Autonomous Control Method, Device and Computer Equipment for Nuclear Reactor

Abstract

The present application relates to an autonomous control method, device, computer device, storage medium, and computer program product for a nuclear reactor. The method includes: determining a target state trajectory and a full reactor state trajectory of the nuclear reactor; performing trajectory optimization on the full reactor state trajectory according to the target state trajectory to obtain a control action combination for the nuclear reactor; the control action combination is used to ensure that the deviation between the target state trajectory and the full reactor state trajectory reaches a preset value and satisfies the dynamic evolution characteristics of the reactor; and performing autonomous control on the nuclear reactor based on the control action combination. Using this method can improve the control accuracy of the nuclear reactor.

Description

Technical Field

[0001] The present application relates to the field of nuclear reaction technology, and particularly to an autonomous control method, device, computer equipment, storage medium and computer program product for a nuclear reactor. Background Art

[0002] During the operation of a reactor, uncertainties from multiple sources, such as measurements, control effects, and dynamic models of reactor state transitions, are usually involved. Therefore, the reactor usually needs to be controlled within the full state confidence space. The measurement signals of the reactor, including neutron detector signals, thermocouple readings, or loop coolant flow rates, pressurizer pressures, etc., can be directly measured. However, there are also a large number of safety-related state parameters that cannot be directly measured, such as the effective neutron multiplication factor keff, macroscopic burnup, microscopic nucleon density, etc.

[0003] The autonomous control of a nuclear reactor requires the reactor to have more flexible operation and autonomous control capabilities to ensure the long-term reliable operation of the system in harsh or remote service environments. However, the existing reactor power control mainly adopts the PID control method. The output signal of the PID controller drives the reactor control rod drive mechanism, thereby introducing reactivity for power control of the reactor, resulting in low control accuracy of the autonomous control of the nuclear reactor. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide an autonomous control method, device, computer equipment, computer-readable storage medium and computer program product for a nuclear reactor that can improve the control accuracy of the autonomous control of the nuclear reactor.

[0005] In a first aspect, the present application provides an autonomous control method for a nuclear reactor. The method includes:

[0006] Determine the target state trajectory and the full state trajectory of the nuclear reactor;

[0007] Perform trajectory optimization on the full state trajectory of the reactor according to the target state trajectory to obtain a control action combination for the nuclear reactor; the control action combination is used to ensure that the deviation between the target state trajectory and the full state trajectory of the reactor reaches a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0008] Autonomously control the nuclear reactor based on the control action combination.

[0009] In one embodiment, the determining the target state trajectory and the full state trajectory of the nuclear reactor includes:

[0010] Determine the target state trajectory of the nuclear reactor;

[0011] Input the full state parameters of the nuclear reactor into a pre-constructed integrated prediction model for prediction to obtain the full state trajectory of the reactor.

[0012] In one embodiment, before determining the target state trajectory of the nuclear reactor, the method further includes:

[0013] Pre-construct the mathematical model and machine learning model of the nuclear reactor;

[0014] Obtain the integrated prediction model by performing integrated processing on the mathematical model and the machine learning model.

[0015] In one embodiment, the pre-constructing the mathematical model and machine learning model of the nuclear reactor includes:

[0016] Construct the mathematical model of the nuclear reactor; the mathematical model at least includes a core neutron point kinetic model, a thermal-hydraulic model, a reactivity feedback model, and a nuclide decay model;

[0017] Obtain the reactor operation samples of the nuclear reactor;

[0018] Train different types of machine learning models constructed according to the reactor operation samples to obtain trained machine learning models.

[0019] In one embodiment, the optimizing the trajectory of the reactor full state trajectory according to the target state trajectory to obtain the control action combination of the nuclear reactor includes:

[0020] Determine the reaction state and corresponding control actions in the reactor full state trajectory;

[0021] Based on the dynamic evolution characteristics of the reactor, determine a first trajectory including the reactor state and control actions;

[0022] Perform linearization processing on the first trajectory through Taylor approximation to obtain a second trajectory under the optimal strategy;

[0023] Perform iterative processing on the second trajectory until it converges to the target state trajectory to obtain the control action combination of the nuclear reactor.

[0024] In one embodiment, before performing iterative processing on the second trajectory until it converges to the target state trajectory to obtain the control action combination of the nuclear reactor, the method further includes:

[0025] Update the next reactor state of the first trajectory according to the actual state transition function, and perform the step of performing linearization processing on the first trajectory through Taylor approximation to obtain a second trajectory under the optimal strategy.

[0026] In one embodiment, the machine learning model at least includes a DNN prediction model, an RNN prediction model, and a prediction model based on Gaussian regression.

[0027] In a second aspect, the present application also provides an autonomous control device for a nuclear reactor. The device includes:

[0028] A determination module, configured to determine the target state trajectory of the nuclear reactor and the full state trajectory of the reactor;

[0029] A trajectory optimization module, configured to optimize the full state trajectory of the reactor according to the target state trajectory to obtain a control action combination of the nuclear reactor; the control action combination enables the deviation between the target state trajectory and the full state trajectory of the reactor to reach a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0030] A control module, configured to perform autonomous control on the nuclear reactor based on the control action combination.

[0031] In a third aspect, the present application also provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0032] Determine the target state trajectory of the nuclear reactor and the full state trajectory of the reactor;

[0033] Optimize the full state trajectory of the reactor according to the target state trajectory to obtain a control action combination of the nuclear reactor; the control action combination is used to ensure that the deviation between the target state trajectory and the full state trajectory of the reactor reaches a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0034] Perform autonomous control on the nuclear reactor based on the control action combination.

[0035] In a fourth aspect, the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0036] Determine the target state trajectory of the nuclear reactor and the full state trajectory of the reactor;

[0037] Optimize the full state trajectory of the reactor according to the target state trajectory to obtain a control action combination of the nuclear reactor; the control action combination is used to ensure that the deviation between the target state trajectory and the full state trajectory of the reactor reaches a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0038] Based on the control action combination, perform autonomous control on the nuclear reactor.

[0039] In a fifth aspect, the present application also provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0040] Determine the target state trajectory and the full state trajectory of the reactor of the nuclear reactor;

[0041] Perform trajectory optimization on the full state trajectory of the reactor according to the target state trajectory to obtain the control action combination of the nuclear reactor; the control action combination is used to ensure that the deviation between the target state trajectory and the full state trajectory of the reactor reaches a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0042] Based on the control action combination, perform autonomous control on the nuclear reactor.

[0043] The above-mentioned autonomous control method, device, computer device, storage medium, and computer program product of the nuclear reactor obtain the control action combination of the nuclear reactor by performing trajectory optimization according to the target state trajectory and the full state trajectory of the reactor; the control action combination enables the deviation between the target state trajectory and the full state trajectory of the reactor to reach a preset value and satisfies the dynamic evolution characteristics of the reactor; perform autonomous control on the nuclear reactor according to the control action combination; that is, realize multi-objective operation optimization in a large state space, solve the control deficiencies of PID single-input single-output or multi-input single-output, and adopt the dynamic programming method, effectively improving the accuracy of autonomous control of the nuclear reactor. Description of the Drawings

[0044] Figure 1 It is a schematic flow chart of the autonomous control method of the nuclear reactor in an embodiment of the present application;

[0045] Figure 2 It is a schematic diagram of the integrated processing of the prediction model in an embodiment of the present application;

[0046] Figure 3 It is a schematic structural diagram of the DNN prediction model in an embodiment of the present application;

[0047] Figure 4 It is a schematic structural diagram of the RNN prediction model in an embodiment of the present application;

[0048] Figure 5 It is a schematic structural diagram of the prediction model based on Gaussian regression in an embodiment of the present application;

[0049] Figure 6 It is a schematic flow chart of the autonomous control method of the nuclear reactor in another embodiment of the present application;

[0050] Figure 7 The effect diagram of the autonomous control of the nuclear reactor in an embodiment of the present application;

[0051] Figure 8 The structural block diagram of the autonomous control device of the nuclear reactor in an embodiment of the present application;

[0052] Figure 9 The internal structure diagram of the computer device in an embodiment of the present application. Detailed implementation manners

[0053] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0054] In one embodiment, as Figure 1 shown, an autonomous control method for a nuclear reactor is provided. In this embodiment, the method is described by taking its application to a terminal as an example. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0055] Step 102, determine the target state trajectory and the full state trajectory of the reactor of the nuclear reactor.

[0056] Among them, the nuclear reactor includes different types of reactors such as second-generation reactors, third-generation reactors (pressurized water reactors, boiling water reactors, etc.), and fourth-generation reactors (metal-cooled fast reactors, gas-cooled fast reactors, molten salt fast reactors).

[0057] The target state trajectory is the reactor operating power level of the nuclear reaction at different predetermined times, and the target state trajectory is represented as s t,target ; the full state trajectory of the reactor is the state trajectory determined based on the full state parameters of the nuclear reactor; the full state trajectory of the reactor can be predicted based on the full state parameters of the nuclear reactor through an integrated prediction model.

[0058] The integrated prediction model is obtained by integrating the pre-constructed mathematical model and machine learning model of the nuclear reactor based on the Stacking technology framework; by constructing the mathematical model and machine learning model, the difficulty of building the reactor state transition model is reduced, and the problems such as difficult modeling and slow convergence of the traditional reactor mathematical model are avoided; the integrated processing steps are as Figure 2As shown in the figure, the mathematical physics models (including the point reactor model and the high-precision model, etc.) and the machine learning models (including machine learning model 1... machine learning model N) are determined as the basic learning models, and linear regression is used as the second-level learning model (including multiple linear regression, random forest regression, etc.). They are trained through the Stacking learning strategy, and the prediction results of each model are averaged or weighted averaged to obtain an integrated prediction model.

[0059] The mathematical physics models can be but are not limited to the point reactor model, one-dimensional model, three-dimensional model, and fractional-order model; in this embodiment, the point reactor model is taken as an example for illustration. The point reactor model includes the core neutron point reactor dynamics model, thermal-hydraulic model, reactivity feedback model, and nuclide decay model;

[0060] The point reactor dynamics model is expressed as:

[0061]

[0062] The thermal-hydraulic model can be expressed as:

[0063]

[0064]

[0065] The nuclide decay model can be expressed as:

[0066]

[0067]

[0068]

[0069] The reactivity feedback model can be expressed as:

[0070] ρ = ρ0 + ρ Xe +ρ Sm +ρ T +ρ rod (9) ρ T =α f (T f -T f0 )+α c (T cav -T cav0 ) (10)

[0071]

[0072] The following are the meanings of the parameters in the above point reactor dynamics model, thermal-hydraulic model, nuclide decay model, and reactivity feedback model:

[0073]

[0074]

[0075]

[0076] Among them, the parameters in the state space of the nuclear reactor include:

[0077] s t = [n r , C r,1 , C r,2 , C r,3 , C r,4 , C r,5 , C r,6 , T f , T cav , Xe, I od , Pm, Sm], and the parameters in the action space of the reactor are: a t = [ρ rod , T cin . The control actions of the nuclear reactor include reactivity insertion, including control rod or drum movement, soluble boron or other neutron absorbers; changing the core inlet temperature by adjusting the pressure and heat exchange of the primary and secondary nuclear reactor systems.

[0078] The machine learning model is obtained by learning based on the determined reactor state change samples, which include the parameters in the reactor state space and action space; the acquisition method of the reactor state change samples can be but is not limited to being determined based on a high-precision numerical reactor by simulating a large amount of operation data. The machine learning model at least includes DNN (Deep Neural Networks) prediction models (as shown in Figure 3 ), RNN (Recurrent Neural Network) prediction models (as shown in Figure 4 ), and prediction models based on Gaussian regression (as shown in Figure 5 ), etc.

[0079] As shown in Figure 3 , for the DNN prediction model of the nuclear reactor in an embodiment, the reactor operation sample at time t includes a state space vector and an action space vector, and is used to predict the reactor state at time t + 1 based on the operation sample data at the past T time instants (t, t - 1, t - T + 1). The DNN prediction model includes an input layer, hidden layers 1 - N, and an output layer; when obtaining the complete state space parameters S t , the historical information S t-1 before the time step t is not required in the DNN.

[0080] Figure 4 The network structure in the medium reactor's RNN prediction model does not need to be very deep, that is, it does not require many hidden layers. Based on the operating sample data at the past T moments (t, t - 1, t - T + 1), it is used to predict the reactor state at the (t + 1)th moment. Figure 5 In the prediction model based on Gaussian regression, based on the operating sample data at the past T moments (t, t - 1, t - T + 1), Gaussian process regression is performed to predict the reactor state at the (t + 1)th moment.

[0081] Step 104: Optimize the full-state trajectory of the reactor according to the target state trajectory to obtain the control action combination of the nuclear reactor.

[0082] Among them, optimizing the full-state trajectory of the reactor according to the target state trajectory is to ensure that the deviation between the target state trajectory and the full-state trajectory of the reactor is minimized during the power operation process, meeting the reaction requirements of the nuclear reaction and satisfying the dynamic evolution characteristics of the reactor. The dynamic evolution characteristics refer to the relationship (which can be a linear relationship or a non-linear relationship) satisfied between the reactor state at the tth moment and the control action at the tth moment.

[0083] The control action combination includes the reactor state and control actions that meet the target operation trajectory. The reactor state includes relative neutron density ratio, the density of the i-th group of delayed neutron precursors (i is a positive integer), average fuel temperature in the core, average coolant temperature, xenon density, iodine density, promethium density, and samarium density, etc.; the control actions (which can be understood as control quantities) are used for control rod reactivity negative feedback and average coolant temperature at the core inlet; it can be understood that the target operation trajectory is an operating trajectory under ideal conditions, and the control action combination controls the nuclear reactor to ensure that the deviation between the actual operation state and the target operation state of the nuclear reactor is minimized. It can be understood that the control action combination makes the deviation between the target state trajectory and the full-state trajectory of the reactor reach a preset value and satisfies the dynamic evolution characteristics of the reactor.

[0084] Specifically, according to the pre-determined target operation trajectory, combined with the integrated prediction model, optimize the full-state trajectory of the reactor to obtain the control action combination at each moment of the reactor state.

[0085] Step 106: Autonomously control the nuclear reactor based on the control action combination.

[0086] Specifically, autonomously control the nuclear reactor according to the control action combination at each moment; that is, drive the reactor control rod drive mechanism according to the control action combination, thereby introducing reactivity for power control of the reactor.

[0087] In the above-mentioned autonomous control method of the nuclear reactor, through trajectory optimization based on the target state trajectory and the full-state trajectory of the reactor, a control action combination of the nuclear reactor is obtained; the control action combination enables the deviation between the target state trajectory and the full-state trajectory of the reactor to reach a preset value and satisfies the dynamic evolution characteristics of the reactor; the nuclear reactor is autonomously controlled according to the control action combination; that is, multi-objective operation optimization in a large state space is realized, the control deficiencies of PID single-input single-output or multi-input single-output are solved, and the dynamic programming method is adopted, effectively improving the accuracy of the autonomous control of the nuclear reactor.

[0088] In another embodiment, as Figure 6 shown, an autonomous control method of a nuclear reactor is provided. In this embodiment, taking the application of this method to a terminal as an example, it can be understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is realized through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0089] Step 602, pre-construct a mathematical model and a machine learning model of the nuclear reactor.

[0090] Step 604, obtain an integrated prediction model by performing integrated processing on the mathematical model and the machine learning model.

[0091] Specifically, the pre-constructed mathematical model and machine learning model of the nuclear reactor are integrated based on the Stacking technology framework to obtain an integrated prediction model.

[0092] Step 606, determine the target state trajectory of the nuclear reactor.

[0093] Specifically, obtain the reactor operation power level of the nuclear reaction at different times to obtain the target state trajectory of the nuclear reactor.

[0094] Step 608, input the full-state parameters of the nuclear reactor into the pre-constructed integrated prediction model for prediction to obtain the full-state trajectory of the reactor.

[0095] Step 610, determine the reaction state and the corresponding control action in the full-state trajectory of the reactor.

[0096] Specifically, determine the reaction state s t of the nuclear reactor system according to the full-state trajectory of the reactor, t and take action a t (i.e., the control action).

[0097] Step 612, based on the dynamic evolution characteristics of the reactor, determine a first trajectory including the reactor state and the control action.

[0098] Among them, in the state transition of the reactor, the dynamic evolution characteristic x of the reactor t+1 = f(τ t ) is a highly non-linear function.

[0099] Specifically, based on the dynamic evolution characteristics of the reactor and the model predictive control (MPC) method, a first trajectory including the reactor state and control actions is determined.

[0100] MPC can be expressed as:

[0101]

[0102] Among them, w is the weight of each variable (including measurable or unmeasurable hidden variables) in the state space, s is the reactor power level, s t,target is the reactor target power level, τ t is the control action at the t-th moment, C t is the diagonal coefficient of the negative weight, x t is the reactor state at the t-th moment.

[0103] Determine the reaction state s of the nuclear reactor system according to the full state trajectory of the reactor t , and take action a t at the nominal operating point of the reaction state s t (i.e., control action), to obtain a first trajectory including the reactor state and control actions The nominal operating point is a fixed point assumed in advance for linear approximation.

[0104] Step 614, linearize the first trajectory through Taylor approximation to obtain a second trajectory under the optimal strategy.

[0105] Among them, the optimal strategy refers to determining the reactor state and control actions that satisfy the target operating state.

[0106] Specifically, based on the control problem of the reactor, linearize the dynamic evolution characteristic x of the reactor t+1 = f(τ t ) (dynamic model) to obtain the linearized model:

[0107]

[0108] It can be seen that s t+1 is a linear function of s t and a t : s t+1 ≈ A t · s t + B t·a t , that is, the next reactor state of the reactor, the reactor state at the previous moment, and the control action satisfy a linear relationship; thus, it can be determined

[0109] the function τ in t =(s t , a t ) is approximated by Taylor expansion as:

[0110]

[0111] wherein, H sa is the Hessian matrix of R t at the point . R t is written as:

[0112] According to to determine the second trajectory when x t+1 =f(τ t ) is a highly nonlinear function to satisfy the target operating state:

[0113]

[0114] Furthermore, by performing linearization processing on the first trajectory through Taylor approximation to obtain the second trajectory under the optimal strategy, and updating the next reactor state of the first trajectory according to the actual state transition function, it can be expressed as Continue to execute the step of performing linearization processing on the first trajectory through Taylor approximation to obtain the second trajectory under the optimal strategy.

[0115] Step 616, perform iterative processing on the second trajectory until it converges to the target state trajectory to obtain the control action combination of the nuclear reactor.

[0116] Specifically, iterate the obtained second trajectory until it converges to the target state trajectory to obtain the control action combination of the nuclear reactor; it can be understood that when determining the actual operating trajectory of the nuclear reactor, based on the actual state transition function of the reaction, determine the nominal trajectory that satisfies the preset conditions to obtain the preset reactor state (i.e., the approximate state) of each control action, and perform iterative processing based on the above linearization function to generate the next nominal trajectory until the nominal trajectory converges to the target trajectory to obtain the control action combination of the nuclear reactor.

[0117] Step 618, perform autonomous control on the nuclear reactor based on the control action combination.

[0118] Such as Figure 7As shown, the effect diagram of the autonomous control of the nuclear reactor by the control action combination determined based on the integrated prediction model and the target operation trajectory.

[0119] The above-mentioned autonomous control method of the nuclear reactor integrates the mathematical model and the machine learning model; that is, it combines the advantages of different reactor models to improve the control accuracy and reduce the influence of detection noise pollution; it optimizes the trajectory according to the integrated prediction model and the target state trajectory to obtain the control action combination of the nuclear reactor; the control action combination makes the deviation between the target state trajectory and the full state trajectory of the reactor reach the preset value and satisfies the dynamic evolution characteristics of the reactor; it autonomously controls the nuclear reactor according to the control action combination; that is, it realizes the multi-objective operation optimization in the large state space, solves the control deficiencies of PID single-input single-output or multi-input single-output, and adopts the dynamic programming method, effectively improving the accuracy of the autonomous control of the nuclear reactor and realizing the integrated model predictive control of the reactor.

[0120] It should be understood that although each step in the flowcharts involved in the above-described embodiments is displayed sequentially as indicated by the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless otherwise clearly stated in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or alternately with at least a part of other steps or steps in other steps.

[0121] Based on the same inventive concept, the embodiment of the present application also provides an autonomous control device for a nuclear reactor for implementing the above-mentioned autonomous control method of the nuclear reactor. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the autonomous control device for a nuclear reactor provided below can refer to the limitations on the autonomous control method of the nuclear reactor in the above text, and will not be repeated here.

[0122] In one embodiment, as Figure 8 shown, an autonomous control device for a nuclear reactor is provided, including: a determination module 802, a trajectory optimization module 804, and a control module 806, where:

[0123] The determination module 802 is used to determine the target state trajectory and the full state trajectory of the nuclear reactor;

[0124] A trajectory optimization module 804, configured to optimize the full-state trajectory of the reactor according to the target state trajectory to obtain a control action combination for the nuclear reactor; the control action combination enables the deviation between the target state trajectory and the full-state trajectory of the reactor to reach a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0125] A control module 806, configured to autonomously control the nuclear reactor based on the control action combination.

[0126] The above-mentioned autonomous control device for the nuclear reactor obtains a control action combination for the nuclear reactor by optimizing the trajectory according to the target state trajectory and the full-state trajectory of the reactor; the control action combination enables the deviation between the target state trajectory and the full-state trajectory of the reactor to reach a preset value and satisfies the dynamic evolution characteristics of the reactor; autonomously controls the nuclear reactor according to the control action combination; that is, it realizes multi-objective operation optimization in a large state space, solves the control deficiencies of PID single-input single-output or multi-input single-output, and effectively improves the accuracy of the autonomous control of the nuclear reactor by using the dynamic programming method.

[0127] Optionally, in an embodiment, the determination module 802 is further configured to determine the target state trajectory of the nuclear reactor.

[0128] Optionally, in an embodiment, the autonomous control device for the nuclear reactor further includes a prediction module, configured to input the full-state parameters of the nuclear reactor into a pre-constructed integrated prediction model for prediction to obtain the full-state trajectory of the reactor.

[0129] Optionally, in an embodiment, the autonomous control device for the nuclear reactor further includes a construction module and an integration module, where:

[0130] The construction module is configured to pre-construct a mathematical model and a machine learning model of the nuclear reactor; the machine learning model at least includes a DNN prediction model, an RNN prediction model, and a prediction model based on Gaussian regression.

[0131] The integration module is configured to obtain an integrated prediction model by performing integration processing on the mathematical model and the machine learning model.

[0132] Optionally, in an embodiment, the construction module is configured to construct a mathematical model of the nuclear reactor; the mathematical model at least includes a core neutron point kinetic model, a thermal-hydraulic model, a reactivity feedback model, and a nuclide decay model.

[0133] Optionally, in an embodiment, the autonomous control device for the nuclear reactor further includes a training module, configured to obtain reactor operation samples of the nuclear reactor; train different types of constructed machine learning models according to the reactor operation samples to obtain trained machine learning models.

[0134] Optionally, in one embodiment, the trajectory optimization module 804 is further configured to determine the reaction state and corresponding control actions in the full-state trajectory of the reactor;

[0135] Based on the dynamic evolution characteristics of the reactor, determine a first trajectory including the reactor state and control actions;

[0136] Perform linearization processing on the first trajectory through Taylor approximation to obtain a second trajectory under the optimal strategy;

[0137] Perform iterative processing on the second trajectory until it converges to the target state trajectory to obtain a combination of control actions for the nuclear reactor.

[0138] Optionally, in one embodiment, the trajectory optimization module 804 is further configured to update the next reactor state of the first trajectory according to the actual state transition function.

[0139] Each module in the above-mentioned autonomous control device of the nuclear reactor can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0140] In one embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 9 shown. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be achieved through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements an autonomous control method for a nuclear reactor. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, a touchpad, or a mouse, etc.

[0141] Those skilled in the art can understand, Figure 9The structure shown is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have a different component layout.

[0142] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the following steps are implemented:

[0143] Determine the target state trajectory of the nuclear reactor and the full-state trajectory of the reactor;

[0144] Perform trajectory optimization on the full-state trajectory of the reactor according to the target state trajectory to obtain a control action combination for the nuclear reactor; the control action combination is used to ensure that the deviation between the target state trajectory and the full-state trajectory of the reactor reaches a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0145] Autonomously control the nuclear reactor based on the control action combination.

[0146] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0147] Determine the target state trajectory of the nuclear reactor;

[0148] Input the full-state parameters of the nuclear reactor into a pre-constructed integrated prediction model for prediction to obtain the full-state trajectory of the reactor.

[0149] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0150] Pre-construct a mathematical model and a machine learning model of the nuclear reactor;

[0151] Obtain an integrated prediction model by performing integrated processing on the mathematical model and the machine learning model.

[0152] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0153] Construct a mathematical model of the nuclear reactor; the mathematical model at least includes a core neutron point kinetics model, a thermal-hydraulic model, a reactivity feedback model, and a nuclide decay model;

[0154] Obtain the reactor operation samples of the nuclear reactor;

[0155] Train different types of machine learning models constructed according to the reactor operation samples to obtain trained machine learning models.

[0156] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0157] Determine the reaction state and corresponding control actions in the full state trajectory of the reactor;

[0158] Based on the dynamic evolution characteristics of the reactor, determine the first trajectory including the reactor state and control actions;

[0159] Perform linearization processing on the first trajectory through Taylor approximation to obtain the second trajectory under the optimal strategy;

[0160] Iteratively process the second trajectory until it converges to the target state trajectory to obtain the control action combination of the nuclear reactor.

[0161] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0162] Update the next reactor state of the first trajectory according to the actual state transition function, and perform the step of performing linearization processing on the first trajectory through Taylor approximation to obtain the second trajectory under the optimal strategy.

[0163] In one embodiment, when the processor executes the computer program, the following are further implemented:

[0164] The machine learning model at least includes a DNN prediction model, an RNN prediction model, and a prediction model based on Gaussian regression.

[0165] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0166] Determine the target state trajectory of the nuclear reactor and the full state trajectory of the reactor;

[0167] Perform trajectory optimization on the full state trajectory of the reactor according to the target state trajectory to obtain the control action combination of the nuclear reactor; the control action combination is used to ensure that the deviation between the target state trajectory and the full state trajectory of the reactor reaches a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0168] Autonomously control the nuclear reactor based on the control action combination.

[0169] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0170] Determine the target state trajectory of the nuclear reactor;

[0171] Input the full state parameters of the nuclear reactor into a pre-constructed integrated prediction model for prediction to obtain the full state trajectory of the reactor. In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0172] Pre - construct the mathematical model and machine learning model of the nuclear reactor;

[0173] Through the integrated processing of the mathematical model and the machine learning model, an integrated prediction model is obtained.

[0174] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0175] Construct the mathematical model of the nuclear reactor; the mathematical model at least includes the core neutron point - reactor kinetics model, the thermal - hydraulic model, the reactivity feedback model, and the nuclide decay model;

[0176] Obtain the reactor operation samples of the nuclear reactor;

[0177] Train different types of machine learning models constructed according to the reactor operation samples to obtain the trained machine learning models.

[0178] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0179] Determine the reaction state and corresponding control actions in the full - state trajectory of the reactor;

[0180] Based on the dynamic evolution characteristics of the reactor, determine the first trajectory including the reactor state and control actions;

[0181] Perform linearization processing on the first trajectory through Taylor approximation to obtain the second trajectory under the optimal strategy;

[0182] Perform iterative processing on the second trajectory until it converges to the target state trajectory to obtain the control action combination of the nuclear reactor.

[0183] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0184] Update the next reactor state of the first trajectory according to the actual state - transfer function, and execute the step of performing linearization processing on the first trajectory through Taylor approximation to obtain the second trajectory under the optimal strategy.

[0185] In one embodiment, when the computer program is executed by a processor, the following is also implemented:

[0186] The machine learning model at least includes a DNN prediction model, an RNN prediction model, and a prediction model based on Gaussian regression.

[0187] In one embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0188] Determine the target state trajectory and the full-state trajectory of the nuclear reactor;

[0189] Optimize the full-state trajectory of the reactor according to the target state trajectory to obtain the control action combination of the nuclear reactor; the control action combination is used to ensure that the deviation between the target state trajectory and the full-state trajectory of the reactor reaches a preset value and satisfies the dynamic evolution characteristics of the reactor;

[0190] Autonomously control the nuclear reactor based on the control action combination.

[0191] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0192] Determine the target state trajectory of the nuclear reactor;

[0193] Input the full-state parameters of the nuclear reactor into a pre-constructed integrated prediction model for prediction to obtain the full-state trajectory of the reactor. In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0194] Pre-construct a mathematical model and a machine learning model of the nuclear reactor;

[0195] Obtain an integrated prediction model by performing integrated processing on the mathematical model and the machine learning model.

[0196] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0197] Construct a mathematical model of the nuclear reactor; the mathematical model at least includes a core neutron point kinetics model, a thermal-hydraulic model, a reactivity feedback model, and a nuclide decay model;

[0198] Obtain the reactor operation samples of the nuclear reactor;

[0199] Train different types of machine learning models constructed according to the reactor operation samples to obtain the trained machine learning models.

[0200] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0201] Determine the reaction state and the corresponding control actions in the full-state trajectory of the reactor;

[0202] Based on the dynamic evolution characteristics of the reactor, determine the first trajectory including the reactor state and the control actions;

[0203] Perform linearization processing on the first trajectory through Taylor approximation to obtain the second trajectory under the optimal strategy;

[0204] Iteratively process the second trajectory until it converges to the target state trajectory to obtain the control action combination of the nuclear reactor.

[0205] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0206] Update the next reactor state of the first trajectory according to the actual state transition function, and perform linearization processing on the first trajectory through Taylor approximation to obtain the second trajectory step under the optimal strategy.

[0207] In one embodiment, when the computer program is executed by a processor, the following is further implemented:

[0208] The machine learning model at least includes a DNN prediction model, an RNN prediction model, and a prediction model based on Gaussian regression.

[0209] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0210] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0211] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0212] The above-described embodiments merely represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. An autonomous control method for a nuclear reactor, characterized in that, The method includes: Determining a target state trajectory and a full reactor state trajectory of a nuclear reactor; Determining a reaction state of the nuclear reactor according to the full reactor state trajectory, and taking actions at a nominal operating point in the reaction state to obtain a first trajectory including reactor states and control actions; the nominal operating point is a fixed point assumed in advance for linear approximation; Performing linearization processing on the first trajectory through Taylor approximation to obtain a second trajectory under an optimal strategy; Performing iterative processing on the second trajectory until it converges to the target state trajectory to obtain a control action combination of the nuclear reactor; the control action combination is used to ensure that a deviation between the target state trajectory and the full reactor state trajectory reaches a preset value and satisfies the dynamic evolution characteristics of the reactor; Autonomously controlling the nuclear reactor based on the control action combination.

2. The method according to claim 1, wherein The determining a target state trajectory and a full reactor state trajectory of a nuclear reactor includes: Determining a target state trajectory of the nuclear reactor; Inputting full state parameters of the nuclear reactor into a pre-constructed integrated prediction model for prediction to obtain a full reactor state trajectory.

3. The method according to claim 2, wherein Before the determining a target state trajectory of the nuclear reactor, the method further includes: Pre-constructing a mathematical model and a machine learning model of the nuclear reactor; Obtaining the integrated prediction model by performing integrated processing on the mathematical model and the machine learning model.

4. The method according to claim 3, wherein The pre-constructing a mathematical model and a machine learning model of the nuclear reactor includes: Constructing a mathematical model of the nuclear reactor; the mathematical model at least includes a core neutron point kinetic model, a thermal-hydraulic model, a reactivity feedback model, and a nuclide decay model; Obtaining reactor operation samples of the nuclear reactor; Training different types of machine learning models constructed according to the reactor operation samples to obtain trained machine learning models.

5. The method according to claim 1, wherein Before the performing iterative processing on the second trajectory until it converges to the target state trajectory to obtain a control action combination of the nuclear reactor, the method further includes: Updating the next reactor state of the first trajectory according to an actual state transition function, and performing the step of performing linearization processing on the first trajectory through Taylor approximation to obtain a second trajectory under an optimal strategy.

6. The method according to claim 3, wherein The machine learning model at least includes a DNN prediction model, an RNN prediction model, and a prediction model based on Gaussian regression.

7. An autonomous control device for a nuclear reactor, characterized in that The device includes: A determining module, configured to determine a target state trajectory and a full reactor state trajectory of a nuclear reactor; A trajectory optimization module, configured to determine the reaction state and corresponding control actions in the full-state trajectory of the reactor; determine the reaction state of the nuclear reactor according to the full-state trajectory of the reactor, and take actions at the nominal operating point in the reaction state to obtain a first trajectory including the reactor state and control actions; the nominal operating point is a fixed point assumed in advance for linear approximation; perform linearization processing on the first trajectory through Taylor approximation to obtain a second trajectory under the optimal strategy; perform iterative processing on the second trajectory until it converges to the target state trajectory to obtain a control action combination of the nuclear reactor; the control action combination enables the deviation between the target state trajectory and the full-state trajectory of the reactor to reach a preset value and satisfies the dynamic evolution characteristics of the reactor; A control module, configured to perform autonomous control on the nuclear reactor based on the control action combination.

8. The device according to claim 7, wherein: The determination module is further configured to determine the target state trajectory of the nuclear reactor; The autonomous control device of the nuclear reactor further includes a prediction module, configured to input the full-state parameters of the nuclear reactor into a pre-constructed integrated prediction model for prediction to obtain the full-state trajectory of the reactor.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images