Reactor component transfer positioning method based on neural network

Abstract

The invention discloses a reactor component transfer positioning method based on a neural network. The method comprises the following steps: constructing an original positioning model of a neural network structure based on multi-source sensor data; and establishing a three-dimensional rectangular coordinate system, and planning an optimal transfer path by combining the assembly safety requirement, the initial position and the target position of the reactor fuel assembly. And training the original positioning model through historical transfer data, and optimizing parameters of the original positioning model based on the optimal transfer path to obtain a transfer positioning model. Multi-source sensor data obtained by transfer equipment in real time are input into the transfer positioning model, so that the transfer positioning model outputs and obtains positioning data of the reactor fuel assembly in real time. According to the invention, operators can be assisted in high-precision real-time positioning of the reactor fuel assembly in a blind operation area, and the positioning accuracy of the reactor fuel assembly in the operation process is ensured.

Description

Technical Field

[0001] The invention relates to the technical field of reactor fuel assembly transfer and positioning, and in particular to a reactor assembly transfer and positioning method based on a neural network. Background Art

[0002] The transfer and transportation of reactor fuel assemblies in nuclear power plants is a basic operation carried out in nuclear power plants and is also a necessary condition for the operation of nuclear power plants. In order to improve the efficiency of assembly transportation, nuclear power plants generally transport new and spent assemblies simultaneously in the off-reactor process transportation system. However, after the reactor fuel assemblies enter the preheating box, they are all blindly operated during transportation and use. Therefore, there must be a suitable monitoring system to ensure that the reactor fuel assemblies are in the correct position, otherwise the reactor fuel assemblies may be damaged.

[0003] Therefore, the existing technology has a high reliance on the operating experience of the operator, and the operator needs to pay continuous attention to the monitoring system. When the operator operates the reactor fuel assembly transfer for a long time, it is easy for the operator to be distracted or make an error, which may cause a safety accident in serious cases. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of reactor fuel assembly transfer in the prior art and to provide a reactor assembly transfer positioning method based on a neural network. The method constructs an original positioning model through a neural network, and trains the original positioning model through historical transfer data to obtain a transfer positioning model. The transfer positioning model displays the operation trajectory of the transfer equipment in real time to achieve high-precision positioning of the reactor fuel assembly, and automatically corrects the position according to the deviation when the reactor fuel assembly deviates, thereby ensuring that the reactor fuel assembly is aligned with the reactor core.

[0005] The technical solution of the present invention provides a method for transporting and positioning a reactor component based on a neural network, which is characterized by comprising: Constructing an original positioning model with a neural network structure based on multi-source sensor data; Establish a three-dimensional rectangular coordinate system to plan the optimal transfer path based on the component safety requirements, starting position, and target position of the reactor fuel assembly; The original positioning model is trained by using historical transfer data, and then the parameters of the original positioning model are optimized based on the optimal transfer path to obtain a transfer positioning model; By inputting the multi-source sensor data acquired in real time by the transfer equipment into the transfer positioning model, the transfer positioning model outputs the positioning data of the reactor fuel assembly in real time. The positioning data includes position information, speed information, acceleration information, temperature information, rotation angle information relative to the original position, and height position information.

[0006] In one of the optional technical solutions, it also includes: displaying the operation trajectory of the transfer equipment in real time, and controlling the transfer equipment to make corresponding positioning corrections if the positioning data meets the risk conditions.

[0007] In one of the optional technical solutions, the multi-source sensor data includes laser ranging data, height position sensor data and ultrasonic measurement data; The method of constructing an original positioning model based on a neural network, taking multi-source sensor data as input of the original positioning model, and outputting positioning data of a reactor fuel assembly includes: Integrate laser ranging data, height position sensor data and ultrasonic measurement data into a multi-source sensor data vector as input data for the original positioning model; Normalize the input data and unify the value ranges of different types of sensor data to improve the stability and efficiency of model training; Constructing the original positioning model of the deep neural network structure including input layer, multiple hidden layers and output layer; receiving a normalized multi-source sensor data vector through an input layer; The hidden layer uses nonlinear transformation and feature extraction on the input data. The neurons in each hidden layer perform weighted summation and activation operations on the input according to the weights and biases, and gradually extract more advanced features. The output layer outputs the positioning data of the reactor fuel assembly including position information, velocity information, acceleration information, temperature information, rotation angle information relative to the original position and height position information through linear transformation.

[0008] In one of the optional technical solutions, the establishing of a three-dimensional rectangular coordinate system includes: The farthest top corner of the fuel assembly preheating box is taken as the origin, the transport channel direction is the X-axis, the horizontal and vertical direction is the Y-axis, and the height direction is the Z-axis, which are used to determine the spatial position of the fuel assembly.

[0009] In one of the optional technical solutions, the planning of the optimal transfer path in combination with the safety requirements of the reactor fuel assembly, the starting position, the target position, the starting azimuth and the target azimuth includes: Combined with the component safety requirements, the height that does not meet the requirements in the transfer path is interlocked to prohibit operations in the X or Y direction; the path cost function is constructed by comprehensively considering the path length and the risk assessment value of passing through the dangerous area; The path length is obtained by calculating the distance between the starting position and the target position in a three-dimensional rectangular coordinate system; Risk assessment values ​​are quantified based on the degree of danger; The A* search algorithm is used. On the basis of considering the component safety requirements and path costs, it starts from the starting position, continuously expands the adjacent nodes, calculates the path cost of each node, and gradually searches to the target position to determine the optimal transfer path.

[0010] In one of the optional technical solutions, the training of the original positioning model by using historical transport data includes: Collect historical multi-source sensor data during historical transportation and corresponding historical positioning data of reactor fuel assemblies to form a historical transportation data set; The historical transshipment data set is divided into a training set and a validation set according to a preset ratio for the training and performance verification of the original positioning model; The mean square error loss function is used to calculate the average of the sum of square errors between the positioning data predicted by the model and the actual positioning data. The mean square error loss function is used to measure the degree of deviation between the predicted value of the original positioning model and the true value, so as to optimize the parameters of the original positioning model. Use the stochastic gradient descent optimization algorithm to iteratively train the original positioning model on the training set; In each iteration, the gradient of the model's weights and biases is calculated according to the mean square error loss function, and the weights and biases are adjusted according to the learning rate to reduce the value of the loss function.

[0011] In one of the optional technical solutions, the optimizing the parameters of the original positioning model based on the optimal transfer path to obtain the transfer positioning model includes: On the basis of the original mean square error loss function, the deviation term between the predicted path and the optimal transfer path is added. The deviation term is obtained by calculating the difference between the fuel assembly position and azimuth predicted by the model in the three-dimensional rectangular coordinate system and the azimuth coordinate system and the corresponding point on the optimal transfer path. The balance coefficient is set to balance the impact of the original loss function and the path deviation term. According to the needs of the actual transfer task, the size of the balance coefficient is adjusted to determine the emphasis on positioning data accuracy and path compliance during the model optimization process; Using optimization algorithms such as stochastic gradient descent, and based on a new loss function that includes a path deviation term, the parameters of the original positioning model are iteratively updated again to make the path output by the model closer to the optimal transfer path, thus obtaining a transfer positioning model.

[0012] In one of the optional technical solutions, the multi-source sensor data acquired in real time by the transfer equipment is input into the transfer positioning model, so that the transfer positioning model outputs the positioning data of the reactor fuel assembly in real time, including: During operation, the transfer equipment collects laser ranging data, height position sensor data and ultrasonic measurement data in real time and integrates them into a multi-source sensor data vector; Normalize the multi-source sensor data vectors collected in real time to make them consistent with the input data format and value range when training the transfer positioning model, and then input them into the transfer positioning model; The transfer positioning model outputs the current position information, velocity information, acceleration information, temperature information, rotation angle information relative to the original position and height position information of the reactor fuel assembly based on the input real-time data through forward propagation calculation.

[0013] In one of the optional technical solutions, if the positioning data meets the risk condition, controlling the transfer equipment to make corresponding positioning corrections includes: Construct a risk condition judgment function, comprehensively consider the position deviation, azimuth deviation and speed deviation, and when the value of the risk condition judgment function exceeds the preset risk threshold, it is determined that the positioning data meets the risk condition; The PID control algorithm is used to calculate the correction values ​​of position, azimuth and speed respectively according to the position deviation, azimuth deviation and speed deviation; The correction amount is sent to the transfer equipment, which controls the transfer equipment to adjust its own operating state according to the correction amount, including adjusting the position, azimuth and speed, thereby realizing positioning correction.

[0014] The technical solution of the present invention also provides an electronic device, including a memory, a processor and an electronic device program on the memory, wherein the processor executes the electronic device program to implement any step of the aforementioned neural network-based reactor component transfer and positioning method.

[0015] The above technical solution has the following beneficial effects: The reactor assembly transfer positioning method based on neural network provided by the present invention constructs an original positioning model through a neural network, and trains the original positioning model through historical transfer data to obtain a transfer positioning model, and displays the operation track of the transfer equipment in real time through the transfer positioning model, so as to realize high-precision positioning of the reactor fuel assembly, and automatically correct the position according to the deviation when the reactor fuel assembly deviates, so as to ensure that the reactor fuel assembly is aligned with the operating position, thereby reducing the requirement of the assembly for the operating experience of the operator during the transfer process. After the reactor fuel assembly enters the preheating box, the operator ensures that the reactor fuel assembly is in the correct position and the azimuth angle is correct during the blind operation, so as to avoid the burning of the reactor fuel assembly or jamming when entering the reactor core. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The disclosure of the present invention will become more easily understood with reference to the accompanying drawings. 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 1A flowchart of a method for transporting and positioning a reactor component based on a neural network according to an embodiment of the present invention; Figure 2 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings.

[0018] It should be noted that the reactor fuel assembly transfer control method provided in the embodiment of the present invention is applied to the reactor fuel assembly transfer system, and various data of the reactor fuel assembly are obtained through the transfer equipment. For example, when the transportation device is a transfer cart, multiple sensors can be set on the transfer cart to measure laser ranging data, height position sensor data and ultrasonic measurement data. In the process of transferring the reactor fuel assembly from the temporary 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 cart is docked with the preheating box, and the assembly is positioned by the preheating box assembly manipulator, and inserted into the designated position of the transport container, the azimuth angle is determined to be correct, and the assembly coding and position are recorded. Then the reactor fuel assembly is transported to the reactor for reaction. During the transfer process, the reactor fuel assembly transfer control method provided by the present invention can ensure the safety of the reactor fuel assembly under the premise of transferring the reactor fuel assembly as soon as possible.

[0019] like Figure 1 A method for transporting and positioning a reactor component based on a neural network is shown in an embodiment of the present invention, and is characterized by comprising: Step S101: constructing an original positioning model of a neural network structure based on multi-source sensor data.

[0020] Step S102: Establish a three-dimensional rectangular coordinate system and an azimuth coordinate system, and plan an optimal transfer path in combination with the component safety requirements, starting position, and target position of the reactor fuel assembly.

[0021] Step S103: The original positioning model is trained by using historical transfer data, and then the parameters of the original positioning model are optimized based on the optimal transfer path to obtain a transfer positioning model.

[0022] Step S104: By inputting the multi-source sensor data acquired by the transfer equipment in real time into the transfer positioning model, the transfer positioning model is enabled to output the positioning data of the reactor fuel assembly in real time. The positioning data includes position information, speed information, acceleration information, temperature information, rotation angle information relative to the original position, and height position information.

[0023] Among them, step S101 realizes the deep fusion processing of multi-source heterogeneous sensor data by constructing the original positioning model based on neural network. After model feature extraction and nonlinear mapping, the multi-source sensor data outputs high-precision three-dimensional spatial coordinates and posture information of the fuel assembly. The adaptive compensation of the deep learning algorithm lays a reliable data foundation for subsequent path planning. Step S102 constructs a standardized kinematic model for fuel assembly transportation by establishing a spatial representation system of a three-dimensional rectangular coordinate system. It can generate an optimal transportation path that takes into account both safety and efficiency. Step S103 uses historical transportation data to perform transfer learning training on the original positioning model. The historical transportation data can include sensor readings and verification posture data under typical working conditions, and then focuses on optimizing the feature weight distribution of the deep neural network. Step S104 inputs laser ranging data, height position sensor data, and ultrasonic measurement data from the multi-source sensor data obtained in real time by the transportation equipment into the transportation positioning model, so that the model can output the positioning data of the reactor fuel assembly in real time, realize the real-time application of the positioning model, ensure that the system can dynamically perceive the position and posture of the fuel assembly, and the system can monitor the transportation process in real time to provide data support for subsequent risk judgment and correction operations.

[0024] In summary, the reactor assembly transfer positioning method based on neural network provided by the embodiment of the present invention constructs an original positioning model through a neural network, and trains the original positioning model through historical transfer data to obtain a transfer positioning model, and displays the operation trajectory of the transfer equipment in real time through the transfer positioning model, so as to achieve high-precision positioning of the reactor fuel assembly, and automatically correct the position according to the deviation when the reactor fuel assembly deviates, so as to ensure that the reactor fuel assembly is aligned with the reactor core, thereby reducing the demand for the operating experience of the operator during the transfer of the reactor fuel assembly. After the reactor fuel assembly enters the preheating box, the operator ensures that the reactor fuel assembly is in the correct position and the azimuth is correct during the blind operation, so as to avoid the burning of the reactor fuel assembly or jamming when entering the reactor core.

[0025] In one embodiment, the multi-source sensor data includes laser ranging data, height position sensor data, and ultrasonic measurement data.

[0026] Furthermore, step S101 includes the following sub-steps: Integrate laser ranging data, height position sensor data and ultrasonic measurement data into a multi-source sensor data vector as input data for the original positioning model; Normalize the input data and unify the value ranges of different types of sensor data to improve the stability and efficiency of model training; Constructing the original positioning model of the deep neural network structure including input layer, multiple hidden layers and output layer; receiving a normalized multi-source sensor data vector through an input layer; The hidden layer uses nonlinear transformation and feature extraction on the input data. The neurons in each hidden layer perform weighted summation and activation operations on the input according to the weights and biases, and gradually extract more advanced features. The output layer outputs the positioning data of the reactor fuel assembly including position information, velocity information, acceleration information, temperature information, rotation angle information relative to the original position and height position information through linear transformation.

[0027] In this embodiment, a deep neural network (DNN) is used to construct the original positioning model. The network includes an input layer, several hidden layers and an output layer. Assume that the number of hidden layers is L, the number of neurons in the input layer is n, corresponding to the multi-source sensor data dimension, and the number of neurons in the output layer is 3, corresponding to the location information , speed information and altitude information . T in the following text represents the transpose of a matrix or vector.

[0028] The multi-source sensor data sources used as input are: in, is the laser ranging data, is the height position sensor data, Ultrasonic measurement data.

[0029] The input of the lth hidden layer (l=1,2,…,L) is , when the input layer , the weight matrix is , the bias vector is , then the weighted input for: After activation function , such as the ReLU function , get the activation output ; Output of the output layer for: In summary, the original positioning model is: in, , , and are model parameters, The current azimuth of the reactor fuel assembly can be obtained through the rotation angle information relative to the original position, and specifically the starting azimuth, target azimuth and azimuth deviation can be obtained. Therefore, when planning the optimal conversion path, the starting azimuth and target azimuth of the reactor fuel assembly also need to be obtained.

[0030] In one embodiment, step S102 includes the following sub-steps: Determine the positive direction of a three-dimensional rectangular coordinate system with a specific fixed point in the transfer area as the origin, which is used to represent the position of the reactor fuel assembly; An azimuth coordinate system is established based on the geographic north or a specific reference direction to determine the azimuth of the fuel assembly.

[0031] Combined with the component safety requirements, the height in the transfer path that does not meet the requirements is interlocked to prohibit operations in the X or Y direction; the path cost function is constructed by comprehensively considering the path length, azimuth deviation and risk assessment value of passing through dangerous areas; The path length is obtained by calculating the distance between the starting position and the target position in a three-dimensional rectangular coordinate system; The azimuth deviation is the difference between the starting azimuth and the target azimuth; the risk assessment value is quantified according to the area and degree of danger of the dangerous area passed by the path; The A* search algorithm is used. On the basis of considering the component safety requirements and path costs, it starts from the starting position, continuously expands the adjacent nodes, calculates the path cost of each node, and gradually searches to the target position to determine the optimal transfer path.

[0032] Among them, the A* algorithm can take into account the time complexity and space complexity of the path. The time complexity can reach exponential complexity in the worst case in theory, but in practical applications, due to the role of heuristic functions, the optimal solution can usually be found in a shorter time. The space complexity is similar to the time complexity. In the worst case, an exponential number of nodes may need to be stored, but a good heuristic function can greatly reduce the number of nodes that need to be stored.

[0033] In this embodiment, a three-dimensional rectangular coordinate system is established. and azimuth coordinate system , assuming the starting position of the reactor fuel assembly is , the target location is Component safety requirements can be determined through the safety zone matrix Indicates that the elements Representing coordinates Whether it is a safe area, 1 is safe, 0 is dangerous.

[0034] Define the path cost function for: in, and is the position and direction angle of a point on the path, To estimate the risk of a path passing through a dangerous area, the safety area matrix calculate, and is the weight coefficient, and .

[0035] The matrix S is calculated as follows: in, is the static weight coefficient, is the baseline safety static matrix (set according to industry safety regulations), is the reference security dynamic matrix (usually a diagonal matrix is ​​selected). Get the coordinate point corresponding to the point , and then according to the security area matrix calculate , the specific calculation method is as follows: Where N is the total number of sampling points on the path.

[0036] The path cost can be calculated through the above method, and then the optimal transfer path can be determined.

[0037] In one embodiment, step S103 includes the following sub-steps: Collect historical multi-source sensor data during historical transportation and corresponding historical positioning data of reactor fuel assemblies to form a historical transportation data set; The historical transshipment data set is divided into a training set and a validation set according to a preset ratio for the training and performance verification of the original positioning model; The mean square error loss function is used to calculate the average of the sum of square errors between the positioning data predicted by the model and the actual positioning data. The mean square error loss function is used to measure the degree of deviation between the predicted value of the original positioning model and the true value, so as to optimize the parameters of the original positioning model. Use the stochastic gradient descent optimization algorithm to iteratively train the original positioning model on the training set; In each iteration, the gradient of the model's weights and biases is calculated according to the mean square error loss function, and the weights and biases are adjusted according to the learning rate to reduce the value of the loss function.

[0038] Furthermore, step S103 also includes the following sub-steps: On the basis of the original mean square error loss function, the deviation term between the predicted path and the optimal transfer path is added. The deviation term is obtained by calculating the difference between the fuel assembly position and azimuth predicted by the model in the three-dimensional rectangular coordinate system and the corresponding point on the optimal transfer path. The balance coefficient is set to balance the impact of the original loss function and the path deviation term. According to the needs of the actual transfer task, the size of the balance coefficient is adjusted to determine the emphasis on positioning data accuracy and path compliance during the model optimization process; Using optimization algorithms such as stochastic gradient descent, and based on a new loss function that includes a path deviation term, the parameters of the original positioning model are iteratively updated again to make the path output by the model closer to the optimal transfer path, thus obtaining a transfer positioning model.

[0039] In this example, historical transport data is used The original positioning model is trained, where N is the number of samples. The mean square error loss function is used. : in, For real positioning data, The positioning data predicted by the model. Then use optimization algorithms such as stochastic gradient descent (SGD) to update the model parameters.

[0040] Add path constraints to the loss function To further optimize the model based on the optimal transport path: in, is the balance coefficient, for: in, and are the position and direction angle of the i-th point on the optimal transport path, and The model predicts the position and direction of the corresponding point on the path respectively.

[0041] The optimization algorithm is used again to update the model parameters, and the transfer positioning model can be obtained based on the original positioning model.

[0042] In one embodiment, step S104 includes the following sub-steps: During operation, the transfer equipment collects laser ranging data, height position sensor data and ultrasonic measurement data in real time and integrates them into a multi-source sensor data vector; Normalize the multi-source sensor data vectors collected in real time to make them consistent with the input data format and value range when training the transfer positioning model, and then input them into the transfer positioning model; The transfer positioning model outputs the current position information, speed information and altitude position information of the reactor fuel assembly based on the input real-time data through forward propagation calculation.

[0043] In this embodiment, the multi-source sensor data acquired by the transport equipment in real time is Input into the trained transfer positioning model, the model outputs the positioning data of the reactor fuel assembly in real time .

[0044] In order to avoid uncertainty in the predicted positioning data, this embodiment further performs risk assessment on the real-time predicted positioning data based on the operating trajectory of the transfer equipment, and controls the transfer equipment to make corresponding positioning corrections when the positioning data meets the risk conditions.

[0045] Preferably, the method further includes step S105: displaying the running track of the transfer equipment in real time, and controlling the transfer equipment to make corresponding positioning corrections if the positioning data meets the risk conditions.

[0046] Step S105 enables the operator to intuitively observe the position and posture of the reactor fuel assembly by displaying the operation trajectory of the transfer equipment in real time, and controls the transfer equipment to make corresponding positioning corrections when the positioning data meets the risk conditions, thereby realizing visual monitoring and risk warning of the transfer process, ensuring that the system can discover and correct potential problems in a timely manner, and enabling the system to dynamically adjust operations during the transfer process, no longer overly relying on manual monitoring and operation by operators, thereby ensuring the safety and reliability of the continuous operation of the transfer process.

[0047] In one embodiment, step S105 includes the following sub-steps: Construct a risk condition judgment function, comprehensively consider the position deviation, azimuth deviation and speed deviation, and when the value of the risk condition judgment function exceeds the preset risk threshold, it is determined that the positioning data meets the risk condition; The PID control algorithm is used to calculate the correction values ​​of position, azimuth and speed respectively according to the position deviation, azimuth deviation and speed deviation; The correction amount is sent to the transfer equipment, which controls the transfer equipment to adjust its own operating state according to the correction amount, including adjusting the position, azimuth and speed, thereby realizing positioning correction.

[0048] In this embodiment, the risk condition judgment function is defined as ,when When the preset risk threshold is met, the positioning data is considered to meet the risk condition.

[0049] in, , and is the weight coefficient, is the correction amount.

[0050] Set the risk threshold to ,when When the transfer equipment is controlled, it will make corresponding positioning corrections. It can be calculated through feedback control algorithms such as PID control. The specific calculation method is: in, is the reference speed, , and The proportional coefficient, the integral coefficient and the differential coefficient of the PID controller are used to adjust the transfer equipment according to the correction value. Adjust its own status to achieve real-time positioning correction and posture correction.

[0051] During the entire process of reactor fuel assembly transportation, a unique identifier is assigned to each reactor fuel assembly, and its position, azimuth and operation node information is recorded in real time to ensure that each reactor fuel assembly can be correctly distinguished and recorded when multiple reactor fuel assemblies are transported at the same time. Risk monitoring and position correction are also performed for each reactor fuel assembly to ensure the safety and efficiency of the overall system.

[0052] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0053] like Figure 2 The figure shows a schematic diagram of the hardware structure of an electronic device of the present invention, including a memory 202, a processor 201 and an electronic device program on the memory 202, wherein the processor 201 executes the electronic device program to implement the steps of the reactor component transfer positioning method based on neural network in any of the above-mentioned embodiments.

[0054] Figure 2 A processor 201 is taken as an example.

[0055] The electronic device may further include: an input device 203 and a display device 204 .

[0056] The processor 201, the memory 202, the input device 203 and the display device 204 may be connected via a bus or other means, and the figure takes the connection via a bus as an example.

[0057] The memory 202 is a non-volatile electronic device readable storage medium, which can be used to store non-volatile software programs, non-volatile electronic device executable programs and modules, such as program instructions / modules corresponding to the reactor assembly transfer positioning method based on neural network in the embodiment of the present application. The processor 201 executes various functional applications and data processing by running the non-volatile software programs, instructions and modules stored in the memory 202, that is, the reactor assembly transfer positioning method based on neural network in the above embodiment is realized.

[0058] The memory 202 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required by at least one function; the data storage area may store data created according to the use of the reactor assembly transfer positioning method based on a neural network, etc. In addition, the memory 202 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 202 may optionally include a memory remotely arranged relative to the processor 201, and these remote memories may be connected to a device for executing the reactor assembly transfer positioning method based on a neural network through a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0059] The input device 203 can receive user clicks and generate signal inputs related to user settings and function control of the reactor component transport positioning method based on neural network. The display device 204 can include display devices such as a display screen.

[0060] The one or more modules are stored in the memory 202, and when executed by the one or more processors 201, the reactor component transfer positioning method based on neural network in any of the above method embodiments is executed.

[0061] When the electronic device disclosed in the present invention is in operation, it can execute all the steps of the above-mentioned reactor assembly transfer positioning method based on neural network, construct an original positioning model through a neural network, and train the original positioning model through historical transfer data to obtain a transfer positioning model. The operation trajectory of the transfer equipment is displayed in real time through the transfer positioning model to achieve high-precision positioning of the reactor fuel assembly, and automatically correct the position according to the deviation when the reactor fuel assembly deviates, to ensure that the reactor fuel assembly is aligned with the reactor core.

[0062] An embodiment of the present invention provides an electronic device readable storage medium, which stores an electronic device program / instruction. When the electronic device program / instruction is executed by the processor 201, all steps of the reactor component transfer positioning method based on neural network as described above are implemented.

[0063] In the context of the present disclosure, a storage medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or apparatus. The storage medium may be a machine-readable signal medium or a machine-readable storage medium. Alternatively, the storage medium may be a non-temporary electronic device readable storage medium, for example, a non-temporary electronic device readable storage medium may be a ROM, a random access memory (Random Access Memory, RAM), a compact disc read-only memory (Compact Disc ROM, CDROM), a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0064] An embodiment of the present invention provides an electronic device program product, including an electronic device program / instruction, which, when executed by a processor, implements the steps of the reactor component transfer positioning method based on neural network as described above.

[0065] By running the above-mentioned electronic device program product, all steps of the reactor assembly transfer positioning method based on neural network as mentioned above can be executed, the original positioning model is constructed by the neural network, and the original positioning model is trained by historical transfer data to obtain the transfer positioning model. The operation trajectory of the transfer equipment is displayed in real time through the transfer positioning model, and high-precision positioning of the reactor fuel assembly is achieved. When the reactor fuel assembly deviates, the position is automatically corrected according to the deviation to ensure that the reactor fuel assembly is aligned with the reactor core.

[0066] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for transporting and locating reactor components based on a neural network, characterized in that: include: Constructing an original positioning model with a neural network structure based on multi-source sensor data; Establish a one-dimensional rectangular coordinate system to plan the optimal transfer path based on the component safety requirements, starting position, and target position of the reactor fuel assembly; The original positioning model is trained by using historical transfer data, and then the parameters of the original positioning model are optimized based on the optimal transfer path to obtain a transfer positioning model; By inputting the multi-source sensor data acquired in real time by the transfer equipment into the transfer positioning model, the transfer positioning model outputs the positioning data of the reactor fuel assembly in real time. The positioning data includes position information, speed information, acceleration information, temperature information, rotation angle information relative to the original position, and height position information.

2. The method for transporting and locating reactor components based on a neural network according to claim 1, characterized in that: Also includes: The operation track of the transfer equipment is displayed in real time, and if the positioning data meets the risk conditions, the transfer equipment is controlled to make corresponding positioning corrections.

3. The method for transporting and positioning a reactor component based on a neural network according to claim 2, characterized in that: The multi-source sensor data includes laser ranging data, height position sensor data and ultrasonic measurement data; The method of constructing an original positioning model based on a neural network, taking multi-source sensor data as input of the original positioning model, and outputting positioning data of a reactor fuel assembly includes: Integrate laser ranging data, height position sensor data and ultrasonic measurement data into a multi-source sensor data vector as input data for the original positioning model; Normalize the input data and unify the value ranges of different types of sensor data to improve the stability and efficiency of model training; Constructing the original positioning model of the deep neural network structure including input layer, multiple hidden layers and output layer; receiving a normalized multi-source sensor data vector through an input layer; The hidden layer uses nonlinear transformation and feature extraction on the input data. The neurons in each hidden layer perform weighted summation and activation operations on the input according to the weights and biases, and gradually extract more advanced features. The output layer outputs the positioning data of the reactor fuel assembly including position information, velocity information, acceleration information, temperature information, rotation angle information relative to the original position and height position information through linear transformation.

4. The method for transporting and locating reactor components based on a neural network according to claim 1, characterized in that: The step of establishing a one-dimensional rectangular coordinate system comprises: The farthest top corner of the fuel assembly preheating box is taken as the origin, the transport channel direction is the X-axis, the horizontal and vertical direction is the Y-axis, and the height direction is the Z-axis, which are used to determine the spatial position of the fuel assembly.

5. The method for transporting and locating reactor components based on a neural network according to claim 4, characterized in that: The optimal transport path is planned in combination with the safety requirements, the starting position and the target position of the reactor fuel assembly, including: Combined with component safety requirements, operations in the X or Y direction are prohibited by interlocking when the height in the transfer path does not meet the requirements; The path cost function is constructed by comprehensively considering the path length and the risk assessment value of passing through the dangerous area; The path length is obtained by calculating the distance between the starting position and the target position in a three-dimensional rectangular coordinate system; Risk assessment values ​​are quantified based on the degree of danger; The A* search algorithm is used. On the basis of considering the component safety requirements and path costs, it starts from the starting position, continuously expands the adjacent nodes, calculates the path cost of each node, and gradually searches to the target position to determine the optimal transfer path.

6. The method for transporting and locating reactor components based on a neural network according to claim 1, characterized in that: The method of training the original positioning model by using historical transport data comprises: Collect historical multi-source sensor data during historical transportation and corresponding historical positioning data of reactor fuel assemblies to form a historical transportation data set; The historical transshipment data set is divided into a training set and a validation set according to a preset ratio for the training and performance verification of the original positioning model; The mean square error loss function is used to calculate the average of the sum of square errors between the positioning data predicted by the model and the actual positioning data. The mean square error loss function is used to measure the degree of deviation between the predicted value of the original positioning model and the true value, so as to optimize the parameters of the original positioning model. Use the stochastic gradient descent optimization algorithm to iteratively train the original positioning model on the training set; In each iteration, the gradient of the model's weights and biases is calculated according to the mean square error loss function, and the weights and biases are adjusted according to the learning rate to reduce the value of the loss function.

7. The method for transporting and locating reactor components based on a neural network according to claim 6, characterized in that: The step of optimizing the parameters of the original positioning model based on the optimal transfer path to obtain a transfer positioning model includes: On the basis of the original mean square error loss function, the deviation term between the predicted path and the optimal transfer path is added. The deviation term is obtained by calculating the difference between the fuel assembly position and azimuth predicted by the model in the three-dimensional rectangular coordinate system and the azimuth coordinate system and the corresponding point on the optimal transfer path. The balance coefficient is set to balance the impact of the original loss function and the path deviation term. According to the needs of the actual transfer task, the size of the balance coefficient is adjusted to determine the emphasis on positioning data accuracy and path compliance during the model optimization process; Using optimization algorithms such as stochastic gradient descent, and based on a new loss function that includes a path deviation term, the parameters of the original positioning model are iteratively updated again to make the path output by the model closer to the optimal transfer path, thus obtaining a transfer positioning model.

8. The method for transporting and locating reactor components based on a neural network according to claim 1, characterized in that: The method comprises: inputting multi-source sensor data acquired in real time by the transfer equipment into the transfer positioning model, so that the transfer positioning model outputs and acquires the positioning data of the reactor fuel assembly in real time, comprising: During operation, the transfer equipment collects laser ranging data, height position sensor data and ultrasonic measurement data in real time and integrates them into a multi-source sensor data vector; Normalize the multi-source sensor data vectors collected in real time to make them consistent with the input data format and value range when training the transfer positioning model, and then input them into the transfer positioning model; The transfer positioning model outputs the current position information, velocity information, acceleration information, rotation angle information relative to the original position and height position information of the reactor fuel assembly based on the input real-time data through forward propagation calculation.

9. The method for transporting and locating reactor components based on a neural network according to claim 2, characterized in that: If the positioning data meets the risk condition, controlling the transfer equipment to make corresponding positioning corrections includes: Construct a risk condition judgment function, comprehensively consider the position deviation, azimuth deviation and speed deviation, and when the value of the risk condition judgment function exceeds the preset risk threshold, it is determined that the positioning data meets the risk condition; The PID control algorithm is used to calculate the correction values ​​of position, azimuth and speed respectively according to the position deviation, azimuth deviation and speed deviation; The correction amount is sent to the transfer equipment, which controls the transfer equipment to adjust its own operating state according to the correction amount, including adjusting the position, azimuth and speed, thereby realizing positioning correction.

10. An electronic device, 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 component transfer positioning method described in any one of claims 1-9.

Images