolution as the simulation grid using the cubic spline interpolation algorithm, and then the deviation of each grid point is calculated. When the root mean square error exceeds 5%, the first level of optimization is triggered; when it exceeds 10%, the second level is triggered; and when it exceeds 15%, the third level is triggered. During the optimization process, the automatic differentiation function of TensorFlow is used to calculate the gradient, and the parameters are adjusted by the Adam optimizer with a learning rate of 0.01, and the accuracy improvement is evaluated every 50 iterations. For example, in one optimization, the root mean square error dropped from 12% to 3%, improving the accuracy by 75%. The optimized model parameters and measured data are stored in JSON format, including {"timestamp":"2024-09-27T10:30:00Z","density":1.2,"shape":"cylinder","dimensions":[100,50],"irradiation_time":3600,"intensity":500,"method":"dual_side","dose_data":[...]}, and a unique identifier such as "f47ac10b-58cc-4372-a567-0e02b2c3d479" is generated through the UUIDv4 algorithm. These data are packaged into blocks, the hash value of the previous block is calculated using the SHA-256 algorithm, and then a new block is generated through the proof-of-work algorithm (the difficulty value is set to 0 in the first 4 digits).Smart contracts automatically check the integrity and continuity of dosage data in each block, ensuring the reliability and immutability of the records.

[0040] S107. Conduct sampling inspection on the goods after actual irradiation, measure the irradiation dose at different positions inside the goods, evaluate the actual irradiation effect, verify the accuracy of the simulation model and optimization parameters, and modify the simulation model and optimization algorithm based on the verification results.

[0041] According to the volume and shape of the cargo, the importance sampling method of Monte Carlo simulation is used for automatic sampling to obtain the spatial coordinates of the key measurement points; according to the spatial coordinates of the key measurement points, the preset position is located to obtain the actual dose data after irradiation; the actual dose data is compared with the prediction results of the simulation model to determine the accuracy of the simulation model; if the accuracy of the simulation model does not meet the preset threshold, the genetic algorithm is used to optimize the simulation model parameters; the optimized simulation model is applied to the new irradiation task, and by comparing the accuracy improvement before and after optimization, it is determined whether the optimized simulation model meets the update conditions; if the accuracy improvement exceeds the preset threshold, the optimized simulation model parameters and algorithm are updated to the time series database.

[0042] Specifically, according to the volume and shape of the cargo, an automated sampling scheme was designed using the importance sampling method of Monte Carlo simulation to determine 20 key measurement points. A six-axis robotic arm equipped with a thermoluminescent dosimeter was used to accurately locate the preset position, perform dose measurement, and collect dose data after actual irradiation. The actual dose data collected was compared with the prediction results of the simulation model, and the three-dimensional error distribution was constructed using the Kriging interpolation method to quantitatively evaluate the accuracy of the simulation model. Based on the error analysis results, the simulation model parameters were optimized using a genetic algorithm, with the population size set to 100, the crossover rate to 0.8, and the mutation rate to 0.1. New parameter combinations were generated through crossover and mutation operations, and iterative calculations were performed until the optimal solution changed by less than 1% for 10 consecutive generations or reached 1,000 generations. The optimized model was applied to the new irradiation task to verify the accuracy improvement. If the accuracy improvement exceeded 10%, the optimized model parameters and algorithm were updated to the time series database, and the records included information such as timestamps, original parameters, optimized parameters, and accuracy improvement. A weekly automatic evaluation trigger mechanism was set to analyze historical optimization records and predict parameter change trends to guide parameter settings and continuous optimization in subsequent irradiation processes. In a 2m×2m×3m irradiation container, the system determines 20 key measurement points through Monte Carlo simulation, including 8 corner points, 6 surface center points and 6 internal high-gradient points. The ABBI RB6700 six-axis robotic arm is equipped with a TLD-100 thermoluminescent dosimeter, which is precisely positioned according to the preset path and stays at each point for 10 seconds to complete the measurement. The actual dose data collected and the simulation prediction results are used to construct a three-dimensional error distribution of 100×100×100 grids through Kriging interpolation, with an average error of 7.5%. The genetic algorithm sets the population size to 100 and optimizes 10 key model parameters within the range of ±20%. After 738 generations of iterations, the optimal solution changes by less than 1% for 10 consecutive generations, and the operation is terminated. The optimized model is applied to new batches of irradiation tasks, and the accuracy is improved by 15.3%. The system records the optimization results to the InfluxDB time series database, including the "2024-09-27T15:30:00Z" timestamp, the original parameter set, the optimized parameter set, the 15.3% accuracy improvement, etc. The evaluation program is automatically triggered at 3 a.m. every Sunday to analyze 86 optimization records in the past four weeks, use the ARIMA model to predict parameter change trends, and generate parameter recommendations for the next week.

[0043] The above description is merely a preferred embodiment of one or more embodiments of the present specification and is not intended to limit one or more embodiments of the present specification. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present specification shall be included in the scope of protection of one or more embodiments of the present specification.

Claims

1. A method for predicting the irradiation of irradiated goods, characterized in that: The method comprises: According to the characteristic parameters of the irradiated goods, including the density, shape and size of the goods, a three-dimensional model of the irradiated goods is established to simulate the radiation dose distribution during the irradiation process, obtain the radiation dose values ​​at various positions inside the irradiated goods, and generate a radiation dose distribution map to identify the irradiation blind spots or over-irradiated areas and evaluate the irradiation uniformity; According to the irradiation uniformity evaluation results, optimize the irradiation parameters, including adjusting the irradiation time, irradiation intensity and irradiation mode, and obtain the optimal irradiation parameter combination through iterative optimization method; Input the optimized irradiation parameters into the computer simulation model, re-simulate the irradiation process, obtain the optimized irradiation dose distribution diagram, and evaluate the optimization effect. If the expected irradiation uniformity and efficiency targets are not achieved, continue to iterate and optimize to obtain the optimal irradiation parameter combination until the requirements are met; According to the optimized irradiation parameters, formulate an irradiation plan, including irradiation time, irradiation intensity and irradiation method, and input the irradiation plan into the irradiation equipment control system to automatically execute the irradiation process; During the irradiation process, sensors or detectors arranged inside the cargo are used to collect irradiation dose data in real time and monitor the dose distribution during the irradiation process. If an abnormal deviation from the expected dose distribution is found, an early warning is triggered, the irradiation process is interrupted, and the irradiation parameters are readjusted; Compare the real-time collected irradiation dose data with the simulated predicted dose distribution map, calculate the deviation value, and if the deviation exceeds the preset threshold, update the simulation model parameters to improve the simulation accuracy. At the same time, associate the measured data, cargo information and irradiation parameters to form a traceable irradiation quality record; Sampling inspection is carried out on the goods after actual irradiation, the radiation dose at different positions inside the goods is measured, the actual irradiation effect is evaluated, the accuracy of the simulation model and optimization parameters is verified, and the simulation model and optimization algorithm are corrected according to the verification results.

2. The method according to claim 1, characterized in that The method includes establishing a three-dimensional model of the irradiated cargo based on the characteristic parameters of the irradiated cargo, including the density, shape and size of the cargo, simulating the radiation dose distribution during the irradiation process, obtaining the radiation dose values ​​at various positions inside the irradiated cargo, generating a radiation dose distribution map, identifying irradiation blind spots or over-irradiated areas, and evaluating the irradiation uniformity, including: Generate the corresponding 3D mesh model according to the density, shape and size parameters of the cargo, perform adaptive mesh refinement on complex shape areas, evaluate mesh quality and optimize mesh structure, and obtain a high-precision 3D model database of irradiated cargo; Set the energy spectrum and initial ray parameters of the radiation source and use the Monte Carlo method to simulate the transmission process of photons in the three-dimensional model, including random number generation, photon path tracing and energy deposition calculation; Record the energy absorbed by the grid cells and accumulate statistics to obtain the radiation dose distribution data; The dose data of discrete grid points were spatially reconstructed using a three-dimensional interpolation algorithm, and the dose distribution data was mapped onto a three-dimensional visualization model. The HSV color space was used to map dose values ​​to colors. Based on the dose distribution map, the DBSCAN density clustering algorithm is used to identify continuous areas with higher or lower doses than the expected dose, obtain irradiation dead angles and over-irradiation areas, calculate the volume proportion of each area and the degree of dose deviation, and quantitatively evaluate the overall irradiation uniformity index by calculating the dose variation coefficient; It also includes: using the calculation method of the overall average absorbed dose of the dynamic gamma irradiation device with automatic surface replacement of a single plate source to analyze the total energy of ionizing radiation absorbed by the material in a given irradiated product, combined with the characteristics of the dynamic gamma irradiation device with automatic surface replacement of a single plate source and the dose distribution law in the irradiation container.

3. The method according to claim 2, characterized in that The method for calculating the overall average absorbed dose of the dynamic gamma irradiation device using the single-plate source automatic surface replacement is used to analyze the total energy of ionizing radiation absorbed by the material in a given irradiated product by combining the characteristics of the dynamic gamma irradiation device using the single-plate source automatic surface replacement and the dose distribution law in the irradiation container, including: A three-dimensional geometric model of the irradiation device is established according to the size of the irradiation container and the activity of the plate source. The Monte Carlo method is used to simulate the transmission process of gamma rays in the container to obtain the spatial dose rate distribution data in the container. The dose rate distribution data is processed by a three-dimensional spline interpolation algorithm to generate a continuous dose rate distribution function. The cumulative absorbed dose at any point in the irradiation container is calculated in combination with the plate source switching frequency and the total irradiation time. Based on the cumulative absorbed dose, the dose distribution is corrected in combination with the attenuation coefficient of the irradiated product to obtain the actual dose distribution. Based on the actual dose distribution, a dosimeter layout plan is set up, and dosimeters are virtually placed at key positions in the irradiated product. The maximum dose point, minimum dose point and average dose point are determined through dose gradient analysis. The virtual dosimeter measurement values ​​are converted into actual measurement values ​​in combination with the dosimeter response characteristics and calibration factors, and the overall average absorbed dose is calculated using the volume weighted average method.

4. The method according to claim 1, characterized in that: The irradiation parameters are optimized according to the irradiation uniformity evaluation result, including adjusting the irradiation time, irradiation intensity and irradiation mode, and the optimal irradiation parameter combination is obtained through an iterative optimization method, including: Establish an irradiation parameter optimization model, which includes irradiation time, irradiation intensity and irradiation mode as independent variables and dose uniformity as the objective function; According to the irradiation parameter optimization model, the genetic algorithm is used to generate the initial irradiation parameter combination population, in which the irradiation time and irradiation intensity are coded with real numbers, and the irradiation mode is coded with integers. The dose distribution under each combination of irradiation parameters is calculated by Monte Carlo simulation to obtain the dose uniformity; Perform arithmetic crossover and Gaussian mutation operations on the irradiation parameter combinations in the population to obtain a new generation of irradiation parameter combinations; Determine whether the optimization result of the new generation irradiation parameter combination meets the preset convergence condition. If the preset convergence condition is met, output the optimal irradiation parameter combination; If the preset convergence condition is not met, the iteration is continued until the optimal irradiation parameter combination is obtained or the preset maximum number of iterations is reached.

5. The method according to claim 1, characterized in that The optimized irradiation parameters are input into the computer simulation model, the irradiation process is re-simulated, the optimized irradiation dose distribution diagram is obtained, and the optimization effect is evaluated. If the expected irradiation uniformity and efficiency targets are not achieved, the optimal irradiation parameter combination is obtained by iterative optimization until the requirements are met, including: According to the optimized irradiation time, irradiation intensity and irradiation mode parameters, the parameters are input into a Monte Carlo simulation program, wherein the Monte Carlo simulation program is provided with the geometric structure of the irradiation container, the characteristics of the radiation source and the properties of the irradiated object; The Mersenne Twister algorithm is used to generate random numbers, which are used to simulate photon trajectories and obtain energy deposition distribution data; Generate a three-dimensional irradiation dose distribution map based on the energy deposition distribution data, and improve the resolution of the voxelized data by a trilinear interpolation algorithm; Obtain the irradiation uniformity index of the three-dimensional irradiation dose distribution map, which includes the ratio of maximum dose to minimum dose and the dose variation coefficient, and calculate the irradiation efficiency; If the irradiation uniformity index and irradiation efficiency do not reach the preset targets, the optimization algorithm parameters including the search range and the number of iterations are adjusted, and the simulated annealing algorithm is used for re-optimization until the requirements are met or the maximum number of iterations is reached.

6. The method according to claim 1, characterized in that The method comprises formulating an irradiation plan according to the optimized irradiation parameters, including irradiation time, irradiation intensity and irradiation mode, and inputting the irradiation plan into the irradiation equipment control system to automatically execute the irradiation process, including: Generate a standard irradiation program file including irradiation time, irradiation intensity and irradiation mode according to the optimized irradiation parameters. The standard irradiation program file converts the parameters into an instruction set recognizable by the equipment through a lookup table and a linear interpolation algorithm; Call the equipment compatibility checker to verify the compatibility of the standard irradiation plan file and the irradiation equipment. The compatibility checker determines whether the irradiation parameters are within the allowable range of the equipment. If they are exceeded, they are adjusted to the nearest feasible value. Initiate the safety limit check procedure to conduct a safety assessment on the standard irradiation plan documents. The safety assessment uses the threshold comparison method to ensure that the plan meets safety specifications, including the maximum dose rate, cumulative dose upper limit and radiation protection requirements; The checked and adjusted irradiation plan is transmitted to the irradiation equipment control system through the data interface, and the real-time monitoring module is activated to monitor the fluctuation of irradiation intensity; Start the automatic execution program and continuously record the irradiation process data. When the dose rate deviates from the set value and exceeds the preset threshold, the automatic execution program triggers the PID control algorithm for real-time adjustment. When the cumulative dose reaches the preset percentage of the target value, the slow power reduction program is started. When equipment failure occurs or the dose rate fluctuates abnormally and exceeds the preset threshold, the emergency processing subroutine is called to execute the safe shutdown process.

7. The method according to claim 1, characterized in that During the irradiation process, the irradiation dose data is collected in real time through sensors or detectors arranged inside the cargo to monitor the dose distribution during the irradiation process. If an abnormal deviation from the expected dose distribution is found, an early warning is triggered, the irradiation process is interrupted, and the irradiation parameters are readjusted, including: Determine key locations inside the cargo based on cargo volume and shape information, and install thermoluminescent dosimeters at key locations; Collect real-time dose data through thermoluminescent dosimeter and transmit the real-time dose data to the central control unit; According to the received real-time dose data, the dose rate and cumulative dose of each measuring point are calculated to obtain the dose distribution result; Determine whether the dose distribution result triggers an early warning according to a preset dose deviation threshold. If the dose distribution result exceeds the preset dose deviation threshold, a corresponding level of early warning signal is triggered; After receiving the early warning signal, the adaptive adjustment program is started to automatically adjust the irradiation parameters according to the dose distribution results to generate an optimized irradiation plan.

8. The method according to claim 1, characterized in that The real-time collected irradiation dose data is compared with the simulated predicted dose distribution map, and the deviation value is calculated. If the deviation exceeds the preset threshold, the simulation model parameters are updated to improve the simulation accuracy. At the same time, the measured data, cargo information and irradiation parameters are associated and stored to form a traceable irradiation quality record, including: Acquire irradiation dose data and generate predicted dose distribution map; According to the irradiation dose data and the predicted dose distribution diagram, the cubic spline interpolation algorithm is used to align the data points to obtain the deviation values ​​of the corresponding points; For the deviation values ​​of the corresponding points, the overall deviation distribution is calculated, and a three-level deviation threshold is set, where the three-level deviation threshold includes a first threshold, a second threshold, and a third threshold; If the overall deviation distribution exceeds any of the three-level deviation thresholds, the simulation model parameter automatic optimization program is triggered, and the simulation model parameter automatic optimization program uses the gradient descent method to adjust the model parameters; Integrate irradiation dose data, basic cargo information and irradiation parameters into a unified data structure, and assign a unique identification code through the UUID algorithm; The unified data structure is received and transmitted to a distributed database system for storing data, wherein each data block contains a hash value, a timestamp, and an irradiation record of the previous block.

9. The method according to claim 1, characterized in that: The actual irradiated goods are sampled and tested, the irradiation dose at different positions inside the goods is measured, the actual irradiation effect is evaluated, the accuracy of the simulation model and the optimization parameters is verified, and the simulation model and the optimization algorithm are corrected according to the verification results, including: According to the volume and shape of the cargo, the importance sampling method of Monte Carlo simulation is used for automated sampling to obtain the spatial coordinates of key measurement points; According to the spatial coordinates of key measurement points, locate to the preset position and obtain the actual dose data after irradiation; Compare the actual dose data with the prediction results of the simulation model to determine the accuracy of the simulation model; If the accuracy of the simulation model does not meet the preset threshold, a genetic algorithm is used to optimize the simulation model parameters; Apply the optimized simulation model to a new irradiation task, and determine whether the optimized simulation model meets the update conditions by comparing the accuracy improvement before and after optimization; If the accuracy improvement exceeds a preset threshold, the optimized simulation model parameters and algorithms are updated to the timing database.

Images