CN-121997526-A - Performance prediction method and system for two-stage compression refrigeration cycle system

CN121997526ACN 121997526 ACN121997526 ACN 121997526ACN-121997526-A

Abstract

The application provides a performance prediction method and system of a two-stage compression refrigeration system, and relates to the technical field of refrigeration system performance prediction. The method comprises the steps of constructing a multi-physical field coupling simulation model of a two-stage compression refrigeration cycle, obtaining a sample set of input parameters through Latin hypercube sampling, inputting the sample set into the multi-physical field coupling simulation model, obtaining refrigeration power and performance coefficients of a corresponding sample set in a stable refrigeration process, combining the refrigeration power and the performance coefficients with the sample set to obtain a prediction data set, training the prediction data set and a neural network to obtain a prediction model, optimizing the prediction model by utilizing a particle swarm optimization algorithm, inputting real-time input parameters into the optimized prediction model, and predicting the refrigeration power and the performance coefficients of the two-stage compression refrigeration cycle system. The application improves the accuracy of the refrigeration power and the performance coefficient prediction, and ensures that the performance of the two-stage compression refrigeration cycle in a wide working condition range is accurately predicted and estimated.

Inventors

WANG ZHIHENG
YUAN JINGZHI
LI GUOHAO
Deng Jinshuo
XI GUANG

Assignees

西安交通大学

Dates

Publication Date: 20260508
Application Date: 20251202

Claims (10)

1. A performance prediction method for a two-stage compression refrigeration cycle system, comprising: constructing a multi-physical field coupling simulation model of the two-stage compression refrigeration cycle according to the structure of the two-stage compression refrigeration cycle and the energy transmission flow of the two-stage compression refrigeration cycle; Uniformly sampling from the numerical value input range of the integral input parameters by Latin hypercube sampling to obtain a sample set of the input parameters; Inputting the sample set into the multi-physical field coupling simulation model to obtain the refrigeration power and the performance coefficient of the two-stage compression refrigeration cycle system corresponding to the sample set in the stable refrigeration process; combining the refrigeration power and the performance coefficient with the sample set to construct a prediction data set; Training by using the prediction data set and the neural network, and constructing to obtain the prediction model, wherein the prediction model is used for predicting the refrigeration power and the performance coefficient of the two-stage compression refrigeration cycle system under the condition of inputting parameters in real time; optimizing the prediction model by using a particle swarm optimization algorithm according to the prediction data set; And inputting the real-time input parameters into the optimized prediction model, and predicting the refrigeration power and the performance coefficient of the two-stage compression refrigeration cycle system.
2. The performance prediction method according to claim 1, wherein constructing a multi-physical field coupling simulation model of the two-stage compression refrigeration cycle according to a structure of the two-stage compression refrigeration cycle and an energy transmission flow thereof comprises: Determining various components required in the two-stage compression refrigeration cycle system, wherein the various components comprise an evaporator, a low-pressure compressor, a high-pressure compressor, a condenser, a subcooler and an expansion valve; designing an energy transmission flow based on each component, and further determining the connection relation of each component to obtain the structure of the two-stage compression refrigeration cycle system; establishing a mathematical model of each component; Based on the structure of the two-stage compression refrigeration cycle system and the energy transmission flow, the multi-physical field coupling simulation model is constructed by using a simulation technology.
3. The performance prediction method according to claim 2, wherein in the process of constructing the multi-physical field coupling simulation model, each component in the two-stage compression refrigeration cycle system is regarded as an independent control body, the energy transmission flow is simulated through parameter transmission between each independent control body, coupling simulation is realized, and physical parameters of the refrigerant are obtained in real time through a refplop software interface.
4. The method of claim 1, wherein uniformly sampling from the numerical input range of the overall input parameter by latin hypercube sampling, obtaining a sample set of the input parameter, comprises: combining the requirements of a preset standard on the maximum load working condition and the low-temperature working condition of the vapor compression cycle, and determining the numerical value input range of the integral input parameters; Based on the numerical value input range of the integral input parameter, uniformly adopting the numerical value input range of the integral input parameter by utilizing Latin hypercube sampling to obtain a sample set of the input parameter; the integral input parameters comprise chilled water mass flow, chilled water temperature, low-pressure compressor rotating speed, low-pressure compressor guide vane opening, cooling water mass flow and cooling water temperature.
5. The performance prediction method according to claim 1, wherein inputting the sample set into the multi-physical-field-coupling simulation model, obtaining the refrigeration power and the performance coefficient of the two-stage compression refrigeration cycle system corresponding to the sample set in a stable refrigeration process, comprises: Inputting the sample set into the multi-physical field coupling simulation model, and outputting the current refrigeration power and performance coefficient of the two-stage compression refrigeration cycle system by the multi-physical field coupling simulation model; And continuously operating the multi-physical field coupling simulation model, and determining the result output when the multi-physical field coupling simulation model operates stably as the refrigeration power and the performance coefficient of the two-stage compression refrigeration cycle system corresponding to the sample set in the stable refrigeration process.
6. The method of claim 1, wherein combining the refrigeration power and coefficient of performance with the sample set to construct a predicted data set comprises: Combining the refrigeration power and the performance coefficient with a sample set of the input parameters, wherein the sample set comprises six inputs including refrigeration power and performance coefficients corresponding to the six inputs, namely refrigeration water mass flow, refrigeration water temperature, low-pressure compressor rotating speed, low-pressure compressor guide vane opening, refrigeration water mass flow and refrigeration water temperature; Based on the combination, the six inputs and the refrigeration power and performance coefficients corresponding to the six inputs are stored as Excel files, wherein A is the number of a predicted data set, B-G is the input, H, I is the output, and the formed predicted data set is in an eight-dimensional shape.
7. The performance prediction method according to claim 6, wherein training using the prediction dataset and a neural network, constructing the prediction model, comprises: based on the prediction data set, the prediction model is constructed by utilizing a BP neural network, wherein the BP neural network comprises an input layer, an hidden layer and an output layer, and the construction of the prediction model comprises the following specific steps: the predicted data set is disturbed, a training set and a testing set are divided according to a set proportion, wherein the six inputs are input, and the six inputs correspond to refrigeration power and performance coefficients; normalizing the training data set and the test data set; setting weight coefficients of refrigeration power and performance coefficients, selecting the node number with the minimum mean square error of a training set as an optimal value by traversing the experience range of hidden layer nodes, and carrying out hidden layer node optimization: Training the BP neural network based on the input layer after normalization processing and the hidden layer after hidden layer node optimizing; predicting and inversely normalizing the test set by using the trained BP neural network to obtain the prediction model; The expression for selecting the node number with the minimum mean square error of the training set as the optimal value by combining the weight coefficients is as follows: In the above-mentioned method, the step of, Respectively represent the weight coefficients of the refrigeration power and the performance coefficient, Representing the total number of samples, Representing the corresponding real values in the training set of refrigeration power and coefficient of performance, A predicted value indicating the cooling power and the coefficient of performance.
8. The method of claim 6, wherein optimizing the predictive model using a particle swarm optimization algorithm based on the predictive dataset comprises: Enabling each particle to represent a weight and a threshold value of a group of BP neural networks, and simulating the motion of the particle to perform optimizing; Setting PSO population, iteration number and the weight coefficient, and setting the initial position of particles through random initialization, so as to calculate fitness according to a fitness function, wherein the optimal position of an initial individual is the position of the initial individual, and the fitness function is a double-output mean square error between predicted output and a true value; searching the calculated particle fitness, finding a particle index with the minimum fitness, and setting the particle index as a global optimal position; Traversing each particle, and updating the speed and the position of the particle through inertia, individual cognition and social cognition; updating the initial individual optimal position and the global optimal position by calculating the fitness of the current position and comparing with the history record; and when the precision requirement is met or the iteration times are maximum, the iteration is stopped, and the obtained optimal weight and optimal threshold value are substituted into the BP neural network for training.
9. The method of claim 8, wherein the PSO population is 10; The number of iterations is 50; The weight coefficient is as follows: ; the individual cognition and the social cognition are both set to 2; the inertia is set to 0.9.
10. A two-stage compression refrigeration cycle system, characterized in that the two-stage compression refrigeration cycle system performs performance prediction by using the performance prediction method of the two-stage compression refrigeration cycle system according to any one of claims 1 to 9, and the two-stage compression refrigeration cycle system is a one-time throttling and intermediate incomplete cooling two-stage compression refrigeration cycle system, and comprises an evaporator, a low-pressure compressor, a high-pressure compressor, a condenser, a subcooler and an expansion valve; the evaporator is respectively connected with the low-pressure compressor, the expansion valve and the subcooler, and the low-pressure compressor is also connected with the subcooler; the high-pressure compressor is respectively connected with the subcooler and the condenser, the condenser is also connected with the subcooler, and the subcooler is also connected with the expansion valve; the low-pressure steam from the evaporator firstly enters the low-pressure compressor, is compressed to an intermediate pressure pm, is mixed with saturated steam from the subcooler in a pipeline, then enters the high-pressure compressor, is further compressed to a condensing pressure pk, and then enters the condenser to be condensed into liquid; The liquid flowing out from the condenser is divided into two paths, wherein one path of liquid flows through a coil pipe in the subcooler, is supercooled by refrigerant liquid outside the coil pipe and flows to the expansion valve, is throttled by the expansion valve to the evaporation pressure p0 and then enters the evaporator to be evaporated to generate refrigeration effect, the other path of liquid is throttled by the air supplementing valve to the middle pressure pm and enters the subcooler to be evaporated to generate saturated steam, the high-pressure liquid in the coil pipe is supercooled, and the saturated steam is mixed with low-pressure steam from the low-pressure compressor and then enters the second stage.

Description

Performance prediction method and system for two-stage compression refrigeration cycle system Technical Field The application relates to the technical field of performance prediction of refrigeration systems, in particular to a performance prediction method of a two-stage compression refrigeration system and the two-stage compression refrigeration system. Background The design of the refrigerating system generally only depends on the rated working condition to make single-point design, and better performance can be achieved at the design point. However, when the water chiller runs under the non-rated working condition, because the system has the characteristics of multiple loops, nonlinearity and strong coupling, the matching among all the components is difficult to achieve accurate matching, and the high-efficiency running of the water chiller is influenced. The existing refrigeration system has large energy consumption and insufficient regulation precision during operation, and has remarkable energy-saving potential. If the operation control of the refrigeration system is required to be optimized and adjusted timely and effectively, the accurate prediction of the performance of the refrigeration system is required to be realized first. For realizing accurate prediction of the performance of the refrigeration system, the current commonly used prediction methods include a physical modeling method and a data driving method. The physical modeling method relies on thermodynamic principles to perform energy consumption modeling and analysis, so that prediction logic is difficult to find for a more complex system such as two-stage compression refrigeration, application difficulty is high, and the data driving method does not depend on the internal structure and mechanism of the system, predicts through historical data, and is suitable for a more complex refrigeration system. However, the existing data driving method depends on real-time monitoring data of the refrigerating system, the sampling period is long, the monitoring working condition range is narrow, and accurate prediction of the whole working condition wide range of the refrigerating system is difficult. Disclosure of Invention In view of the above problems, the present application provides a performance prediction method of a two-stage compression refrigeration system and a two-stage compression refrigeration system, so as to solve the technical problem that the existing method lacks of performing accurate prediction for the rapid prediction and the full-working-condition wide range of the refrigeration system. In a first aspect, an embodiment of the present application provides a performance prediction method of a two-stage compression refrigeration cycle system, including: constructing a multi-physical field coupling simulation model of the two-stage compression refrigeration cycle according to the structure of the two-stage compression refrigeration cycle and the energy transmission flow of the two-stage compression refrigeration cycle; Uniformly sampling from the numerical value input range of the integral input parameters by Latin hypercube sampling to obtain a sample set of the input parameters; Inputting the sample set into the multi-physical field coupling simulation model to obtain the refrigeration power and the performance coefficient of the two-stage compression refrigeration cycle system corresponding to the sample set in the stable refrigeration process; combining the refrigeration power and the performance coefficient with the sample set to construct a prediction data set; Training by using the prediction data set and the neural network, and constructing to obtain the prediction model, wherein the prediction model is used for predicting the refrigeration power and the performance coefficient of the two-stage compression refrigeration cycle system under the condition of inputting parameters in real time; optimizing the prediction model by using a particle swarm optimization algorithm according to the prediction data set; And inputting the real-time input parameters into the optimized prediction model, and predicting the refrigeration power and the performance coefficient of the two-stage compression refrigeration cycle system. Optionally, constructing a multi-physical field coupling simulation model of the two-stage compression refrigeration cycle according to the structure of the two-stage compression refrigeration cycle and the energy transmission flow thereof, including: Determining various components required in the two-stage compression refrigeration cycle system, wherein the various components comprise an evaporator, a low-pressure compressor, a high-pressure compressor, a condenser, a subcooler and an expansion valve; designing an energy transmission flow based on each component, and further determining the connection relation of each component to obtain the structure of the two-stage compression refrigeration cycle system; establ