A thermodynamic coupling optimization and parameter matching method and system for a CO2 hybrid heat pump drying system

Abstract

This invention discloses a thermodynamic coupling optimization and parameter matching method and system for a CO2 composite heat pump drying system, belonging to the field of refrigeration and heat pump technology. The method first collects and preprocesses system operating parameters in real time, then inputs the preprocessed parameters into a multi-layer perceptron neural network model to minimize system energy consumption and product outlet moisture content as multi-objective functions. Next, the SPEA2 algorithm is used to iteratively optimize the multi-objective functions, and the trained neural network model calculates the objective function value corresponding to each set of parameters. After population initialization, constraint processing, non-dominated sorting, environment selection, and crossover mutation, the Pareto optimal solution set is obtained. Finally, the optimal combination of objective functions and corresponding operating parameters are selected according to actual production needs. This invention achieves multi-objective collaborative optimization of system energy consumption and drying quality, significantly improving energy utilization efficiency and system adaptability.

Description

Technical Field

[0001] This invention belongs to the field of refrigeration and heat pump technology, and relates to a thermodynamic coupling optimization and parameter matching method and system for a CO2 composite heat pump drying system. Background Technology

[0002] With the global promulgation of policies promoting energy conservation, emission reduction, and green manufacturing, the adoption of green and efficient energy conversion technologies in industrial production is explicitly encouraged, driving the transformation and upgrading of traditional high-energy-consuming industries. As a major energy consumer, the drying industry urgently needs technological innovation, and seeking more energy-efficient and environmentally friendly methods has become an inevitable trend in its development.

[0003] Traditional drying methods, such as electric heating and coal-fired heating, have many drawbacks. Electric heating relies entirely on the conversion of electrical energy into heat energy, resulting in low energy efficiency and high operating costs. While coal-fired heating is low-cost, it emits large amounts of pollutants, causing serious environmental pollution and failing to meet current environmental protection requirements. Heat pump technology, as a new type of energy-saving technology, transports energy from a low-temperature heat source to a high-temperature heat source by consuming a small amount of electricity, achieving highly efficient energy utilization. Compared to traditional drying methods, it features lower energy consumption, lower drying costs, and less environmental pollution, gradually becoming a key development direction for drying method transformation. However, single working fluids have significant limitations in meeting the demands of high-temperature drying. Due to the fixed boiling point and critical temperature of the working fluid, the evaporation pressure of the system becomes excessively high under high-temperature conditions, leading to increased compressor load. This not only increases the operating load of the equipment but also significantly reduces the performance of the entire system. As drying processes increasingly demand higher temperature ranges and energy efficiency, single working fluids can no longer meet diverse production needs.

[0004] Currently, to balance low-temperature performance and environmental requirements, CO2 composite heat pump drying systems, composed of R134a and CO2, have found some application in the drying field. These systems combine an R134a heat pump system with a CO2 heat pump system, specifically integrating an R134a evaporator with a CO2 gas cooler. This system combines the excellent low-temperature performance of R134a with the environmentally friendly and high-temperature advantages of CO2, broadening the temperature adaptability range and improving drying efficiency and quality. However, in practical applications, the system faces multi-objective optimization challenges under specific operating conditions, including reducing energy consumption, increasing system dehumidification rate, improving system energy efficiency, and ensuring product drying quality. Furthermore, these objectives are interdependent. Therefore, optimizing a single objective is insufficient to achieve overall performance improvement; a comprehensive balance between multiple objectives is necessary. For these reasons, there is an urgent need to provide an optimization method for CO2 composite heat pump drying systems. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a thermodynamic coupling optimization and parameter matching method for a CO2 composite heat pump drying system, so as to solve the problems of single optimization target, incomplete parameter matching and inaccurate control logic in the prior art.

[0006] To achieve the above objectives, the present invention employs the following technical solution: A thermodynamic coupling optimization and parameter matching method for a CO2 hybrid heat pump drying system includes: The operating parameters of the CO2 composite heat pump drying system are collected and preprocessed in real time to obtain the preprocessed operating parameters. The operating parameters include the exhaust pressure of the CO2 compressor, the evaporation temperature of the R134a heat pump system, the exhaust temperature of the R134a heat pump system, the evaporation temperature of the CO2 heat pump system, the exhaust temperature of the CO2 heat pump system, the supply air temperature and humidity of the upper layer of the drying chamber, the supply air temperature and humidity of the lower layer of the drying chamber, the return air temperature and humidity of the drying chamber, the supply air volume of the upper layer of the drying chamber, and the supply air volume of the lower layer of the drying chamber. The input layer data includes the evaporation superheat of the R134a heat pump system, the exhaust pressure of the CO2 heat pump system, the air volume ratio of the upper and lower layers in the three-layer sludge drying chamber, and the heat loss of the return air in the lower layer. The input layer data is fed into a multilayer perceptron neural network model. The output of the multilayer perceptron neural network model includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, the power consumption of the fan, and the product outlet moisture content. The multilayer perceptron neural network model's multi-objective function includes minimizing system energy consumption and minimizing product outlet moisture content. The system energy consumption includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, and the power consumption of the fan. The multi-objective function is iteratively optimized using the SPEA2 algorithm. The trained multilayer perceptron neural network model is used to calculate the objective function value corresponding to each set of operating parameters. Based on the constraints of the operating parameters, population initialization, objective function calculation, constraint processing, non-dominated sorting, and environment selection are performed. After crossover mutation and iterative optimization, the Pareto optimal solution set is obtained. The optimal combination of objective functions and the corresponding operating parameters are selected from the Pareto optimal solution set based on actual production needs.

[0007] A further improvement of the present invention is that: Preferably, the multilayer perceptron neural network model is trained using a training set and validated using a test set.

[0008] Preferably, during training, the multilayer perceptron neural network model is initialized using the Xavier initialization method.

[0009] Preferably, the system energy consumption to be minimized in the multi-objective function is to minimize the power consumption of the CO2 heat pump system compressor, the R134a heat pump system compressor, and the fan.

[0010] Preferably, the constraints on the system operating parameters include: (5) in, For the superheat at the evaporator outlet of the R134a system, The discharge pressure of the CO2 heat pump system compressor. These are the relative humidity values ​​of the inlet and outlet air in the upper and lower layers of the drying chamber, respectively. This refers to the air volume supplied to the upper part of the drying chamber. This refers to the air volume supplied to the lower part of the drying chamber. The air supply ratio between the upper and lower layers of the drying chamber. Set the total air supply volume for the system. To dissipate heat from the lower-level return air, To calculate the upper limit value.

[0011] Preferably, the specific process of iterating the multilayer perceptron neural network model using the SPEA2 algorithm is as follows: setting the population size, number of iterations, crossover probability, and mutation probability, and randomly generating an initial parameter set; calculating the objective function value of each set of parameters based on the trained multilayer perceptron neural network model; eliminating parameter combinations that exceed the set limit for product export moisture content; extracting the current optimal solution set by non-dominated sorting of the remaining solution set, and performing environment selection through density estimation to obtain elite solutions; the elite solutions form a new population through crossover and mutation operations, and the iteration is repeated until the set number of iterations is reached.

[0012] Preferably, the step of selecting the optimal objective function combination from the Pareto optimal solution set according to actual production needs includes: based on the trade-off between dehumidification rate (SMER) and product outlet moisture content (MR), selecting a parameter configuration scheme that balances system energy efficiency and drying quality according to actual production needs.

[0013] Preferably, the formula for calculating the performance evaluation index SMER is: (6) (7) in, The quality of water removed from the product after passing through the drying chamber. Input quality into the product, This refers to the moisture content of the product before drying. This refers to the moisture content of the product after drying. For the compressor power consumption of a transcritical CO2 heat pump system, For the compressor power consumption of the R134a heat pump system, For wind turbine power consumption, This refers to the drying time of the materials in the drying chamber.

[0014] A thermodynamic coupling optimization and parameter matching device for a CO2 composite heat pump drying system, comprising: The data acquisition unit is used to collect and preprocess the operating parameters of the CO2 composite heat pump drying system in real time to obtain the preprocessed operating parameters. These operating parameters include the CO2 compressor exhaust pressure, R134a heat pump system evaporation temperature, R134a heat pump system exhaust temperature, CO2 heat pump system evaporation temperature, CO2 heat pump system exhaust temperature, upper-layer supply air temperature and humidity, lower-layer supply air temperature and humidity, dryer return air temperature and humidity, upper-layer supply air volume, and lower-layer supply air volume. Based on these operating parameters, input layer data is acquired. This input layer data includes the R134a heat pump system evaporation superheat, CO2 heat pump system exhaust pressure, the ratio of upper to lower air volume in the three-layer sludge drying chamber, and the heat loss from the lower-layer return air. The model processing unit is used to input the input layer data into a multilayer perceptron neural network model. The output of the multilayer perceptron neural network model includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, the power consumption of the fan, and the product outlet moisture content. The multilayer perceptron neural network model has multiple objective functions, including minimizing system energy consumption and minimizing product outlet moisture content. The system energy consumption includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, and the power consumption of the fan. The multi-objective function is iteratively optimized using the SPEA2 algorithm. Specifically, the objective function value corresponding to each set of operating parameters is calculated using the trained multilayer perceptron neural network model. Based on the constraints of the operating parameters, population initialization, objective function calculation, constraint processing, non-dominated sorting, and environment selection are performed. After crossover mutation and iterative optimization, the Pareto optimal solution set is obtained. The running parameter generation unit is used to select the optimal combination of objective functions from the Pareto optimal solution set according to actual production needs, as well as the running parameters corresponding to the optimal combination of objective functions.

[0015] A CO2 composite heat pump drying system includes a CO2 heat pump system, an R134a heat pump system, a drying air system, and a heat dissipation circulation system; This includes transcritical CO2 heat pump cycle systems, R134 heat pump cycle systems, and drying air systems; The transcritical CO2 heat pump cycle system includes a CO2 compressor, a primary CO2 gas cooler, a CO2 regenerator, and a secondary CO2 gas cooler; After the refrigerant is output from the CO2 compressor, it is cooled by the CO2 primary gas cooler, CO2 regenerator and secondary CO2 gas cooler, and then evaporated and heated by the CO2 evaporator before entering the CO2 gas-liquid separator. The refrigerant separated by the CO2 gas-liquid separator is returned to the CO2 compressor through the CO2 regenerator. The refrigerant in the R134a heat pump cycle system is heated in the secondary CO2 gas cooler and then flows back to the R134a heat pump cycle system. The drying air system includes a three-layer sludge drying chamber, a heat exchanger, and a drying air regenerator. The return air duct of the three-layer sludge drying chamber is divided into two paths: one path is heated by an R134 heat pump circulation system and then flows back to the three-layer sludge drying chamber; the other path is cooled by the heat exchanger and the drying air regenerator, then heated by the transcritical CO2 heat pump circulation system and flows back to the three-layer sludge drying chamber.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention discloses a thermodynamic coupling optimization and parameter matching method for a CO2 composite heat pump drying system. The method first collects and preprocesses system operating parameters, then defines system energy consumption and product outlet moisture content as multi-objective optimization functions. Subsequently, a trained neural network model is combined with the SPEA2 intelligent algorithm to iteratively optimize the objective functions, obtaining the optimal set of function combinations. Finally, the best combination of objective functions is selected based on production needs. Compared to traditional optimization methods, this invention effectively solves the energy waste problems caused by insufficient thermodynamic performance and coarse parameter matching in existing systems, achieving adaptive adjustment and efficient stable operation of the system under fluctuating operating conditions, significantly improving energy utilization efficiency. By combining thermodynamic coupling analysis and multi-objective optimization algorithms, and comprehensively considering factors such as energy consumption, dehumidification rate, and product drying quality, the optimal combination of key system parameters is determined, addressing the problems of single optimization objectives, incomplete parameter matching, and inaccurate control logic in existing technologies.

[0017] This invention accurately maps the input-output relationship of a system using a neural network model, significantly improving prediction accuracy and providing a reliable basis for optimization decisions. It utilizes the SPEA2 algorithm to achieve multi-objective collaborative optimization of system energy consumption and product outlet moisture content in a CO2 composite heat pump drying system. This reduces energy consumption while ensuring system performance, achieving energy conservation and emission reduction goals with significant economic and environmental benefits. The Pareto optimal solution set generated by this invention can be flexibly selected according to actual needs, applicable to different operating conditions, enhancing the system's adaptability and practicality. Attached Figure Description

[0018] Figure 1 This is a structural diagram of a CO2 composite heat pump drying system according to the present invention. 1. CO2 compressor; 2. Primary gas cooler; 3. CO2 regenerator; 4. Secondary gas cooler; 5. CO2 throttle valve; 6. CO2 evaporator; 7. CO2 gas-liquid separator; 8. R134a compressor; 9. R134a condenser; 10. R134a throttle valve; 11. R134a gas-liquid separator; 12. Three-layer sludge drying chamber; 13. Cooling tower; 14. Cooling water pump; 15. Heat exchanger; 16. Drying air regenerator; 17. Upper air supply fan; 18. Lower air supply fan; Figure 2 This is a block diagram of an optimized process structure for a CO2 composite heat pump drying system according to the present invention; Figure 3 This is a schematic diagram of the neural network structure of a CO2 composite heat pump drying system according to the present invention; Figure 4 This is a flowchart of an optimization algorithm for a CO2 composite heat pump drying system according to the present invention. Detailed Implementation

[0019] Hereinafter, the terms "first," "second," "third," and "fourth" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," "third," or "fourth" may explicitly or implicitly include one or more of that feature.

[0020] The method provided in this application can be applied to mobile phones, tablets, wearable devices, in-vehicle devices, augmented reality (AR) / virtual reality (VR) devices, laptops, and ultra-mobile personal computers. In this application, the specific type of terminal device is not limited to terminal devices such as mobile personal computers (UMPCs), netbooks, and personal digital assistants (PDAs).

[0021] It should be noted that the terms "first," "second," etc., used in the specification and drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0022] Please see Figure 1 The first aspect of the present invention discloses a CO2 composite heat pump drying system, including a CO2 heat pump system, an R134a heat pump system, a drying air system, and a heat dissipation circulation system.

[0023] In the CO2 heat pump system, the outlet of CO2 compressor 1 is connected to the inlet of primary gas cooler 2, the outlet of primary gas cooler 2 is connected to the high-pressure side inlet of CO2 regenerator 3, the high-pressure side outlet of CO2 regenerator 3 is connected to the high-pressure side inlet of secondary gas cooler 4, the high-pressure side outlet of secondary gas cooler 4 is connected to the high-pressure side inlet of CO2 expansion valve 5, the low-pressure side outlet of CO2 expansion valve 5 is connected to the cold side inlet of CO2 evaporator 6, the cold side outlet of CO2 evaporator 6 is connected to the inlet of CO2 gas-liquid separator 7, the outlet of CO2 gas-liquid separator 7 is connected to the low-pressure side inlet of regenerator 3, and the low-pressure side outlet of regenerator 3 is connected to the inlet of compressor 1. In this process, the refrigerant is compressed to a high temperature and high pressure state in CO2 compressor 1 and discharged into CO2 primary gas cooler 2 and CO2 regenerator 3. As a hot-side gas, it is cooled for pre-cooling. Then, it releases heat and cools down again in the high-pressure side of secondary gas cooler 4. At this time, the temperature has been significantly reduced but is still in a supercritical state. The high-pressure CO2 is throttled and depressurized by CO2 expansion valve 5, becoming a low-temperature and low-pressure gas-liquid two-phase mixture. It then enters CO2 evaporator 6 to absorb heat and evaporate, transforming into low-pressure gas. After that, the gas enters CO2 gas-liquid separator 7 to ensure that pure gas without liquid droplets enters the low-pressure side of CO2 regenerator 3. In CO2 regenerator 3, it absorbs heat from the high-pressure side and becomes superheated gas. Finally, it is drawn into CO2 compressor 1 and compressed again to complete the entire cycle.

[0024] R134a heat pump system: The outlet of R134a compressor 8 is connected to the hot side inlet of condenser 9, the hot side outlet of condenser 9 is connected to the high-pressure side inlet of expansion valve 10, the low-pressure side outlet of expansion valve 10 is connected to the low-pressure side inlet of secondary gas cooler 4, the low-pressure side outlet of secondary gas cooler 4 is connected to the inlet of R134a gas-liquid separator 11, and the outlet of R134a gas-liquid separator 11 is connected to the inlet of R134a compressor 8. In the above process, the high-temperature, high-pressure gaseous refrigerant discharged from the R134a compressor first enters the condenser 9, where it releases heat and condenses into a liquid state. Then, the liquid R134a enters the expansion valve 10 for throttling and pressure reduction, becoming a low-temperature, low-pressure gas-liquid two-phase mixture. Next, this mixture enters the low-pressure side channel of the secondary gas cooler 4, where it absorbs heat from the high-pressure side of CO2 and evaporates into a low-pressure gas. The gaseous R134a exiting from the low-pressure side of the secondary gas cooler 4 then enters the R134a gas-liquid separator 11, ensuring that the pure gas without liquid droplets finally enters the suction port of the R134a compressor 8, completing the entire R134a cycle.

[0025] The drying air system includes upper air supply and lower air supply, with the lower air supply utilizing a heat exchanger.

[0026] The air outlet of the three-layer sludge drying chamber 12 is located between the first and second layer conveyor belts, and is connected to the inlet of the upper-layer air supply fan 17 and the heat exchanger 15. Air supplied via the upper-layer air supply fan 17 is the upper-layer air supply method, and air supplied via the heat exchanger 15 is the lower-layer air supply method. Detailed analysis: Upper-level air supply method: The outlet of the upper-level air supply fan 17 is connected to the cold-side inlet of the R134a condenser 9, and the cold-side outlet of the R134a condenser 9 is connected to the air outlet at the top of the three-layer sludge drying chamber 12. During this process, the upper-level air supply fan 17 pressurizes and supplies air to the cold side of the R134a condenser 9. In the R134a condenser 9, the low-temperature return air absorbs the heat released during the condensation of the high-temperature gaseous R134a, resulting in a significant increase in temperature and a decrease in relative humidity, transforming into high-temperature, low-humidity dry air. Subsequently, this dry air enters from the air outlet at the top of the three-layer sludge drying chamber 12, passes through the sludge layer from top to bottom, and carries away the moisture from the sludge.

[0027] Lower-level air supply method: The hot-side outlet of the heat exchanger 15 is connected to the high-temperature inlet of the drying air regenerator 16; the high-temperature outlet of the drying air regenerator 16 is connected to the hot-side inlet of the CO2 evaporator 6; the hot-side outlet of the CO2 evaporator 6 is connected to the low-temperature inlet of the drying air regenerator 16; the low-temperature outlet of the drying air regenerator 16 is connected to the inlet of the lower-level air supply fan 18; the outlet of the lower-level air supply fan 18 is connected to the cold-side inlet of the first-stage gas cooler 2; and the cold-side outlet of the first-stage gas cooler 2 is connected to the air outlet at the bottom of the third-level sludge drying chamber 12.

[0028] Heat dissipation circulation system: The outlet of cooling water pump 14 is connected to the cold side inlet of heat dissipation heat exchanger 15, the cold side outlet of heat dissipation heat exchanger 15 is connected to the inlet of cooling water tower 13, and the outlet of cooling water tower 13 is connected to the inlet of cooling water pump 14.

[0029] During the lower-level air supply process, another portion of the hot and humid air from the outlet first enters the hot side of the heat exchanger 15, where it exchanges heat with the circulating cooling water from the cooling tower. The air is initially cooled and some moisture is released. Next, the air enters the high-temperature side of the drying air regenerator 16, where it exchanges sensible heat with the low-temperature air returning from the CO2 evaporator 6, further reducing its temperature. Then, the air enters the hot side of the CO2 evaporator 6, acting as a heat source to heat the low-temperature, low-pressure CO2 refrigerant, further decreasing its own temperature and moisture content. Finally, it leaves the CO2 evaporator 6... After evaporator 6, the air re-enters the low-temperature side of the drying air regenerator 16, absorbing the heat from the previously high-temperature inlet air, achieving partial recovery of cold energy. The temperature rises slightly, but the moisture content remains unchanged. Finally, the air is sent to the cold side of the first-stage gas cooler 2 by the lower-level air supply fan 18, absorbing the heat released by the cooling of the high-temperature and high-pressure CO2 gas in the CO2 heat pump system, and transforming into high-temperature and low-humidity dry air. It is then sent in from the air supply port at the bottom of the three-layer sludge drying chamber 12, passing through the sludge layer from bottom to top, and together with the upper-level air supply, achieves efficient and uniform drying.

[0030] See Figure 2 , Figure 3 and Figure 4 The second aspect of this invention discloses a thermodynamic coupling optimization and parameter matching method for a CO2 composite heat pump drying system, comprising the following steps: S1, real-time acquisition and preprocessing of the operating parameters of the CO2 composite heat pump drying system to obtain the preprocessed operating parameters; the operating parameters include the evaporation superheat of the R134a heat pump system, the exhaust pressure of the CO2 heat pump system, the air volume ratio of the upper and lower layers in the drying chamber, and the heat loss of the return air in the lower layer. Specifically, it includes: S1.1 Data Acquisition and Recording: Based on several sensors deployed in the CO2 composite heat pump system, the operating status of each component of the system and environmental conditions are collected in real time, and the collected data is stored in the database. Input parameters include CO2 compressor exhaust pressure, R134a heat pump system evaporation and exhaust temperatures, CO2 heat pump system evaporation and exhaust temperatures, supply and return air temperatures and humidity in the upper and lower layers of the drying chamber, and supply air volume in the upper and lower layers of the drying chamber.

[0031] S1.2, Data Preprocessing: First, the collected data undergoes data cleaning to address any issues and improve data quality. Specific data cleaning operations include missing value handling, outlier detection and handling, deduplication, noise smoothing, and data consistency checks. Then, the cleaned data is normalized to transform it into a distribution with a mean of 0 and a standard deviation of 1. (1) in, This represents the mean of the cleaned data. This represents the standard deviation of the cleaned data.

[0032] S1.3 Calculate the input layer data of the multilayer perceptron neural network model using the preprocessed operating parameters. The input layer data includes the evaporation superheat of the R134a heat pump system, the exhaust pressure of the CO2 heat pump system, the air volume ratio of the upper and lower layers in the three-layer sludge drying chamber 12, and the heat loss of the return air in the lower layer.

[0033] S2, the preprocessed operating parameters are input into a multilayer perceptron neural network model. The output of the multilayer perceptron neural network model includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, the total power consumption of the upper and lower fans, and the product outlet moisture content. The multilayer perceptron neural network model's multi-objective function includes minimizing system energy consumption and minimizing product outlet moisture content. The system energy consumption includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, and the fan power consumption. The multilayer perceptron neural network model is iterated using the SPEA2 algorithm. Based on the constraints of the operating parameters, population initialization, objective function calculation, constraint processing, non-dominated sorting, and environment selection are performed. After crossover mutation and iterative optimization, the Pareto optimal solution set is obtained.

[0034] The specific process includes: S2.1, Neural Network Model Construction and Training: A multilayer perceptron neural network model is constructed, with the system parameters R134a heat pump system evaporative superheat, CO2 heat pump system exhaust pressure, upper and lower air volume ratio in the three-layer sludge drying chamber 12, and lower layer return air heat loss as the input layer. The CO2 heat pump system compressor power consumption, R134a heat pump system compressor power consumption, upper and lower layer fan total power consumption, and product outlet moisture content as the output layer. The hidden layer is connected to the input layer and the output layer in a fully connected structure.

[0035] S2.2 defines a multi-objective optimization function using system energy consumption and product outlet moisture content. In a CO2 composite heat pump system, the optimization objectives are minimizing system energy consumption and minimizing the product outlet moisture content (MR). Specifically, system energy consumption includes the power consumption of the CO2 heat pump system compressor. R134a heat pump system compressor power consumption Fan power consumption The objective function is defined as follows: (2) in, The system parameters include the evaporative superheat θ of the R134a heat pump system, the exhaust pressure P of the CO2 heat pump system, the air volume ratio w between the upper and lower layers in the drying room, and the heat loss of the return air Q in the lower layer.

[0036] Each term in the objective function has the following form: (3) in, The function predicted by the neural network model.

[0037] S2.3 combines the trained neural network model with the SPEA2 intelligent algorithm and iteratively optimizes the objective function to obtain the optimal set of function combinations, including: S2.31, Initialize the population: Set the population size N (50≤N≤100, and N is a positive integer), iteration number i, crossover probability Pc, and mutation probability Pm. Based on the constraints of the operating parameters of the CO2 composite heat pump drying system, randomly generate N sets of initial parameters, including the R134a system evaporation superheat. CO2 system compressor discharge pressure Air volume ratio between upper and lower layers in the drying chamber And the heat loss Q from the lower-level return air.

[0038] The constraints are as follows: (4) in, For the superheat at the evaporator outlet of the R134a system, The discharge pressure of the CO2 heat pump system compressor. These are the relative humidity values ​​of the inlet and outlet air in the upper and lower layers of the drying chamber, respectively. This refers to the air volume supplied to the upper part of the drying chamber. This refers to the air volume supplied to the lower part of the drying chamber. The air supply ratio between the upper and lower layers of the drying chamber. Set the total air supply volume for the system. To dissipate heat from the lower-level return air, To calculate the upper limit value.

[0039] S2.32, Calculation of objective function: Based on the randomly generated initial population, the verified neural network model is connected and the objective function value of each group is calculated through the multilayer perceptron neural network model, i.e., the calculation of formula (3) and formula (4).

[0040] S2.33, Constraint Treatment: Feasibility screening is performed on the solution set of the objective function calculated by the neural network model. After the actual drying process, the moisture content (MR) of the product needs to be maintained below a set limit, i.e., MR... 实际 ≤MR 需求 The solution set is filtered by removing parameter combinations that exceed the limits, thereby selecting a reasonable solution set.

[0041] S2.34, Non-dominated sorting and environment selection: The remaining objective function solution set after constraint processing is non-dominated sorted, all non-dominated solutions are extracted as the current optimal solution set, and solutions that achieve the best trade-off among these objectives are selected first to optimize the performance of the drying system; then, solutions with higher fitness are selected from the non-dominated solution set to enter the next generation population. If the size of the non-dominated solution set exceeds the preset value, solutions with uniform distribution are retained by density estimation. If the size is insufficient, solutions with lower fitness are selected from the remaining solutions to supplement them, thereby ensuring population diversity and the retention of elite solutions, and obtaining the final elite solution.

[0042] S2.35, Performance Evaluation Metric (SMER) Calculation: The processed elite solution is subjected to the performance evaluation metric (SMER) calculation, and the data is recorded.

[0043] The dehumidification rate (SMER) is calculated as follows: (5) (6) in, The quality of water removed from the product after passing through the drying chamber. Input quality into the product, This refers to the moisture content of the product before drying. This refers to the moisture content of the product after drying. For the compressor power consumption of a transcritical CO2 heat pump system, For the compressor power consumption of the R134a heat pump system, For wind turbine power consumption, This refers to the drying time of the materials in the drying chamber.

[0044] SMER calculations are performed on the elite solutions to select the SMERs. 实际 ≥SMER 设定 The elite solution.

[0045] S2.36, Crossover and Mutation: After selection, the elite population forms a new population through crossover and mutation, and the new population satisfies the system parameter constraints. Crossover operation: For each pair of individuals in the initial population, a random probability Rc is generated. If this probability is less than the crossover probability Pc, crossover recombination is performed on their structure; otherwise, it is not. Mutation operation: For each individual in the initial population, a random probability Rm is generated. If this probability is less than the mutation probability Pm, genetic mutation is performed on their structure; otherwise, it is not, resulting in a new population.

[0046] S2.37, Iterative optimization: The new population obtained in S2.36 is merged with the elite solutions obtained in S2.35 before evolution. The objective function value is calculated using the trained neural network model. Then, the objective function solution set is constrained, non-dominated sorting is performed, and environmental selection is carried out to select individuals with higher fitness as the solution set.

[0047] S2.38 Convergence Judgment: Determine whether the current iteration number t has reached i. If not, repeat the above S2.36 and S2.37 processes. Otherwise, output the solution set of the objective function and end.

[0048] S2.39, Based on production requirements, select the optimal combination of objective functions, including: Based on demand-driven solution selection: The Pareto optimal solution set generated through iterative optimization covers all non-dominated solutions for the CO2 composite heat pump drying system among multiple objectives of dehumidification rate (SMER) and product outlet moisture content (MR). Users can select the appropriate parameter configuration scheme according to their actual production needs.

[0049] Generally speaking, a higher SMER (Self-Moisture Ratio) is better, indicating a more energy-efficient and high-performance system; a lower MR (Moisture Content) value at the product outlet is better, indicating more thorough drying. However, these two are often contradictory. If the MR is very low, the drying time needs to be extended or the temperature increased, which leads to increased power consumption and a lower SMER. Conversely, to achieve greater energy savings and a higher SMER, the drying process needs to be terminated earlier, resulting in a higher MR. Multiple non-dominated solution combinations are obtained by iteratively optimizing the generated Pareto optimal solution set.

[0050] In some embodiments of the present invention, a pre-training process for the multilayer perceptron neural network model is included before S1, comprising the following steps: (1) Data acquisition and recording: The working status of each component of the system and environmental conditions are collected in real time by several sensors arranged in the CO2 composite heat pump system, and the collected data is stored in the database. The input parameters include CO2 compressor exhaust pressure, R134a and CO2 heat pump system evaporation temperature and exhaust temperature, upper and lower air supply and return air temperature and humidity in drying chamber 12, and upper and lower air supply air volume in drying chamber 12; the output parameters include CO2 compressor energy consumption, R134a compressor energy consumption, total power consumption of upper and lower fans, and product outlet moisture content.

[0051] Among them, the evaporation temperature of the R134a heat pump system is the low-pressure side outlet temperature of the secondary gas cooler 4, the exhaust temperature of the R134a heat pump system is the high-pressure side outlet temperature of the R134a compressor 8, the evaporation temperature of the CO2 heat pump system is the outlet temperature of the CO2 evaporator 6, the exhaust temperature of the CO2 heat pump system is the high-pressure side outlet temperature of the CO2 compressor, the supply air temperature is the average of the upper and lower inlet air temperatures of the drying chamber 12, and the average humidity of the upper and lower inlets, the return air temperature is the outlet temperature of the drying chamber 12, the return air humidity is the outlet humidity of the drying chamber 12, and the total power consumption of the upper and lower fans is the sum of the upper supply air fan 17 and the lower supply air fan 18.

[0052] (2) Data Preprocessing: First, the collected data is cleaned to address any issues and improve data quality. Specific data cleaning operations include missing value handling, outlier detection and handling, deduplication, noise smoothing, and data consistency checks. Then, the cleaned data is normalized to a distribution with a mean of 0 and a standard deviation of 1. (1) in, This represents the mean of the cleaned data. This represents the standard deviation of the cleaned data.

[0053] (3) Classify the processed data and construct and train the neural network model, including: (3.1) Data classification: The preprocessed dataset will be randomly divided into training set, validation set and test set.

[0054] (3.2) Neural Network Model Construction and Training: A multilayer perceptron neural network model was constructed. The system parameters R134a heat pump system evaporative superheat, CO2 heat pump system exhaust pressure, upper and lower air volume ratio in the three-layer sludge drying chamber 12, and lower layer return air heat loss were used as the input layer (based on collected operating data or calculated from collected operating data). The CO2 heat pump system compressor power consumption, R134a heat pump system compressor power consumption, fan power consumption, and product outlet moisture content were used as the output layer. The hidden layers were connected to the input and output layers using a fully connected structure. First, the model was initialized using the Xavier initialization method. Then, the model was trained using the training set data, and the accuracy of the model was verified using the test set. The trained neural network model was then saved and output.

[0055] The Xavier initialization method is as follows: (7) in and These represent the number of input nodes and the number of output nodes for each layer.

[0056] The third aspect of this invention discloses a thermodynamic coupling optimization and parameter matching device for a CO2 composite heat pump drying system, comprising: The data acquisition unit is used to collect and preprocess the operating parameters of the CO2 composite heat pump drying system in real time to obtain the preprocessed operating parameters. The operating parameters include the evaporation superheat of the R134a heat pump system, the exhaust pressure of the CO2 heat pump system, the air volume ratio between the upper and lower layers in the drying chamber, and the heat loss of the return air in the lower layer. The model processing unit is used to input the preprocessed operating parameters into a multilayer perceptron neural network model. The output of the multilayer perceptron neural network model includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, the power consumption of the fan, and the product outlet moisture content. The multilayer perceptron neural network model has multiple objective functions, including minimizing system energy consumption and minimizing product outlet moisture content. The system energy consumption includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, and the power consumption of the fan. The multilayer perceptron neural network model is iterated using the SPEA2 algorithm. Based on the constraints of the operating parameters, population initialization, objective function calculation, constraint processing, non-dominated sorting, and environment selection are performed. After crossover mutation and iterative optimization, the Pareto optimal solution set is obtained. The running parameter generation unit is used to select the optimal combination of objective functions from the Pareto optimal solution set according to actual production needs, as well as the running parameters corresponding to the optimal combination of objective functions.

[0057] Compared with existing technologies, this invention addresses the complex heat and mass transfer processes and the highly nonlinear characteristics of multi-component coupled operation in CO2 composite heat pump drying systems. It utilizes system parameters to perform deep training on a neural network model, constructing a precise input-output mapping model to achieve high-precision prediction of system performance. Simultaneously, it combines the SPEA2 multi-objective evolutionary algorithm to collaboratively optimize key system performance indicators such as system energy consumption and product outlet moisture content. Furthermore, through SMER (System Merger and Evaluation) calculation, it effectively resolves the contradictory relationships between multiple objectives, providing diverse optimization schemes to meet different operational requirements.

[0058] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus systems, and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0059] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0060] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

[0062] 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, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A thermodynamic coupling optimization and parameter matching method and system for a CO2 compound heat pump drying system, characterized in that, include: The operating parameters of the CO2 composite heat pump drying system are collected and preprocessed in real time to obtain the preprocessed operating parameters. These operating parameters include the CO2 compressor exhaust pressure, R134a heat pump system evaporation temperature, R134a heat pump system exhaust temperature, CO2 heat pump system evaporation temperature, CO2 heat pump system exhaust temperature, upper and lower air supply temperatures and humidity of the drying chamber, drying chamber return air temperature and humidity, upper and lower air supply volumes of the drying chamber. Input layer data is obtained based on these operating parameters. This input layer data includes the R134a heat pump system evaporation superheat, CO2 heat pump system exhaust pressure, the ratio of upper to lower air volume in the three-layer sludge drying chamber, and the heat loss from the lower return air. The input layer data is fed into a multilayer perceptron neural network model. The output of the multilayer perceptron neural network model includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, the power consumption of the fan, and the product outlet moisture content. The multilayer perceptron neural network model's multi-objective function includes minimizing system energy consumption and minimizing product outlet moisture content. The system energy consumption includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, and the power consumption of the fan. The multi-objective function is iteratively optimized using the SPEA2 algorithm. The trained multilayer perceptron neural network model is used to calculate the objective function value corresponding to each set of operating parameters. Based on the constraints of the operating parameters, population initialization, objective function calculation, constraint processing, non-dominated sorting, and environment selection are performed. After crossover mutation and iterative optimization, the Pareto optimal solution set is obtained. The optimal combination of objective functions and the corresponding operating parameters are selected from the Pareto optimal solution set based on actual production needs.

2. The thermodynamic coupling optimization and parameter matching method of a CO2 compound heat pump drying system according to claim 1, characterized in that, The multilayer perceptron neural network model is trained using a training set and validated using a test set.

3. The thermodynamic coupling optimization and parameter matching method of a CO2 compound heat pump drying system according to claim 2, characterized in that, During training, the multilayer perceptron neural network model is initialized using the Xavier initialization method.

4. The thermodynamic coupling optimization and parameter matching method for a CO2 composite heat pump drying system according to claim 1, characterized in that, In the multi-objective function, minimizing system energy consumption means minimizing the power consumption of the CO2 heat pump system compressor, the R134a heat pump system compressor, and the fan.

5. The thermodynamic coupling optimization and parameter matching method for a CO2 composite heat pump drying system according to claim 1, characterized in that, The constraints on the system operating parameters include: in, For the superheat at the evaporator outlet of the R134a system, The discharge pressure of the CO2 heat pump system compressor. These are the relative humidity values ​​of the inlet and outlet air in the upper and lower layers of the drying chamber, respectively. This refers to the air volume supplied to the upper part of the drying chamber. This refers to the air volume supplied to the lower part of the drying chamber. The air supply ratio between the upper and lower layers of the drying chamber. Set the total air supply volume for the system. To dissipate heat from the lower-level return air, To calculate the upper limit value.

6. The thermodynamic coupling optimization and parameter matching method for a CO2 composite heat pump drying system according to claim 1, characterized in that, The specific process of iterating the multilayer perceptron neural network model using the SPEA2 algorithm is as follows: setting the population size, number of iterations, crossover probability, and mutation probability, and randomly generating an initial parameter set; calculating the objective function value of each set of parameters based on the trained multilayer perceptron neural network model; eliminating parameter combinations that exceed the set limit for product export moisture content; extracting the current optimal solution set by non-dominated sorting of the remaining solution set, and obtaining elite solutions by environment selection through density estimation; the elite solutions form a new population through crossover and mutation operations, and the iteration is repeated until the set number of iterations is reached.

7. The thermodynamic coupling optimization and parameter matching method for a CO2 composite heat pump drying system according to claim 1, characterized in that, The selection of the optimal objective function combination from the Pareto optimal solution set based on actual production needs includes: selecting a parameter configuration scheme that balances system energy efficiency and drying quality based on the trade-off between dehumidification rate (SMER) and product outlet moisture content (MR).

8. The thermodynamic coupling optimization and parameter matching method for a CO2 composite heat pump drying system according to claim 7, characterized in that, The formula for calculating the performance evaluation index SMER is as follows: in, The quality of water removed from the product after passing through the drying chamber. Input quality into the product, This refers to the moisture content of the product before drying. This refers to the moisture content of the product after drying. For the compressor power consumption of a transcritical CO2 heat pump system, For the compressor power consumption of the R134a heat pump system, For wind turbine power consumption, This refers to the drying time of the materials in the drying chamber.

9. A thermodynamic coupling optimization and parameter matching device for a CO2 composite heat pump drying system, characterized in that, include The data acquisition unit is used to collect and preprocess the operating parameters of the CO2 composite heat pump drying system in real time to obtain the preprocessed operating parameters. These operating parameters include the CO2 compressor exhaust pressure, R134a heat pump system evaporation temperature, R134a heat pump system exhaust temperature, CO2 heat pump system evaporation temperature, CO2 heat pump system exhaust temperature, upper-layer supply air temperature and humidity, lower-layer supply air temperature and humidity, dryer return air temperature and humidity, upper-layer supply air volume, and lower-layer supply air volume. Based on these operating parameters, input layer data is acquired. This input layer data includes the R134a heat pump system evaporation superheat, CO2 heat pump system exhaust pressure, the ratio of upper to lower air volume in the three-layer sludge drying chamber, and the heat loss from the lower-layer return air. The model processing unit is used to input the input layer data into a multilayer perceptron neural network model. The output of the multilayer perceptron neural network model includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, the power consumption of the fan, and the product outlet moisture content. The multilayer perceptron neural network model has multiple objective functions, including minimizing system energy consumption and minimizing product outlet moisture content. The system energy consumption includes the power consumption of the CO2 heat pump system compressor, the power consumption of the R134a heat pump system compressor, and the power consumption of the fan. The multi-objective function is iteratively optimized using the SPEA2 algorithm. Specifically, the objective function value corresponding to each set of operating parameters is calculated using the trained multilayer perceptron neural network model. Based on the constraints of the operating parameters, population initialization, objective function calculation, constraint processing, non-dominated sorting, and environment selection are performed. After crossover mutation and iterative optimization, the Pareto optimal solution set is obtained. The running parameter generation unit is used to select the optimal combination of objective functions from the Pareto optimal solution set according to actual production needs, as well as the running parameters corresponding to the optimal combination of objective functions.

10. A CO2 composite heat pump drying system, characterized in that, This includes CO2 heat pump systems, R134a heat pump systems, dry air systems, and heat dissipation circulation systems; This includes a transcritical CO2 heat pump cycle system, an R134 heat pump cycle system, and a drying air system; The transcritical CO2 heat pump cycle system includes a CO2 compressor (1), a CO2 primary gas cooler (2), a CO2 regenerator (3), and a secondary CO2 gas cooler (4). After the refrigerant is output from the CO2 compressor, it is cooled by the CO2 primary gas cooler (2), CO2 regenerator (3) and secondary CO2 gas cooler (4), and then evaporated and heated by the CO2 evaporator (6) before entering the CO2 gas-liquid separator (7). The refrigerant separated by the CO2 gas-liquid separator (7) is returned to the CO2 compressor (1) through the CO2 regenerator (3). The refrigerant in the R134a heat pump cycle system is heated in the secondary CO2 gas cooler (4) and then flows back to the R134a heat pump cycle system. The drying air system includes a three-layer sludge drying chamber (12), a heat exchanger (15), and a drying air regenerator (16). The return air duct of the three-layer sludge drying chamber (12) is divided into two paths. One path is heated by the R134 heat pump circulation system and then flows back to the three-layer sludge drying chamber (12). The other path is cooled by the heat exchanger (15) and the drying air regenerator (16) and then heated by the transcritical CO2 heat pump circulation system before flowing back to the three-layer sludge drying chamber (12).

Images