Reactor fuel assembly transfer control method and system based on neural network

Abstract

The invention discloses a reactor fuel assembly transfer control method and system based on a neural network. The method comprises the following steps: constructing a simulation platform based on a transportation environment for simulating a reactor fuel assembly; simulating the transportation process of the reactor fuel assembly through the simulation platform and recording transportation process parameters; analyzing a coupling relationship of various data in the transportation process parameters, and constructing a multi-modal transportation model according to the coupling relationship; inputting the experimental data into the multi-modal transportation model for training to obtain a transportation safety model; and inputting the transportation process parameters into the transportation safety model in real time in the transportation process of the reactor fuel assembly, and executing corresponding transfer control according to the safety index output by the transportation safety model.

Description

Technical Field

[0001] The present invention relates to the technical field of reactor transportation, and particularly to a method and system for controlling the transfer of reactor fuel assemblies based on a neural network. Background Art

[0002] The transfer and transportation of reactor fuel assemblies in nuclear power plants are basic operations carried out within nuclear power plants and are also important parameters affecting the load factor of nuclear power plants. In existing fuel assembly transfer systems, in order to achieve the transfer of fuel assemblies, special machinery is usually provided, and the transfer operation is carried out along a specified path according to a preset process. During the transfer process, operators of the refueling controller need to continuously monitor manually to avoid accidents and ensure the transportation safety of reactor fuel assemblies.

[0003] With the increase in the design power of new nuclear power plants and the number of reactor clusters, the number of fuel assemblies that need to be replaced during the refueling stage has increased significantly, and the dependence on the experience of operators is relatively serious. If the operators lack experience, it will cause the refueling time to be prolonged, reducing the operating efficiency and economic benefits of nuclear power plants. At the same time, the high-intensity operation will increase the possibility of mistakes by on-site operators, resulting in the inability to ensure the safety of reactor fuel assemblies during the high-frequency and high-efficiency transfer process. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a method and system for controlling the transfer of reactor fuel assemblies based on a neural network. The method constructs a multi-modal transportation model by building a simulation platform, then trains a transportation safety model with experimental data, and adjusts the transfer control strategy in real time through the transportation safety model during the actual transfer process of reactor fuel assemblies, ensuring the safety and efficiency of reactor fuel assemblies during the transfer process while reducing the burden on operators.

[0005] The technical solution of the present invention provides a method for controlling the transfer of reactor fuel assemblies based on a neural network, including: Constructing a simulation platform based on the transportation environment simulating reactor fuel assemblies; Simulating the transportation process of reactor fuel assemblies through the simulation platform and recording the transportation process parameters; Analyzing the coupling relationship of various data in the transportation process parameters and constructing a multi-modal transportation model accordingly; Inputting experimental data into the multi-modal transportation model for training to obtain a transportation safety model; During the transportation process of reactor fuel assemblies, inputting the transportation process parameters into the transportation safety model in real time, and performing corresponding transfer control according to the safety index output by the transportation safety model.

[0006] In one of the alternative technical solutions, the transportation process parameters include a vibration safety index, a temperature safety index, a pressure safety index, and a position safety index; The multi-modal transportation model and the transportation safety model include dynamic weights and correction factors.

[0007] In one of the alternative technical solutions, the vibration safety index is determined by the impact intensity and vibration intensity collected by an acceleration sensor and a vibration sensor; the temperature safety index is determined by the temperature value collected by a temperature sensor; the pressure safety index is determined by the pressure value collected by a pressure sensor; the position safety index is determined by the position information collected by an ultrasonic detection sensor.

[0008] In one of the alternative technical solutions, the construction of the simulation platform based on the multi-physical-field coupled transportation environment of the simulated reactor fuel assembly includes: Simulating the vibration response of the reactor fuel assembly based on finite element analysis to generate a vibration data set; Calculating the temperature and pressure inside the container of the transfer equipment by the lumped parameter method to generate a temperature data set and a pressure data set; Generating multi-configuration transportation scenarios corresponding to multiple stop positions through a path planning algorithm and forming a path data set, where the transportation scenarios include a start-up acceleration scenario, a deceleration stop scenario, and an emergency braking scenario; Integrating the vibration data set, the temperature data set, the pressure data set, and the path data set into a multi-physical-field coupled transportation environment to construct a simulation platform.

[0009] In one of the alternative technical solutions, the inputting of experimental data into the multi-modal transportation model for training to obtain a transportation safety model includes: Inputting the transportation process parameters corresponding to each stage in the experiment into the simulation platform for simulation; Extracting the time-series data of the vibration intensity, the temperature value, the pressure value, and the position information from the simulation platform, performing dimensionality reduction through principal component analysis, and extracting key features; Inputting the key features into the multi-modal transportation model and training through a neural network; Determining the dynamic weights and the correction factors according to the training results to obtain a transportation safety model.

[0010] In one of the alternative technical solutions, the transportation safety model is: Wherein, is the transportation safety index, is the vibration safety index, is the temperature safety index, is the pressure safety index, is the position safety index, and are the dynamic weights, is the position correction factor, is the path ruggedness correction factor, is the time correction factor, 、 、 and are the normalization coefficients, is the path ruggedness, is the maximum allowable path ruggedness, is the transportation time, is the maximum allowable transportation time.

[0011] In one alternative technical solution, the corresponding transfer control is performed in a hierarchical response according to the safety index output by the transportation safety model, specifically including: If the safety index is less than or equal to the first safety threshold, the current transportation speed is maintained; If the safety index is greater than the first safety threshold and less than or equal to the second safety threshold, the transportation speed is reduced according to the safety index; If the safety index is greater than the second safety threshold, emergency braking is initiated.

[0012] The technical solution of the present invention provides a transfer control system for reactor fuel assemblies based on a neural network, including a memory, a processor, and an electronic device program on the memory. The processor executes the electronic device program to implement the steps of any of the foregoing transfer control methods for reactor fuel assemblies based on a neural network.

[0013] The technical solution of the present invention provides a computer-readable storage medium having stored thereon an electronic device program / instructions, which when executed by a processor implement the steps of any of the foregoing transfer control methods for reactor fuel assemblies based on a neural network.

[0014] The technical solution of the present invention provides an electronic device program product, including electronic device program / instructions, which when executed by a processor implement the steps of any of the foregoing transfer control methods for reactor fuel assemblies based on a neural network.

[0015] Adopting the above technical solution, the following beneficial effects are achieved: The reactor fuel assembly transfer control method based on neural network provided by the present invention constructs a multi-modal transportation model by building a simulation platform, and then trains a transportation safety model with experimental data. During the actual transfer process of the reactor fuel assembly, the transfer control strategy is adjusted in real time through the transportation safety model, reducing the burden on operators while ensuring the safety and efficiency of the reactor fuel assembly during transfer, enabling the transfer process of the reactor fuel assembly to adapt to the continuously increasing design power of nuclear power plants. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Referring to the drawings, the disclosure of the present invention will become more understandable. It should be understood that these drawings are only for illustrative purposes and are not intended to limit the scope of protection of the present invention. In the drawings: Figure 1 is a flowchart of the working process of the reactor fuel assembly transfer control method based on neural network provided by an embodiment of the present invention; Figure 2 is a schematic structural diagram of the reactor fuel assembly transfer control system provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The following further describes the specific embodiments of the present invention with reference to the drawings.

[0018] It should be noted that the reactor fuel assembly transfer control method based on neural network provided by the embodiments of the present invention is applied to the reactor fuel assembly transfer system. Various data of the reactor fuel assembly are obtained through transfer equipment. For example, when the transfer equipment is a transfer trolley, multiple sensors can be provided on the transfer trolley, such as acceleration sensors, vibration sensors, temperature sensors, pressure sensors, and ultrasonic detection sensors, etc. During the process of transferring the reactor fuel assembly from the fuel storage container in the fuel depot to the preheating box, the coding, position change, and various data of the reactor fuel assembly are recorded. The transport container on the transfer trolley is docked with the preheating box, and the positioning component of the preheating box assembly is used to insert it into the specified position of the transport container. After determining that the azimuth angle is correct, the component coding and position are recorded. The transport container is transported to the lower part of the transfer chamber by the transport trolley, and the transfer chamber transfer machine inserts the component into the intermediate position in the reactor. After determining that the azimuth angle is correct, the component coding is recorded; finally, the refueling machine transfers the component from the intermediate position to the reaction in the reactor. Then the reactor fuel assembly is transported to the reactor for reaction. During the transfer process, the reactor fuel assembly transfer control method based on neural network provided by the present invention can ensure the safety of the reactor fuel assembly on the premise of transferring the reactor fuel assembly as soon as possible. As Figure 1 shown is a reactor fuel assembly transfer control method based on neural network provided by an embodiment of the present invention, including: Step S101: Construct a simulation platform based on the transportation environment simulating the reactor fuel assembly.

[0019] Step S102: Simulate the transportation process of the reactor fuel assembly through a simulation platform and record the transportation process parameters.

[0020] Step S103: Analyze the coupling relationship of each item of data in the transportation process parameters, and construct a multimodal transportation model accordingly.

[0021] Step S104: Input the experimental data into the multimodal transportation model for training to obtain a transportation safety model.

[0022] Step S105: During the transportation process of the reactor fuel assembly, input the transportation process parameters into the transportation safety model in real time, and perform corresponding transfer control according to the safety index output by the transportation safety model.

[0023] In this embodiment, through the combination of multi-physical field coupling simulation and neural network model, a closed-loop system from virtual training to actual control is constructed. Traditional methods usually rely on single-physical field simulation or fixed-threshold control, and it is difficult to comprehensively reflect the complex interaction effects during the transportation process.

[0024] Preferably, the simulation platform can perform joint simulation through finite element analysis and lumped parameter method to accurately reproduce the multi-field coupling effects such as mechanical vibration, mechanical shock, thermal stress, air pressure fluctuation, and sudden acceleration change of the reactor fuel assembly during transportation, and accurately simulate the physical state of the reactor fuel assembly during the transportation process.

[0025] In addition, in the embodiment of the present invention, by injecting experimental data, the model can learn the parameter correlation law in extreme scenarios, such as the non-linear coupling relationship between vibration and pressure during emergency braking, which improves the generalization ability of the model. Based on the hierarchical response mechanism of the safety index, the traditional fixed-threshold control is upgraded to dynamic fuzzy control to avoid efficiency loss caused by excessive braking. During the entire transportation process of the reactor fuel assembly, corresponding transfer control is performed according to the safety index output by the transportation safety model. The transfer control can be to change the specific control parameters of the transfer control or to select one of the transfer strategies for switching.

[0026] In summary, the transfer control method for reactor fuel assembly based on neural network provided by the present invention constructs a multimodal transportation model by building a simulation platform, then trains a transportation safety model through experimental data, and adjusts the transfer control strategy in real time through the transportation safety model during the actual transfer process of the reactor fuel assembly, reducing the burden on operators while ensuring the safety and efficiency of the reactor fuel assembly during the transfer process, and enabling the transfer process of the reactor fuel assembly to adapt to the continuously increasing design power of nuclear power plants.

[0027] In one embodiment, the transportation process parameters include a vibration safety index, a temperature safety index, a pressure safety index, and a position safety index.

[0028] The multimodal transportation model and the transportation safety model include dynamic weights and correction factors.

[0029] It should be noted that the transportation of reactor fuel assemblies includes two transportation scenarios, namely the transportation of new fuel and the transportation of spent fuel. During the transportation of new fuel, a thermal insulation container needs to be carried, and the temperature of the new fuel assembly needs to be detected to prevent excessive heat loss. For the transportation of spent fuel, infrared temperature measurement is directly set in the transportation equipment and the spent fuel transportation channel to prevent the spent fuel from overheating and avoid radioactive leakage caused by the overheating of the spent fuel. All pressure-related parameters mentioned in the present invention refer to the pressure of the thermal insulation container. For example, the internal pressure is the pressure measured in the internal space of the thermal insulation container, and the external pressure is the pressure measured on the surface of the thermal insulation container.

[0030] In this embodiment, the vibration safety index (VSI) is determined by the impact intensity and vibration intensity collected by an acceleration sensor and a vibration sensor, the temperature safety index (TSI) is determined by the temperature value collected by a temperature sensor, the pressure safety index (PSI) is determined by the pressure value inside the thermal insulation container collected by a pressure sensor, and the position safety index (LSI) is determined by the position information collected by an ultrasonic detection sensor. The design of the four types of safety indexes is complementary: the vibration index (VSI) reflects the mechanical impact risk, the temperature index (TSI) monitors the thermal stress, the pressure index (PSI) evaluates the sealing performance, and the position index (LSI) ensures the positioning accuracy. The four cover the core dimensions of transportation safety. The dynamic weight design enables the weight coefficients and to be adjusted dynamically according to the environment. For example, the weight of VSI is increased on bumpy roads, and the weight of TSI is increased in a low-temperature environment. The introduction of correction factors λ, γ, and β performs non-linear compensation for variables such as the irradiation environment, path complexity, and transportation duration, avoiding misjudgment dominated by a single parameter. Through the synergistic effect of the dynamic weight and the correction factor in this embodiment, the comprehensive false alarm rate of safety assessment is significantly reduced.

[0031] In one embodiment, the vibration safety index is determined by the impact intensity and vibration intensity collected by an acceleration sensor and a vibration sensor, the temperature safety index is determined by the temperature value collected by a temperature sensor, the pressure safety index is determined by the pressure value inside the thermal insulation container collected by a pressure sensor, and the position safety index is determined by the position information collected by an ultrasonic detection sensor.

[0032] In this embodiment, the sensor selection and parameter mapping relationship are designed more precisely. The acceleration sensor uses a MEMS triaxial accelerometer, which has high sensitivity and good frequency response characteristics. Installed at the four corners of the transfer equipment, it can capture the acceleration changes during transportation in all directions. The acceleration data is converted into vibration intensity through a frequency-domain integration algorithm, which can effectively filter out noise interference and accurately reflect the true intensity of vibration. The temperature sensors are symmetrically arranged on the inner and outer walls of the transfer equipment and use high-precision thermocouple temperature sensors. Environmental interference is eliminated through differential measurement to ensure that the temperature difference between the inside and outside is within the preset range. The differential measurement method can effectively improve the accuracy and stability of temperature measurement and maintain reliable data output even in a complex electromagnetic environment. The ultrasonic sensor uses a pulse array, combines time-of-flight and phase-difference dual positioning to achieve high positioning accuracy, and has good radiation resistance. It can work stably in a strong radiation environment during the transfer of reactor fuel assemblies. The design of its pulse array makes the positioning signal stronger and significantly improves the anti-interference ability. At the same time, the time-of-flight and phase-difference dual positioning method can effectively improve the accuracy and reliability of positioning, providing a strong guarantee for the precise positioning of reactor fuel assemblies.

[0033] In one of the embodiments, a simulation platform is constructed based on the multi-physical-field coupled transportation environment of a simulated reactor fuel assembly, including: Based on finite element analysis, simulate the vibration response of the reactor fuel assembly to generate a vibration data set.

[0034] Use the lumped parameter method to calculate the temperature and pressure inside the container of the transfer equipment to generate a temperature data set and a pressure data set.

[0035] Generate multi-configuration transportation scenarios corresponding to multiple stop positions through a path planning algorithm and form a path data set. The transportation scenarios include start-up acceleration scenarios, deceleration stop scenarios, and emergency braking scenarios.

[0036] Integrate the vibration data set, temperature data set, pressure data set, and path data set into a multi-physical-field coupled transportation environment to construct a simulation platform.

[0037] In this embodiment, a simulation platform is constructed based on the multi-physical-field coupling transportation environment of the simulated reactor fuel assembly, including: simulating the vibration response of the reactor fuel assembly based on finite element analysis. An advanced finite element software is used to finely model the structure of the reactor fuel assembly, considering factors such as the nonlinear characteristics of materials, geometric nonlinear effects, and changes in boundary conditions, to generate a vibration data set. This data set can truly reflect the vibration characteristics of the reactor fuel assembly during transportation, providing accurate data support for subsequent analysis and model construction; using the lumped parameter method to calculate the temperature and pressure inside the container of the transfer equipment, generating a temperature data set and a pressure data set. These data sets can detail the distribution laws of the temperature field and pressure field, providing an important basis for analyzing the thermal stress and sealing performance of the reactor fuel assembly during transportation. Through the combined simulation of finite element analysis and the lumped parameter method, the multi-field coupling effects such as mechanical vibration, mechanical shock, thermal stress, air pressure fluctuation, and sudden acceleration change during the transportation of the reactor fuel assembly are accurately reproduced, and the physical state of the reactor fuel assembly during transportation is precisely simulated, providing a highly realistic virtual experimental environment for the research and development of the transfer control method of the reactor fuel assembly based on neural network.

[0038] When constructing the component model, the specific coupling situation of the multi-field coupling effect can also be used as the basis for model construction, making some data with higher coupling degrees be correspondingly associated, or even sharing the same dynamic weight index.

[0039] In one of the embodiments, the experimental data is input into the multi-modal transportation model for training to obtain a transportation safety model, including: Input the transportation process parameters corresponding to each stage in the experiment into the simulation platform for simulation.

[0040] Extract the time-series data of vibration intensity, temperature value, pressure value, and position information from the simulation platform, perform dimensionality reduction through principal component analysis, and extract key features.

[0041] Input the key features into the multi-modal transportation model and train through neural network.

[0042] Determine the dynamic weight and correction factor according to the training results to obtain the transportation safety model.

[0043] In this embodiment, the experimental data is input into the simulation platform for simulation, thereby outputting corresponding data, and then key features are extracted from the output data to analyze the critical value of danger. The multi-modal transportation model is trained through neural network to obtain an accurate transportation safety model applicable to this transportation environment.

[0044] In one of the embodiments, the transportation safety model is: Among them, is the transportation safety index, is the vibration safety index, is the temperature safety index, is the pressure safety index, is the position safety index, and are the dynamic weights, is the position correction factor, is the path ruggedness correction factor, is the time correction factor, , , and are the normalization coefficients, is the path ruggedness, is the maximum allowable path ruggedness, is the transportation time, is the maximum allowable transportation time.

[0045] Preferably, the transportation safety index , the temperature safety index , the pressure safety index and the position safety index are calculated as follows: The acceleration data collected is converted into vibration intensity by the frequency domain integration algorithm, and a maximum allowable vibration intensity is set.

[0046] During the calculation, for example, the vibration intensity obtained by the frequency domain integration algorithm, while the maximum allowable vibration intensity , at this time, the transportation safety index is 0.5.

[0047] The temperature safety index is related to the temperature difference inside and outside the transfer equipment. The internal temperature of the transfer equipment is , the external temperature of the transfer equipment is , and the maximum allowable temperature difference is . At this time, For example, when the internal temperature of the transfer equipment is 30°C, the external temperature is 28°C, and the maximum allowable temperature difference is 5°C, at this time is 0.6.

[0048] Pressure safety index Depends on the currently collected pressure value , the maximum allowable pressure value And the minimum allowable pressure value . If the current pressure value Is greater than the maximum allowable pressure value , the pressure safety index Takes 1; if the current pressure value Is less than the minimum allowable pressure value , the pressure safety index Takes 0.

[0049] If the current pressure value Is between the minimum allowable pressure value And the maximum allowable pressure value , For example, when the current pressure value Is 1.02 , the maximum allowable pressure value Is 1.05 , the minimum allowable pressure value Is 1.0 , at this time, Is 0.4.

[0050] Position safety index Is related to the deviation of the reactor fuel assembly from the target position And the maximum allowable deviation , specifically: For example, if the measured deviation of the reactor fuel assembly from the target position Is 0.1m and the maximum allowable deviation Is 0.2m, then Is 0.5.

[0051] The first dynamic weight And the second dynamic weight Respectively represent the dynamic weight ratios of the vibration safety index , temperature safety index And pressure safety index In the transportation safety index. When constructing the transportation safety model, the coupling of the temperature safety index And pressure safety index Is also considered, and the multiplied temperature safety index Is uniformly passed through the second dynamic weight and the pressure safety index A unified weight distribution is carried out. Moreover, the normalization coefficient serves as the divisor for each safety index, and the dynamic weight serves as the exponent, which can make the result of the finally output transportation safety index more accurate and flexible in change, can calculate in real time the accurate safety situation of the reactor fuel assembly during transportation, and is convenient for the system to make an automated response in a timely manner.

[0052] Vibration safety index The determination of is not only based on the impact intensity and vibration intensity collected by the acceleration sensor and vibration sensor, but also comprehensively considers factors such as the frequency range and duration of vibration. Through advanced signal processing algorithms, the collected data is analyzed and processed to accurately evaluate the mechanical shock risk; Temperature safety index It is determined by the temperature value collected by the temperature sensor. At the same time, by combining parameters such as the temperature change rate and the temperature fluctuation amplitude, the thermal stress situation is comprehensively monitored, and the potential impact of abnormal temperature changes on the reactor fuel assembly can be detected in a timely manner; Pressure safety index It is determined by the pressure value inside the heat preservation container collected by the pressure sensor, and also considers the instantaneous change of pressure and the frequency characteristics of pressure fluctuation to accurately evaluate the sealing performance and ensure that the sealing performance of the reactor fuel assembly is always in a safe state during transportation; Position safety index It is determined by the position information collected by the ultrasonic detection sensor. At the same time, by combining information such as the speed and acceleration of position change, the positioning accuracy of the reactor fuel assembly is monitored in real time to ensure that its position is accurate during transportation. The design of the four types of safety indexes is complementary. Through multi-dimensional data fusion and analysis, it comprehensively covers the core dimensions of transportation safety and provides all-round guarantee for the transfer safety of the reactor fuel assembly. The dynamic weight design enables the weight coefficients and to be adjusted dynamically according to the environment. For example, on a bumpy road section, by real-time monitoring the road condition information and the operation status of the transfer equipment, the weight is automatically increased to make the system pay more attention to the mechanical shock risk; in a low-temperature environment, according to the real-time data of the temperature sensor, the weight is automatically increased to strengthen the monitoring of thermal stress. The introduction of the correction factors λ, γ, and β performs non-linear compensation for variables such as the irradiation environment, path complexity, and transportation duration. Advanced mathematical models and algorithms are used to accurately calculate and compensate these variables to avoid misjudgment dominated by a single parameter and ensure the accuracy and reliability of safety assessment. In this embodiment, through the synergistic effect of the dynamic weight and the correction factor, the comprehensive false alarm rate of safety assessment is significantly reduced, and the reliability and effectiveness of the reactor fuel assembly transfer control method based on the neural network are improved.

[0053] Preferably, the dynamic weight It is calculated by computing the information entropy of each parameter, specifically as follows: Among them, is the information entropy, is , , and one of them, representing vibration information entropy, temperature information entropy, pressure information entropy, and position information entropy respectively. is the dynamic weight, is a constant such as 1, 2, 3, and 4. According to requirements, a dynamic weight can be independently set for each safety index. In the foregoing embodiments, only the first dynamic weight and the second dynamic weight are set, and some safety indices share the dynamic weight. The transportation safety model can also be dynamically changed according to the actual application scenario. Only one example is provided here.

[0054] Furthermore, the calculation method of the path ruggedness is as follows: First, the position acquisition and dynamic change acquisition of the transfer device are realized through one or a combination of ultrasonic positioning, GPS, and inertial positioning. According to the time change, a three-dimensional coordinate sequence S(x, y, z) of multiple coordinate points is obtained, and then the elevation change rate between adjacent coordinate points is calculated as the path ruggedness .

[0055] Among them, , and are the change amounts in the direction, direction, and direction between adjacent coordinate points. The calculated is the path ruggedness of the current section.

[0056] Preferably, the position correction factor is positively correlated with the distance between the positioning unit of the transfer device and the center point of the transfer device. The path ruggedness correction factor is negatively correlated with the suspension stiffness of the transfer device. The time correction factor is positively correlated with the length of the transfer device.

[0057] In one of the embodiments, corresponding transfer control is performed according to the safety index output by the transportation safety model, adopting hierarchical response, specifically including: If the safety index is less than or equal to the first safety threshold, the current transportation speed is maintained.

[0058] If the safety index is greater than the first safety threshold and less than or equal to the second safety threshold, the transportation speed is reduced according to the safety index.

[0059] If the safety index is greater than the second safety threshold, emergency braking is activated.

[0060] Furthermore, in connection with the foregoing embodiments, the first safety threshold is defined as 0.6 and the second safety threshold is defined as 0.8.

[0061] When reducing the transportation speed according to the safety index, the calculation formula for the speed change is: wherein, is the speed after speed reduction, is the original speed. As needed, the constants in the calculation formula can be modified to make the speed change calculation method more suitable for the transportation path of the reactor fuel assembly in the current nuclear power plant.

[0062] The selection of the first safety threshold and the second safety threshold can generate simulation data through the Monte Carlo simulation method, and determine the threshold based on the statistical distribution division. For example, the safety area of the simulation result corresponds to a 90% confidence interval, and a dynamic threshold can also be introduced to dynamically modify the threshold according to the actual environment.

[0063] In this embodiment, through the setting of the safety threshold, the safety index output by the transportation safety model can accurately control the specific safety measures in various situations during the transfer process of the reactor fuel assembly, with a high degree of automation. According to the safety index output by the transportation safety model, measures such as preset speed reduction, parking, or activation of emergency protection are automatically triggered, reducing the dependence on operators during the entire transfer process of the reactor fuel assembly. As needed, more hierarchical response strategies can also be set to meet more detailed transfer control for different transportation safety indices.

[0064] In summary of all the above embodiments, a process of a specific application scenario is provided. Assume that in a newly built nuclear power plant, during the process of transferring the reactor fuel assembly from the fuel storage area to the reactor, the transfer control method of the present invention needs to be applied. The transfer path in this nuclear power plant includes a flat straight section with a length of 1000 meters, connecting the fuel storage area and the reactor area, and there is a turning section with a radius of 50 meters before approaching the reactor area. The maximum allowable transportation time for the entire transfer process is 30 minutes.

[0065] At this time, first build a simulation platform, and model the structure of the reactor fuel assembly through finite element analysis software. Set the material parameters, such as the elastic modulus is 2.1 10 11The Poisson's ratio is 0.3. Simulate its vibration response during transportation to generate a vibration dataset.

[0066] Then, use computational fluid dynamics software to establish a fluid model inside the thermal insulation container. Set the ambient temperature to 25°C and the pressure to 1 standard atmosphere. Considering the turbulent characteristics of the fluid, simulate the distribution of the temperature field and pressure field to generate a temperature dataset and a pressure dataset.

[0067] Generate multiple transportation scenarios through a path planning algorithm combined with the actual layout inside the nuclear power plant. For example, in a flat straight road scenario, set the transportation speed to 2 m / s; in a turning path scenario, the speed drops to 1 m / s to form a path dataset. Integrate the vibration, temperature, pressure, and path datasets to build a simulation platform.

[0068] Then, collect transportation process parameters. Install sensors on the transfer equipment. The acceleration sensor uses a triaxial accelerometer. At a certain moment, the acceleration values of the x, y, and z axes collected are 0.5 m / s, 0.3 m / s, and 0.2 m / s respectively. Calculate the vibration intensity through the frequency domain integration algorithm, and then determine the vibration safety index. The temperature sensor uses a K-type thermocouple temperature sensor, which is symmetrically arranged on the inner and outer walls of the transfer equipment. The measured internal temperature is 30°C and the external temperature is 28°C. Calculate the temperature difference and determine the temperature safety index. The pressure value collected by the pressure sensor is 1.02 10 5 pa, determine the pressure safety index. The ultrasonic detection sensor uses an HC-SR04 pulse array, combines time-of-flight and phase difference dual positioning, measures the deviation of the fuel assembly from the target position, and determines the position safety index. .

[0069] Then, build a multimodal transportation model and train a transportation safety model. Analyze the coupling relationship of various data in the collected transportation process parameters to build a multimodal transportation model, which includes dynamic weights and correction factors. Collect experimental data from previous similar transportation processes in this nuclear power plant. For example, in an emergency braking accident, record transportation process parameters such as the vibration intensity, temperature value, pressure value, and position information at that time. Input these experimental data into the simulation platform for simulation. Extract relevant time series data from the simulation platform, perform dimensionality reduction through principal component analysis, and extract key features. Input the key features into the multimodal transportation model and use neural networks for deep learning training. Assume that after multiple trainings, the dynamic weight is 0.4, is 0.6; the position correction factor is 0.5; the path roughness correction factor is 0.3; the time correction factor is 0.2; the normalization coefficient 、 、 and are all 1, to obtain the transportation safety model.

[0070] Finally, when the transportation safety index of the reactor fuel assembly is calculated in real time through the transportation safety model , the transfer control automatic control strategy is executed. During the actual transportation process, the transportation process parameters are collected in real time and input into the transportation safety model. Suppose that at 10 minutes of transportation, the transportation safety index is calculated. At this time, the first safety threshold is set to 0.6, and the second safety threshold is set to 0.8. If the calculated transportation safety index is 0.5, which is less than the first safety threshold, the current transportation speed is maintained. When transporting to a turning section, due to factors such as changes in path ruggedness, the calculated transportation safety index is 0.7, which is greater than the first safety threshold and less than the second safety threshold. According to the speed change calculation formula , the original speed is 2 m / s, then the speed after deceleration , and the turning transportation is carried out at this speed. If during the transportation process, due to emergencies, the transportation safety index is 0.9, which is greater than the second safety threshold, emergency braking is immediately started, and at the same time, inert gas (such as nitrogen) is released to cover the reactor fuel assembly to ensure its safety.

[0071] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0072] Such as Figure 2 shown is a schematic diagram of the hardware structure of a transfer control system for a reactor fuel assembly based on a neural network provided by an embodiment of the present invention, including a memory 202, a processor 201, and an electronic device program on the memory 202. The processor 201 executes the electronic device program to implement the steps of the transfer control method for a reactor fuel assembly based on a neural network in any of the above embodiments.

[0073] Figure 2 Take one processor 201 as an example in

[0074] The electronic device may further include: an input device 203 and a display device 204.

[0075] The processor 201, the memory 202, the input device 203, and the display device 204 may be connected via a bus or other means. In the figure, the connection via the bus is taken as an example.

[0076] As a non-volatile electronic device-readable storage medium, the memory 202 can be used to store non-volatile software programs, non-volatile electronic device-executable programs, and modules, such as the program instructions / modules corresponding to the neural network-based reactor fuel assembly transfer control method in the embodiments of the present application. By running the non-volatile software programs, instructions, and modules stored in the memory 202, the processor 201 executes various functional applications and data processing, that is, implements the neural network-based reactor fuel assembly transfer control method in the above embodiments.

[0077] The memory 202 may include a program storage area and a data storage area. Among them, the program storage area can store an operating system and application programs required for at least one function; the data storage area can store data created according to the use of the neural network-based reactor fuel assembly transfer control method, etc. In addition, the memory 202 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the memory 202 may optionally include a memory remotely provided relative to the processor 201, and these remote memories can be connected to the device for executing the neural network-based reactor fuel assembly transfer control method through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0078] The input device 203 can receive input user clicks and generate signal inputs related to user settings and function controls of the neural network-based reactor fuel assembly transfer control method. The display device 204 may include a display device such as a display screen.

[0079] When the one or more modules are stored in the memory 202 and run by the one or more processors 201, they execute the neural network-based reactor fuel assembly transfer control method in any of the above method embodiments.

[0080] When the electronic device disclosed in the present invention is running, it can execute all steps of the above-mentioned neural network-based reactor fuel assembly transfer control method. By building a simulation platform to build a multi-modal transportation model, and then training a transportation safety model with experimental data, the transportation control strategy is adjusted in real time through the transportation safety model during the actual transfer process of the reactor fuel assembly, reducing the burden on operators while ensuring the safety and efficiency of the reactor fuel assembly during the transfer process, so that the transfer process of the reactor fuel assembly can adapt to the continuously increasing design power of nuclear power plants.

[0081] An embodiment of the present invention provides a computer-readable storage medium storing a computer program / instructions, and when the computer program / instructions are executed by a processor 201, all steps of the above-mentioned neural network-based reactor fuel assembly transfer control method are implemented.

[0082] In the context of the present disclosure, the storage medium may be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. The storage medium may be a machine-readable signal medium or a machine-readable storage medium. Optionally, the storage medium may be a non-transitory computer-readable storage medium. For example, the non-transitory computer-readable storage medium may be a ROM, a random access memory (RAM), a compact disc read-only memory (CDROM), magnetic tape, a floppy disk, and an optical data storage device, etc.

[0083] An embodiment of the present invention provides a computer program product including a computer program / instructions, and when the computer program / instructions are executed by a processor, the steps of the above-mentioned neural network-based reactor fuel assembly transfer control method are implemented.

[0084] By running the above computer program product, all steps of the above-mentioned neural network-based reactor fuel assembly transfer control method can be executed. A multi-modal transportation model is constructed by building a simulation platform, and then a transportation safety model is trained with experimental data. During the actual transfer process of the reactor fuel assembly, the transfer control strategy is adjusted in real time through the transportation safety model, reducing the burden on operators while ensuring the safety and efficiency of the reactor fuel assembly during the transfer process, so that the transfer process of the reactor fuel assembly can adapt to the continuously increasing design power of the nuclear power plant.

[0085] The above-described embodiments merely represent several implementation manners of the present invention, and their descriptions are relatively specific and detailed, but should not be construed as limiting the scope of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several variations and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the appended claims.

Claims

1. A method for controlling the transport of reactor fuel assemblies based on a neural network, characterized in that: include: Build a simulation platform based on the transport environment of simulated reactor fuel assemblies; Simulate the reactor fuel assembly transportation process through the simulation platform and record the transportation process parameters; Analyze the coupling relationship between various data in the transportation process parameters, and construct a multimodal transportation model based on the coupling relationship; Inputting the experimental data into the multimodal transportation model for training to obtain a transportation safety model; During the transportation of the reactor fuel assembly, the transportation process parameters are input into the transportation safety model in real time, and corresponding transfer control is performed according to the safety index output by the transportation safety model.

2. The method for controlling the transport of reactor fuel assemblies based on a neural network according to claim 1, characterized in that: The transportation process parameters include a vibration safety index, a temperature safety index, a pressure safety index and a position safety index; The multimodal transportation model and the transportation safety model include dynamic weights and correction factors.

3. The method for controlling the transport of reactor fuel assemblies based on a neural network according to claim 2, characterized in that: The vibration safety index is determined by the impact intensity and vibration intensity collected by the acceleration sensor and the vibration sensor; the temperature safety index is determined by the temperature value collected by the temperature sensor; the pressure safety index is determined by the pressure value collected by the pressure sensor; and the position safety index is determined by the position information collected by the ultrasonic detection sensor.

4. The method for controlling the transport of reactor fuel assemblies based on a neural network according to claim 3, characterized in that: The simulation platform is constructed based on a multi-physics field coupled transportation environment simulating a reactor fuel assembly, including: Simulate the vibration response of reactor fuel assemblies based on finite element analysis to generate vibration data sets; The temperature and pressure inside the container of the transfer equipment are calculated using the lumped parameter method to generate a temperature data set and a pressure data set; Generate a multi-configuration transportation scenario corresponding to a plurality of stop positions through a path planning algorithm and form a path data set, wherein the transportation scenario includes a start-up acceleration scenario, a deceleration stop scenario, and an emergency braking scenario; The vibration data set, the temperature data set, the pressure data set and the path data set are integrated into a multi-physics field coupled transportation environment to construct a simulation platform.

5. The method for controlling the transport of reactor fuel assemblies based on a neural network according to claim 4, characterized in that: The step of inputting the experimental data into the multimodal transport model for training to obtain a transport safety model comprises: Input the transportation process parameters corresponding to each stage in the experiment into the simulation platform for simulation; Extracting the time series data of the vibration intensity, the temperature value, the pressure value and the position information from the simulation platform, performing dimension reduction through principal component analysis, and extracting key features; Inputting the key features into the multimodal transport model through neural network training; The dynamic weight and the correction factor are determined according to the training results to obtain a transportation safety model.

6. The method for controlling the transport of reactor fuel assemblies based on a neural network according to claim 5, characterized in that: The transport safety model is: in, is the transport safety index, is the vibration safety index, is the temperature safety index, is the pressure safety index, is the location security index, and is the dynamic weight, is the position correction factor, is the path roughness correction factor, is the time correction factor, , , and is the normalization coefficient, is the path roughness, is the maximum allowable path roughness, For transportation time, The maximum permissible shipping time.

7. The method for controlling the transport of reactor fuel assemblies based on a neural network according to claim 1, characterized in that: The step of executing corresponding transshipment control according to the safety index output by the transportation safety model adopts a graded response, specifically including: If the safety index is less than or equal to the first safety threshold, maintaining the current transportation speed; If the safety index is greater than a first safety threshold and less than or equal to a second safety threshold, reducing the transport speed according to the safety index; If the safety index is greater than the second safety threshold, emergency braking is initiated.

8. A reactor fuel assembly transport control system based on a neural network, comprising a memory, a processor and an electronic device program on the memory, characterized in that: The processor executes the electronic device program to implement the steps of the neural network-based reactor fuel assembly transport control method described in any one of claims 1-7.

9. An electronic device readable storage medium having an electronic device program / instruction stored thereon, characterized in that: When the electronic device program / instruction is executed by the processor, the steps of the neural network-based reactor fuel assembly transport control method described in any one of claims 1-7 are implemented.

10. An electronic device program product, comprising an electronic device program / instruction, characterized in that: When the electronic device program / instruction is executed by the processor, the steps of the neural network-based reactor fuel assembly transport control method described in any one of claims 1-7 are implemented.

Images