Performance optimization and evaluation method and system for supercritical carbon dioxide nuclear power cycle system

Abstract

The invention belongs to the field of nuclear power circulation systems, and particularly discloses a supercritical carbon dioxide nuclear power circulation system performance optimization and evaluation method and system, and the method comprises the steps: preliminarily selecting optimization variables and performance indexes of a thermodynamic model of a supercritical carbon dioxide nuclear power circulation system; the preliminarily selected performance indexes comprise a thermodynamic index, a compactness index and an economic index, based on a thermodynamic model of the supercritical carbon dioxide nuclear power cycle system, analyzing the change trend of each performance index when an optimization variable changes, and only reserving one of the performance indexes with the same change trend; and constructing an objective function according to the reserved performance indexes, and solving the objective function based on the thermodynamic model to obtain an optimal solution about the optimization variables so as to realize system performance optimization. According to the invention, rapid performance optimization and evaluation of the supercritical carbon dioxide nuclear power cycle system under different loads can be realized.

Description

Technical Field

[0001] The present invention belongs to the field of nuclear power cycle systems, and more specifically, relates to a performance optimization and evaluation method and system for a supercritical carbon dioxide nuclear power cycle system. Background Art

[0002] The supercritical carbon dioxide (S-CO2) nuclear power cycle system uses the core as the direct heat source of the cycle, and S-CO2 serves as the reactor coolant and the flowing working fluid of the circulation loop. Its large density and drastic changes in physical properties near the critical point make the S-CO2 nuclear power cycle system have the advantages of high efficiency, compact structure, and environmental friendliness, bringing broad application prospects to the fields of ships, space, and vehicle power systems.

[0003] At present, the research on S-CO2 nuclear power cycle system mainly focuses on the overall system concept design, cycle construction analysis and key component development. At the same time, in the process of system performance evaluation and optimization target selection, the existing research mainly focuses on the power generation system with thermal performance and economic cost as the target, and the decision weights of each optimization target are mostly based on subjective judgment and experience, lacking objective weight assignment and decision-making methods based on sample data. For application scenarios such as ships, space, and vehicle power, it is necessary to consider the comprehensive performance optimization analysis of more dimensions such as mass, volume, efficiency, and specific power, and it is necessary to balance the weights of each dimension objectively. The introduction of these additional dimensions and the potential conflicts between the dimensions significantly increase the complexity and computational cost of the optimization problem. In addition, during the optimization process, when the optimization variables fluctuate within a large range, it is easy to encounter problems such as unstable solution of physical property parameters when entering the model calculation, which further increases the difficulty of analyzing different optimization schemes and brings great challenges to the entire optimization and performance evaluation. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides a performance optimization and evaluation method and system for a supercritical carbon dioxide nuclear power cycle system, the purpose of which is to achieve comprehensive performance optimization of the S-CO2 nuclear power cycle system and reduce the complexity and computational cost during optimization.

[0005] To achieve the above object, according to a first aspect of the present invention, a method for optimizing the performance of a supercritical carbon dioxide nuclear power cycle system is proposed, comprising the following steps:

[0006] Based on the thermodynamic model of the supercritical carbon dioxide nuclear power cycle system, the changing trends of various performance indicators when the optimization variables of the thermodynamic model are changed are analyzed. If the changing trends of several performance indicators are the same when all the optimization variables are changed, only one of the performance indicators is retained to complete the performance indicator screening; the performance indicators to be screened include thermal indicators, compact indicators and economic indicators, among which the thermal indicators include thermal efficiency, Efficiency, net output power and specific power; compactness indicators include system mass, system volume and unit power heat exchanger area; economic indicators include levelized cost and payback period;

[0007] The objective function is constructed according to the screened performance indicators, and the objective function is solved based on the thermodynamic model to obtain the optimal solution of the optimization variables and achieve system performance optimization.

[0008] As a further preferred method, solving the objective function based on the thermodynamic model to obtain the optimal solution of the optimization variable includes:

[0009] Based on the thermodynamic model, the objective function is solved by NSGA-Ⅱ to obtain the Pareto front solution set; the optimal solution is determined from the Pareto front solution set by the superior and inferior solution distance method.

[0010] As a further preferred method, the optimal solution is determined from the Pareto front solution set by using a superior-inferior solution distance method, including:

[0011] The performance indicator matrix corresponding to the Pareto frontier solution set is forward normalized to obtain a decision matrix, and the decision matrix is ​​weighted according to the indicator weight; then the proximity of the Pareto solution is calculated based on the weighted decision matrix, and the Pareto solution with the highest proximity is taken as the optimal solution; the indicator weight is determined by the CRITIC objective weighting method.

[0012] As a further preferred method, the optimization variable-performance index data is obtained in advance based on the thermodynamic model to construct a sample set; the neural network model is trained by the sample set to obtain a trained neural network model; and in the process of solving the objective function by NSGA-Ⅱ, the objective function value is determined by the trained neural network model.

[0013] As a further preference, the neural network model adopts a ReLU activation function and an Adam optimizer; when training the neural network model, a mean square error is used as a loss function.

[0014] According to a second aspect of the present invention, there is provided a performance optimization system for a supercritical carbon dioxide nuclear power cycle system, comprising a processor for executing the above-mentioned performance optimization method for a supercritical carbon dioxide nuclear power cycle system.

[0015] According to a third aspect of the present invention, there is provided a method for evaluating the performance of a supercritical carbon dioxide nuclear power cycle system, comprising the following steps:

[0016] Preset a series of system loads, including a rated load and multiple different non-rated loads;

[0017] For each set system load, the above-mentioned supercritical carbon dioxide nuclear power cycle system performance optimization method is used to obtain the corresponding optimal solution;

[0018] According to the optimal solutions corresponding to different system loads, the system mass flow corresponding to different system loads is calculated; the system load-mass flow is fitted to obtain the mass flow change rate when the system load changes, so as to evaluate and determine the ability of the supercritical carbon dioxide nuclear power cycle system to adapt to load changes.

[0019] As a further preference, under rated load, the optimized variables include the main compressor inlet pressure and temperature, the turbine inlet pressure and temperature, the total thermal conductivity of the heat exchanger, the thermal conductivity ratio of the low-temperature regenerator, the thermal conductivity ratio of the high-temperature regenerator, the mass flow split ratio, the low-temperature regenerator mass flux, the high-temperature regenerator mass flux, and the cooler mass flux.

[0020] As a further preference, under non-rated load, the optimization variables include the main compressor inlet pressure and the mass flow split ratio; and when determining the optimal solution under non-rated load, the values ​​of the compressor isentropic efficiency, turbine isentropic efficiency, low-temperature regenerator thermal conductivity, high-temperature regenerator thermal conductivity and cooler thermal conductivity in the thermodynamic model are determined according to the corresponding parameter values ​​and component characteristic curves under rated load.

[0021] According to a fourth aspect of the present invention, there is provided a performance evaluation system for a supercritical carbon dioxide nuclear power cycle system, comprising a processor for executing the above-mentioned performance evaluation method for a supercritical carbon dioxide nuclear power cycle system.

[0022] In general, the above technical solution conceived by the present invention has the following technical advantages compared with the prior art:

[0023] 1. The present invention provides a multi-objective optimization method for the S-CO2 nuclear power cycle system that considers comprehensive performance analysis in multiple dimensions, including thermal performance, compactness, and economy. Compared with the existing research on power generation systems that mainly focuses on thermal performance and economy, the present invention also designs a compactness index for evaluating performance for the nuclear power cycle system, thereby providing a more comprehensive system comprehensive performance optimization solution, which can lay the foundation for the design of nuclear power cycle systems such as ships or space that have high requirements for compactness and mass size. At the same time, for more performance indicators, the performance indicators are screened according to their changing trends when the optimization variables change, thereby reducing the complexity and calculation cost of subsequent target optimization and improving the overall performance of the S-CO2 nuclear power cycle system.

[0024] 2. After obtaining the Pareto frontier solution set, the optimal solution is determined by combining the CRITIC objective weighting method and the TOPSIS superior and inferior solution distance method. This can objectively balance the weights of the objectives in each dimension, ensure the objectivity and accuracy of the evaluation results as much as possible, and reduce the intervention of human factors and experience evaluation.

[0025] 3. The present invention replaces the initial thermodynamic model with a pre-trained neural network model to quickly predict the performance indicators of the system; the neural network model can greatly improve the calculation efficiency while ensuring the prediction accuracy, shorten the time cost of the optimization process, and make the optimization process of complex systems more efficient and convenient.

[0026] 4. The present invention has good universality and is suitable for system optimization and evaluation under different power and operating parameters, and can also be extended to other power cycles and systems. At the same time, the present invention covers the optimization of the system under design conditions (rated load) and non-design conditions, which can fully ensure that the system has good overall performance under various operating conditions, and can further effectively evaluate the system performance based on the system's mobility (ability to adapt to load changes). BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a cycle flow chart of a supercritical carbon dioxide nuclear power cycle system according to an embodiment of the present invention;

[0028] Figure 2 The present invention is a flowchart of a method for evaluating the performance of a supercritical carbon dioxide nuclear power cycle system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0030] The embodiment of the present invention provides a method for optimizing the performance of a supercritical carbon dioxide nuclear power cycle system, such as Figure 2 As shown, the following steps are included:

[0031] S1. Construct a thermodynamic model of the supercritical carbon dioxide nuclear power cycle system and preliminarily select optimization variables and performance indicators.

[0032] The recompression cycle configuration in the S-CO2 nuclear power cycle system in this embodiment is as follows: Figure 1 As shown, CO2 enters turbine C to generate electricity, enters high-temperature regenerator F to release heat after doing work, and then enters low-temperature regenerator E to continue releasing heat; one part of CO2 enters recompressor B to increase pressure, and the other part of CO2 enters cooler H to be cooled by external cooling medium, enters main compressor A for compression, and absorbs heat through low-temperature regenerator E; the two parts of working fluid are combined and enter high-temperature regenerator F to absorb heat, and the working fluid absorbs heat through reactor D.

[0033] Set the input parameters: system power P, main compressor A inlet pressure p1, main compressor A inlet temperature T1, turbine C inlet pressure p6, turbine C inlet temperature T6 (T0), cooler H cold side inlet pressure p 13 、Cooler H cold side inlet temperature T 13 , the total thermal conductivity UA of low temperature regenerator E / high temperature regenerator F / cooler H, the thermal conductivity ratio of low temperature regenerator E High temperature regenerator F thermal conductivity ratio Mass flow split ratio x, turbine isentropic efficiency η T , Isentropic efficiency η of main compressor A MC , isentropic efficiency η of recompressor B RC 、Low temperature regenerator E / high temperature regenerator F / cooler H / reactor D pressure loss pr、Low temperature regenerator E / high temperature regenerator F / cooler H mass flux G x .

[0034] The nonlinear equations are constructed by inputting the equations corresponding to the parameters to form a thermodynamic model. The main equations are listed as follows:

[0035] The power consumption of the turbine, main compressor and recompressor is calculated by the following formula:

[0036]

[0037] Where W T is the work done by the turbine, MW; W MC With W RC are the power consumption of the main compressor and the recompressor, MW; is the total mass flow rate of the cycle, kg / s; h i is the specific enthalpy of the working fluid at node i, kJ / kg; h i,s is the specific enthalpy under isentropic conditions, kJ / kg; η T , η MC With η RC are the isentropic efficiencies of the turbine, main compressor and recompressor, respectively, %; x is the split ratio.

[0038] For the heat exchanger, the logarithmic mean temperature difference method is used for modeling, and the formula is as follows:

[0039]

[0040] In the formula, Q LTR , Q HTR , Q Cooler Heat exchange of low and high temperature regenerator, high temperature regenerator and cooler, kW; UA Total is the total thermal conductivity, kW / K; is the distribution ratio of the total thermal conductivity, indicating the proportion of the low-temperature regenerator and the high-temperature regenerator in the total thermal conductivity UA; T i is the temperature of each node, K.

[0041] For the reactor, its model is regarded as a simple heat exchanger, and the heat absorbed by the loop is expressed as follows:

[0042]

[0043] For the pressure drop pr between the regenerator and the reactor, the value is the ratio of the component outlet pressure to the inlet pressure, that is, pr = p out / p in .

[0044] A thermodynamic model is constructed based on the above nonlinear equations. When the thermodynamic model is used subsequently, after obtaining the input parameter values, the Newton iteration method can be used to solve the characteristics of each component and the node temperature / pressure values, and then determine the values ​​of each performance indicator. The physical properties of CO2 and coolant are called by the open source physical property library CoolProp.

[0045] The optional performance indicators are classified into multiple dimensions, including thermal indicators, compactness indicators and economic indicators, as shown in Table 1. The indicators can be selected as needed to initially form the performance indicator system of the S-CO2 nuclear power cycle system.

[0046] Table 1 Multi-dimensional comprehensive performance evaluation index system

[0047]

[0048] S2. Performance index screening.

[0049] Analyze the impact of system optimization variables on performance indicators: First, use the control variable method to analyze the impact trend of each optimization variable to be optimized on performance indicators; that is, keep other parameters unchanged, change individual optimization variables separately, and observe their impact on system performance indicators.

[0050] Conduct conflict analysis between indicators, that is, compare the impact trend of the same variable on different indicators, and choose indicators with the same change trend. Specifically, when two or more performance indicators have the same change trend when all variables change, only one of the performance indicators is retained to reduce the optimization scale and form the final system performance indicator system.

[0051] S3. Build and solve the objective function based on the performance indicators to achieve system performance optimization.

[0052] The specific steps include:

[0053] S31. Calculation and analysis of objective weights of performance indicators.

[0054] The CRITIC objective weighting method is used to calculate the indicator weights. This method uses standard deviation and correlation coefficient to represent the volatility and correlation of indicators within the optimization variable range, thereby calculating the proportion of information contained in each indicator to represent its weight.

[0055] First, based on the thermodynamic model, m sample points are generated within the selected optimization variable range. Assuming there are n performance indicators, the sample matrix X is constructed as shown in formula (9).

[0056]

[0057] Where x ij Represents the value of the jth indicator of the ith sample.

[0058] In order to eliminate the influence of different dimensions on the calculation results, the sample matrix X is dimensionless processed to obtain the normalized matrix M, as shown in equations (10) and (11).

[0059]

[0060] Next, calculate the standard deviation s of each dimensionless indicator j And the correlation coefficient ρ ij , to show the differential fluctuation of the same indicator and the correlation between different indicators;

[0061]

[0062] In the formula, is the mean; ρ jk represents the correlation coefficient between the jth indicator and the kth indicator, cov(yj ,y k ) represents the covariance between the j-th index and the k-th index.

[0063] The correlation coefficient is used to represent the conflict between indicators. The stronger the correlation with other indicators, the smaller the conflict between the indicator and other indicators. This reflects that the more the same information is, the higher the repetitiveness of the evaluation content can be reflected, which weakens the evaluation strength of the indicator to a certain extent, thereby reducing the weight of the indicator.

[0064] Finally, the weight value w of the indicator is obtained by calculating the proportion of information contained in each indicator. j , as shown in formula (14) and formula (15), the weights of all indicators w j The indicator weight matrix W is formed.

[0065]

[0066] In the formula, I j Represents the amount of information contained in target j.

[0067] S32. Multi-objective optimization decision analysis.

[0068] Multi-objective genetic algorithm and multi-criteria decision-making method are used for optimization and decision-making. Different from single-objective optimization problems, multi-objective optimization problems involve conflicts and incommensurability between multiple objective functions, so their optimal solution is usually a set of solutions, rather than a unique solution. This solution set is called the Pareto front, which contains a series of non-dominated solutions. To obtain the Pareto front, perform the following steps:

[0069] First, define the optimization problem, including the objective function, constraints, and the range of optimization variables. Taking three objectives as an example, the objective function can be expressed as:

[0070] minF(X)=[max(F1),max(F2),min(F3)]

[0071] =[min(-F1),min(-F2),min(F3)]#(16)

[0072] In the formula, the selected indicators F1 and F2 are positive indicators, such as thermal efficiency, output power, etc.; F3 is a negative indicator, such as mass, volume, cost, etc.

[0073] Secondly, an algorithm is selected, preferably a fast non-dominated sorting genetic algorithm (NSGA-Ⅱ) based on an elite selection strategy, and appropriate algorithm parameters are selected, including the population size and the number of optimization iterations; then the population is initialized and a population object is created, which contains a series of possible solutions (individuals), which will gradually evolve during the optimization process; a genetic algorithm is used to iteratively optimize the population to obtain a Pareto front solution set.

[0074] S33. Select the best feasible solution from the Pareto front.

[0075] In actual engineering, usually only one or several sets of parameters can be selected as design parameters, so the best feasible solution must be selected from the Pareto front. The superior and inferior solution distance method is used for multi-criteria decision making. According to the degree of proximity between a limited number of non-dominated solutions and the ideal solution, if this non-dominated solution is closest to the positive ideal solution and farthest from the negative ideal solution, it is the best, otherwise it is the worst. The calculation of the proximity function is as follows:

[0076] First, the indicator matrix corresponding to the optimized Pareto frontier is forward normalized; secondly, the indicator weight matrix W obtained in step S31 is used to weight the normalized matrix M to obtain the decision matrix G, as shown in formula (17):

[0077] G=W·M#(17)

[0078] Then calculate each group of solutions G in the decision matrix G ij Distance to positive ideal solution and negative ideal solution As shown in formula (18):

[0079]

[0080] In the formula, represents the positive / negative ideal solution, that is, the best and worst hypothetical solutions under the jth performance indicator.

[0081] Finally, the proximity between the alternatives and the ideal point in the decision matrix is ​​calculated and ranked, as shown in formula (19):

[0082]

[0083] Where n is the number of optimization objectives; m is the number of optimal individuals (samples) in the Pareto solution set after multi-objective optimization; is the distance from the i-th sample point to the positive ideal solution and the negative ideal solution; M is the normalized matrix.

[0084] Select proximity C i The highest Pareto solution is taken as the optimal decision design point.

[0085] Furthermore, a neural network model is used to replace the initial thermodynamic model by training, testing and verifying the model, and then used for objective index calculation analysis and multi-objective optimization. That is, when the objective function is solved by a genetic algorithm in step S3, the optimization variables are passed into the pre-trained neural network model, the neural network model outputs the corresponding index values, and then these results are passed into the optimization algorithm for evolution. With the iteration of the optimization algorithm, the generated new scheme (i.e., the new combination of optimization variables) is passed into the neural network model again for prediction. This process is cyclical, and information is transmitted back and forth between the neural network model and the optimization algorithm, and gradually evolves to obtain the Pareto frontier solution set.

[0086] The pre-training neural network model includes the following steps:

[0087] First, prepare the sample set. Based on the initial thermodynamic model and the performance evaluation index system, randomly combine the selected optimization variables within the range, and use the calculation of the thermodynamic model to generate a sample set containing a large number of sample points. Divide the sample set into training set, test set and validation set. Usually, the training set accounts for 70% of the total number of samples and is used to train the neural network model; the test set and validation set each account for 15%, which are used to detect the effect of the model in real time during the training process and to finally verify and evaluate the performance of the model. Such a division can ensure the stability and generalization ability of model training.

[0088] Secondly, the neural network model is constructed, including designing the network architecture and setting hyperparameters. The network architecture design involves the number of neural network layers and the number of neuron nodes in each layer, including the configuration of the input layer, hidden layer, and output layer. Among them, the optimization variables selected in the performance evaluation index system are used as the input of the input layer, and the evaluation index is used as the output of the neural network output layer. The number of nodes corresponds to the dimensions of the selected optimization variables and evaluation indicators respectively; the number of layers and nodes of the hidden layer will affect the learning and expression ability of the network. According to the scale and complexity of the data, multiple models are gradually set up from small-scale networks to large-scale networks, and the optimal network architecture that takes into account both training speed and accuracy is determined through the final training effect. In terms of hyperparameter setting, activation function, optimization algorithm, learning rate, and weight and bias initialization methods all play an important role. The ReLU activation function is selected because of its excellent performance in deep networks; the Adam optimizer is widely used because of its high efficiency in dealing with non-stationary objective functions; the learning rate is usually set to 1e-3 to balance the convergence speed and accuracy of the model; weight initialization usually uses He initialization to adapt to the characteristics of the ReLU activation function, thereby ensuring the stability and efficiency of model training.

[0089] Next, start training the neural network model. Input the preprocessed training set into the neural network model, and adjust the weights and biases of each layer through the back-propagation algorithm to minimize the loss function. In regression tasks, the mean square error (MSE) is usually selected as the loss function. During the training process, the performance of the model will gradually improve with each round of iteration, but in order to prevent the model from overfitting, it is necessary to consider early stopping or introduce regularization technology. Early stopping monitors the loss changes through the test set to prevent the model from performing too well on the training set and performing poorly on the test set. Regularization techniques, such as L2 regularization or Dropout, can effectively reduce the complexity of the model and enhance the generalization ability of the model.

[0090] Finally, evaluate the neural network model after training. By evaluating the performance of the model on the validation set, you can determine the generalization ability and actual application effect of the model. Common evaluation indicators include mean square error (MSE), accuracy, F1 score, etc. According to the evaluation results, if the model performance is not ideal, you need to readjust the model architecture or hyperparameters, or even increase the training data, to further optimize the model performance.

[0091] The embodiment of the present invention also provides a performance evaluation method for a supercritical carbon dioxide nuclear power cycle system, comprising the following steps:

[0092] A series of system loads are preset, namely, the system power P in the thermodynamic model; the system load includes rated load and non-rated load. In this embodiment, the non-rated load includes 90% rated load, 80% rated load, 70% rated load, 60% rated load, 50% rated load, 40% rated load, 30% rated load, 20% rated load and 10% rated load.

[0093] For each set system load, the above performance optimization method is used to obtain the corresponding optimal solution, thereby obtaining the optimal solution corresponding to each preset system load; preferably, when determining the optimal solution corresponding to the non-rated load, the same performance indicators as those for the rated load can be directly used.

[0094] According to the optimal solution corresponding to different system loads, the system mass flow corresponding to different system loads is calculated through the thermodynamic model; the system load-mass flow is fitted to determine the ability of the supercritical carbon dioxide nuclear power cycle system to adapt to load changes. Specifically, the mass flow change rate of unit percentage load change after fitting is used to evaluate the system's ability to adapt to load changes, thereby completing the system's variable load evaluation.

[0095] Specifically:

[0096] In the actual operation of the S-CO2 nuclear power cycle system, the system may often be in non-design conditions such as partial load. In this case, the performance of components such as compressors, turbines, and heat exchangers will change with the flow rate of the working fluid flowing through them. Therefore, when analyzing the non-design conditions of the system, it is necessary to consider the characteristic curves of each component and select appropriate optimization parameters to optimize and evaluate the non-design state of the system and further improve the global performance of the system. Therefore, when determining the optimal solution corresponding to different system loads, the steps for selecting optimization variables and non-design condition parameters are as follows:

[0097] (1) The optimal decision point corresponding to the rated load is taken as the rated design condition and used as the benchmark for the non-design state. Under the rated load, for the comprehensive performance of the entire system, the selected optimization variables include: main compressor inlet pressure p1, main compressor inlet temperature T1, turbine inlet pressure p6, turbine inlet temperature T6 (T0), total thermal conductivity UA of the three heat exchangers (low-temperature regenerator, high-temperature regenerator, cooler), and thermal conductivity ratio of the low-temperature regenerator. High temperature regenerator F thermal conductivity ratio Mass flow split ratio x, low temperature regenerator mass flux G LTR , high temperature regenerator mass flux G HTR and the cooler mass flux G Cooler The remaining thermodynamic model input parameters can be set based on experience.

[0098] (2) For each non-rated load (i.e., non-design operating condition), the selected optimization variables include: the main compressor inlet pressure p1 and the mass flow split ratio x.

[0099] At the same time, the non-design operating parameters are set, including the isentropic efficiency of the compressor / turbine, the thermal conductivity of the heat exchanger, the pressure drop of the heat exchanger, and the turbine inlet pressure. The calculation method is as follows:

[0100] Isentropic efficiency η of compressor / turbine under off-design conditions C / T,off and heat exchanger thermal conductivity UA LTR / HTR / Cooler,off The values ​​of are based on the relative size of the system mass flow under the non-design condition and the rated design condition. The corresponding size relationship between the non-design condition parameters is obtained through the characteristic curve of the component, which is shown in equations (20) and (21):

[0101] η C / T,off =η C / T ·f C / T (x m )#(20)

[0102] UA LTR / HTR / Cooler,off =UA LTR / HTR / Cooler ·g(x m )#(twenty one)

[0103] Where η C / T Indicates the isentropic efficiency of the compressor and turbine under rated design conditions, %; UA LTR / HTR / Cooler Indicates the thermal conductivity of the heat exchanger (low temperature regenerator / high temperature regenerator / cooler) under rated design conditions, kW / K; f C / T (x m ) and g(x m ) represent the values ​​obtained from the compressor / turbine and heat exchanger characteristic curves, respectively, x m It represents the ratio of the system mass flow rate under non-design conditions to the rated design conditions, that is, x m =m off / m.

[0104] For non-design point conditions, the pressure drop of the heat exchanger is variable. The friction coefficient f, which is independent of the heat exchanger pipe size, is used to consider the changes in excess resistance of the system under different conditions, so as to better simulate the pressure drop of the heat exchanger in actual operation. The calculation formula of the friction coefficient f is as follows:

[0105]

[0106] In the formula, Δp represents the change in the inlet and outlet pressure of the heat exchanger, Pa; m is the mass flow rate of the working fluid, kg / s; v is the average specific volume of the working fluid at the inlet and outlet of the heat exchanger, m 3 / kg.

[0107] The Stodola law is used to describe the relationship between the mass flow rate into the turbine and the inlet pressure, which can be expressed as:

[0108]

[0109] In the formula, m in,off 、m in They represent the mass flow rate of the working fluid flowing through the turbine under non-design conditions and design conditions, kg / s; p in,off 、p out,off Respectively represent the inlet and outlet pressure values ​​of the turbine under non-design conditions, MPa; p in 、p out Respectively represent the inlet and outlet pressure values ​​of the turbine under design conditions, MPa; v in,off 、v in are the specific volumes at the turbine inlet under off-design and design conditions, respectively, m 3 / kg.

[0110] For other thermodynamic model parameters that are not optimization variables and non-design operating condition parameters, the same parameter values ​​as those under rated design conditions are directly adopted.

[0111] The embodiment of the present invention also provides a performance evaluation visualization platform for an S-CO2 nuclear power cycle system, which specifically includes the following steps:

[0112] The first step is interface design and layout: use Qt tools to arrange buttons, input boxes, context menus and other controls. The main window provides an overview interface to guide users to select the required functional modules; secondly, design the sub-window interface of each functional module, including the loop configuration selection area, loop flow chart demonstration area, input parameter area, calculation / reset / return buttons, result display area, etc.;

[0113] The second step is to build functional modules and connect signals: for each functional module, build corresponding classes, functions and parameter transfer interfaces according to its method flow, and use Qt's signal and slot mechanism to connect the controls on the interface with the corresponding methods of the functional modules to realize the interaction between the interface and the functions;

[0114] The third step is status monitoring and fault handling: considering the errors and faults that may be triggered when users use the visualization platform, a pop-up window is used to remind users of system problems and provide corresponding handling suggestions, such as unreasonable input parameters, calculation divergence, etc.; secondly, in the typical working condition multi-objective optimization decision-making and circulating system variable load evaluation and analysis module, the key to ensuring that the system interface responds to user operations and avoids the long optimization time of the genetic algorithm blocking the main thread is to use multi-threading for algorithm optimization, specifically: create a sub-thread and execute the calculation task in the sub-thread; secondly, display the optimization progress bar and progress information on the interface, and pop up a window to display the optimization results of each generation of the population during the optimization execution, so that users can understand the dynamic progress of the optimization process.

[0115] The fourth step is testing and packaging. According to the process steps of cycle system performance calculation, input variable and performance indicator screening, multi-objective optimization decision-making under typical operating conditions, and variable load evaluation and analysis, the integrated system is tested and packaged to ensure the normal use of each function, thereby completing the development of a comprehensive performance evaluation visualization integrated system for the S-CO2 nuclear power cycle system.

[0116] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing the performance of a supercritical carbon dioxide nuclear power cycle system, characterized in that: The steps include: Based on the thermodynamic model of the supercritical carbon dioxide nuclear power cycle system, the changing trends of various performance indicators when the optimization variables of the thermodynamic model are changed are analyzed. If the changing trends of several performance indicators are the same when all the optimization variables are changed, only one of the performance indicators is retained to complete the performance indicator screening; the performance indicators to be screened include thermal indicators, compact indicators and economic indicators, among which the thermal indicators include thermal efficiency, Efficiency, net output power and specific power; compactness indicators include system mass, system volume and unit power heat exchanger area; economic indicators include levelized cost and payback period; The objective function is constructed according to the screened performance indicators, and the objective function is solved based on the thermodynamic model to obtain the optimal solution of the optimization variables and achieve system performance optimization.

2. The performance optimization method of the supercritical carbon dioxide nuclear power cycle system according to claim 1, characterized in that: Solve the objective function based on the thermodynamic model to obtain the optimal solution of the optimization variables, including: Based on the thermodynamic model, the objective function is solved by NSGA-Ⅱ to obtain the Pareto front solution set; the optimal solution is determined from the Pareto front solution set by the superior and inferior solution distance method.

3. The performance optimization method of the supercritical carbon dioxide nuclear power cycle system according to claim 2, characterized in that: The optimal solution is determined from the Pareto frontier solution set by the superior and inferior solution distance method, including: The performance indicator matrix corresponding to the Pareto frontier solution set is forward normalized to obtain a decision matrix, and the decision matrix is ​​weighted according to the indicator weight; then the proximity of the Pareto solution is calculated based on the weighted decision matrix, and the Pareto solution with the highest proximity is taken as the optimal solution; the indicator weight is determined by the CRITIC objective weighting method.

4. The performance optimization method of the supercritical carbon dioxide nuclear power cycle system according to claim 2, characterized in that: The optimization variable-performance index data is obtained in advance based on the thermodynamic model to construct a sample set; the neural network model is trained through the sample set to obtain a trained neural network model; and in the process of solving the objective function through NSGA-Ⅱ, the objective function value is determined through the trained neural network model.

5. The performance optimization method of the supercritical carbon dioxide nuclear power cycle system according to claim 4, characterized in that: The neural network model adopts the ReLU activation function and the Adam optimizer; when training the neural network model, the mean square error is used as the loss function.

6. A performance optimization system for a supercritical carbon dioxide nuclear power cycle system, characterized in that: It comprises a processor, wherein the processor is used to execute the performance optimization method of the supercritical carbon dioxide nuclear power cycle system as described in any one of claims 1 to 5.

7. A method for evaluating the performance of a supercritical carbon dioxide nuclear power cycle system, characterized in that: The steps include: Preset a series of system loads, including a rated load and multiple different non-rated loads; For each set system load, the performance optimization method of the supercritical carbon dioxide nuclear power cycle system as described in any one of claims 1 to 5 is used to obtain a corresponding optimal solution; According to the optimal solutions corresponding to different system loads, the system mass flow corresponding to different system loads is calculated; The system load-mass flow rate is fitted to obtain the mass flow rate change rate when the system load changes, so as to evaluate the ability of the supercritical carbon dioxide nuclear power cycle system to adapt to load changes.

8. The performance evaluation method of the supercritical carbon dioxide nuclear power cycle system according to claim 7, characterized in that: Under rated load, the optimization variables include the main compressor inlet pressure and temperature, turbine inlet pressure and temperature, total thermal conductivity of the heat exchanger, thermal conductivity ratio of the low-temperature regenerator, thermal conductivity ratio of the high-temperature regenerator, mass flow split ratio, low-temperature regenerator mass flux, high-temperature regenerator mass flux, and cooler mass flux.

9. The performance evaluation method of the supercritical carbon dioxide nuclear power cycle system according to claim 7, characterized in that: Under non-rated load, the optimization variables include the main compressor inlet pressure and the mass flow split ratio; and when determining the optimal solution under non-rated load, the values ​​of the compressor isentropic efficiency, turbine isentropic efficiency, low-temperature regenerator thermal conductivity, high-temperature regenerator thermal conductivity and cooler thermal conductivity in the thermodynamic model are determined according to the corresponding parameter values ​​and component characteristic curves under rated load.

10. A performance evaluation system for a supercritical carbon dioxide nuclear power cycle system, characterized in that: It includes a processor, which is used to execute the performance evaluation method of the supercritical carbon dioxide nuclear power cycle system as described in any one of claims 7 to 9.

Images