PSO-BPNN-based intelligent parameter setting method and system for air-cooled micro-reactor

Abstract

The invention discloses a PSO-BPNN-based intelligent parameter setting method and system for an air-cooled micro reactor, and belongs to the technical field of nuclear reactor control. The method comprises the following steps: determining an initial PID parameter by adopting a frequency domain analysis method; an air-cooled micro-reactor state space model is obtained through linearization, and PID parameters are optimized and set off line by using a PSO algorithm; a BP neural network of a 3-3-3 structure is constructed, the weight is updated in combination with a gradient descent method, and a BPNN-PID controller is constructed; the controller is connected to a nonlinear air-cooled micro-reactor control system, and PID parameter real-time adjustment is realized based on an off-line optimal parameter reference and on-line self-learning. The method effectively deals with the strong nonlinearity of the air-cooled micro-reactor, shortens the electric power adjustment time by 48.4%, reduces the steady-state error by 80%, improves the control precision and response speed, and adapts to a complex operation scene.

Description

Technical Field

[0001] This invention belongs to the field of nuclear reactor control technology, specifically relating to an intelligent parameter tuning method and system for a gas-cooled microreactor based on PSO-BPNN. Background Technology

[0002] According to the OECD definition, nuclear reactors with a power output between 10MW and 300MW are small modular reactors (MMRs), while reactors with a power output below 10MW are called micro modular reactors (MMRs). MMRs can provide long-term and stable power, and can also provide steam heating. Due to their small size, they can be mobilely deployed, making them promising for applications in remote, harsh environments, coastal islands, or disaster reconstruction. With the deepening research on MMRs both domestically and internationally, China Nuclear Power Engineering Co., Ltd. has proposed a mobile nuclear power source (gas-cooled microreactor) based on a prism-shaped high-temperature gas-cooled microreactor. This system combines a direct Brayton cycle energy conversion system with a passive safety modular high-temperature gas-cooled reactor, possessing many of the advantages of micro modular reactors.

[0003] The graphite core in a gas-cooled microreactor (GRM) possesses high heat capacity and high-temperature resistance. The ceramic particle-dispersed fuel encapsulates fission products, and the unique fuel assembly design results in a lower power density compared to other reactor types, thus providing inherent safety. However, due to its high-temperature gas-cooled reactor integrated with a Brayton cycle structure, the GRM exhibits strong nonlinear characteristics, requiring appropriate parameter tuning for the PID control system to cope with external load variations.

[0004] This invention aims to solve this problem by proposing an intelligent parameter tuning method for gas-cooled microreactors based on particle swarm optimization algorithm-backpropagation neural network (PSO-BPNN). Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an intelligent parameter tuning method and system for gas-cooled microreactors based on PSO-BPNN, which addresses the shortcomings of the prior art. This method can change PID parameters in real time according to changes in operating conditions and model, and can also cope with the strong nonlinearity of gas-cooled microreactors to achieve better control results. This invention solves the technical problem that traditional PID parameter tuning methods cannot achieve good results due to their reliance on experience and weak adaptive ability.

[0006] The present invention adopts the following technical solution: A method for intelligent parameter tuning of a gas-cooled microreactor based on PSO-BPNN includes the following steps: S1. The initial PID parameters of the gas-cooled microreactor control system are determined using the frequency domain analysis method. S2. Obtain the state-space model of the gas-cooled microreactor through linearization, and optimize and tune the PID parameters offline using the PSO algorithm based on the state-space model of the gas-cooled microreactor. S3. Construct a BP neural network, output the parameters to be optimized for the PID controller through the self-learning function of the BP neural network, and update the weights of the BP neural network by combining the gradient descent method to construct a BPNN-PID controller. S4. Connect the BPNN-PID controller to the nonlinear gas-cooled microreactor control system. Based on the initial benchmark of the offline optimal PID parameters obtained in step S2 and the online self-learning of the BP neural network, realize the online adjustment of the PID parameters and complete the intelligent parameter tuning of the gas-cooled microreactor.

[0007] Preferably, in step S1, the time-domain performance indicators of the frequency domain analysis method include overshoot, settling time, and steady-state error. The phase margin calculated based on the overshoot is selected as the core performance indicator, and the initial PID parameters are determined by combining the crossover frequency and the damping ratio.

[0008] Preferably, in step S2, the operation flow of the PSO algorithm includes: S201. Initialize the parameters of the PSO algorithm, including population size, maximum number of iterations, inertia weight, acceleration constant, and upper and lower limits of particle velocity and position. S202. Determine the fitness function and the optimal solution criterion. The fitness function serves as the basis for particle update iteration in order to find the optimal solution for each individual. p Best and group optimal g Best ; S203. Update the particle's velocity and position, calculate the updated particle's fitness value, and determine the individual's optimal fitness value. p i The optimal fitness value of the particle swarm g i The corresponding individual optimal p Best and group optimal g Best ; S204. Determine if the termination condition has been met. If the fitness value is optimal or the maximum number of iterations has been reached, output the global optimal solution. g i and the corresponding group optimal g Best Otherwise, return to step S203 to continue the iteration.

[0009] Preferably, in step S202, the fitness function The integral of time multiplied by the absolute error is as follows:

[0010] in, e ( t ) represents systematic bias; t For time.

[0011] Preferably, in step S203, the particles i In the t The velocity and position update formulas for the +1st iteration are as follows:

[0012]

[0013] in, ω For inertial weights, , A random number between 0 and 1 For particles i No. t The speed of each iteration For particles i No. t The position of the next iteration. For particles i In the t The optimal position of the individual in the next iteration. For the particle swarm in the th t The optimal position of the population in the next iteration.

[0014] Preferably, in step S3, the BP neural network has a 3-3-3 structure, that is, the input layer has 3 neurons, the hidden layer has 3 neurons, and the output layer has 3 neurons; the three inputs of the input layer are system set values. r(t) Compared with the actual output value of the system y(t) deviation E(k) The three outputs of the output layer—the rate of change of deviation and the integral of deviation—correspond to the parameters of the PID controller. K p , K i , K d .

[0015] Preferably, the activation function of the neurons in the hidden layer of the BP neural network is the tanh function, and its calculation formula is as follows: .

[0016] Preferably, in step S3, the process of updating the weights of the BP neural network using the gradient descent method includes: setting a performance index function, calculating the updated weight values ​​of the output layer and the hidden layer, and compensating for the error caused by the replacement of unknown terms during the weight update process using the learning rate and the inertia coefficient.

[0017] Preferably, in step S4, the PID controller includes a PI controller and a PD controller; the BPNN-PID controller achieves online adjustment of PID parameters under load variation conditions by replacing the parameters of the original PI controller or PD controller in the nonlinear gas-cooled microreactor control system.

[0018] Secondly, embodiments of the present invention provide an intelligent parameter tuning system for a gas-cooled microreactor based on PSO-BPNN, comprising: The initial parameter determination module is used to determine the initial PID parameters for the gas-cooled microreactor control system using the frequency domain analytical method. The offline parameter optimization module is used to obtain the state-space model of the gas-cooled microreactor through linearization, and to optimize and tune the PID parameters offline using the PSO algorithm based on the state-space model of the gas-cooled microreactor, and output the offline optimal PID parameters. The neural network construction module is used to construct a BP neural network, output the parameters to be optimized for the PID controller through the self-learning function of the BP neural network, and update the weights of the BP neural network by combining the gradient descent method to construct a BPNN-PID controller. The online parameter tuning module is used to connect the BPNN-PID controller to the nonlinear gas-cooled microreactor control system. Based on the initial benchmark of the offline optimal PID parameters output by the offline parameter optimization module and the online self-learning of the BP neural network, the PID parameters are adjusted online. The system control module coordinates the working sequence of the initial parameter determination module, offline parameter optimization module, neural network construction module, and online parameter tuning module to ensure that the modules are connected consistently and to complete the intelligent parameter tuning and operation control of the gas-cooled microreactor.

[0019] Thirdly, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described intelligent parameter tuning method for gas-cooled microreactors based on PSO-BPNN.

[0020] Fourthly, embodiments of the present invention provide a computer-readable storage medium including a computer program, which, when executed by a processor, implements the steps of the above-described intelligent parameter tuning method for a gas-cooled microreactor based on PSO-BPNN.

[0021] Fifthly, a chip includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described intelligent parameter tuning method for gas-cooled microreactors based on PSO-BPNN.

[0022] In a sixth aspect, embodiments of the present invention provide an electronic device, including a computer program, which, when executed by the electronic device, implements the steps of the above-described intelligent parameter tuning method for gas-cooled microreactors based on PSO-BPNN.

[0023] Compared with the prior art, the present invention has at least the following beneficial effects: An intelligent parameter tuning method for gas-cooled microreactors based on PSO-BPNN combines offline optimization with online learning. Initial PID parameters are determined using frequency domain analytical methods, offline optimization is performed in the state-space model using the PSO algorithm, and online self-tuning is achieved through a BP neural network. This method effectively solves the control challenges of highly nonlinear gas-cooled microreactor systems, significantly improving the system's response speed and control accuracy under load changes, and overcoming the shortcomings of traditional PID parameter tuning, which relies on experience and has poor adaptive capabilities.

[0024] Furthermore, the performance indicators for the frequency domain analytical method are limited, combining time-domain indicators such as overshoot, settling time, and steady-state error with phase margin, and accurately calculating the initial PID parameters through crossover frequency and damping ratio. This method improves the scientific rigor and reliability of parameter initialization, lays a solid foundation for subsequent optimization, avoids blind trial and error, and shortens the tuning time.

[0025] Furthermore, by clearly defining parameter initialization, fitness function setting, particle update iteration, and termination condition judgment in steps, a closed-loop optimization logic is formed to ensure the stability and operability of the algorithm. Explicitly setting parameters such as population size and maximum number of iterations avoids the algorithm from getting stuck in infinite iteration or premature convergence; the dual selection mechanism of individual optimality and swarm optimality ensures the global optimality of the optimization results and effectively avoids local optima. It fully leverages the simple structure and high optimization efficiency of the PSO algorithm, and for the linearized state-space model of the gas-cooled microreactor, it can quickly select suitable PID parameters, providing a high-quality benchmark for subsequent online adjustments. At the same time, the modular design of the process facilitates engineering implementation and parameter debugging.

[0026] Furthermore, the ITAE function comprehensively considers the system's response speed and steady-state accuracy by integrating time and absolute error. Compared to single indicators such as overshoot and steady-state error, it can more comprehensively reflect the control performance of the gas-cooled microreactor. Addressing the dual requirements of rapid response to load changes and reduced steady-state deviation in gas-cooled microreactors, this function effectively suppresses the one-sidedness of prioritizing response speed over accuracy or vice versa during parameter optimization. This ensures that the PID parameters obtained through offline optimization can both shorten the settling time and reduce steady-state error. Simulation data shows that ITAE-based optimization reduces the steady-state error of electrical power by 80%, fully demonstrating its advantages in improving parameter adaptability and laying a high-precision foundation for subsequent online adjustments.

[0027] Furthermore, by introducing inertial weights and random numbers, the global exploration and local exploitation capabilities of the particles are balanced: inertial weights ensure that particles inherit historical velocities, avoiding blind optimization; random numbers in the 0-1 range increase the diversity of optimization and reduce the probability of getting trapped in local optima. Compared to the PSO algorithm with fixed update rules, this formula allows particles to both follow the direction of the optimal solution and explore new parameter spaces during iteration. For the complex characteristics of the gas-cooled microreactor state-space model, it can quickly converge to the globally optimal PID parameters. The quantitative expression of the update formula standardizes the algorithm implementation process, facilitating programming implementation and parameter fine-tuning, while ensuring the consistency of optimization results under different operating conditions, thus improving the engineering practicality of the technical solution.

[0028] Furthermore, the input layer selects deviation, deviation change rate, and deviation integral to comprehensively reflect the dynamic error characteristics of the system; the output layer directly corresponds to the Kp, Ki, and Kd parameters, realizing direct optimization of control parameters and avoiding complex mapping transformations. The simplified 3-3-3 structure reduces network computation while maintaining fitting accuracy, meeting the real-time requirements of online control of the gas-cooled microreactor. Compared to deep neural networks, it effectively reduces weight update time, ensuring rapid output of optimized parameters when operating conditions change. Precise matching of input and output makes the neural network's self-learning more targeted, quickly capturing the correlation between error and PID parameters, improving the sensitivity of online parameter adjustment, and adapting to the strong nonlinear disturbances of the gas-cooled microreactor.

[0029] Furthermore, the tanh function has an output range of [-1, 1]. Compared to the sigmoid function, it exhibits a less pronounced gradient vanishing problem, effectively transmitting gradient information from deep networks and improving weight update efficiency. For the dynamic characteristics of the air-cooled microreactor error signal, the tanh function demonstrates stronger nonlinear mapping capabilities, accurately fitting the complex nonlinear relationship between the error and PID parameters, making the Kp, Ki, and Kd parameters output by the neural network more closely match actual control requirements. Simultaneously, the symmetry and smoothness of this function make the weight update process more stable, avoiding oscillations during training, shortening the convergence time of the neural network, ensuring rapid response to changes in operating conditions during online adjustments, and improving the dynamic performance of the control closed loop.

[0030] Furthermore, by setting a performance index function, the deviation between the network output and the ideal parameters is quantified to guide the direction of weight updates. The introduction of the learning rate and inertia coefficient ensures the rationality of the weight update step size and compensates for the error caused by the replacement of unknown terms, avoiding divergence or slow convergence during training. Compared with the simple gradient descent method, the error compensation mechanism improves the accuracy of weight updates, enabling the neural network to quickly approach the optimal weight configuration, and the output PID parameters are more suitable for the nonlinear characteristics of the gas-cooled microreactor. The quantization process of weight updates makes the training process adjustable, facilitating the adjustment of the learning rate and inertia coefficient according to the actual control effect, enhancing the flexibility and adaptability of the technical solution.

[0031] Furthermore, it explicitly supports PI and PD controller types, breaking the application limitations of a single PID controller and adapting to the design requirements of different control systems in gas-cooled microreactors (e.g., using a PI controller for temperature control and a PD controller for speed control). By replacing the original parameters with a BPNN-PID controller, there is no need to reconstruct the control hardware, reducing the difficulty and cost of engineering modifications and facilitating its widespread application in existing gas-cooled microreactor systems. Simultaneously, the online parameter adjustment mechanism for load-changing conditions ensures low overshoot, fast response, and high-precision control effects under different controller types. Simulation data shows that after replacing the PI controller parameters, the reactor outlet helium temperature settling time is reduced by 39.9%, fully demonstrating its adaptability and effectiveness among various PID controllers.

[0032] It is understood that the beneficial effects of the second to sixth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0033] In summary, this invention combines offline PSO optimization with online BPNN learning to achieve intelligent tuning of PID parameters for air-cooled microreactors, significantly improving the dynamic performance and adaptability of the control system and solving the problem of poor performance of traditional methods in strongly nonlinear systems.

[0034] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0035] Figure 1 This is a flowchart of the intelligent parameter tuning process for the PSO-BPNN of the present invention; Figure 2 This describes the parameter tuning process for a PID controller based on the PSO algorithm. Figure 3 This is a diagram of the BP neural network structure. Figure 4 This is a BPNN-based online parameter tuning structure; Figure 5 Flowchart for BP-PID parameter optimization; Figure 6 The response diagram of the gas-cooled microreactor system is shown for a step drop from 100%FP to 90%FP load, where (a) is the rotational speed, (b) is the reactor outlet helium temperature, (c) is the reactor power, (d) is the electrical power, (e) is the charging flow rate, (f) is the exhaust flow rate, (g) is the bypass flow rate, (h) is the turbine inlet helium pressure, (i) is the precooler outlet helium temperature, and (j) is the intercooler outlet helium temperature. Figure 7 A schematic diagram of a computer device provided in an embodiment of the present invention; Figure 8 This is a block diagram of a chip provided according to an embodiment of the present invention.

[0036] Among them, 60. Computer equipment; 61. Processor; 62. Memory; 63. Computer program; 600. Electronic device; 610. Processing unit; 620. Storage unit; 6201. Random access memory unit; 6202. Cache memory unit; 6203. Read-only memory unit; 6204. Program / utility; 6205. Program module; 630. Bus; 640. Display unit; 650. Input / output interface; 660. Network adapter; 700. External device. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0039] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0040] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this invention generally indicates that the preceding and following objects have an "or" relationship.

[0041] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0042] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0043] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0044] This invention provides an intelligent parameter tuning method for gas-cooled microreactors based on PSO-BPNN. The parameter tuning is divided into offline and online parts. First, the PSO algorithm is used for offline tuning to obtain PID controller parameters that meet the control requirements. During the operation of the gas-cooled microreactor, BPNN is used to adjust the PID controller parameters online to ensure that the control system can achieve timely and effective control during load changes.

[0045] Please see Figure 1 The present invention discloses an intelligent parameter tuning method for a gas-cooled microreactor based on PSO-BPNN, comprising the following steps: S1. The initial PID parameters of the gas-cooled microreactor control system are determined using the frequency domain analysis method, laying the foundation for the optimization of initial parameters in the subsequent step S2. The time-domain performance indicators in the frequency domain analytical method include overshoot (OS), settling time, and steady-state error. The phase margin (PM) obtained after calculating OS is selected as the performance indicator.

[0046] (1) (2) in, t To adjust the time; ω For crossing frequency; ζ The damping ratio is denoted as .

[0047] In the known OS and t When the target value is reached, the corresponding PM and ω Thus, the controller parameters are calculated; when the OS and t When the target value is unknown, the corresponding time-domain indicators, such as OS and error square integral, can be obtained through iterative simulation calculations, and the control system parameters can be further determined in combination with the control requirements.

[0048] S2. The state-space model of the gas-cooled microreactor is obtained by linearization, and the PID parameters are optimized and tuned offline using the PSO algorithm based on the state-space model of the gas-cooled microreactor. The PSO algorithm is used to further optimize PID parameters due to its simple structure, high optimization efficiency, and high accuracy.

[0049] The specific PSO algorithm execution flow is as follows: S201. Initialize parameters: Set the parameters required for the PSO algorithm, set the population size (i.e., the number of particles), set the maximum number of iterations to prevent infinite iterations, and determine the inertia weights. 、 The magnitudes of parameters such as the acceleration constant are determined, and the upper and lower limits of the particle's velocity and position are clearly defined. S202. Determine the fitness function and the optimal fitness value used in the PSO algorithm, and use these as particles for iterative updates. Find the optimal individual fitness value by calculating the fitness value. p Best and group optimal g Best In PID controller parameter tuning, the commonly used fitness functions include overshoot, settling time, steady-state error, and time-integral-absolute-error (ITAE). S203. Perform particle update iterations, updating the particle's velocity and position using equations (7) and (8). The updated particle's fitness value is calculated in the fitness function, and the optimal individual fitness value of the particle is obtained by comparison. p i The optimal fitness value of the particle swarm g i ,at this time p i and g i The corresponding position is the individual optimal position of the particle. p Best and group optimal g Best ; S204. Determine if the termination condition has been met. If the particle obtains the global optimal solution after updating its velocity and position... g i The algorithm terminates when the fitness value is found to be optimal, or when the update iteration reaches the maximum number of iterations, and outputs the global optimum. g i and g Best If neither of these conditions is met, return to step S203 to continue updating.

[0050] Assume each particle is in D The particles move in 1-dimensional space, and the swarm size is N ,particle i ( i =1,2,3,…, N The information generated during PSO operation includes location information. x i ,speed v i Individual optimal solution p i and the group optimal solution g i The following are respectively: (3) (4) (5) (6) particle i In the t The update formula for +1 iterations is: (7) (8) in, ω The inertial weight represents the degree to which a particle inherits its velocity from the previous iteration; r 1. r 2 is a random number between 0 and 1.

[0051] This method selects ITAE as the fitness function for the PSO algorithm, and its calculation method is as follows: (9) in, e ( t ) represents systematic bias; t Time (in seconds).

[0052] S3. Construct a BP neural network; Combining the parallel computing and adaptive learning capabilities of BP neural networks, which theoretically can approximate nonlinear functions with arbitrary precision, this network is selected for online parameter optimization. The network is constructed as follows: Figure 3 The BP neural network structure.

[0053] A backpropagation (BP) neural network typically consists of three layers: an input layer, hidden layers, and an output layer. The input layer is the interface between the neural network and the external world; it usually doesn't process information, and the number of neurons depends on the size of the input data. The hidden layers are the most important part of the neural network, enabling it to fit any function. The output layer outputs the processed data, and the number of neurons is usually determined by the dimension of the output data.

[0054] After constructing the BP neural network in step S2, it is necessary to determine the number of input and output layers and the parameters of the neural network. Let... Figure 3 In this diagram, M is the input layer, I is the hidden layer, and J is the output layer. Therefore, we set M=3, I=3, and J=3 to construct a 3-3-3 BP neural network for online optimization of PID parameters. The three inputs are the system setpoints. r(t) Actual output value of the system y(t) The deviation between the two E(k) Through the self-learning of the neural network, the optimal control parameters are output. K p , K i , Kd The BP neural network is then combined with the PID controller, and the specific parameter tuning formula is as follows: Let the first m The input neuron, the first i The hidden layer neuron, the first j Each output neuron is used with x m , x m , k i , y j express, x m and k i The weights between ω mi , k i and y j Weights between ω ij Therefore, the input to the M layer of the BP neural network is: (10) in, m =1, 2, 3.

[0055] The inputs and outputs of hidden layer I of the network are: (11) (12) Where is the weighting coefficient of the hidden layer; superscripts (1), (2), and (3) represent the input layer, hidden layer, and output layer, respectively. The activation function of the neurons in hidden layer I is taken as... tanh function: (13) The input and output of network output layer J are:

[0056] (14)

[0057]

[0058]

[0059] The three tuning parameters of the PID controller correspond to the three outputs of the output layer of the BP neural network.

[0060] The performance index function is: (15) The weights of a backpropagation (BP) neural network are iteratively modified using gradient descent. The weight update method for the neural network is as follows: (16) (17) in, Update the output layer weights. This is the update value for the hidden layer weights. For learning rate, The inertia coefficient.

[0061] by For example, then (18) Among them, due to Unknown, therefore symbolic functions can be used. sgn Instead of formula (10), that is The resulting error is due to the learning rate. η Compensation will be provided.

[0062] (19) The above method combines a BP neural network with a PID controller to construct a BPNN-PID controller.

[0063] S4. The BPNN-PID controller constructed in step S3 is then incorporated into the nonlinear gas-cooled microreactor control system.

[0064] Combining the BPNN-PID parameter optimization principle in step S3, this algorithm is integrated with the gas-cooled microreactor control system to construct, as follows: Figure 4 The structure shown.

[0065] The gas-cooled microreactor control system uses a single PI controller for control. This method is also applicable to all types of PID controllers, including PI and PD controllers. Therefore, the diagram shows the PI controller as the target, and the specific BPNN-PID controller parameter change flow is as follows: Figure 5 .

[0066] Finally, by combining the optimized parameters of S3 with the original PI parameters, a better effect was obtained compared to the original PI controller.

[0067] In another embodiment of the present invention, an intelligent parameter tuning system for a gas-cooled microreactor based on PSO-BPNN is provided. This system can be used to implement the above-mentioned intelligent parameter tuning method for a gas-cooled microreactor based on PSO-BPNN. Specifically, the intelligent parameter tuning system for a gas-cooled microreactor based on PSO-BPNN includes an initial parameter determination module, an offline parameter optimization module, a neural network construction module, an online parameter tuning module, and a system control module.

[0068] Among them, the initial parameter determination module is used to determine the initial PID parameters for the gas-cooled microreactor control system using the frequency domain analysis method. The offline parameter optimization module is used to obtain the state-space model of the gas-cooled microreactor through linearization, and to optimize and tune the PID parameters offline using the PSO algorithm based on the state-space model of the gas-cooled microreactor, and output the offline optimal PID parameters. The neural network construction module is used to construct a BP neural network, output the parameters to be optimized for the PID controller through the self-learning function of the BP neural network, and update the weights of the BP neural network by combining the gradient descent method to construct a BPNN-PID controller. The online parameter tuning module is used to connect the BPNN-PID controller to the nonlinear gas-cooled microreactor control system. Based on the initial benchmark of the offline optimal PID parameters output by the offline parameter optimization module and the online self-learning of the BP neural network, the PID parameters are adjusted online. The system control module coordinates the working sequence of the initial parameter determination module, offline parameter optimization module, neural network construction module, and online parameter tuning module to ensure that the modules are connected consistently and to complete the intelligent parameter tuning and operation control of the gas-cooled microreactor.

[0069] This invention provides a terminal device comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, graphics processing units (GPUs), tensor processing units (TPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or function. The processor described in this embodiment can be used for the operation of an intelligent parameter tuning method for a gas-cooled microreactor based on PSO-BPNN, including: Initial PID parameters for the gas-cooled microreactor control system are determined using the frequency domain analytical method. A state-space model of the gas-cooled microreactor is obtained through linearization, and the PID parameters are offline optimized and tuned based on this model using the PSO algorithm. A BP neural network is constructed, and its self-learning function outputs the parameters to be optimized for the PID controller. The weights of the BP neural network are updated using gradient descent, thus constructing a BPNN-PID controller. This BPNN-PID controller is then connected to the nonlinear gas-cooled microreactor control system. Based on the initial baseline of the offline optimal PID parameters obtained in step S2 and the online self-learning of the BP neural network, online adjustment of the PID parameters is achieved, completing the intelligent parameter tuning of the gas-cooled microreactor.

[0070] Please see Figure 7 The terminal device is a computer device. In this embodiment, the computer device 60 includes a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When executed by the processor 61, the computer program 63 implements the intelligent parameter tuning method for the gas-cooled microreactor based on PSO-BPNN in this embodiment. To avoid repetition, these details are not elaborated here. Alternatively, when executed by the processor 61, the computer program 63 implements the functions of each model / unit in the intelligent parameter tuning system for the gas-cooled microreactor based on PSO-BPNN in this embodiment. To avoid repetition, these details are not elaborated here.

[0071] Computer device 60 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. Computer device 60 may include, but is not limited to, a processor 61 and a memory 62. Those skilled in the art will understand that... Figure 7 This is merely an example of computer device 60 and does not constitute a limitation on computer device 60. It may include more or fewer components than shown, or combine certain components, or different components. For example, computer device may also include input / output devices, network access devices, buses, etc.

[0072] The processor 61 may be a Central Processing Unit (CPU), or other general-purpose processors, graphics processing units (GPUs), tensor processing units (TPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0073] The memory 62 can be an internal storage unit of the computer device 60, such as a hard disk or RAM of the computer device 60. The memory 62 can also be an external storage device of the computer device 60, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the computer device 60.

[0074] Furthermore, the memory 62 may include both internal storage units of the computer device 60 and external storage devices. The memory 62 is used to store computer programs and other programs and data required by the computer device. The memory 62 can also be used to temporarily store data that has been output or will be output.

[0075] Please see Figure 8The terminal device is an electronic device 600, which is manifested in the form of a general-purpose computing device. The components of the electronic device may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including storage unit 620 and processing unit 610), a display unit 640, etc.

[0076] The storage unit stores program code, which can be executed by the processing unit 610 to perform the steps described in the method section of this specification according to various exemplary embodiments of the present invention. For example, the processing unit 610 can perform actions such as... Figure 1 The steps are shown in the figure.

[0077] Storage unit 620 may include a readable medium in the form of a volatile storage unit, such as random access memory (RAM) 6201 and / or cache memory 6202, and may further include a read-only memory (ROM) 6203.

[0078] Storage unit 620 may also include a program / utility 6204 having a set (at least one) program module 6205, such program module 6205 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0079] Bus 630 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the multiple bus structures.

[0080] Electronic device 600 can also communicate with one or more external devices 700 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 600, and / or with any device that enables electronic device 600 to communicate with one or more other computing devices (e.g., router, modem). This communication can be performed via input / output interface 650. Furthermore, electronic device 600 can also communicate with one or more networks (e.g., local area network, wide area network, and / or public network, such as the Internet) via network adapter 660. Network adapter 660 can communicate with other modules of electronic device 600 via bus 630. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0081] Example 4 This invention also provides a storage medium, specifically a computer-readable storage medium, which is a memory device in a terminal device for storing programs and data. It is understood that the computer-readable storage medium here can include both built-in storage media in the terminal device and extended storage media supported by the terminal device; it can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor, which can be one or more computer programs (including program code). More specific examples of the computer-readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical fiber, portable compact disk read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0082] Computer-readable storage media also include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable storage medium can also be any readable medium other than a readable storage medium that can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium can be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, radio frequency, etc., or any suitable combination thereof.

[0083] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0084] One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the intelligent parameter tuning method for the PSO-BPNN-based gas-cooled microreactor in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor to perform the following steps: Initial PID parameters for the gas-cooled microreactor control system are determined using the frequency domain analytical method. A state-space model of the gas-cooled microreactor is obtained through linearization, and the PID parameters are offline optimized and tuned based on this model using the PSO algorithm. A BP neural network is constructed, and its self-learning function outputs the parameters to be optimized for the PID controller. The weights of the BP neural network are updated using gradient descent, thus constructing a BPNN-PID controller. This BPNN-PID controller is then connected to the nonlinear gas-cooled microreactor control system. Based on the initial baseline of the offline optimal PID parameters obtained in step S2 and the online self-learning of the BP neural network, online adjustment of the PID parameters is achieved, completing the intelligent parameter tuning of the gas-cooled microreactor.

[0085] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0086] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0087] To demonstrate the advantages of this method over traditional PID control methods for gas-cooled microreactors, a transient dynamic simulation analysis was conducted under 100% power operation conditions, with the load step reduced by 10% FP. The simulation lasted 500 seconds of steady-state operation, followed by a 5000-second simulation of the load stepping from 100% FP to 90% FP. A comparison of the system dynamic responses based on frequency domain analysis and PSO-BPNN-based intelligent parameter tuning for gas-cooled microreactors is provided. Figure 6 As shown in the figure, the results are compared in Table 1.

[0088] Table 1 Performance indicators of gas-cooled microreactor system from 100% FP to 90% FP

[0089] As shown in Table 1, the intelligent parameter tuning method using PSO-BPNN can effectively optimize various performance indicators. Compared with the frequency domain analytical method, the power adjustment time is shortened by about 48.4%, the power steady-state error is reduced by about 80%, the power overshoot is reduced by about 63.2%, the reactor power steady-state error is reduced by about 38.9%, the reactor outlet helium temperature adjustment time is shortened by about 39.9%, the maximum deviation of the reactor outlet helium temperature is reduced by 16.3%, the precooler outlet helium temperature adjustment time is shortened by about 44.7%, the maximum deviation of the precooler outlet helium temperature is reduced by 10.2%, the intercooler outlet helium temperature adjustment time is shortened by about 43.2%, and the maximum deviation of the precooler outlet helium temperature is reduced by 8.3%.

[0090] In summary, this invention presents an intelligent parameter tuning method and system for gas-cooled microreactors based on PSO-BPNN. Through an optimization strategy combining offline and online methods, it significantly improves the performance of the control system. Specifically, the power regulation time is reduced by approximately 48.4%, the steady-state error by approximately 80%, and the overshoot by approximately 63.2%; the reactor power steady-state error is reduced by 38.9%, the outlet helium temperature regulation time is reduced by 39.9%, and the maximum deviation is reduced by 16.3%; the temperature regulation times of the precooler and intercooler are reduced by 44.7% and 43.2%, respectively, and the maximum deviations are reduced by 10.2% and 8.3%. This effectively overcomes the control challenges posed by the strong nonlinear characteristics of gas-cooled microreactors, achieves adaptive tuning of PID parameters, and greatly improves the system's response speed and control accuracy under load changes, providing reliable technical support for the safe and stable operation of micro modular reactors.

[0091] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0092] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0093] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0094] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0095] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0096] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0097] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random-access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0098] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0099] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0100] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0101] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A PSO-BPNN-based intelligent parameter setting method for air-cooled microreactors, characterized in that, The method comprises the following steps: S1, using a frequency domain analysis method to determine initial PID parameters for a gas-cooled micro reactor control system; S2, obtaining a gas-cooled micro reactor state space model through linearization, and using a PSO algorithm to optimize and set PID parameters offline based on the gas-cooled micro reactor state space model; S3, constructing a BP neural network, outputting parameters to be optimized of a PID controller through a self-learning function of the BP neural network, updating weights of the BP neural network in combination with a gradient descent method, and constructing a BPNN-PID controller; S4, connecting the BPNN-PID controller to a nonlinear gas-cooled micro reactor control system, realizing online adjustment of PID parameters based on initial criteria of offline optimal PID parameters obtained in step S2 and online self-learning of the BP neural network, and completing intelligent parameter setting of the gas-cooled micro reactor.

2. The PSO-BPNN-based intelligent parameter setting method for air-cooled microreactors according to claim 1, characterized in that, In step S1, time domain performance indicators of the frequency domain analysis method include overshoot, regulation time, and steady-state error, a phase margin calculated based on the overshoot is selected as a core performance indicator, and the initial PID parameters are determined in combination with a crossover frequency and a damping ratio.

3. The PSO-BPNN-based intelligent parameter setting method for air-cooled microreactors according to claim 1, characterized in that, In step S2, an operation process of the PSO algorithm comprises: S201, initializing parameters of the PSO algorithm, including population size, maximum iteration number, inertia weight, acceleration constant, upper and lower limits of particle velocity and position; S202, determining a fitness function and an optimal solution criterion, the fitness function being used as a basis for particle update iteration to find an individual optimal solution p Best and a population optimal solution g Best ; S203, updating the speed and position of the particle, calculating the fitness value of the updated particle, determining the individual optimal fitness value p i and the group optimal fitness value of the particle group g i respectively corresponding individual optimal p Best and group optimal g Best ; S204, judging whether a termination condition is reached, if the fitness value optimal solution or the maximum iteration number is reached, outputting a global optimal solution g i and the corresponding population optimal g Best , otherwise returning to step S203 for continuing iteration.

4. The PSO-BPNN-based intelligent parameter setting method for air-cooled microreactors according to claim 3, characterized in that, In step S202, the fitness function is the time multiplied by the absolute error integral, in particular: wherein e t is the system bias; t is time.​ 5. The PSO-BPNN-based intelligent parameter setting method for air-cooled microreactors according to claim 3, characterized in that, In step S203, the particles i In the first t The velocity and position update equations at the first iteration are in, ω For inertial weights, , A random number between 0 and 1 For particles i No. t The speed of each iteration For particles i No. t The position of the next iteration. For particles i In the t The optimal position of the individual in the next iteration. For the particle swarm in the th t The optimal position of the population in the next iteration.

6. The PSO-BPNN based intelligent parameter setting method for air-cooled microreactor according to claim 1, characterized in that, In step S3, the structure of the BP neural network is 3-3-3, that is, 3 neurons are set in the input layer, 3 neurons are set in the hidden layer, and 3 neurons are set in the output layer; the 3 inputs of the input layer are respectively the system set value r(t) , the deviation y(t) , the deviation change rate, and the deviation integral of the actual output value of the system from the set value, and the 3 outputs of the output layer correspond to the parameters of the PID controller E(k) K p , K i , K d .​ 7. The PSO-BPNN-based intelligent parameter setting method for air-cooled microreactors according to claim 6, characterized in that, A neuron activation function of a hidden layer of the BP neural network is a tanh function, and a calculation formula thereof is: 。 8. The PSO-BPNN based intelligent parameter setting method for air-cooled microreactor according to claim 1, characterized in that, In step S3, a process of updating weights of the BP neural network by the gradient descent method comprises: setting a performance indicator function, calculating output layer weight update values and hidden layer weight update values, and compensating errors generated in a weight update process due to replacement of unknown items through a learning rate and an inertia coefficient.

9. The PSO-BPNN based intelligent parameter setting method for air-cooled microreactor according to claim 1, characterized in that, In step S4, the PID controller comprises a PI controller and a PD controller; and the BPNN-PID controller realizes online adjustment of PID parameters under a load change condition by replacing parameters of the original PI controller or the PD controller in the nonlinear gas-cooled micro reactor control system.

10. A PSO-BPNN-based intelligent parameter setting system for air-cooled microreactors, characterized in that, The method comprises: An initial parameter determination module configured to determine initial PID parameters for a gas-cooled micro reactor control system by using a frequency domain analysis method; An offline parameter optimization module configured to obtain a gas-cooled micro reactor state space model through linearization, and to optimize and set PID parameters offline based on the gas-cooled micro reactor state space model by using a PSO algorithm, and to output offline optimal PID parameters; A neural network construction module configured to construct a BP neural network, to output parameters to be optimized of a PID controller through a self-learning function of the BP neural network, to update weights of the BP neural network in combination with a gradient descent method, and to construct a BPNN-PID controller; An online parameter setting module configured to connect the BPNN-PID controller to a nonlinear gas-cooled micro reactor control system, to realize online adjustment of PID parameters based on initial criteria of offline optimal PID parameters output by the offline parameter optimization module and online self-learning of the BP neural network, and to complete intelligent parameter setting of the gas-cooled micro reactor. The system control module is used for coordinating the working time sequence of the initial parameter determination module, the offline parameter optimization module, the neural network construction module and the online parameter setting module, ensuring the consistency of the modules, and completing the intelligent parameter setting and operation control of the air-cooled micro reactor.

Images