A method for controlling the position of a reactor control rod and related apparatus

Abstract

This invention discloses a method and related apparatus for controlling the positions of reactor control rods, comprising: acquiring the original position sequence of the reactor control rods; decomposing the original position sequence to obtain M IMF components; and inputting the M IMF components into a trained Long Short-Term Memory (LSTM) network model to obtain predicted values ​​X1, X2, ..., X... of the M IMF components. M Based on the predicted values ​​X1, X2, ..., X of the M IMF components M An error evaluation index, D-SMAPE, is established with its actual value to obtain the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence. M The error coefficients ΔY1, ΔY2, ..., ΔY M The inputs are fed into the trained ResNet prediction model to obtain the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence. M Based on the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence M The method and related devices for controlling the movement of control rods can accurately control the position of reactor control rods.

Description

Technical Field

[0001] This invention belongs to the field of nuclear power engineering technology and relates to a method and related device for controlling the position of reactor control rods. Background Technology

[0002] Control rods, as a core and critical component of a nuclear power plant, are essential for its safe and stable operation. Their main functions include: when the unit triggers reactor protection actions for an emergency shutdown, the control rod drive mechanism loses power, allowing the control rods to rapidly fall into the reactor core under gravity. This rapid descent effectively absorbs neutrons within the reactor, quickly suppressing the rate of nuclear fission reactions and preventing dangerous situations such as over-powering, thus constructing a robust safety barrier for the reactor. During normal operation, the control rod drive mechanism's pawls must precisely and accurately transmit magnetic pole movement commands according to the correct timing, providing precise control rod positioning. This process demands extremely high precision and stability to prevent significant power drops or even unexpected reactor shutdowns due to control rod misalignment.

[0003] With the continuous increase in the total installed capacity of nuclear power units, the demand for nuclear power plants to participate in peak shaving and frequency regulation in line with grid load will increase. During the operation of a nuclear power plant, the position of reactor control rods will be continuously adjusted according to the unit load, which places higher demands on control rod position control. It is necessary to develop more accurate control rod position prediction methods, which is crucial for the safe and stable operation of nuclear power plants. Currently developed reactor control rod prediction algorithms have problems with poor prediction accuracy. For example, Particle Swarm Optimization (PSO) is prone to getting trapped in local optima and has a fast convergence speed, resulting in low convergence accuracy. Wavelet Energy Method (WEM) lacks adaptability and is sensitive to noise, especially when processing complex signals, which may lead to edge blurring or loss. Local Mean Decomposition (LMD) also has some limitations. For example, when processing some extremely complex or noisy signals, the decomposition results may be unstable or inaccurate. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and related apparatus for controlling the position of reactor control rods, which can accurately control the position of reactor control rods.

[0005] To achieve the above objectives, this invention discloses a method for controlling the position of reactor control rods, comprising:

[0006] Obtain the original rod position sequence of the reactor control rods, and decompose the original rod position sequence to obtain M IMF components;

[0007] The M IMF components are input into the trained Long Short-Term Memory (LSTM) network model to obtain the predicted values ​​X1, X2, ..., X of the M IMF components. M Based on the predicted values ​​X1, X2, ..., X of the M IMF components M An error evaluation index, D-SMAPE, is established with its actual value to obtain the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence. M ;

[0008] The error coefficients ΔY1, ΔY2, ..., ΔY M The inputs are fed into the trained ResNet prediction model to obtain the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar position sequence. M ;

[0009] Based on the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence M Control the movement of the control rod.

[0010] A further improvement of the reactor control rod position control method of the present invention is that:

[0011] Furthermore, the process of decomposing the original bar sequence to obtain M IMF components is as follows:

[0012] The original bar sequence was subjected to SWT-LMD mode decomposition to obtain M IMF components.

[0013] Furthermore, the loss function of the ResNet prediction model during training is:

[0014]

[0015] Where N is the sample size. Corrected predicted values ​​for the control rod position.

[0016] Furthermore, the final IMF component prediction values ​​Y1, Y2, ..., Y based on the original bar sequence... M The process of controlling the movement of the control stick is as follows:

[0017] The final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence are calculated. M The signal is converted into an electrical signal, and the control rod drive mechanism is driven by the electrical signal to push the control rod to move up or down.

[0018] Furthermore, based on the dynamic resolution adjustment-adaptive digital-to-analog conversion algorithm, the final IMF component prediction values ​​Y1, Y2, ..., Y... based on the original bar position sequence are... M It is converted into an electrical signal.

[0019] This invention discloses a control system for the position of reactor control rods, comprising:

[0020] The acquisition module is used to acquire the original rod position sequence of the reactor control rods, and decompose the original rod position sequence to obtain M IMF components;

[0021] The first prediction module is used to input the M IMF components into the trained Long Short-Term Memory network model to obtain the predicted values ​​X1, X2, ..., X of the M IMF components. M Based on the predicted values ​​X1, X2, ..., X of the M IMF components M An error evaluation index, D-SMAPE, is established with its actual value to obtain the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence. M ;

[0022] The second prediction module is used to calculate the error coefficients ΔY1, ΔY2, ..., ΔY. M The inputs are fed into the trained ResNet prediction model to obtain the final IMF component prediction values ​​Y1, Y2, ..., YM of the original bar position sequence;

[0023] The control module is used to predict the final IMF component values ​​Y1, Y2, ..., Y based on the original bar position sequence. M Control the movement of the control rod.

[0024] A further improvement of the reactor control rod position control system of the present invention is that:

[0025] Furthermore, the process of decomposing the original bar sequence to obtain M IMF components is as follows:

[0026] The original bar sequence was subjected to SWT-LMD mode decomposition to obtain M IMF components.

[0027] Furthermore, the final IMF component prediction values ​​Y1, Y2, ..., Y based on the original bar sequence... M The process of controlling the movement of the control stick is as follows:

[0028] The final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence are... M The signal is converted into an electrical signal, and the control rod drive mechanism is driven by the electrical signal to push the control rod to move up or down.

[0029] The present invention discloses a computer device, including 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 a method for controlling the position of a reactor control rod.

[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 control method for the position of the reactor control rod.

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

[0032] The reactor control rod position control method and related device described in this invention, in specific operation, employs a complementary approach of stationary wavelet transform (SWT) and LMD data preprocessing techniques to decompose the original rod position sequence. A hierarchical decomposition method is used to couple SWT and LMD techniques, overcoming the shortcomings of each method and avoiding the modal aliasing and insufficient frequency resolution problems of traditional methods, thus improving prediction accuracy. Finally, a long short-term memory network model and a ResNet prediction model are used to predict the final IMF component prediction values ​​Y1, Y2, ..., Y of the original rod position sequence. M This allows for highly accurate control of the reactor control rod positions. 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 controlling the position of reactor control rods according to the present invention includes the following steps:

[0045] 1) Obtain the original rod position sequence of the reactor control rods, and perform SWT-LMD mode decomposition on the original rod position sequence to obtain M IMF components;

[0046] The original bar sequence is decomposed by SWT by selecting a wavelet basis and a decomposition level to obtain sub-signals of different scales, such as approximate components and detail components; each sub-signal is decomposed by LMD, and the results of LMD decomposition are reconstructed to obtain M IMF components.

[0047] 2) Input the M IMF components into the trained Long Short-Term Memory (LSTM) network model to obtain the predicted values ​​X1, X2, ..., X of the M IMF components. M Wherein, the actual values ​​of the IMF components corresponding to each predicted value are M1, M2, ..., M M An error evaluation index, D-SMA.PE, is established based on the predicted and actual values ​​of each IMF component, and the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence are obtained. M ;

[0048] 3) The error coefficients ΔY1, ΔY2, ..., AY M The inputs are fed into the trained ResNet prediction model to obtain the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence. M ;

[0049] 4) Based on the Dynamic Resolution Adaptive Digital-to-Analog Converter (D-ADC) algorithm, the final IMF component prediction values ​​Y1, Y2, ..., Y... based on the original bar position sequence are... MThe signal is converted into an electrical signal, and the control rod drive mechanism is driven by the electrical signal to push the control rod to move up or down.

[0050] The process of performing SWT-LMD mode decomposition on the original bar sequence in step 1) is as follows:

[0051] 11) Perform SWT decomposition on the original bar sequence. The specific process is as follows:

[0052] 111) Wavelet transform;

[0053] Denoising the reactor rod position signal Continuous wavelet transform is performed, and based on the characteristics of the reactor rod position signal, the Daubechies wavelet basis function method is selected to obtain the wavelet coefficients W. x (a, b) is:

[0054]

[0055] Where a is the scaling parameter, b is the translation parameter, ψ(t) is the wavelet basis function, and ψ* represents complex conjugate.

[0056] 112) For wavelet coefficients W x Synchronous compression of (a, b) yields a high-resolution time-frequency representation T. x (a, b) is:

[0057]

[0058] Where, ω x (a, b) are the instantaneous frequencies of the wavelet coefficients, and δ(·) is the Daubechies wavelet basis function.

[0059] 113) Modality extraction;

[0060] According to the time-frequency representation T x (a, b), extract the modal component M of the reactor rod control signal. k (t) is:

[0061]

[0062] Where, ω k and ω k+1 Let be the frequency range of the k-th modal component.

[0063] 12) For modal component M k (t) Perform LMD decomposition, the specific process is as follows:

[0064] 121) Calculate the local mean and envelope function:

[0065] For the modal component M of the reactor rod control signalk (t), the initial residual signal r(t) is:

[0066] r(t) = M k (t)

[0067] Obtain all local extreme points of the initial residual control rod position signal r(t), and then obtain the upper envelope max(r(t)) and lower envelope min(r(t)) through linear interpolation between adjacent extreme points;

[0068] The average value m(t) of the upper and lower envelopes is calculated as follows:

[0069]

[0070] The half of the difference between the upper and lower envelopes, a(t), is calculated as follows:

[0071]

[0072] 122) Separate the amplitude modulation and frequency modulation components;

[0073] Subtracting the local mean function m(t) from the control rod position signal r(t), and then dividing by the envelope estimation function a(t), we obtain the pure frequency modulation function F(t) as follows:

[0074]

[0075] 123) Construct the product function;

[0076] By integrating the pure frequency modulation function F(t), the phase function is obtained. Differentiating the phase function yields the instantaneous frequency, leading to the product function PF. n (t) is:

[0077] PF n (t)=a n (t)·cos(∫F n (t)dt)

[0078] 124) Iterative decomposition;

[0079] reactor rod position signal Subtract the first product function PF n (t), to obtain the residual control rod position signal r1(t); repeat the above steps for r1(t) to obtain the second product function PF2(t) and the residual signal r2(t), and so on, until the residual control rod position signal r is obtained. n (t) The termination condition is met, wherein the amplitude of the remaining signal is less than a preset amplitude or the preset number of decomposition layers is reached.

[0080] 13) Modal reconstruction combining SWT and LMD;

[0081] The components after LMD decomposition and reconstruction are subjected to inverse wavelet transform to obtain the final control rod position reconstruction signal. for:

[0082]

[0083] This algorithm provides a high-resolution time-frequency representation through Synchronous Compressed Wavelet Transform (SWT) and combines it with the adaptive decomposition capability of Local Mean Decomposition (LMD). By using inverse wavelet transform, the signal is converted from the time-frequency domain back to the time domain, completing the entire combination process. This results in a signal decomposition result that has both the multi-resolution analysis characteristics of wavelet transform and the adaptive decomposition characteristics of LMD, enabling more accurate extraction of the modal components of non-stationary signals.

[0084] Step 2) involves feeding the M IMF components into the trained Long Short-Term Memory (LSTM) network model for prediction.

[0085] 21) Initialize the IMF parameters of the network control rod position;

[0086] 211) Initialize the control bar IMF parameters of the LSTM network, including the weight matrix W. f W i W O and W c and bias term b f b i b0, b C ;

[0087] 212) Initialize the hidden state h0 and the memory unit state C0.

[0088] 22) Adaptive gating mechanism;

[0089] 221) Forget Gate:

[0090]

[0091] Among them, f t The output of the forget gate at time t is σ, where σ is the sigmoid function and its output value is between 0 and 1; W f Here is the weight matrix for the forget gate; This indicates that the hidden state h from the previous time step is... t-1 and the input x at the current time t spliced ​​together; b f α is the bias term for the forgetting gate; g(t) is a time-dependent adaptive function used to dynamically adjust the weights of the forgetting gate; f These are adaptive coefficients.

[0092] 222) Input Gate:

[0093] The input gate outputs i at time t t for:

[0094]

[0095] Among them, W i Let b be the weight matrix of the input gate. i For the corresponding bias term, α i These are adaptive coefficients.

[0096] 223) Memory unit state;

[0097] Calculate the state values ​​of candidate memory cells for:

[0098]

[0099] Calculate the state value C of the candidate memory cell t for:

[0100]

[0101] The tanh function maps the output value to the range of -1 to 1. c Let b be the weight matrix of the candidate memory cell states. c This is the corresponding bias term.

[0102] 224) Output Gate:

[0103] The output gate outputs o at time t t for

[0104]

[0105] Among them, W o Let b be the weight matrix of the output gate. o For the bias term, α o These are adaptive coefficients.

[0106] 225) Update hidden status:

[0107] h t =o t ·tanh(C t )

[0108] The tanh function maps the output value to the range of -1 to 1.

[0109] 23) Multi-level memory units;

[0110] For each level 1, update the memory cell state.

[0111]

[0112] in, and These are the forget gate and input gate outputs of the first layer, respectively; This represents the candidate memory cell state of the first layer.

[0113] Calculate multi-level hidden states for:

[0114]

[0115] in, This is the output gate of the first layer.

[0116] Determine the final concealment state h t for:

[0117]

[0118] Where, β l These are the weighting coefficients for the first layer.

[0119] 24) Output the control rod position prediction result;

[0120] The final hidden state h is hidden through a fully connected layer. t Mapping to the helium flow output space, the control rod position prediction results are obtained. for:

[0121]

[0122] Among them, w y and b y These are the weight matrix and bias term of the fully connected layer, respectively.

[0123] Step 2) involves establishing the error evaluation index D-SMAPE based on the predicted and actual values ​​of each IMF component.

[0124] 2a1) Calculate the dynamic error e of the control rod position. i for:

[0125]

[0126] 2a2) Error aggregation;

[0127] Average all errors of the control bar position sequence:

[0128]

[0129] Segmented aggregation:

[0130]

[0131] 2a3) Output error coefficient;

[0132] Multiply the average value obtained in step 2a2) by 100% to obtain the final SMAPE value, and then express the predicted control rod position error coefficients ΔY1, ΔY2, ..., ΔY as percentages. M .

[0133] In step 3), the error coefficients ΔY1, ΔY2, ..., ΔY are... M The inputs are fed into the trained ResNet prediction model to obtain the final IMF component predictions Y1, Y2, ... Y of the original bar sequence. M The process is as follows:

[0134] 31) Residual block construction;

[0135] Introducing dynamic residual connection, based on the input control rod position error coefficient ΔY i The dynamic characteristics adjust the weighting coefficient α(Δy) of the residual connection. i ):

[0136] y i =F(ΔY) i , {w i})+α(ΔY i )·ΔY i

[0137] Among them, y i To control the predicted value of the rod position, F(ΔY) i , {w i}) is the residual function; σ is the Sigmoid activation function, which sets α(ΔY) to the residual function. i The range is limited to [0, 1].

[0138] α(ΔY i )for:

[0139] α(ΔY i )=σ(W α ·ΔY i +b α )

[0140] 32) Construct the residual network;

[0141] A residual network for controlling the position of the control rod is constructed by stacking multiple residual blocks;

[0142] Assume the residual network has I control bar position residual blocks, and the input of the first control bar position residual block is X.1-1 If the output is X1, then:

[0143] x l =ResBlock(x 1-1 )

[0144] Here, ResBlock represents the operation of the residual block.

[0145] Let the output of the last residual block be X1. Passing through a fully connected layer, we obtain the predicted value y for the control rod position. i for:

[0146] y i =w out x l +b out

[0147] Among them, w out Let b be the weight matrix of the output layer. out This is the bias vector.

[0148] 33) Model training;

[0149] 331) Define the loss function;

[0150] The loss function is:

[0151]

[0152] Where N is the sample size. Corrected predicted values ​​for the control rod position.

[0153] 332) Select the optimizer;

[0154] Calculate the gradient: Let the parameter be θ, and the gradient g of the loss function with respect to the parameter be θ. t for:

[0155]

[0156] The first-order moment estimate and the second-order moment estimate are calculated as follows:

[0157]

[0158] Wherein, β1 and β2 are the attenuation rates, which are usually set to 0.9 and 0.999, respectively.

[0159] Update parameters:

[0160]

[0161] Where α is the learning rate. ∈ is a factor constant.

[0162] 333) Cyclic training;

[0163] Forward propagation: x i The data is input into the model, and through calculations at each layer, the predicted values ​​of the control rod positions are obtained.

[0164] Calculate the loss: Calculate the loss L for the current batch based on the loss function.

[0165] Backpropagation: Calculate the gradient g of the loss with respect to the model parameters. t .

[0166] Parameter update: Update the model's parameters θ using the optimizer.

[0167] 334) Output the prediction results;

[0168] The final features are mapped to the output space through a fully connected layer to obtain the final predicted value of the control rod position. for:

[0169]

[0170] Among them, w yi and b yi These are the weight matrix and bias term of the fully connected layer, respectively.

[0171] The operation process for step 4) is as follows:

[0172] 41) Adjust dynamic resolution;

[0173] The frequency f of the input control rod position signal is detected. k ;

[0174] The control rod position resolution N is dynamically adjusted according to the frequency. k for:

[0175]

[0176] The control bar position signal is requantized using the adjusted resolution.

[0177] 42) The generated electrical signal is:

[0178]

[0179] Among them, v out(n) f is the analog voltage signal of the control rod position output at the nth sampling time. i，n N is the conversion factor. k，n The digital signal value of the control rod position input at the nth sampling time.

[0180] Example 2

[0181] The reactor control rod position control system of the present invention includes:

[0182] The acquisition module is used to acquire the original rod position sequence of the reactor control rods, and decompose the original rod position sequence to obtain M IMF components;

[0183] The first prediction module is used to input the M IMF components into the trained Long Short-Term Memory network model to obtain the predicted values ​​X1, X2, ..., X of the M IMF components. M Based on the predicted values ​​X1, X2, ..., X of the M IMF components M An error evaluation index, D-SMAPE, is established with its actual value to obtain the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence. M ;

[0184] The second prediction module is used to calculate the error coefficients ΔY1, ΔY2, ..., ΔY. M The inputs are fed into the trained ResNet prediction model to obtain the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar position sequence. M ;

[0185] The control module is used to predict the final IMF component values ​​Y1, Y2, ..., Y based on the original bar position sequence. M Control the movement of the control rod.

[0186] In this embodiment, the process of decomposing the original bar sequence to obtain M IMF components is as follows:

[0187] The original bar sequence was subjected to SWT-LMD mode decomposition to obtain M IMF components.

[0188] In this embodiment, the predicted values ​​Y1, Y2, ..., Y based on the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence are... M The process of controlling the movement of the control stick is as follows:

[0189] The final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence are... M The signal is converted into an electrical signal, and the control rod drive mechanism is driven by the electrical signal to push the control rod to move up or down.

[0190] 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.

[0191] Example 3

[0192] 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 control method for reactor control rod positions. For example, the method includes: acquiring an original sequence of reactor control rod positions; decomposing the original sequence to obtain M IMF components; and inputting the M IMF components into a trained Long Short-Term Memory (LSTM) network model to obtain predicted values ​​X1, X2, ..., X... for the M IMF components. M Based on the predicted values ​​X1, X2, ..., X of the M IMF components M An error evaluation index, D-SMAPE, is established with its actual value to obtain the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence. M The error coefficients ΔY1, ΔY2, ..., ΔY M The inputs are fed into the trained ResNet prediction model to obtain the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar position sequence. M Based on the final IMF component prediction values ​​Y1, Y2, ... of the original rod position sequence, YM controls the rod's movement. The memory may include main memory, such as high-speed random access memory, or it may also include non-volatile memory, such as at least one disk storage device. The processor, network interface, and memory are interconnected via an internal bus, which can be an industry-standard architecture bus, a peripheral component interconnection standard bus, an extended industry-standard architecture bus, etc. The bus can be divided into address bus, data bus, control bus, etc. The memory is used to store 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.

[0193] Example 4

[0194] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a control method for reactor control rod positions, for example including: obtaining an original rod position sequence of the reactor control rods; decomposing the original rod position sequence to obtain M IMF components; and inputting the M IMF components into a trained Long Short-Term Memory (LSTM) network model to obtain predicted values ​​X1, X2, ..., X... of the M IMF components. M Based on the predicted values ​​X1, X2, ..., X of the M IMF components M An error evaluation index, D-SMAPE, is established with its actual value to obtain the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence. M The error coefficients ΔY1, ΔY2, ..., ΔY M The inputs are fed into the trained ResNet prediction model to obtain the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar position sequence. M Based on the final IMF component prediction values ​​Y1, Y2, ..., Y of the original bar sequence M The control stick's movement is controlled. 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 and / or cache memory, etc. The non-volatile memory may include read-only memory, hard disk, flash memory, optical disk, magnetic disk, etc.

[0195] 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 implemented on one or more computer-usable storage media containing computer-usable program code.

[0196] 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.

[0197] 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.

[0198] 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.

[0199] 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.

[0200] 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.

[0201] 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 method of controlling a position of a control rod of a nuclear reactor, characterized by, The method comprises the following steps: obtaining an original rod position sequence of a reactor control rod, decomposing the original rod position sequence to obtain M IMF components; The M IMF components are input into the trained Long Short-Term Memory (LSTM) network model to obtain the predicted values ​​X1, X2, ..., X of the M IMF components. M Based on the predicted values ​​X1, X2, ..., X of the M IMF components M An error evaluation index, D-SMAPE, is established with its actual value to obtain the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence. M ; The error coefficients ΔY1, ΔY2,..., ΔY M are input into the trained ResNet prediction model to obtain the final IMF component prediction values Y1, Y2,..., Y M of the original stick position sequence. According to the final IMF component prediction values Y1, Y2,..., Y M Controls control rod motion; the process of decomposing the original rod position sequence to obtain M IMF components comprises the following steps: performing SWT-LMD modal decomposition on the original rod position sequence to obtain M IMF components; performing SWT decomposition on the original rod position sequence by selecting a wavelet base and a decomposition level to obtain sub-signals of different scales, wherein the sub-signals of different scales comprise an approximate component and a detail component; performing LMD decomposition on each sub-signal, and reconstructing the LMD decomposition result to obtain M IMF components.

2. The method of controlling the position of a control rod of a nuclear reactor according to claim 1, characterized in that, In the training process of the ResNet prediction model, a loss function is: wherein, N is the number of samples ， is the corrected prediction of the control rod position, is the control rod position prediction.

3. The method of controlling the position of a control rod of a nuclear reactor according to claim 1, characterized in that, The final IMF component prediction values Y1, Y2,..., Yn according to the original rod position sequence are calculated as follows: M The process of controlling the movement of the control rod is as follows: The final IMF component prediction values Y1, Y2,..., Y M The electrical signal is converted into a control signal for controlling the rod drive mechanism to push the control rods up or down.

4. The method of controlling the position of a control rod of a nuclear reactor according to claim 3, characterized in that, Based on a dynamic resolution adjustment-adaptive digital-to-analog conversion algorithm, the final IMF component prediction values Y1, Y2,..., Y M are converted into electrical signals.

5. A control system for control of the position of a control rod of a nuclear reactor, characterized in that The method comprises the following steps: a obtaining module, configured to obtain an original rod position sequence of a reactor control rod, and decompose the original rod position sequence to obtain M IMF components; The first prediction module is used to input the M IMF components into the trained Long Short-Term Memory network model to obtain the predicted values ​​X1, X2, ..., X of the M IMF components. M Based on the predicted values ​​X1, X2, ..., X of the M IMF components M An error evaluation index, D-SMAPE, is established with its actual value to obtain the error coefficients ΔY1, ΔY2, ..., ΔY corresponding to the SWT-LMD subsequence. M ; a second prediction module configured to input the error coefficients ΔY1, ΔY2,..., ΔY M into the trained ResNet prediction model to obtain final IMF component prediction values Y1, Y2,..., Y M of the original stick position sequence. a control module configured to predict values Y1, Y2,..., Yn of the final IMF components of the original rod position sequence based on the initial IMF components of the original rod position sequence M controlling the control rod motion; the process of decomposing the original rod position sequence to obtain M IMF components comprises the following steps: performing SWT-LMD modal decomposition on the original rod position sequence to obtain M IMF components; performing SWT decomposition on the original rod position sequence by selecting a wavelet base and a decomposition level to obtain sub-signals of different scales, wherein the sub-signals of different scales comprise an approximate component and a detail component; performing LMD decomposition on each sub-signal, and reconstructing the LMD decomposition result to obtain M IMF components.

6. The control system for control rod position of a nuclear reactor as recited in claim 5, wherein, said final IMF component prediction values Y1, Y2,..., Yn according to said original stick position sequence M The process of controlling the movement of the control rod is: The final IMF component prediction values Y1, Y2,..., Y M The electrical signal is converted into a control signal for controlling the rod drive mechanism to push the control rods up or down.

7. 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 processor executes the computer program to realize the steps of the control method of the reactor control rod rod position according to any one of claims 1-4.

8. 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 7. The computer program is executed by the processor to realize the steps of the control method of the reactor control rod rod position according to any one of claims 1-4.

Images