Feedwater system for evaporator of reactor nuclear power unit

Abstract

Disclosed in the present application is a feedwater system for an evaporator of a reactor nuclear power unit. By means of sensors, the feedwater system monitors in real time the operating state of an evaporator, such as water level, temperature and pressure parameters, and data processing and analysis algorithms are introduced to a backend to perform time-series collaborative correlation analysis on these operating parameters of the evaporator, so as to monitor in real time the operating state of the evaporator. In this way, an opening value of a regulation valve can be adaptively controlled in real time on the basis of the operating state of the evaporator, such that automatic control over the feedwater system for the evaporator of the reactor nuclear power unit is implemented, thereby meeting the requirements of the flow and pressure of feedwater, ensuring the stability and safety of the feedwater system, and improving the safety and efficiency of nuclear power stations.

Description

A water supply system for an evaporator of a reactor nuclear power unit Technical Field

[0001] The present application relates to the field of intelligent control, and more specifically, to an evaporator water feed system for a reactor nuclear power unit. Background Art

[0002] The evaporator feedwater system of a nuclear power plant is a crucial component of a nuclear power plant. It condenses steam into water and returns it to the reactor core to maintain stable operation. The stability of the feedwater system is crucial to the safety and efficiency of the nuclear power plant.

[0003] Traditional water supply systems typically rely on manual control of feedwater flow and pressure, but this approach presents several challenges. For example, manual operation is susceptible to operator bias and experience, potentially leading to inaccurate or delayed responses, impacting system stability and safety. Furthermore, because feedwater system operation requires manual intervention, there is a risk of human error. This means that operators could cause system failures or anomalies due to negligence, misoperation, or other factors, potentially impacting the normal operation of the nuclear power plant.

[0004] Furthermore, the operating status of a water supply system may be affected by multiple parameters, such as water level, temperature, and pressure. Traditional water supply systems can typically only monitor and control a single piece of data in real time, making it difficult to monitor and adjust these parameters simultaneously. This can result in delayed system responses or failure to meet actual needs.

[0005] Therefore, an optimized evaporator feedwater system for a nuclear power reactor is desired. Technical issues

[0006] The present application is proposed to solve the above-mentioned technical problems. The embodiments of the present application provide a reactor nuclear power unit evaporator water feed system, which uses sensors to monitor the working status of the evaporator in real time, such as water level, temperature, and pressure parameters, and introduces data processing and analysis algorithms at the back end to perform time-series collaborative correlation analysis of these evaporator working parameters, thereby monitoring the working status of the evaporator in real time. In this way, the opening value of the regulating valve can be adaptively controlled in real time based on the working status of the evaporator, thereby realizing automatic control of the reactor nuclear power unit evaporator water feed system to meet the requirements of water feed flow and pressure, ensure the stability and safety of the water feed system, and improve the safety and efficiency of the nuclear power plant. Technical Solutions

[0007] According to one aspect of the present application, a reactor nuclear power unit evaporator water supply system is provided, comprising:

[0008] An evaporator data acquisition module is used to obtain the water level value, pressure value and temperature value of the evaporator at multiple predetermined time points within a predetermined time period;

[0009] an evaporator data parameter time series arrangement module, configured to arrange the water level values, pressure values, and temperature values ​​of the evaporator at the plurality of predetermined time points into a water level value time series input vector, a pressure value time series input vector, and a temperature value time series input vector, respectively, according to a time dimension;

[0010] an evaporator parameter time series characteristic analysis module, configured to perform time series analysis on the water level value time series input vector, the pressure value time series input vector, and the temperature value time series input vector to obtain a water level value time series characteristic vector, a pressure value time series characteristic vector, and a temperature value time series characteristic vector;

[0011] an evaporator multi-parameter time series feature fusion module, configured to fuse the water level value time series feature vector, the pressure value time series feature vector, and the temperature value time series feature vector to obtain an evaporator parameter time series collaborative fusion feature;

[0012] The regulating valve opening control module is used to determine whether the valve opening value of the regulating valve at the current time point should be increased, decreased or maintained based on the time series collaborative fusion characteristics of the evaporator parameters. Beneficial effects

[0013] Compared with the prior art, the present application provides a reactor nuclear power unit evaporator water feed system, which uses sensors to monitor the working status of the evaporator in real time, such as water level, temperature and pressure parameters, and introduces data processing and analysis algorithms at the back end to perform time-series collaborative correlation analysis of these evaporator working parameters, so as to monitor the working status of the evaporator in real time. In this way, the opening value of the regulating valve can be adaptively controlled in real time based on the working status of the evaporator, thereby realizing automatic control of the reactor nuclear power unit evaporator water feed system to adapt to the requirements of water feed flow and pressure, ensure the stability and safety of the water feed system, and improve the safety and efficiency of the nuclear power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0015] FIG1 is a block diagram of an evaporator water feed system for a reactor nuclear power unit according to an embodiment of the present application;

[0016] FIG2 is a system architecture diagram of an evaporator water feed system for a reactor nuclear power unit according to an embodiment of the present application;

[0017] FIG3 is a block diagram of a training phase of an evaporator feedwater system for a nuclear power plant in a reactor according to an embodiment of the present application;

[0018] FIG4 is a block diagram of an evaporator parameter timing characteristic analysis module in an evaporator water feed system of a reactor nuclear power unit according to an embodiment of the present application. Modes for Carrying Out the Invention

[0019] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0020] As used in this application and the claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not intended to refer to the singular but may include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0021] Although the present application makes various references to certain modules in the system according to embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are illustrative only, and different aspects of the system and method can use different modules.

[0022] Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the various steps may be processed in reverse order or simultaneously, as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0023] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0024] Specifically, the technical solution of this application proposes an evaporator water feed system for a nuclear power reactor unit, capable of supplying water to the steam generator to maintain stable operation of the nuclear reactor. The evaporator water feed system comprises: a steam generator, a key component in a nuclear power plant, used to convert heat energy generated in the nuclear reactor into steam. The steam generator contains a set of tube bundles through which the nuclear reactor coolant (usually water) flows, exchanging heat between the tube walls and the water / steam interface within the tubes. An evaporator, steam from the steam generator condenses in the evaporator, transferring heat energy to the water surrounding the evaporator. The evaporator is typically a large heat exchanger consisting of a series of parallel tube bundles. Steam condenses outside the tube bundles, releasing heat, causing the water inside the bundles to evaporate. A water feed system, the main function of which is to supply water to the evaporator to replace water lost during evaporation. In particular, the water supply system generally includes the following main components: a. Feed water pump: The feed water pump is used to draw water from the water treatment system or water storage tank and provide sufficient pressure to transport the water to the evaporator; b. Deaerator: The deaerator is used to remove oxygen from the water to prevent corrosion and bubble formation in the evaporator; c. Desalter: The desalter is used to remove salt from the water to prevent scaling and corrosion in the evaporator; d. Regulating valve: The regulating valve is used to control the feed water flow and pressure to meet the needs of the evaporator; e. Water level control system: The water level control system is used to monitor the water level in the evaporator and automatically adjust the operation of the water supply system as needed to maintain the appropriate water level.

[0025] Working Principle: The feedwater system operates by pumping water to the evaporator via a feedwater pump. A regulating valve and water level control system control the flow and pressure of the feedwater. In the evaporator, the water is heated by the steam, partially evaporating into steam. The remaining water cools the steam and returns to the feedwater system. This circulating water supply maintains the water level in the evaporator to meet the needs of the steam generator.

[0026] In the technical solution of the present application, a reactor nuclear power unit evaporator water supply system is proposed. FIG1 is a block diagram of a reactor nuclear power unit evaporator water supply system according to an embodiment of the present application. FIG2 is a system architecture diagram of a reactor nuclear power unit evaporator water supply system according to an embodiment of the present application. As shown in FIG1 and FIG2, the reactor nuclear power unit evaporator water supply system 300 according to an embodiment of the present application includes: an evaporator data acquisition module 310, for acquiring the water level value, pressure value and temperature value of the evaporator at a plurality of predetermined time points within a predetermined time period; an evaporator data parameter timing arrangement module 320, for arranging the water level value, pressure value and temperature value of the evaporator at the plurality of predetermined time points into a water level value timing input vector, a pressure value timing input vector and a temperature value timing input vector according to the time dimension; an evaporator parameter timing feature analysis module 330, for analyzing the water level value timing input vectors respectively. The input vector, the pressure value time series input vector and the temperature value time series input vector are subjected to time series analysis to obtain the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector; the evaporator multi-parameter time series feature fusion module 340 is used to fuse the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector to obtain the evaporator parameter time series collaborative fusion feature; the regulating valve valve opening control module 350 is used to determine whether the valve opening value of the regulating valve at the current time point should be increased, decreased or maintained based on the evaporator parameter time series collaborative fusion feature.

[0027] Specifically, the evaporator data acquisition module 310 is configured to acquire the water level, pressure, and temperature of the evaporator at multiple predetermined time points within a predetermined time period. In one example, the water level of the evaporator at multiple predetermined time points within the predetermined time period may be acquired using a water level sensor, the pressure of the evaporator at multiple predetermined time points within the predetermined time period may be acquired using a pressure sensor, and the temperature of the evaporator at multiple predetermined time points within the predetermined time period may be acquired using a temperature sensor.

[0028] It's worth noting that a water level sensor is a device used to measure the height of a liquid. It's typically connected to a data acquisition system or control system, transmitting the measured water level information to monitoring equipment or actuators. This allows real-time monitoring of water level changes and the implementation of appropriate control measures. A pressure sensor is a device used to measure pressure. It converts a physical quantity into an electrical signal to monitor and measure pressure changes. A pressure sensor typically converts the measured pressure signal into a standard electrical output, such as an analog voltage or digital signal. This allows for convenient transmission of the pressure data to monitoring equipment, control systems, or data acquisition systems for processing and analysis. A temperature sensor is a device used to measure temperature. It converts a physical quantity into an electrical signal to monitor and measure temperature changes. A temperature sensor typically converts the measured temperature signal into a standard electrical output, such as an analog voltage or digital signal. This allows for convenient transmission of the temperature data to monitoring equipment, control systems, or data acquisition systems for processing and analysis.

[0029] In particular, the evaporator data parameter time series arrangement module 320 is used to arrange the water level values, pressure values, and temperature values ​​of the evaporator at the multiple predetermined time points into a water level value time series input vector, a pressure value time series input vector, and a temperature value time series input vector, respectively, according to the time dimension. Considering that the water level value, pressure value, and temperature value of the evaporator all have a time series dynamic change law in the time dimension, that is, the water level value, pressure value, and temperature value of the evaporator at the multiple predetermined time points all have a time series correlation relationship in the sample dimension. Therefore, in the technical solution of the present application, the water level value, pressure value, and temperature value of the evaporator at the multiple predetermined time points are further arranged into a water level value time series input vector, a pressure value time series input vector, and a temperature value time series input vector, respectively, according to the time dimension, so as to respectively integrate the time series distribution information of the water level value, pressure value, and temperature value of the evaporator.

[0030] In particular, the evaporator parameter timing characteristic analysis module 330 is used to perform timing analysis on the water level value timing input vector, the pressure value timing input vector and the temperature value timing input vector respectively to obtain the water level value timing characteristic vector, the pressure value timing characteristic vector and the temperature value timing characteristic vector. In particular, in a specific example of the present application, as shown in Figure 4, the evaporator parameter timing feature analysis module 330 includes: an evaporator parameter timing data preprocessing unit 331, which is used to pass the water level value timing input vector, the pressure value timing input vector and the temperature value timing input vector through a preprocessor including an upsampling layer and a vector-image converter to obtain a water level value timing image, a pressure value timing image and a temperature value timing image; an evaporator multi-parameter timing feature extraction unit 332, which is used to perform feature extraction on the water level value timing image, the pressure value timing image and the temperature value timing image respectively through a timing feature extractor based on a deep neural network model to obtain the water level value timing feature vector, the pressure value timing feature vector and the temperature value timing feature vector.

[0031] Specifically, the evaporator parameter time series data preprocessing unit 331 is configured to process the water level value time series input vector, the pressure value time series input vector, and the temperature value time series input vector, respectively, through a preprocessor comprising an upsampling layer and a vector-to-image converter to obtain a water level value time series image, a pressure value time series image, and a temperature value time series image. It should be understood that in the technical solution of the present application, in order to improve the ability to capture subtle changes in the time series of the evaporator's water level value, pressure value, and temperature value, the technical solution of the present application further performs upsampling processing on the water level value time series input vector, the pressure value time series input vector, and the temperature value time series input vector to increase the density and smoothness of the data parameters in the time series, thereby facilitating a subsequent better representation of the evaporator's operating state characteristics. Furthermore, considering that compared to simple vector representation, time series images can provide more information, including the temporal relationships, fluctuations, and trend changes of the data, etc. Therefore, after upsampling the water level value time series input vector, the pressure value time series input vector, and the temperature value time series input vector, respectively, the upsampled time series input vector is further subjected to vector-to-image conversion. Such conversion can better capture the time series dynamic changes of the water level, pressure, and temperature in the water supply system. Based on this, in the technical solution of the present application, the water level value time series input vector, the pressure value time series input vector, and the temperature value time series input vector are respectively passed through a preprocessor including an upsampling layer and a vector-to-image converter to obtain a water level value time series image, a pressure value time series image, and a temperature value time series image.

[0032] It's worth noting that upsampling refers to increasing the sampling rate of a signal, that is, increasing the density of sampling points. In digital signal processing, upsampling is often used to convert a low-sampling-rate signal to a higher-sampling-rate signal for more refined analysis, processing, or restoration. Upsampling can be used in a variety of applications, such as resampling audio signals, amplifying images, and interpolating signals.

[0033] Specifically, the evaporator multi-parameter time series feature extraction unit 332 is used to perform feature extraction on the water level value time series image, the pressure value time series image, and the temperature value time series image respectively through a time series feature extractor based on a deep neural network model to obtain the water level value time series feature vector, the pressure value time series feature vector, and the temperature value time series feature vector. In the technical solution of the application, a time series feature extractor based on a convolutional neural network model with excellent performance in implicit feature extraction of images is used to perform feature mining on the water level value time series image, the pressure value time series image, and the temperature value time series image respectively, so as to extract the time series implicit distribution feature information of the water level value, pressure value, and temperature value of the evaporator in the time dimension, thereby obtaining the water level value time series feature vector, the pressure value time series feature vector, and the temperature value time series feature vector. More specifically, each layer of the time series feature extractor based on the convolutional neural network model is used to perform the following operations on the input data in the forward pass of the layer: convolution processing is performed on the input data to obtain a convolution feature map; global mean pooling based on the feature matrix is ​​performed on the convolution feature map to obtain a pooled feature map; and nonlinear activation is performed on the pooled feature map to obtain an activation feature map; wherein the output of the last layer of the time series feature extractor based on the convolutional neural network model is the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector, and the input of the first layer of the time series feature extractor based on the convolutional neural network model is the water level value time series image, the pressure value time series image and the temperature value time series image.

[0034] It's worth noting that the Convolutional Neural Network (CNN) is a deep learning model primarily designed for processing grid-structured data, such as images and time series data. CNNs have achieved significant breakthroughs in computer vision and have also been widely applied in other fields, such as natural language processing and medical image analysis. The core concept of a CNN is to automatically extract features from input data using convolutional and pooling layers, and then perform tasks such as classification or regression using fully connected layers. The following are the main components of a CNN: Convolutional layer: The convolutional layer is the core component of a CNN. It performs a convolution operation on the input data by applying a series of convolutional kernels (also called filters). The convolution operation extracts local features of the input data and performs calculations on the input data using a sliding window. Each convolution kernel learns different features, such as edges and textures. The output of a convolutional layer is called a feature map. Activation function: A nonlinear activation function, such as ReLU, is applied to the output of the convolutional layer to introduce nonlinear transformations and increase the network's expressive power. Pooling layer: Pooling layers are used to reduce the size of the feature map and reduce the network's spatial sensitivity to the input data. Common pooling operations include max pooling and average pooling, which represent the features of a local area by selecting the maximum or average value in that area. Fully connected layers: A fully connected layer connects the output of the pooling layer to one or more fully connected layers to map the extracted features to the final output category. Neurons in a fully connected layer are connected to all neurons in the previous layer. Each connection has a weight, which is learned through backpropagation during training. Dropout: To reduce overfitting, CNNs often use the dropout technique. Dropout randomly drops a subset of neurons with a certain probability during training, thereby reducing the dependencies between neurons and improving the model's generalization ability. The CNN training process typically uses the backpropagation algorithm and gradient descent optimization to update network parameters to minimize the loss function. During training, CNNs learn the feature representation of the input data by continuously adjusting the weights and biases of the convolution kernels.

[0035] It is worth mentioning that in other specific examples of the present application, the water level value time series input vector, the pressure value time series input vector and the temperature value time series input vector can also be subjected to time series analysis in other ways to obtain a water level value time series feature vector, a pressure value time series feature vector and a temperature value time series feature vector, for example: the time series analysis step of the water level value time series input vector: collecting time series data of the water level sensor, including the water level value and the corresponding timestamp; preprocessing the timestamp, such as converting it into an appropriate time format; preprocessing the water level value, such as removing noise, filling missing values, etc.; calculating statistical features, such as mean, maximum, minimum, standard deviation, etc.; calculating time domain features, such as autocorrelation, difference, etc.; calculating frequency domain features, such as Fourier transform, wavelet transform, etc.; combining the calculated feature values ​​into a water level value time series feature vector; the time series analysis step of the pressure value time series input vector: collecting time series data of the pressure sensor, including Pressure value and corresponding timestamp; preprocess the timestamp, such as converting it into an appropriate time format; preprocess the pressure value, such as removing noise, filling missing values, etc.; calculate statistical features, such as average, maximum, minimum, standard deviation, etc.; calculate time domain features, such as autocorrelation, difference, etc.; calculate frequency domain features, such as Fourier transform, wavelet transform, etc.; combine the calculated eigenvalues ​​into a pressure value time series feature vector; time series analysis steps for temperature value time series input vector: collect time series data from the temperature sensor, including temperature values ​​and corresponding timestamps; preprocess the timestamp, such as converting it into an appropriate time format; preprocess the temperature value, such as removing noise, filling missing values, etc.; calculate statistical features, such as average, maximum, minimum, standard deviation, etc.; calculate time domain features, such as autocorrelation, difference, etc.; calculate frequency domain features, such as Fourier transform, wavelet transform, etc.; combine the calculated eigenvalues ​​into a temperature value time series feature vector.

[0036] In particular, the evaporator multi-parameter time-series feature fusion module 340 is configured to fuse the water level time-series feature vector, the pressure time-series feature vector, and the temperature time-series feature vector to obtain a coordinated fusion feature of the evaporator's parameter time-series. It should be understood that the water level, pressure, and temperature values ​​are important parameters in the water supply system, and their time-series feature vectors can provide information about the evaporator's state and changes. Therefore, it is necessary to fuse the evaporator's water level time-series feature, pressure time-series feature, and temperature time-series feature to comprehensively monitor the evaporator's state and control the valve opening value of the regulating valve. Based on this, the technical solution of the present application further utilizes a Bayesian probability model to fuse the water level time-series feature vector, the pressure time-series feature vector, and the temperature time-series feature vector to obtain a posterior feature vector for regulating valve operation. By using this Bayesian probability model, the characteristic information in the water level time-series feature vector, the pressure time-series feature vector, and the temperature time-series feature vector can be integrated to obtain a more comprehensive and accurate representation of the evaporator's state. Furthermore, the Bayesian model can also take into account the correlation and weight among the time series characteristics of the water level value, pressure value, and temperature value of the evaporator, thereby better reflecting the actual situation of the evaporator.

[0037] It is worth noting that the Bayesian probability model is a probabilistic modeling method based on Bayes' theorem, used to infer and predict the probability of unknown events. Based on Bayes' theorem, it calculates the posterior probability from the prior probability and observed data, thereby updating the probability estimate of the unknown event. The Bayesian probability model focuses on the conditional dependencies between events. It consists of two important components: the prior probability: The prior probability is an estimate of the probability of an event before considering any observed data. It is based on previous experience, domain knowledge, or assumptions. The prior probability represents a subjective or objective estimate of the probability of an event; and the posterior probability: The posterior probability is the probability of an event calculated according to Bayes' theorem after considering the observed data. It is obtained by combining the prior probability with information from the observed data and updating it based on the conditional dependencies between events. Bayesian probability models have been widely used in many fields, including statistics, machine learning, artificial intelligence, and bioinformatics.

[0038] Specifically, the regulating valve opening control module 350 is configured to determine, based on the evaporator parameter time series collaborative fusion features, whether the regulating valve opening value at the current time point should be increased, decreased, or maintained. In the technical solution of this application, the regulating valve operation posterior feature vector is passed through a classifier to obtain a classification result. The classification result indicates whether the regulating valve opening value at the current time point should be increased, decreased, or maintained. In other words, classification processing is performed using the evaporator's actual operating state feature information to accurately control the regulating valve opening value in real time. In this way, the regulating valve opening value can be adaptively controlled in real time based on the evaporator's operating state, thereby achieving automatic control of the reactor nuclear power unit's evaporator feedwater system to meet feedwater flow and pressure requirements. Specifically, the regulating valve operation posterior feature vector is fully connected encoded using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and the encoded classification feature vector is passed through the classifier's Softmax classification function to obtain the classification result.

[0039] A classifier is a machine learning model or algorithm that is used to classify input data into different categories or labels. Classifiers are part of supervised learning and perform classification tasks by learning a mapping from input data to output categories.

[0040] A fully connected layer is a common layer type in neural networks. In a fully connected layer, each neuron is connected to all neurons in the previous layer, and each connection has a weight. This means that each neuron in a fully connected layer receives input from all neurons in the previous layer, performs a weighted sum of these inputs using the weights, and then passes the result to the next layer.

[0041] The Softmax classification function is a commonly used activation function for multi-classification problems. It converts each element of the input vector into a probability value between 0 and 1, where the sum of these probabilities equals 1. The Softmax function is often used in the output layer of neural networks and is particularly well-suited for multi-classification problems because it maps the network output into a probability distribution for each category. During training, the output of the Softmax function is used to calculate the loss function and update the network parameters through the backpropagation algorithm. It is worth noting that the output of the Softmax function does not change the relative size of the elements; it simply normalizes them. Therefore, the Softmax function does not change the characteristics of the input vector; it simply converts it into a probability distribution.

[0042] It should be understood that before using the above-mentioned neural network model for inference, it is necessary to train the preprocessor including the upsampling layer and the vector-to-image converter, the time series feature extractor based on the convolutional neural network model, the Bayesian probability model, and the classifier. In other words, the reactor nuclear power unit evaporator feedwater system 300 according to the present application also includes a training phase 400 for training the preprocessor including the upsampling layer and the vector-to-image converter, the time series feature extractor based on the convolutional neural network model, the Bayesian probability model, and the classifier.

[0043] FIG3 is a block diagram of the training phase of the evaporator water supply system of a reactor nuclear power unit according to an embodiment of the present application. As shown in FIG3 , the evaporator water supply system 300 of the reactor nuclear power unit according to an embodiment of the present application includes: a training phase 400, including: a training data acquisition unit 410, for obtaining the training water level value, training pressure value and training temperature value of the evaporator at multiple predetermined time points within a predetermined time period, and whether the valve opening value of the regulating valve at the current time point should be increased, decreased or maintained; a training data time series arrangement unit 420, for arranging the training water level value, training pressure value and training temperature value of the evaporator at the multiple predetermined time points into training water level values ​​according to the time dimension. The training data preprocessing unit 430 is used to pass the training water level value time series input vector, the training pressure value time series input vector and the training temperature value time series input vector through the preprocessor including the upsampling layer and the vector-image converter to obtain the training water level value time series image, the training pressure value time series image and the training temperature value time series image; the training evaporator parameter time series feature extraction unit 440 is used to extract the training water level value time series image, the training pressure value time series image and the training temperature value time series image. The training temperature value time series image is respectively passed through the time series feature extractor based on the convolutional neural network model to obtain a training water level value time series feature vector, a training pressure value time series feature vector and a training temperature value time series feature vector; a training evaporator multi-parameter time series feature fusion unit 450 is used to use the Bayesian probability model to fuse the training water level value time series feature vector, the training pressure value time series feature vector and the training temperature value time series feature vector to obtain a training control valve operation posterior feature vector; a feature distribution optimization unit 460 is used to perform feature accuracy alignment optimization based on dimensional representation and inverse recovery on the training control valve operation posterior feature vector to obtain an optimized training control valve operation posterior feature vector; a classification loss unit 470 is used to pass the optimized training control valve operation posterior feature vector through the classifier to obtain a classification loss function value; a model training unit 480 is used to train the preprocessor including the upsampling layer and the vector-to-image converter, the time series feature extractor based on the convolutional neural network model, the Bayesian probability model and the classifier based on the classification loss function value and through gradient descent direction propagation.

[0044] In particular, in the technical solution of the present application, the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector respectively express the time series correlation characteristics of the water level value, the pressure value and the temperature value with respect to the local time domain under the global time domain determined by vector-image conversion within the local time domain. Thus, since the Bayesian probability model is used to fuse the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector, a point-by-point Bayesian calculation is performed on the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector. Therefore, the obtained posterior feature vector of the regulating valve operation also has a multi-scale time series correlation representation of the time series parameter feature representation within the local time domain and between the local time domains. Therefore, when the posterior feature vector of the regulating valve operation is classified and regressed through a classifier, the multi-scale temporal correlation representation of the posterior feature vector of the regulating valve operation within the local time domain and between the local time domains will also affect the training effect of the posterior feature vector of the regulating valve operation when it is trained by the classifier due to the difference in correlation accuracy of the multi-time domain spatial scale correlation representation of the temporal correlation feature. Therefore, during the training process, the applicant of this application records the posterior feature vector of the regulating valve operation as, for example, Perform feature accuracy alignment based on dimensional representation and inversion recovery, specifically expressed as: ; is the posterior characteristic vector of the regulating valve operation No. The eigenvalues ​​at the positions, Represents the posterior feature vector of the regulating valve operation The zero norm of is the posterior characteristic vector of the regulating valve operation Length, and is a weight hyperparameter. Here, to address the accuracy conflict between high-dimensional feature space encoding of multi-dimensional time series parameter features based on multiple temporal spatial scales and the associated editing of multi-temporal spatial scale time series features, the feature accuracy alignment based on dimensional representation and inverse recovery is achieved by treating the associated editing of multi-temporal spatial scale time series features as an inverse embedding generation of high-dimensional feature space encoding based on the distribution of time series parameter features. This achieves adaptive alignment of accuracy differences during training by equipping the feature values ​​represented by the encoding with sparse distribution balancing of scale representations and performing inverse recovery of associated details based on vector counts. This improves the training effect of the control valve operation posterior feature vector when using a classifier for classification and regression training. In this way, the control valve opening value can be adaptively controlled based on the evaporator's operating status, thereby achieving real-time automatic control of the reactor nuclear power unit evaporator feedwater system to meet the feedwater flow and pressure requirements, ensuring the stability and safety of the feedwater system, and improving the safety and efficiency of the nuclear power plant.

[0045] As described above, the reactor nuclear power unit evaporator water feed system 300 according to an embodiment of the present application can be implemented in various wireless terminals, such as a server having a reactor nuclear power unit evaporator water feed algorithm. In one possible implementation, the reactor nuclear power unit evaporator water feed system 300 according to an embodiment of the present application can be integrated into a wireless terminal as a software module and / or hardware module. For example, the reactor nuclear power unit evaporator water feed system 300 can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the reactor nuclear power unit evaporator water feed system 300 can also be one of the many hardware modules of the wireless terminal.

[0046] Alternatively, in another example, the reactor nuclear power unit evaporator water feed system 300 and the wireless terminal may also be separate devices, and the reactor nuclear power unit evaporator water feed system 300 may be connected to the wireless terminal through a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0047] While various embodiments of the present disclosure have been described above, the above descriptions are illustrative, non-exhaustive, and not intended to be limiting of the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to existing technologies, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A reactor nuclear power unit evaporator water supply system, characterized in that: include: An evaporator data acquisition module is used to obtain the water level value, pressure value and temperature value of the evaporator at multiple predetermined time points within a predetermined time period; an evaporator data parameter time series arrangement module, for arranging the water level values, pressure values ​​and temperature values ​​of the evaporator at the plurality of predetermined time points into a water level value time series input vector, a pressure value time series input vector and a temperature value time series input vector according to the time dimension; an evaporator parameter timing characteristic analysis module, used for performing timing analysis on the water level value timing input vector, the pressure value timing input vector and the temperature value timing input vector respectively to obtain a water level value timing characteristic vector, a pressure value timing characteristic vector and a temperature value timing characteristic vector; An evaporator multi-parameter time series feature fusion module, used to fuse the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector to obtain an evaporator parameter time series collaborative fusion feature; The regulating valve opening control module is used to determine whether the valve opening value of the regulating valve at the current time point should be increased, decreased or maintained based on the evaporator parameter time series coordinated fusion characteristics.

2. The evaporator water supply system for a reactor nuclear power unit according to claim 1, characterized in that: The evaporator parameter time series characteristic analysis module includes: an evaporator parameter time series data preprocessing unit, used for respectively passing the water level value time series input vector, the pressure value time series input vector and the temperature value time series input vector through a preprocessor including an upsampling layer and a vector-to-image converter to obtain a water level value time series image, a pressure value time series image and a temperature value time series image; The evaporator multi-parameter time series feature extraction unit is used to perform feature extraction on the water level value time series image, the pressure value time series image and the temperature value time series image respectively through a time series feature extractor based on a deep neural network model to obtain the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector.

3. The evaporator water supply system for a reactor nuclear power unit according to claim 2, characterized in that: The deep neural network model is a convolutional neural network model.

4. The evaporator water supply system for a reactor nuclear power unit according to claim 3, characterized in that: The evaporator multi-parameter time series feature fusion module is used to: use a Bayesian probability model to fuse the water level value time series feature vector, the pressure value time series feature vector and the temperature value time series feature vector to obtain a regulating valve operation posterior feature vector as the evaporator parameter time series collaborative fusion feature.

5. The evaporator water supply system for a reactor nuclear power unit according to claim 4, characterized in that: The regulating valve opening control module is used to: pass the regulating valve operation posterior feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether the valve opening value of the regulating valve at the current time point should be increased, decreased or maintained.

6. The evaporator water supply system for a reactor nuclear power unit according to claim 5, characterized in that: It also includes a training module for training the preprocessor including the upsampling layer and the vector-to-image converter, the temporal feature extractor based on the convolutional neural network model, the Bayesian probability model and the classifier.

7. The evaporator water supply system for a reactor nuclear power unit according to claim 6, characterized in that: The training module comprises: A training data acquisition unit, used to obtain training water level values, training pressure values ​​and training temperature values ​​of the evaporator at multiple predetermined time points within a predetermined time period, and whether the valve opening value of the regulating valve at the current time point should be increased, decreased or maintained; A training data time series arrangement unit, used to arrange the training water level values, training pressure values ​​and training temperature values ​​of the evaporator at the plurality of predetermined time points into a training water level value time series input vector, a training pressure value time series input vector and a training temperature value time series input vector according to the time dimension respectively; A training data preprocessing unit, used for passing the training water level value time series input vector, the training pressure value time series input vector and the training temperature value time series input vector through the preprocessor including the upsampling layer and the vector-to-image converter to obtain a training water level value time series image, a training pressure value time series image and a training temperature value time series image; A training evaporator parameter time series feature extraction unit, used for respectively passing the training water level value time series image, the training pressure value time series image and the training temperature value time series image through the time series feature extractor based on the convolutional neural network model to obtain a training water level value time series feature vector, a training pressure value time series feature vector and a training temperature value time series feature vector; A training evaporator multi-parameter time series feature fusion unit is used to use the Bayesian probability model to fuse the training water level value time series feature vector, the training pressure value time series feature vector and the training temperature value time series feature vector to obtain a training regulating valve operation posterior feature vector; A feature distribution optimization unit, used for performing feature accuracy alignment optimization based on dimensional characterization and inverse recovery on the training control valve operation posterior feature vector to obtain an optimized training control valve operation posterior feature vector; A classification loss unit, used for passing the optimized training regulating valve operation posterior feature vector through the classifier to obtain a classification loss function value; A model training unit is used to train the preprocessor including the upsampling layer and the vector-to-image converter, the temporal feature extractor based on the convolutional neural network model, the Bayesian probability model and the classifier based on the classification loss function value and through the direction propagation of gradient descent.

8. The evaporator water supply system for a reactor nuclear power unit according to claim 7, characterized in that: The classification loss unit is used to: Using the classifier to process the optimized training regulating valve operation posterior feature vector to obtain a training classification result: and The cross entropy loss function value between the training classification result and the real value of the valve opening value of the regulating valve at the current time point that should be increased, decreased or maintained is calculated as the classification loss function value.

Images