Method and System for Predicting and Controlling the Flow Rate of the Primary Loop Pump of a Pressurized Water Reactor

Abstract

This invention discloses a method and system for predicting and controlling the flow rate of the main pump in the primary loop of a pressurized water reactor (PWR). The method includes: acquiring the original main pump flow rate signal of the PWR primary loop; using the HOFEM module of NL-FCNN to perform multi-scale decomposition on the main pump flow rate signal to obtain multi-scale main pump flow rates; inputting the multi-scale main pump flow rate components into an ANSA dynamic optimization network structure for feature extraction to obtain multi-scale feature vectors; based on the extracted multi-scale feature vectors, using NL-FCNN to predict the main pump flow rate signal to obtain the optimal predicted main pump flow rate value; converting the optimal predicted main pump flow rate value into an electrical signal through a digital-to-analog converter circuit based on a D-ADC algorithm, and then inputting the electrical signal into the main pump frequency converter to drive the main pump to operate at variable frequency, thereby controlling the main pump flow rate of the PWR primary loop. This method and system can accurately control the main pump flow rate of the PWR primary loop.

Description

Technical Field

[0001] This invention belongs to the field of nuclear power engineering technology and relates to a method and system for predicting and controlling the flow rate of the primary loop main pump in a pressurized water reactor. Background Technology

[0002] The primary loop system of a pressurized water reactor nuclear power plant mainly consists of equipment such as the reactor pressure vessel, steam generator, main pump, and pressurizer. The main pump plays a crucial role in the primary loop system; its main function is to transport coolant from the reactor pressure vessel to the steam generator and back to the reactor, forming a closed-loop cycle. The main pump's functions in a nuclear power plant include: 1) Heat transfer: The main pump transfers the heat generated by the reactor to the secondary loop working fluid through the steam generator, producing steam to drive the turbine generator to generate electricity. 2) Pressure control: The operating status of the main pump directly affects the pressure and flow rate of the primary loop system. By adjusting the main pump's flow rate, the system pressure can be kept stable, preventing the leakage of radioactive materials.

[0003] The main pump is one of the key pieces of equipment in the primary loop system of a nuclear power plant. Its reliability and stability are crucial to the safe operation of the plant, and its flow control is paramount. In the future participation of pressurized water reactors (PWRs) in deep peak shaving and frequency regulation of the power grid, existing methods for predicting and controlling the flow of the primary loop main pumps are prone to large fluctuations and poor reliability, necessitating improvements to existing preventative measures. Regarding parameter prediction, existing PSO (Particle Swarm Optimization) algorithms suffer from drawbacks such as being prone to getting trapped in local optima and having fast convergence speeds leading to low convergence accuracy. RNNs (Recurrent Neural Networks) are prone to gradient vanishing and gradient exploding problems, making it difficult for the model to learn long-distance dependencies in sequences. Traditional fully connected neural networks (FCNNs) often rely on simple activation functions (such as ReLU and Sigmoid) and fixed network structures when dealing with complex nonlinear problems, limiting their performance when handling high-dimensional, nonlinear data. Furthermore, existing methods lack dynamic adaptability in parameter optimization and feature extraction, making it difficult to effectively capture high-order nonlinear relationships in the data. Simultaneously, in terms of actuator control, traditional DACs (Digital-to-Analog Converters) struggle to eliminate nonlinear errors caused by component mismatches, resulting in low conversion accuracy. It is necessary to develop more accurate methods for predicting and controlling the flow of the primary loop main pump, which will be beneficial for pressurized water reactor nuclear power plants to participate in the deep peak shaving and frequency regulation of the power grid in the future. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for predicting and controlling the flow rate of the main pump in the primary loop of a pressurized water reactor. This method and system can accurately control the flow rate of the main pump in the primary loop of the pressurized water reactor.

[0005] To achieve the above objectives, this invention discloses a method for predicting and controlling the flow rate of the primary loop main pump in a pressurized water reactor, comprising:

[0006] Obtain the original main pump flow signal of the pressurized water reactor primary loop;

[0007] The HOFEM module of NL-FCNN was used to perform multi-scale decomposition on the main pump flow signal to obtain multi-scale main pump flow components.

[0008] The multi-scale main pump flow components are input into the ANSA dynamic optimization network structure for feature extraction to obtain multi-scale feature vectors.

[0009] Based on the extracted multi-scale feature vectors, NL-FCNN is used to predict the main pump flow signal, and the optimal main pump flow prediction value is obtained.

[0010] The optimal main pump flow prediction value is converted into an electrical signal by a digital-to-analog converter circuit based on the D-ADC algorithm. The electrical signal is then input into the main pump frequency converter to drive the main pump to operate at a variable frequency and control the main pump flow of the pressurized water reactor primary loop.

[0011] The further improvement of the pressurized water reactor primary loop main pump flow prediction and control method described in this invention lies in:

[0012] Furthermore, after acquiring the raw main pump flow signal of the pressurized water reactor primary loop, the process also includes:

[0013] The original main pump flow signal is preprocessed to eliminate differences in dimensions and amplitude.

[0014] Furthermore, in the process of using the HOFEM module of NL-FCNN to perform multi-scale decomposition of the main pump flow signal to obtain multi-scale main pump flow components, DNAF is used to enhance the feature extraction capability of each scale main pump flow component.

[0015] Furthermore, it also includes: optimizing the ANSA dynamic optimization network structure based on the HOA algorithm.

[0016] Furthermore, this also includes optimizing NL-FCNN using RTLM and DSFA.

[0017] Furthermore, the electrical signal is:

[0018]

[0019] in, The analog voltage representing the main pump flow rate output at the nth sampling time. For conversion factors, The input value of the main pump flow rate digital signal at the nth sampling time.

[0020] This invention discloses a flow prediction and control system for the primary loop main pump of a pressurized water reactor, comprising:

[0021] The acquisition module is used to acquire the original main pump flow signal of the pressurized water reactor primary loop;

[0022] The decomposition module is used to perform multi-scale decomposition on the main pump flow signal using the HOFEM module of NL-FCNN to obtain multi-scale main pump flow components.

[0023] The extraction module is used to input the multi-scale main pump flow components into the ANSA dynamic optimization network structure for feature extraction to obtain multi-scale feature vectors;

[0024] The prediction module is used to predict the main pump flow signal based on the extracted multi-scale feature vectors using NL-FCNN, and obtain the optimal main pump flow prediction value.

[0025] The control module is used to convert the optimal main pump flow prediction value into an electrical signal through a digital-to-analog converter circuit based on the D-ADC algorithm, and then input the electrical signal into the main pump frequency converter to drive the main pump to operate at a variable frequency and control the main pump flow of the pressurized water reactor primary loop.

[0026] The further improvement of the pressurized water reactor primary loop main pump flow prediction and control system described in this invention is as follows:

[0027] Furthermore, after acquiring the raw main pump flow signal of the pressurized water reactor primary loop, the process also includes:

[0028] The original main pump flow signal is preprocessed to eliminate differences in dimensions and amplitude.

[0029] This invention discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the pressurized water reactor primary loop main pump flow prediction and control method.

[0030] The present invention discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the pressurized water reactor primary loop main pump flow prediction and control method.

[0031] The present invention has the following beneficial effects:

[0032] The pressurized water reactor primary loop main pump flow prediction and control method and system described in this invention employs a nonlinear fully connected neural network (NL-FCNN) and utilizes a dynamic nonlinear activation function (DNAF) to adaptively adjust the shape and parameters of the activation function according to the distribution characteristics of the input data. An adaptive network structure adjustment strategy (ANSA) based on gradient information feedback is designed to dynamically optimize the number of network layers and neurons. A high-order feature extraction module (HOFEM) is introduced to capture high-order nonlinear relationships in the data through multi-layer nonlinear transformations. A parameter optimization strategy based on a hybrid optimization algorithm (HOA) is proposed, combining gradient descent and global optimization algorithms for parameter updates. A real-time dynamic learning mechanism (RTLM) is developed to dynamically adjust the model parameters and structure according to changes in the input data. By introducing the dynamic nonlinear activation mechanism, adaptive network structure adjustment, and efficient parameter optimization strategy, the modeling capability of the main pump flow model for complex nonlinear data is significantly improved, greatly enhancing the accuracy of pressurized water reactor primary loop main pump flow prediction. Meanwhile, in response to the problems of traditional DAC (digital-to-analog converter) in eliminating nonlinear errors caused by component mismatch and low conversion accuracy in actuator control, this invention utilizes D-ADC (dynamic resolution adjustment-adaptive digital-to-analog conversion algorithm) digital-to-analog conversion technology, which can reduce nonlinear errors caused by component mismatch and effectively improve the control accuracy of the pressurized water reactor primary loop main pump frequency converter. Attached Figure Description

[0033] The accompanying drawings, which form part of this specification, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0034] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0035] 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.

[0036] 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.

[0037] 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.

[0038] 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.

[0039] 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.

[0040] 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)."

[0041] 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.

[0042] 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.

[0043] Example 1

[0044] refer to Figure 1 The method for predicting and controlling the flow rate of the primary loop main pump of a pressurized water reactor as described in this invention includes the following steps:

[0045] 1) Obtain the original main pump flow signal of the pressurized water reactor primary loop, and preprocess the original main pump flow signal to eliminate differences in dimensions and amplitude;

[0046] 2) Use the HOFEM module of NL-FCNN to perform multi-scale decomposition on the main pump flow signal processed in step 1); enhance the feature extraction capability of the main pump flow components at each scale through DNAF to obtain multi-scale main pump flow components.

[0047] 3) Input the multi-scale main pump flow components into the ANSA dynamic optimization network structure for feature extraction, and optimize the parameters of the main pump flow model based on the HOA algorithm to capture the nonlinear features in the main pump flow signal and obtain the multi-scale feature vector.

[0048] 4) Based on the extracted multi-scale feature vectors, NL-FCNN is used to predict the main pump flow signal to obtain the optimal main pump flow prediction value. The model parameters are dynamically adjusted by RTLM to adapt to the changes in the main pump flow signal, and error iterative correction is performed based on DSFA to obtain the optimal main pump flow prediction value.

[0049] 5) Based on the D-ADC algorithm, the optimal main pump flow prediction value is converted into an electrical signal through a digital-to-analog converter circuit, and then the electrical signal is input into the main pump frequency converter to drive the main pump to operate at a variable frequency, thereby controlling the main pump flow of the pressurized water reactor primary loop.

[0050] In step 1), the process of using the HOFEM module of NL-FCNN to perform multi-scale decomposition on the main pump flow signal processed in step 1) is as follows:

[0051] 11) Signal preprocessing;

[0052] 111) Standardized processing;

[0053] The original main pump flow signal x(t) is standardized to eliminate differences in dimensions and amplitude, enabling data to be modeled and analyzed on the same scale. This transforms the original main pump flow signal into a distribution with a mean of 0 and a standard deviation of 1, i.e.:

[0054]

[0055] in, The mean of the signal. The standard deviation of the signal. This is the standardized signal.

[0056] 112) Model training and validation;

[0057] Determine the proportion of the main pump flow training set : Usually take .

[0058] Calculate the length of the training set for the main pump flow rate. : .

[0059] The standardized main pump flow signal is divided into training and testing sets for model training and validation.

[0060] Training set:

[0061] Test set:

[0062] 12) Multiscale signal decomposition;

[0063] The preprocessed main pump flow signal is decomposed into multiple scales using the High-Order Feature Extraction Module (HOFEM) of NL-FCNN, namely:

[0064]

[0065] in, For the main pump flow signal component at the i-th scale, These are the residual components.

[0066] The feature extraction capability of the main pump flow signal components at each scale is enhanced by using a dynamic nonlinear activation function (DNAF), namely:

[0067]

[0068] Where: X represents the input data. This is a dynamic weight parameter, with a value range of [0,1], used to control... functions and The proportion of the function, These are non-linear adjustment parameters used for control. The degree of nonlinearity of the function.

[0069] Dynamic parameter adjustment, i.e.:

[0070]

[0071] Calculate the gradient:

[0072]

[0073] 13) Feature extraction and modeling;

[0074] 131) Feature extraction;

[0075] For each scale of the main pump flow signal component after multi-scale decomposition Perform feature extraction:

[0076] Dynamically optimize the network structure using the Adaptive Network Structure Adjustment (ANSA) strategy of NL-FCNN:

[0077]

[0078] in, This represents the current network layer number. The adjusted number of network layers. The learning rate controls the step size for adjusting the network structure. For loss function The gradient with respect to the number of network layers L.

[0079] The model parameters are optimized using a Hybrid Optimization Algorithm (HOA) to capture the nonlinear characteristics in the main pump flow signal, specifically:

[0080] 1311) Initialize parameters;

[0081] Initialize the model parameters, main pump flow signal, and genetic algorithm population.

[0082] 1312) Gradient descent optimization;

[0083] The main pump flow signal parameters are updated using the gradient descent method as follows:

[0084]

[0085] 1313) Genetic algorithm optimization, specifically:

[0086] Selection: Select individuals with higher fitness from the population.

[0087] Crossover: Generate new individuals through crossover operations.

[0088] Mutation: Introducing new genes through mutation operations.

[0089]

[0090] in, Genetic algorithm for main pump flow signal parameters The optimization results; This is a hybrid weight parameter used to control the ratio of gradient descent to genetic algorithm.

[0091] 1314) Repeat the above steps until the parameters converge or the maximum number of iterations is reached.

[0092] 132) Modeling;

[0093] Constructing a multi-scale main pump flow signal feature vector:

[0094]

[0095] in, fi For the first i Eigenvectors of each scale component.

[0096] 133) Signal prediction;

[0097] 1331) Based on the extracted multi-scale main pump flow feature vector F, signal prediction is performed using NL-FCNN:

[0098]

[0099] in, This is the predicted value of the signal.

[0100] 1332) Dynamic adjustment of model parameters;

[0101] The main pump flow model parameters are dynamically adjusted through a real-time dynamic learning mechanism (RTLM) to adapt to changes in the signal.

[0102] In step 4), the process of dynamically adjusting the model parameters using RTLM to adapt to changes in the main pump flow signal is as follows:

[0103] 21) Monitor signal changes in real time;

[0104] Monitor input main pump flow signal x t Calculate the change ||Δx t ||.

[0105] 22) Calculate the dynamic learning rate;

[0106] RTLM uses dynamic learning rate It adaptively adjusts based on changes in the input main pump flow signal, and its mathematical expression is:

[0107] η t=η0·exp(-γ·||Δx t ||)

[0108] in, η 0 is the initial learning rate. γ The decay coefficient controls the adjustment speed of the learning rate, ||Δx t || represents the change in the input signal, typically calculated as: ||Δx t ||=||x t -x t-1 ||

[0109] 23) Update model parameters;

[0110] RTLM dynamically updates the input main pump flow model parameters using the gradient descent method. θ t ,Right now:

[0111]

[0112] in, θt For the first t The main pump flow model parameters for the next iteration. η t The learning rate is dynamic. L( θ t ;x t ) is the loss function L with respect to parameters θ t The gradient, based on the current input main pump flow signal x t calculate.

[0113] The loss function L is typically chosen as mean squared error (MSE) or cross-entropy, i.e.:

[0114]

[0115] in, y i For the true value, f (x i ; θ t ) is the predicted value of the main pump flow model.

[0116] 24) Repeated real-time learning;

[0117] Repeat steps 21) to 23) to continuously adapt to changes in the main pump flow signal.

[0118] 25) Results evaluation and optimization;

[0119] 251) Calculate the dynamic error of the main pump flow rate. :

[0120]

[0121] 252) Error aggregation;

[0122] Averaging all errors in the main pump flow rate:

[0123]

[0124] Segmented aggregation:

[0125]

[0126] 253) Predicted value output;

[0127] Multiplying the average value obtained from the aggregated errors by 100% yields the final SMAPE value, expressed as a percentage of the model's predicted main pump flow error coefficients ΔY1, ΔY2, ..., ΔY. n Based on the error results, optimize the parameters and structure of the main pump flow model to obtain the optimal predicted value of the main pump flow rate with the minimum error. .

[0128] Step 5) involves converting the optimal main pump flow prediction value into an electrical signal using a digital-to-analog converter circuit based on the D-ADC algorithm.

[0129] 31) Dynamic resolution adjustment;

[0130] Frequency of detecting input main pump flow signal f k ;

[0131] The main pump flow resolution is dynamically adjusted based on frequency. N k :

[0132]

[0133] The main pump flow signal was requantized using the adjusted resolution. N k .

[0134] 32) The output electrical signal is:

[0135]

[0136] in, The analog voltage representing the main pump flow rate output at the nth sampling time. For conversion factors, The input value of the main pump flow rate digital signal at the nth sampling time.

[0137] Simulation Experiment

[0138] To verify the reliability of this invention, the experimental data (i.e., the original main pump flow rate) was obtained from a pressurized water reactor (PWR) with a load regulation range of 60%-100%, and the test period was 45 days. Data from the first 30 days was selected, and a total of 8282 data samples at 15-minute intervals were provided to train the prediction model. After the model's self-learning function was completed, data from the last 15 days were used to test the performance of the proposed model. Simultaneously, to compare the advancement of this invention, the Symmetric Mean Absolute Percentage Error (D-SMAPE) algorithm was used for evaluation, and traditional PSO, RNN, and FCNN were compared with this invention.

[0139] The D-SMAPE error evaluation index was used as the evaluation standard for each model, and the experimental results are shown in Table 1.

[0140] Table 1

[0141]

[0142] As shown in Table 1, this invention exhibits the smallest error coefficient compared to all other benchmark models. Compared to the PSO model, the error coefficients of RNN, FCNN, and this invention are all significantly reduced, demonstrating that the proposed model can improve prediction accuracy better than the PSO model. Furthermore, compared to the RNN and FCNN models, the error coefficients of this invention are smaller, proving that the algorithm model based on the NL-FCNN nonlinear fully connected neural network can improve the prediction accuracy of the primary loop pump flow rate of a pressurized water reactor.

[0143] Example 2

[0144] The pressurized water reactor primary loop main pump flow prediction and control system of the present invention includes:

[0145] The acquisition module is used to acquire the original main pump flow signal of the pressurized water reactor primary loop;

[0146] The decomposition module is used to perform multi-scale decomposition on the main pump flow signal using the HOFEM module of NL-FCNN to obtain multi-scale main pump flow components.

[0147] The extraction module is used to input the multi-scale main pump flow components into the ANSA dynamic optimization network structure for feature extraction to obtain multi-scale feature vectors;

[0148] The prediction module is used to predict the main pump flow signal based on the extracted multi-scale feature vectors using NL-FCNN, and obtain the optimal main pump flow prediction value.

[0149] The control module is used to convert the optimal main pump flow prediction value into an electrical signal through a digital-to-analog converter circuit based on the D-ADC algorithm, and then input the electrical signal into the main pump frequency converter to drive the main pump to operate at a variable frequency and control the main pump flow of the pressurized water reactor primary loop.

[0150] In this embodiment, after acquiring the original main pump flow signal of the pressurized water reactor primary loop, the method further includes:

[0151] The original main pump flow signal is preprocessed to eliminate differences in dimensions and amplitude.

[0152] The module division in this embodiment is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in each embodiment of this application can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0153] Example 3

[0154] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of a pressurized water reactor primary loop main pump flow prediction and control method. For example, the method includes: acquiring the original main pump flow signal of the pressurized water reactor primary loop; performing multi-scale decomposition on the main pump flow signal using the HOFEM module of NL-FCNN to obtain multi-scale main pump flow components; inputting the multi-scale main pump flow components into an ANSA dynamic optimization network structure for feature extraction to obtain multi-scale feature vectors; predicting the main pump flow signal using NL-FCNN based on the extracted multi-scale feature vectors to obtain an optimized main pump flow prediction value; converting the optimized main pump flow prediction value into an electrical signal using a digital-to-analog converter circuit based on a D-ADC algorithm, and then inputting the electrical signal into a main pump frequency converter to drive the main pump to operate at a variable frequency, thereby controlling the main pump flow of the pressurized water reactor primary loop. The memory may include main memory, such as high-speed random access memory (RAM), or non-volatile memory, such as at least one disk storage device. The processor, network interface, and memory are interconnected via an internal bus, which may be an industry-standard architecture bus, a peripheral component interconnection standard bus, or an extended industry-standard architecture bus. The bus can be categorized as an address bus, data bus, or control bus. The memory stores programs; specifically, the program may include program code, which includes computer operation instructions. The memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0155] Example 4

[0156] A computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of a pressurized water reactor primary loop main pump flow prediction and control method. For example, the method includes: acquiring the original main pump flow signal of the pressurized water reactor primary loop; performing multi-scale decomposition on the main pump flow signal using the HOFEM module of NL-FCNN to obtain multi-scale main pump flow components; inputting the multi-scale main pump flow components into an ANSA dynamic optimization network structure for feature extraction to obtain multi-scale feature vectors; predicting the main pump flow signal using NL-FCNN based on the extracted multi-scale feature vectors to obtain an optimized main pump flow prediction value; converting the optimized main pump flow prediction value into an electrical signal using a digital-to-analog converter circuit based on a D-ADC algorithm, and then inputting the electrical signal into a main pump frequency converter to drive the main pump to operate at variable frequency and control the main pump flow of the pressurized water reactor primary loop. Specifically, the computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. The volatile memory may include random access memory (RAM) and / or cache memory, etc. The non-volatile memory may include read-only memory (ROM), hard disk, flash memory, optical disk, magnetic disk, etc.

[0157] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0158] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), 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.

[0159] 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 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0160] 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.

[0161] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and disclosure of the invention. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the following claims.

[0162] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

[0163] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the present invention. Any simple modifications, alterations, or equivalent structural changes made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A pressurized water reactor primary pump flow rate prediction and control method, characterized by, The method comprises the following steps: obtaining the original primary pump flow signal of the pressurized water reactor primary loop; using a high-order feature extraction module (HOFEM) based on a nonlinear full connection neural network (NL-FCNN) to perform multi-scale decomposition on the primary pump flow signal to obtain multi-scale primary pump flow components; inputting the multi-scale primary pump flow components into a gradient information feedback based adaptive network structure adjustment strategy (ANSA) dynamic optimization network structure for feature extraction to obtain multi-scale feature vectors; based on the extracted multi-scale feature vectors, using a nonlinear full connection neural network (NL-FCNN) to predict the primary pump flow signal to obtain an optimized primary pump flow prediction value; based on a dynamic resolution adjustment-adaptive digital-to-analog conversion algorithm (D-ADC), converting the optimized primary pump flow prediction value into an electrical signal through a digital-to-analog conversion circuit, and then inputting the electrical signal into a primary pump frequency converter to drive the primary pump to run at a variable frequency, thereby controlling the primary pump flow of the pressurized water reactor primary loop.

2. The pressurized water reactor primary pump flow rate prediction and control method according to claim 1, characterized by, After obtaining the original primary pump flow signal of the pressurized water reactor primary loop, the method further comprises the following steps: preprocessing the original primary pump flow signal to eliminate dimensional and amplitude differences.

3. The pressurized water reactor primary pump flow rate prediction and control method of claim 1, wherein, In the process of using the high-order feature extraction module (HOFEM) based on the nonlinear full connection neural network (NL-FCNN) to perform multi-scale decomposition on the primary pump flow signal to obtain multi-scale primary pump flow components, a dynamic nonlinear activation function (DNAF) is used to enhance the feature extraction capability of each scale of the primary pump flow components.

4. The pressurized water reactor primary pump flow rate prediction and control method of claim 1, wherein, The method further comprises the following steps: optimizing the gradient information feedback based adaptive network structure adjustment strategy (ANSA) dynamic optimization network structure based on a hybrid optimization algorithm (HOA).

5. The pressurized water reactor primary pump flow prediction and control method of claim 1, wherein, The method further comprises the following steps:

6. The pressurized water reactor primary pump flow prediction and control method of claim 1, wherein, optimizing the nonlinear full connection neural network (NL-FCNN) using a real-time dynamic learning mechanism (RTLM) and a remote sensing image change detection algorithm (DSFA) based on deep networks and slow feature analysis. wherein, is the main pump flow analog voltage output at the nth sampling instant, is the conversion factor, is the main pump flow digital signal value input at the nth sampling instant.

7. A pressurized water reactor primary pump flow rate prediction and control system, characterized by, The electrical signal is represented as: The method comprises the following steps: an acquisition module for acquiring the original primary pump flow signal of the pressurized water reactor primary loop; a decomposition module for using a high-order feature extraction module (HOFEM) based on a nonlinear full connection neural network (NL-FCNN) to perform multi-scale decomposition on the primary pump flow signal to obtain multi-scale primary pump flow components; an extraction module for inputting the multi-scale primary pump flow components into a gradient information feedback based adaptive network structure adjustment strategy (ANSA) dynamic optimization network structure for feature extraction to obtain multi-scale feature vectors; a prediction module for using a nonlinear full connection neural network (NL-FCNN) to predict the primary pump flow signal based on the extracted multi-scale feature vectors to obtain an optimized primary pump flow prediction value; 8. The pressurized water reactor primary pump flow prediction and control system of claim 7, wherein, a control module for converting the optimized primary pump flow prediction value into an electrical signal through a digital-to-analog conversion circuit based on a dynamic resolution adjustment-adaptive digital-to-analog conversion algorithm (D-ADC), and then inputting the electrical signal into a primary pump frequency converter to drive the primary pump to run at a variable frequency, thereby controlling the primary pump flow of the pressurized water reactor primary loop. After acquiring the original primary pump flow signal of the pressurized water reactor primary loop, the method further comprises the following steps: The original main pump flow signal is pre-processed to eliminate dimension and amplitude difference.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The computer program is executed by the processor to realize the steps of the pressurized water reactor primary loop main pump flow prediction and control method in any one of claims 1-6.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to realize the steps of the pressurized water reactor primary loop main pump flow prediction and control method in any one of claims 1-6.

Images