An optimization method and system for a mixed refrigerant refrigeration system

Abstract

This application discloses an optimization method and system for a mixed refrigerant refrigeration system, applicable to the fields of refrigeration and cryogenic engineering technology. The method includes: establishing a process simulation model, determining constraints, refrigerant ratio, and the range of high and low pressure values; generating an initial population using a hybrid generation strategy combining mechanistic seeds, local perturbation, and global random sampling, and evaluating offspring through a population optimization algorithm; determining fitness based on constraint violation information: individuals satisfying all constraints are fitnessd based on performance indicators; individuals violating heat transfer temperature difference constraints but within a feasible range are fitnessd based on partial molar enthalpy difference-corrected ratios; individuals violating compressor discharge temperature or suction phase constraints are fitnessd based on thermodynamic causality-corrected pressure or ratios; and penalizing the remaining individuals; selecting to form a new generation population, iterating to the termination condition, and outputting the optimal parameters; this application embeds thermodynamic mechanism guidance and compressor state causal repair, achieving differentiated processing of various constraints.

Description

Technical Field

[0001] This invention relates to the field of refrigeration and cryogenic engineering technology, and in particular to an optimization method and system for a mixed working fluid refrigeration system based on population iterative optimization and partial molar enthalpy difference repair. Background Technology

[0002] Hybrid refrigerant refrigeration systems, with their advantages of wide-temperature-range refrigeration, high energy efficiency, and simple equipment structure, are widely used in various cryogenic engineering fields such as natural gas liquefaction, cryogenic storage, industrial gas separation / cooling, and superconducting cooling. The refrigeration performance and operational reliability of this system depend primarily on the composition ratio of the mixed refrigerant and the degree of matching between high-pressure and low-pressure operating parameters.

[0003] Parameter optimization of mixed-refrigerant refrigeration systems is a typical complex engineering problem characterized by multiple variables, strong nonlinearity, and multiple coupled constraints. Optimization variables typically include the molar ratios of various refrigerant components, as well as system high-pressure and low-pressure conditions. Constraints cover a number of rigid engineering parameters, such as the minimum allowable heat exchanger temperature difference, the upper limit of compressor discharge temperature, and compressor suction phase requirements. The objective function is often to maximize the system's coefficient of performance (COP). Furthermore, performance evaluation of each parameter combination requires high-precision process simulation models for thermodynamic calculations, resulting in high time costs for each calculation.

[0004] Currently, population-based iterative optimization algorithms such as differential evolution, genetic algorithms, and particle swarm optimization are commonly used for parameter optimization. These algorithms possess global search capabilities, enabling them to find relatively optimal solutions in complex non-convex solution spaces. However, they have significant shortcomings in practical applications: Existing algorithms typically handle infeasible candidates generated during iteration using uniform penalty function weighting, direct elimination, or random resampling, without classifying and distinguishing infeasible individuals. For near-feasible infeasible candidates with only minor local heat exchange temperature differences or slight exhaust temperature exceedances, where the overall situation is close to the feasible region, existing methods fail to effectively extract and utilize their inherent thermodynamic information, resulting in the waste of a large number of valuable intermediate samples. Furthermore, existing methods often rely on purely random sampling during population initialization, easily generating a large number of invalid samples that violate thermodynamic principles, slowing down the optimization process. Simultaneously, for problems threatening system safety such as compressor exhaust overheating or liquid carryover in intake, existing technologies often use penalty functions for direct elimination, failing to effectively utilize the direct thermodynamic response mechanism between pressure and component properties for proactive repair. This defect directly results in slow convergence speed and high computational cost in the optimization process, and the results of multiple independent runs fluctuate greatly, making it difficult to quickly and stably obtain the optimal mixed working fluid ratio scheme that simultaneously satisfies all engineering constraints.

[0005] To overcome these shortcomings, this application proposes an optimization method and system for a hybrid working fluid refrigeration system. Without compromising the global search capability of the swarm algorithm, it effectively utilizes thermodynamic mechanisms to target near-feasible and infeasible solutions, thereby improving optimization efficiency and result stability. Summary of the Invention

[0006] The purpose of this application is to provide an optimization method and system for a mixed refrigerant refrigeration system, which aims to solve the technical problem that the utilization rate of near-feasible and infeasible samples in the existing group iterative optimization method for mixed refrigerant refrigeration systems is low, resulting in slow optimization convergence speed and poor result stability.

[0007] To achieve the above objectives, this application provides the following technical solution: In a first aspect, this application provides a method for optimizing a mixed refrigerant refrigeration system, the steps of which include: S1. Establish a process simulation model of the mixed working fluid refrigeration system, determine the constraints and the range of values ​​for the distribution ratio of each component of the mixed working fluid, as well as the range and step size of the high pressure and the low pressure. S2. Generate an initial population within the range of values. The initial population consists of a combination of mechanism seed individuals, derived individuals generated based on local perturbations of the mechanism seed individuals, and globally randomly sampled individuals. Each individual in the initial population corresponds to a set of optimized variable values, which at least include the distribution ratio of each group of the mixed working fluid, as well as the high pressure and low pressure. S3. The population is iteratively optimized using a population iterative optimization algorithm to generate offspring individuals; wherein, the current population at the first iteration is the initial population, and the offspring individuals are generated through at least one of differential mutation, crossover, position update, and perturbation search. S4. Input the optimization variable values ​​of the offspring individuals into the process simulation model for evaluation, and obtain the performance indicators and constraint violation information of each offspring individual; S5. Based on the constraint violation information, classify and determine the fitness of each offspring individual: if all preset constraints are met, the performance index of the individual is used as the fitness; if the heat exchange temperature difference related constraint is violated and is within the preset near-feasible repair range, the component ratio of the individual is corrected based on the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone; if the compressor exhaust temperature constraint or intake phase constraint is violated, the high pressure, low pressure or component ratio is corrected based on the thermodynamic causal correspondence; otherwise, the performance index is penalized according to the degree of constraint violation, and the penalized result is used as the fitness. S6. Select individuals based on their fitness to form a new generation of the population; S7. Repeat steps S3 to S6 until the preset termination condition is met, and output the optimal mixed working fluid ratio, optimal high pressure, optimal low pressure and corresponding performance indicators.

[0008] Secondly, this application provides an optimized system for a mixed refrigerant refrigeration system, specifically including: The model building module is used to build a process simulation model of the mixed working fluid refrigeration system, determine the constraints and the range of values ​​for the distribution ratio of each component of the mixed working fluid, as well as the range and step size of the high pressure and the low pressure. An initialization module is used to generate an initial population within the range of values. The initial population consists of a combination of mechanism seed individuals, derived individuals generated based on local perturbations of the mechanism seed individuals, and globally randomly sampled individuals. Each individual in the initial population corresponds to a set of optimized variable values, which at least include the distribution ratio of each group of the mixed working fluid, as well as the high pressure and low pressure. The iterative optimization module is used to iteratively optimize the population using a population iterative optimization algorithm to generate offspring individuals; wherein, the current population at the time of the first iteration is the initial population, and the offspring individuals are generated through at least one of differential mutation, crossover, position update, and perturbation search; The evaluation module is used to input the optimization variable values ​​of the offspring individuals into the process simulation model for evaluation, and to obtain the performance indicators and constraint violation information of each offspring individual. The fitness determination module is used to classify and determine the fitness of each offspring individual based on the constraint violation information: if all preset constraints are met, the performance index of the individual is used as the fitness; if the heat exchange temperature difference related constraint is violated and is within the preset near-feasible repair range, the component ratio of the individual is corrected based on the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone; if the compressor exhaust temperature constraint or intake phase constraint is violated, the high pressure, low pressure or component ratio is corrected based on the thermodynamic causal correspondence; otherwise, a penalty is imposed on its performance index according to the degree of constraint violation, and the result after penalty is used as the fitness. The selection module is used to select individuals based on their fitness to form a new generation of the population. The output module is used to repeatedly trigger the iterative optimization module, evaluation module, fitness determination module and selection module until the preset termination condition is met, and output the optimal mixed working fluid ratio, optimal high pressure, optimal low pressure and corresponding performance indicators.

[0009] Thirdly, this application provides a computer device, the computer device including a processor and a memory coupled to the processor, wherein the memory stores program instructions for implementing a method for optimizing a mixed refrigerant refrigeration system; the processor is used to execute the program instructions stored in the memory to implement a method for optimizing a mixed refrigerant refrigeration system.

[0010] Fourthly, this application provides a computer-readable storage medium storing processor-executable program instructions for executing a method for optimizing a mixed refrigerant refrigeration system.

[0011] This application provides an optimization method and system for a mixed refrigerant refrigeration system, which has the following beneficial effects: This application first establishes a process simulation model and determines the range of optimization variables. When generating the initial population, a hybrid generation strategy combining mechanistic seeds, local perturbations, and global random sampling is adopted. This ensures that the initial population possesses both high-quality feasible individuals and global diversity, providing a better search starting point for subsequent iterative optimization. In each iteration, offspring individuals are generated through a population iterative optimization algorithm. After process simulation evaluation, performance indicators and constraint violation information are obtained. Then, constraint violation information is categorized and processed: individuals satisfying all constraints are directly used with performance indicators as fitness; for individuals violating heat transfer temperature difference-related constraints but within the near-feasible repair range... Based on the partial molar enthalpy difference of each component in the critical heat exchange temperature zone, the component ratio is directionally corrected, thereby integrating thermodynamic mechanisms into the local repair process and effectively improving the utilization efficiency of near-feasible and infeasible solutions. For individuals that violate compressor exhaust temperature constraints or intake phase constraints, the high-pressure, low-pressure, or component ratio is directionally corrected based on thermodynamic causal correspondence, achieving differentiated and precise repair for different types of constraint violations. Penalties are applied to other constraint violations. In implementations that include high-pressure and low-pressure optimization, the repair operation only adjusts the component ratio while keeping the pressure variable unchanged, avoiding interference with the global search of pressure parameters. Through the above selection, a new generation of population is formed and iterated repeatedly until the termination condition is met, outputting the optimal ratio, pressure, and performance indicators. This application can retain the global search capability of the swarm iterative optimization algorithm, while using partial molar enthalpy difference to perform directional thermodynamic repair on near-feasible individuals with temperature difference-related constraints, and provides heuristic repair based on causal relationships for compressor state constraints. It realizes differentiated processing of various constraint types, improves constraint processing efficiency and convergence stability, and is applicable to various swarm optimization frameworks such as differential evolution, genetic algorithm, and particle swarm optimization, and has good engineering adaptability. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating an optimization method for a mixed refrigerant refrigeration system according to Embodiment 1 of this application; Figure 2 This is a schematic diagram of the mixed working fluid refrigeration system structure of Embodiment 1 of this application; Figure 3 This is a comparison diagram of ablation results from Example 1 of this application; Figure 4 This is a schematic diagram of the structure of an optimized mixed refrigerant refrigeration system according to Embodiment 2 of this application; Figure 5 This is a schematic diagram of the computer device structure according to Embodiment 3 of this application; Figure 6 This is a schematic diagram of the storage medium structure of Embodiment 4 of this application. Detailed Implementation

[0013] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0015] Example 1 Please see Figure 1 This is a flowchart illustrating an optimization method for a mixed refrigerant refrigeration system according to Embodiment 1 of this application; the steps include: S1. Establish a process simulation model of the mixed working fluid refrigeration system, determine the constraints and the range of values ​​for the distribution ratio of each component of the mixed working fluid, as well as the range and step size of the high pressure and the low pressure.

[0016] In this embodiment, a steady-state process simulation model of the mixed working fluid refrigeration system is first established. This model can be implemented using commercial process simulation software, open-source process simulation programs, or a self-built thermodynamic model. Its core function is to output system performance indicators and constraint violation information under given input conditions such as the mixed working fluid composition ratio, high pressure, and low pressure, providing complete thermodynamic data support for subsequent iterative optimization.

[0017] This embodiment constructs a process simulation model of a mixed refrigerant refrigeration system. The system consists of a screw compressor, a water-cooled condenser, a plate-fin internal regenerator, a capillary expansion valve, and a shell-and-tube evaporator. The mixed refrigerant is pressurized by the compressor and enters the condenser, where it is cooled to saturation or subcooling by 300K cooling water. After counter-current heat exchange with the low-pressure side cold flow in the regenerator, it enters the expansion valve for throttling and cooling. In the evaporator, it absorbs heat from the object being cooled and evaporates to the target temperature of 120K. The low-pressure vapor is preheated by the regenerator and then returned to the compressor inlet. The thermodynamic model uses the Peng-Robinson equation of state to establish thermodynamic parameter calculation models, providing basic physical property parameters such as enthalpy, entropy, and fugacity of each component. The final output parameters of the model include the system coefficient of performance (COP), cooling capacity, compressor discharge temperature, compressor suction gas phase fraction, global minimum temperature difference of the heat exchanger, temperature difference curve along the heat exchange path, and fugacity values ​​of each component under the corresponding conditions on the high and low pressure sides.

[0018] Next, the constraints were determined. Based on the actual operating requirements of the cryogenic refrigeration project and the equipment safety specifications, three types of core constraints were set: The first type is the heat exchange temperature difference constraint, which is a fundamental constraint that must be met. The minimum allowable temperature difference of all heat exchange nodes in the heat exchanger is set to be no less than 3.0K to avoid heat exchange pinch points that could lead to a sharp drop in system heat exchange efficiency or even failure to operate normally. The second type is the compressor safe operation constraint, which sets the upper limit of the compressor exhaust temperature to 380K to prevent the lubricating oil from carbonizing and deteriorating and the internal components of the compressor from overheating and being damaged. The third type is the suction phase constraint, which sets the lower limit of the compressor inlet gas phase fraction to 0.9999 to ensure that there is no risk of liquid slugging in the compressor.

[0019] Finally, the value range and dispersion of each optimization variable were determined. Nitrogen was selected as the candidate component of the mixed working fluid. ), methane ( ), ethane ( ), propane ( ) and isobutane ( There are 5 commonly used working fluids suitable for the 120K temperature range, and the distribution ratio of each group is as follows: The range of values ​​is And the sum of the group allocation ratios satisfies the normalization constraint. The high-pressure value PH is set to a range of 1200 kPa to 2200 kPa, with a distance step size of 100 kPa; the low-pressure value PL is set to a range of 200 kPa to 500 kPa, with a distance step size of 50 kPa; and the maximum pressure ratio constraint PH / PL ≤ 10 is set.

[0020] S2. Generate an initial population within the range of values. The initial population consists of a combination of mechanism seed individuals, derived individuals generated based on local perturbations of the mechanism seed individuals, and globally randomly sampled individuals. Each individual in the initial population corresponds to a set of optimized variable values. The optimized variable values ​​include at least the distribution ratio of each group of the mixed working fluid, as well as the high pressure and low pressure.

[0021] In this embodiment, based on the range of values ​​for the optimization variables determined in step S1, the initial population is generated using a hybrid generation strategy that combines mechanistic seeds, local perturbation, and global random sampling. The specific process is as follows: Generating Mechanism Seeds: First, using a thermodynamic mechanism model, under all constraints, 3 to 4 high-quality feasible solutions are generated as mechanism seed individuals. These seed individuals are already within the feasible region, providing a high starting point for subsequent population evolution.

[0022] Assemble the initial population: The entire initial population consists of three parts: First, the mechanism seed individuals mentioned above are directly retained to ensure that the population contains at least 3 to 4 known feasible solutions; Second, based on the mechanism seed individuals, random perturbations within a preset range are applied near their high pressure, low pressure, and composition ratio to generate a batch of derivative individuals, realizing local mining of the seed neighborhood space; Third, the remaining individuals are generated using a random sampling method within the global value range to fill the preset population size, so as to take into account the global exploration of the solution space.

[0023] Each individual in the initial population corresponds to a complete set of optimization variable values, and the encoding structure is uniformly set as follows: Where PH represents high pressure and PL represents low pressure. to Corresponding to nitrogen gas in sequence ( ), methane ( ), ethane ( ), propane ( ) and isobutane ( The initial population size was set to 30, a value derived from the verification parameters of this application embodiment. This value comprehensively balances global search coverage and process simulation computational cost, and matches the parameters of the subsequent population iterative optimization algorithm. The initial population adopts a hybrid generation strategy combining mechanistic seeds, local perturbation, and global random sampling. This mechanism not only gives the population a high evolutionary starting point but also takes into account the diversity of the understanding space.

[0024] For the generation of globally randomized sampled individuals in the third part mentioned above, a set of legal discrete pressure combinations is first constructed: according to the pressure range and step size determined in step S1, the high pressure (1200 kPa to 2200 kPa) is discretized into 11 high-pressure grid points with a step size of 100 kPa, and the low pressure (200 kPa to 500 kPa) is discretized into 7 low-pressure grid points with a step size of 50 kPa. All combinations of high-pressure and low-pressure grid points are enumerated, and each combination is verified to meet the pressure constraints of PH > PL and PH / PL ≤ 10. Illegal combinations that do not meet the conditions are eliminated, resulting in the final set of legal discrete pressure combinations. Random sampling is performed from this set, and a set of high-pressure and low-pressure values ​​is assigned to each globally randomized sampled individual. This method directly avoids the problem of generating illegal pressure combinations during the group iterative optimization process, reducing invalid computation. For situations where pressure variables deviate from legal discrete grid points during subsequent iterations, they will be automatically absorbed into the nearest legal discrete pressure combination, ensuring that all individuals are always searching within the feasible pressure space.

[0025] For the generation of pressure variables of the second part of the derived individuals, a small perturbation can be applied near the pressure value of the mechanism seed, and then the individual can be absorbed into the nearest legal discrete pressure combination.

[0026] For the component proportions, a method combining uniform random sampling and normalization is used. First, uniform random numbers between 0 and 1 are generated for each of the five components. to Then through the formula The allocation ratios for each group are calculated to ensure that the allocation ratios for all groups meet the requirements. Physical constraints.

[0027] To further improve the spatial uniformity of the initial global random sampling portion of the population, Latin hypercube sampling or low-discrepancy sequence sampling can be used instead of pure random sampling. The core generation steps remain unchanged: first, pressure variables are selected from the set of legal discrete pressure combinations, and then group distribution ratios satisfying non-negativity and normalization constraints are generated. Latin hypercube sampling ensures that sample points are selected for each layer of each variable, effectively avoiding the local sample clustering problem that may occur with pure random sampling, and is more conducive to the population iterative optimization algorithm quickly exploring high-quality solution regions in the early stages. It should be noted that the replacement of the above sampling method only involves the global random sampling portion; the generation methods of mechanistic seed individuals and derived individuals generated based on seed perturbation remain unchanged.

[0028] The final 30 initial individuals constitute a complete initial population, fully covering the preset pressure range and proportion space. This initial population generation method does not rely on any thermodynamic prior information, but only ensures the legitimacy of individuals through constrained sampling. It can fully leverage the global search capability of the population iterative optimization algorithm, providing a diverse initial candidate set for subsequent iterative optimization and selective repair of partial molar enthalpy differences.

[0029] S3. The population is iteratively optimized using a population iterative optimization algorithm to generate offspring individuals; wherein, the current population at the time of the first iteration is the initial population, and the offspring individuals are generated through at least one of differential mutation, crossover, position update, and perturbation search.

[0030] In this embodiment, after the initial population generation is completed, the population iterative optimization process begins. Let the current iteration number be . ,when At this point, the current population is the initial population generated in step S2; in subsequent iterations, the current population is updated to the new generation population formed after selection from the previous generation. For each generation, the individuals in the current population are operated on using a population iterative optimization algorithm to generate offspring individuals. In this embodiment, differential evolution algorithm is preferably used as the population iterative optimization algorithm, but it can also be switched to genetic algorithm, particle swarm optimization algorithm, or evolutionary strategy algorithm as needed. The generation method of offspring individuals includes at least one of differential mutation, crossover, position update, or perturbation search.

[0031] When using the differential evolution algorithm, for each target individual in the current population, three distinct individuals are randomly selected from the population. The difference between the optimized variable values ​​of two of these individuals is multiplied by a scaling factor and added to the third individual to generate a donor vector, i.e., the differential mutation operation. This embodiment employs a phased differential mutation strategy to balance global exploration and local convergence capabilities: in the early stages of iteration, rand / 1 type differential mutation is used to fully maintain population diversity and prevent the algorithm from getting trapped in local optima too early; in the later stages of iteration, best / 1 type differential mutation guided by the current best individual is used to strengthen local search capabilities and improve convergence efficiency. Subsequently, a binomial crossover operation is performed between the donor vector and the optimized variable values ​​of the target individual to generate experimental individuals, i.e., offspring candidate individuals. Uniform constraint processing is applied to the generated experimental individuals: for the mixed working fluid composition ratio dimension, a non-negativity constraint is first applied to the crossover ratio values, correcting ratio values ​​less than 0 to 0, and then all composition ratios are normalized to satisfy the following conditions. Physical constraints; for the pressure variable dimension, if its value does not fall within the legal discrete pressure combination pre-constructed in step S1, it is absorbed into the nearest legal discrete pressure combination to ensure that the pressure variable always satisfies the constraint conditions of PH>PL and PH / PL≤10.

[0032] When using a genetic algorithm, offspring individuals are generated through selection, crossover, and mutation operations. When using a particle swarm optimization algorithm, offspring individuals are generated through position updates, i.e., recording the historical best position of each individual and the global best position of the population, updating the velocity of each individual based on inertia weight, individual learning factor, and social learning factor, and then updating its position to obtain offspring individuals. When using an evolutionary strategy algorithm, offspring individuals are generated through perturbation search, i.e., adding a normally distributed random perturbation to the current individual's optimization variable value to generate offspring individuals, the perturbation amplitude can adaptively decrease with the number of iterations to balance global exploration and local convergence.

[0033] In actual operation, different offspring generation methods can be selected at different iteration stages. For example, differential mutation or perturbation search can be used in the early stages of iteration to maintain population diversity, while particle swarm optimization can be used for position updates to accelerate convergence in the later stages of iteration. Regardless of the algorithm and generation method used, all generated offspring individuals must undergo the above-mentioned constraint processing to ensure that they meet the value range of the optimization variables and physical constraints, thus providing effective candidate inputs for subsequent process simulation and evaluation.

[0034] S4. Input the optimization variable values ​​of the offspring individuals into the process simulation model for evaluation, and obtain the performance indicators and constraint violation information of each offspring individual.

[0035] In this embodiment, all the legally valid offspring individuals generated in step S3 after constraint processing are individually input into the established steady-state process simulation model of the mixed refrigerant refrigeration system for evaluation. The evaluation process of each offspring individual is independent and does not interfere with each other. The input optimization variable values ​​are the codes corresponding to each offspring individual. .

[0036] This embodiment constructs its own process simulation model, based on the Peng-Robinson equation of state and a self-built thermodynamic parameter calculation model to provide physical property parameters. According to the input pressure and component ratio parameters, it sequentially calculates the thermodynamic properties at each state point: compressor inlet / outlet, condenser outlet, regenerator inlet / outlet, before and after the expansion valve, and evaporator inlet / outlet. Through iterative convergence of solving mass conservation, energy conservation, and pressure drop equations, it completes the steady-state simulation calculation of the refrigeration cycle. After the calculation, the process simulation model outputs the system's core performance indicators. In this embodiment, the coefficient of performance (COP) is selected as the performance indicator, which is the ratio of cooling capacity to compressor power consumption. A higher COP value indicates better system energy efficiency. In other embodiments, the performance indicator can also be cooling capacity, cooling capacity per unit power consumption, or a comprehensive economic indicator for subsequent fitness calculations and optimal solution selection.

[0037] Simultaneously, the process simulation model outputs constraint violation information and thermodynamic data required for subsequent repair, specifically including: the heat transfer temperature difference curve along the heat transfer length between the hot and cold flows in the heat exchanger, the global minimum temperature difference in the entire heat transfer temperature difference curve, the minimum temperature difference of the critical temperature zone corresponding to the local minimum point in the heat transfer curve, the discharge temperature at the compressor outlet, the gas phase fraction of the intake gas at the evaporator outlet (i.e., the compressor inlet), and the fugacity values ​​of each component under the corresponding states on the high-pressure and low-pressure sides. Among these, the global minimum temperature difference, compressor discharge temperature, and intake gas phase fraction are used to directly determine whether the progeny individuals meet the preset constraint conditions; the heat transfer temperature difference curve and the minimum temperature difference of the critical temperature zone are used to subsequently extract the critical temperature zone of the heat transfer; and the fugacity values ​​of each component are used to calculate the partial molar enthalpy difference within the critical temperature zone, providing thermodynamic data support for selective repair of the partial molar enthalpy difference. The performance indicators and constraint violation information of all progeny individuals will be structured and stored to form a complete progeny evaluation dataset, serving as the basis for determining the feasibility of progeny individuals, whether repair is triggered, and calculating fitness in the subsequent S5 step.

[0038] S5. Based on the constraint violation information, classify and determine the fitness of each offspring individual: if all preset constraints are met, the performance index of the individual is used as the fitness; if the heat exchange temperature difference related constraint is violated and is within the preset near-feasible repair range, the component ratio of the individual is corrected based on the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone; if the compressor exhaust temperature constraint or intake phase constraint is violated, the high pressure, low pressure or component ratio is corrected based on the thermodynamic causal correspondence; otherwise, the performance index is penalized according to the degree of constraint violation, and the penalized result is used as the fitness.

[0039] In this embodiment, based on the performance indicators and constraint violation information of each offspring individual obtained in step S4, each offspring individual is classified and its fitness is determined. The classification criteria are based on the satisfaction of preset constraints, the type of constraint violation, and the degree of violation.

[0040] First, determine whether the offspring individuals belong to the first category: individuals that satisfy all preset constraints. If an offspring individual simultaneously satisfies the following conditions: global minimum temperature difference ≥ 3.0K, compressor exhaust temperature ≤ 380K, and intake gas phase fraction ≥ 0.9999, that is, simultaneously meets the minimum allowable temperature difference constraint, compressor safe operation constraint, and intake phase state constraint, then the individual's performance index COP is directly used as its fitness. The higher the COP value, the better the fitness, and the greater the probability of being selected in subsequent selections.

[0041] If the offspring individuals do not meet all preset constraints, it is further determined whether they belong to the second category: the offspring individuals have met both the compressor exhaust temperature constraint and the intake phase constraint, but violated the minimum allowable temperature difference constraint, and the degree of violation is within the preset near-feasible repair range. In this embodiment, the near-feasible repair range is determined by the following conditions: First, the global minimum temperature difference is greater than the repair threshold that adapts to the iteration number (1.0K in the early stage of iteration, gradually increasing to 2.5K in the later stage) and less than the minimum allowable temperature difference of 3.0K; Second, the minimum temperature difference in the key temperature zone of heat exchange is less than 3.0K, but the violation magnitude is less than the preset temperature difference tolerance of 1.0K. For individuals that trigger repair, partial molar enthalpy difference selective repair is performed, and after repair, a best-fit acceptance mechanism is implemented: the corrected individuals are re-input into the process simulation model for evaluation. If the newly obtained COP value is better than that before the correction, the corrected COP value is used as the fitness; if the performance does not improve after repair, the fitness before the correction is retained (a penalty has been applied according to the degree of constraint violation before the correction), so as to avoid the repair operation from having a negative impact on the population evolution.

[0042] Specifically, the steps for selective repair of partial molar enthalpy difference include: 1. Extract the region in the heat transfer curve where the temperature difference is lower than the minimum allowable temperature difference as the critical temperature zone for heat transfer.

[0043] The endpoints, local minimum temperature difference points, global minimum temperature difference points, and local temperature difference points with a temperature difference value below 4.0K in the heat transfer curve are selected as candidate key temperature zones. The equivalent enthalpy mismatch of each candidate point is calculated, and the zones are sorted from largest to smallest according to the absolute value of the equivalent enthalpy mismatch. The zone with the worst heat transfer matching ranked first is selected as the repair target.

[0044] 2. Calculate the equivalent enthalpy mismatch in the critical heat exchange temperature zone, and the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone.

[0045] The equivalent enthalpy mismatch is defined as: ; in, For the mixed working fluid at temperature Isothermal enthalpy difference at that point This refers to the enthalpy difference on the low-pressure side under the condition of meeting the minimum allowable temperature difference. For the cooled material at temperature The cumulative heat load at the location. This indicates that the effective cooling capacity provided by the mixed working fluid at this temperature point is insufficient, and the actual heat exchange temperature difference is less than the minimum allowable temperature difference; This indicates that the cooling capacity is relatively redundant, and the temperature point just reaches the minimum allowable temperature difference boundary. When When the temperature is 0, it indicates that there is still a thermodynamic margin at that temperature point, and the actual heat exchange temperature difference is not less than the minimum allowable temperature difference.

[0046] Simultaneously, using the fugacity values ​​of each component on the high and low pressure sides output in step S4, the partial molar enthalpy difference of each component at the corresponding temperature in this critical temperature range is calculated using a finite difference approximation. Specifically, the partial molar enthalpy difference is calculated using a finite difference approximation formula:

[0047] in, The gas constant is... The temperature perturbation step size, For high-pressure eccentric enthalpy, It is a low-pressure partial molar enthalpy. For components The degree of divergence, Assuming high pressure and constant component concentration, To ensure that the low-pressure environment and the concentration of each component remain constant, For temperature. The above formula is only a preferred approximation. In actual implementation, the partial molar enthalpy, partial molar enthalpy difference, or equivalent thermodynamic sensitivity directly output from the thermodynamic model can also be used.

[0048] 3. Based on the sign and magnitude of the equivalent enthalpy loss, and in conjunction with the relative contribution of the partial molar enthalpy difference of each component, determine the component to be increased or decreased in proportion and its adjustment amount.

[0049] like The component with the largest partial molar enthalpy difference is selected as the component to be increased in the ratio; if Then, the selection is reversed. Based on the absolute value of the equivalent enthalpy mismatch, the partial molar enthalpy difference of the selected components, and the preset repair coefficient (0.05 in this embodiment), the ratio adjustment amount is calculated. The adjustment amount is directly proportional to the absolute value of the equivalent enthalpy mismatch and inversely proportional to the magnitude of the partial molar enthalpy difference.

[0050] 4. Normalize the adjusted group allocation ratio to obtain the corrected group allocation ratio.

[0051] The adjusted proportions were sequentially constrained by non-negativity (negative values ​​were set to 0), minimum retention (at least 0.001 of each component was retained), and maximum remediation step size (the adjustment range in a single step did not exceed 0.1). Finally, normalization was performed so that the sum of the proportions of each group was 1. Throughout the remediation process, the high-pressure pH and low-pressure PL of the individual were kept constant.

[0052] For individuals that violate compressor discharge temperature constraints or intake phase constraints, compressor state self-repair is triggered: if the compressor discharge temperature constraint is violated, the pressure ratio is reduced and the highest boiling point group allocation ratio is increased; if the intake phase constraint is violated, the highest boiling point group allocation ratio is reduced first, and when the highest boiling point group allocation ratio has reached the preset lower limit, the low pressure is reduced. After repair, the process simulation model is called again for evaluation. If the fitness of the repaired individual is better than that of the unrepaired individual, the repaired individual is accepted; otherwise, the unrepaired individual is retained.

[0053] For individuals that violate heat exchange temperature difference constraints but are not within the near-feasible repair range, or that still do not meet the constraints after compressor condition repair, a penalty function method is used to apply a penalty term based on the degree of constraint violation. The fitness calculation formula is: Fitness = Original COP - Penalty Term. The penalty term consists of a weighted sum of one or more of the following: exhaust temperature violation, intake phase violation, global minimum temperature difference violation, and critical temperature zone temperature difference violation. The specific value of the penalty coefficient can be set based on engineering experience or adaptively adjusted with the number of iterations.

[0054] After the above classification process, each offspring individual obtains a unique fitness value. All fitness values ​​are structured and stored as a fitness array for use in the population selection in step S6.

[0055] S6. Select individuals based on their fitness to form a new generation of population.

[0056] In this embodiment, based on the final fitness values ​​of all offspring individuals determined in step S5, a population selection operation is performed to select individuals with better fitness from the current population and their corresponding offspring individuals, forming a new generation population for the next round of iteration optimization. The size of the new generation population remains consistent with the initial population, which is 30 individuals in this embodiment. All individuals retain the encoding format defined in step S2, and the values ​​of the optimization variables are always within the legal range determined in step S1, ensuring the stability of subsequent iterations.

[0057] This embodiment corresponds to the differential evolution algorithm selected in step S3, employing a one-to-one greedy selection strategy. Specifically, for each index... The first one generated in step S3 The number of offspring individuals, and the number of offspring individuals currently used to generate that offspring in the population. The fitness of each target individual is compared: if the fitness of the offspring is better than or equal to that of the target individual, the offspring is retained in the new generation of the population; if the fitness of the offspring is worse than that of the target individual, the original target individual is retained in the new generation of the population. This strategy ensures that the overall fitness of the population does not degenerate with iteration. For individuals that triggered selective repair of the partial molar enthalpy difference in step S5, their final fitness is already the better value before and after the repair, and they can directly participate in the above one-to-one comparison without further evaluation.

[0058] When using other population iterative optimization algorithms, a corresponding selection strategy should be adopted. When using genetic algorithms, roulette wheel selection or tournament selection can be used: roulette wheel selection allocates the probability of selection based on the proportion of an individual's fitness to the total fitness of the population, with individuals having higher fitness having a greater probability of being selected; tournament selection randomly selects several individuals to compete, retaining the individual with the best fitness to enter the next generation. When using particle swarm optimization, explicit individual selection is not required; instead, population evolution is driven by updating the historical best position of each individual and the global best position of the population. When using evolutionary strategy algorithms, the following can be used: Choose or Choose from Select from offspring individuals The most fit individuals form a new generation of the population.

[0059] S7. Repeat steps S3 to S6 until the preset termination condition is met, and output the optimal mixed working fluid ratio, optimal high pressure, optimal low pressure and corresponding performance indicators.

[0060] In this embodiment, after completing the population selection in step S6, the generated new generation population is used as the current population for the next iteration, and the complete iterative optimization process from steps S3 to S6 is repeated until a preset termination condition is met. In this embodiment, the preset termination condition is that the current iteration generation reaches the maximum iteration generation of 100.

[0061] Iteration stops when the number of generations reaches 100. The individual with the highest fitness is selected from all individuals recorded in each generation of the population as the final optimization result. The output should include at least: the optimal working fluid ratio, i.e., the nitrogen gas content corresponding to the optimal individual (…). ), methane ( ), ethane ( ), propane ( ) and isobutane ( The output includes the mole fraction of the mixed refrigerant; the optimal high-pressure; the optimal low-pressure; and the corresponding system performance indicators (such as the coefficient of performance, COP). All outputs are based on the final evaluation data of the process simulation model and can be directly used to guide the actual engineering design or operating parameter optimization of the mixed refrigerant refrigeration system.

[0062] Please see Figure 2This is a schematic diagram of the mixed refrigerant refrigeration system according to Embodiment 1 of this application. The mixed refrigerant refrigeration system includes a compressor, a cooler or condenser, a regenerator, a throttling valve, an evaporator, and connecting streams between the components. After being pressurized by the compressor, the mixed refrigerant enters the high-pressure side, where it is cooled or condensed under cooling conditions at an ambient temperature of approximately 300K. After heat exchange through regeneration, it is throttled and cooled, and then provides cooling capacity to the object being cooled in the evaporator. The evaporator outlet temperature is set to 120K. The evaporated mixed refrigerant returns to the compressor inlet via the low-pressure side. In the figure, the high-pressure, low-pressure, and mixed refrigerant composition ratio are considered as optimizable or partially optimizable variables, while the COP, exhaust temperature, suction gas phase fraction, and heat exchange temperature difference curves serve as evaluation and constraint information.

[0063] Please see Figure 3 Figure 1 shows the ablation comparison results of Embodiment 1 of this application; specifically, it shows the ablation comparison results of the method proposed in this application and the pure population iterative optimization algorithm (control group). Under the same initial candidate population, the same random seed, and the same population iterative optimization parameters, the horizontal axis represents the iteration number, and the vertical axis represents the coefficient of performance (COP). In the figure, the solid line represents the method of this application with embedded partial molar enthalpy difference selective repair, and the dashed line represents the pure population iterative optimization algorithm (control group). Experimental data show that: in generation 0, the initial optimal COP of both is 0.027; in generation 1, the COP of the method of this application reaches 0.07496, while that of the control group is 0.06552; in generation 7, the COP of the method of this application increases to 0.11125, while that of the control group is 0.10365; in generation 49, the COP of the method of this application is 0.20001, while that of the control group is 0.18355; in generation 100, the COP of the method of this application converges to 0.20020, while that of the control group is 0.20004. It can be seen that the method in this application is significantly better than the control group in terms of convergence speed, can achieve a higher COP level with fewer iterations, and maintains a stable performance advantage in the final stage.

[0064] In summary, this embodiment provides a specific implementation of an optimization method for a mixed refrigerant refrigeration system. First, a process simulation model is established to determine the refrigerant components, their proportions, and the ranges and step sizes for high and low pressures. The initial population is generated using a hybrid generation strategy combining mechanistic seeds, local perturbations, and global random sampling, ensuring both a high-quality evolutionary starting point and global diversity. In each iteration, a population iterative optimization algorithm performs crossover and mutation operations on the current population to generate offspring individuals, and the performance indicators and constraint violation information of each offspring are obtained through the process simulation model. Individuals are categorized based on constraint violations: those satisfying all constraints are directly assessed using their performance indicators as fitness; individuals violating heat exchange temperature difference constraints but within a near-feasible repair range undergo targeted correction of their component ratios based on the partial molar enthalpy difference of each component in the critical heat exchange temperature zone, are re-evaluated after repair, and are only accepted if performance improves; individuals violating compressor exhaust temperature constraints or intake phase constraints have their high-pressure, low-pressure, or highest boiling point component ratios corrected based on thermodynamic causal relationships, are re-evaluated after repair, and the best performing individual is accepted; the remaining individuals are penalized based on their performance indicators. A new generation of the population is then selected based on fitness, and this process is repeated iteratively until the termination condition is met, outputting the optimal mixed working fluid ratio, high-pressure, low-pressure, and corresponding performance indicators.

[0065] Example 2 Please see Figure 4 This is a schematic diagram of the structure of an optimized mixed refrigerant refrigeration system according to Embodiment 2 of this application; the specific contents include: The model building module 100 is used to build a process simulation model of the mixed working fluid refrigeration system, determine the constraints and the range of values ​​for the distribution ratio of each component of the mixed working fluid, as well as the range and step size of the high pressure and the low pressure. An initialization module 200 is used to generate an initial population within the range of values. The initial population consists of a combination of mechanism seed individuals, derived individuals generated based on local perturbations of the mechanism seed individuals, and globally randomly sampled individuals. Each individual in the initial population corresponds to a set of optimized variable values, which at least include the distribution ratio of each group of the mixed working fluid, as well as the high pressure and low pressure. The iterative optimization module 300 is used to iteratively optimize the population through a population iterative optimization algorithm to generate offspring individuals; wherein, the current population at the time of the first iteration is the initial population, and the offspring individuals are generated through at least one of differential mutation, crossover, position update, and perturbation search. The evaluation module 400 is used to input the optimization variable values ​​of the offspring individuals into the process simulation model for evaluation, and obtain the performance indicators and constraint violation information of each offspring individual; The fitness determination module 500 is used to classify and determine the fitness of each offspring individual based on the constraint violation information: if all preset constraints are met, the performance index of the individual is used as the fitness; if the heat exchange temperature difference related constraint is violated and is within the preset near-feasible repair range, the component ratio of the individual is corrected based on the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone; if the compressor exhaust temperature constraint or intake phase constraint is violated, the high pressure, low pressure or component ratio is corrected based on the thermodynamic causal correspondence; otherwise, the performance index is penalized according to the degree of constraint violation, and the penalized result is used as the fitness. Selection module 600 is used to select individuals based on their fitness to form a new generation of population; The output module 700 is used to repeatedly trigger the iterative optimization module, evaluation module, fitness determination module and selection module until the preset termination condition is met, and output the optimal mixed working fluid ratio, optimal high pressure, optimal low pressure and corresponding performance indicators.

[0066] In this embodiment, the model building module 100 is used to build a steady-state process simulation model of the mixed working fluid refrigeration system. In specific implementation, commercial software or a self-built thermodynamic model can be used, and the input parameters include the types of mixed working fluid components (e.g., ...). The output includes the range of distribution ratios for each group, as well as the high-pressure range (1200-2200 kPa, step size 100 kPa) and the low-pressure range (200-500 kPa, step size 50 kPa). It also outputs the objective function (e.g., COP) and information required for constraint and repair, such as heat exchange temperature difference curves, exhaust temperature, intake gas phase fraction, and fugacity.

[0067] The initialization module 200 generates an initial population within the value range determined by the model building module 100. The initial population employs a hybrid generation strategy combining mechanistic seeds, local perturbations, and global random sampling: first, a set of legal discrete pressure combinations is generated; then, mechanistic seed individuals are generated based on the thermodynamic mechanism model under constrained conditions; local perturbations are applied using the mechanistic seed individuals as a baseline to generate derived individuals; finally, random sampling is performed within the global value range to generate the remaining individuals, filling the preset population size. Each individual's code includes high pressure, low pressure, and the distribution ratio of each component of the mixed working fluid. The global random sampling can employ pure randomness, Latin hypercube, or low-difference sequence sampling to ensure initial population diversity. This hybrid generation strategy provides the population with a high evolutionary starting point while also considering the diversity of the understanding space.

[0068] The iterative optimization module 300 employs a population iterative optimization algorithm to operate on the current population and generate offspring individuals. Generation methods include at least one of differential mutation, crossover, position update, or perturbation search. For example, in differential evolution, three different individuals are randomly selected for each target individual for mutation and crossover; particle swarm optimization updates velocity and position; and evolutionary strategies add normal perturbations. After generation, the matching variables need to be non-negatively corrected and normalized, and the pressure variables are absorbed to the nearest valid discrete pressure combination.

[0069] The evaluation module 400 inputs the optimization variable values ​​of each offspring individual into the process simulation model, through Peng Steady-state cyclic calculations are performed using the Robinson equation of state and the REFPROP property library. Outputs include performance index (COP), heat transfer temperature difference curves, global minimum temperature difference, minimum temperature difference in critical temperature zones, exhaust temperature, intake gas phase fraction, and fugacity values ​​of each component on the high-pressure and low-pressure sides.

[0070] The fitness determination module 500 classifies and processes each offspring individual based on the constraint violation information returned by the evaluation module 400. If the minimum temperature difference is ≥3K, the exhaust temperature is ≤380K, and the gas phase fraction of the inhalation is ≥0.9999, then the COP is directly used as the fitness. If only the minimum temperature difference is violated and it is within the near-feasible range, then the partial molar enthalpy difference repair is triggered: the equivalent enthalpy mismatch is calculated, and the worst temperature range is selected by sorting by absolute value, according to... The component with the largest / smallest partial molar enthalpy difference is selected for adjustment, normalized after constraint processing, and the best repair result is accepted. If the compressor discharge temperature constraint or suction phase constraint is violated, the compressor state self-repair is triggered: if the discharge temperature constraint is violated, the pressure ratio is reduced and the highest boiling point component ratio is increased; if the suction phase constraint is violated, the highest boiling point component ratio is reduced first, and the low pressure is reduced when the component has reached the preset lower limit; after repair, the result is re-evaluated and the best repair result is accepted. In other cases, a penalty is applied, and the result after penalty is used as the fitness.

[0071] The selection module 600 determines the final fitness of each offspring individual obtained by the fitness module 500 based on the fitness, and selects the better offspring individuals from the current population to form a new generation of population.

[0072] Output module 700 is responsible for iteration control and result output. After each selection, it checks whether the termination condition is met (e.g., reaching the maximum iteration number of 100 generations). If not, it triggers iteration optimization module 300 to continue to the next generation; if so, it extracts the optimal working fluid ratio, optimal high-pressure, optimal low-pressure, and corresponding COP and other performance indicators from the best individuals in each generation. If the final individual still does not meet all constraints, it can output a near-optimal feasible solution and mark the violation. All results are based on the final evaluation data of the process simulation model and are directly used for engineering guidance.

[0073] For further details regarding the implementation of the technical solutions for each module in the above-described hybrid refrigerant refrigeration system optimization system, please refer to the description in the above-described hybrid refrigerant refrigeration system optimization method, which will not be repeated here.

[0074] Example 3 Please see Figure 5 This is a schematic diagram of the computer device structure according to Embodiment 3 of this application. The computer device 50 includes a processor 51 and a memory 52 coupled to the processor 51.

[0075] The memory 52 stores program instructions for implementing the above-described method for optimizing a mixed refrigerant refrigeration system.

[0076] The processor 51 is used to execute program instructions stored in the memory 52 to implement a hybrid refrigerant refrigeration system optimization.

[0077] The processor 51 can also be referred to as a CPU (Central Processing Unit).

[0078] Processor 51 may be an integrated circuit chip with signal processing capabilities. Processor 51 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor may be a microprocessor or any conventional processor.

[0079] Example 4 Please see Figure 6 This is a schematic diagram of the storage medium in Embodiment 4 of this application. The storage medium in this embodiment stores a program file 61 capable of implementing all the above methods. This program file 61 can be stored in the storage medium in the form of a software product, including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, or devices such as computers, servers, mobile phones, and tablets.

[0080] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0081] The above description is only a preferred embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural changes made based on the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0082] Although embodiments of this application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the appended claims and their equivalents.

[0083] Of course, the present invention may have many other embodiments. Based on this embodiment, other embodiments obtained by those skilled in the art without any creative effort are all within the scope of protection of the present invention.

Claims

1. A method for optimizing a mixed working fluid refrigeration system, characterized in that, include: S1. Establish a process simulation model of the mixed working fluid refrigeration system, determine the constraints and the range of values ​​for the distribution ratio of each component of the mixed working fluid, as well as the range and step size of the high pressure and the low pressure. S2. Generate an initial population within the range of values. The initial population consists of a combination of mechanism seed individuals, derived individuals generated based on local perturbations of the mechanism seed individuals, and globally randomly sampled individuals. Each individual in the initial population corresponds to a set of optimized variable values, which at least include the distribution ratio of each group of the mixed working fluid, as well as the high pressure and low pressure. S3. The population is iteratively optimized using a population iterative optimization algorithm to generate offspring individuals; wherein, the current population at the first iteration is the initial population, and the offspring individuals are generated through at least one of differential mutation, crossover, position update, and perturbation search. S4. Input the optimization variable values ​​of the offspring individuals into the process simulation model for evaluation, and obtain the performance indicators and constraint violation information of each offspring individual; S5. Based on the constraint violation information, classify and determine the fitness of each offspring individual: if all preset constraints are met, the performance index of the individual is used as the fitness; if the heat exchange temperature difference related constraint is violated and is within the preset near-feasible repair range, the component ratio of the individual is corrected based on the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone; if the compressor exhaust temperature constraint or intake phase constraint is violated, the high pressure, low pressure or component ratio is corrected based on the thermodynamic causal correspondence; otherwise, the performance index is penalized according to the degree of constraint violation, and the penalized result is used as the fitness. S6. Select individuals based on their fitness to form a new generation of the population; S7. Repeat steps S3 to S6 until the preset termination condition is met, and output the optimal mixed working fluid ratio, optimal high pressure, optimal low pressure and corresponding performance indicators.

2. The method for optimizing a mixed working fluid refrigeration system according to claim 1, characterized in that, In step S1, the constraints include at least one or more of the following: minimum allowable temperature difference constraint, compressor exhaust temperature constraint, and intake phase constraint.

3. The method for optimizing a mixed working fluid refrigeration system according to claim 2, characterized in that, In step S5, the preset near-feasible repair range is determined by at least one of the following conditions: The global minimum temperature difference in the constraint violation information is greater than the preset repair threshold, but less than the lower limit of the temperature difference specified by the minimum allowable temperature difference constraint; The minimum temperature difference within the critical temperature zone of the heat exchange is lower than the lower limit of the minimum allowable temperature difference constraint, but the extent of the violation is less than the preset temperature difference tolerance.

4. The method for optimizing a mixed working fluid refrigeration system according to claim 3, characterized in that, In step S5, the correction of the component ratio of the individual sample based on the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone specifically includes: The region in the heat transfer curve with a temperature difference value lower than the minimum allowable temperature difference is extracted as the critical temperature zone for heat transfer. Calculate the equivalent enthalpy mismatch in the critical heat exchange temperature zone, and the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone; Based on the sign and magnitude of the equivalent enthalpy mismatch, and in conjunction with the relative contribution of the partial molar enthalpy difference of each component, determine the component to be increased or decreased in proportion and its adjustment amount. The adjusted group allocation ratios are normalized to obtain the corrected group allocation ratios.

5. The method for optimizing a mixed working fluid refrigeration system according to claim 4, characterized in that, The step of extracting the region in the heat transfer curve where the temperature difference is lower than the minimum allowable temperature difference as the critical temperature zone for heat transfer specifically includes: One or more of the following in the heat transfer curve are selected as candidate key temperature zones: endpoints, local minimum temperature difference points, global minimum temperature difference points, and local temperature difference points with temperature difference values ​​below a preset threshold. The candidate critical temperature zones are sorted according to the absolute value of the equivalent enthalpy mismatch, and one or more of the top-ranked zones are selected as the critical temperature zones for heat exchange.

6. The method for optimizing a mixed working fluid refrigeration system according to claim 5, characterized in that, The step of determining the component to be increased or decreased in proportion and its adjustment amount based on the sign and magnitude of the equivalent enthalpy mismatch, combined with the relative contribution of the partial molar enthalpy difference of each component, specifically includes: When the equivalent enthalpy mismatch is less than zero, the component with the largest partial molar enthalpy difference is selected as the component to be increased in proportion, and the component with the smallest partial molar enthalpy difference is selected as the component to be decreased in proportion. When the equivalent enthalpy mismatch is greater than zero, the component with the smallest partial molar enthalpy difference is selected as the component to be increased in proportion, and the component with the largest partial molar enthalpy difference is selected as the component to be decreased in proportion. The adjustment amount of each component to be adjusted is calculated based on the absolute value of the equivalent enthalpy mismatch, the partial molar enthalpy difference of the selected components, and the preset repair coefficient.

7. The method for optimizing a mixed working fluid refrigeration system according to claim 1, characterized in that, In step S5: The corrected individuals are input into the process simulation model for evaluation. If the newly obtained performance index is better than the original performance index, the newly obtained performance index is used as the fitness for selection; otherwise, the fitness before the correction is retained for selection.

8. The method for optimizing a mixed working fluid refrigeration system according to claim 1, characterized in that, The correction of high-pressure, low-pressure, or component ratio based on thermodynamic causal correspondence specifically includes: If the compressor discharge temperature constraint is violated, the pressure ratio is reduced and the highest boiling point component ratio is increased; If the intake phase constraint is violated, the highest boiling point group distribution ratio will be reduced first. When the highest boiling point group distribution ratio has reached the preset lower limit, the low pressure will be reduced.

9. An optimization system for a mixed working fluid refrigeration system, characterized in that, A method for optimizing a mixed refrigerant refrigeration system according to any one of claims 1 to 8, wherein the mixed refrigerant refrigeration system optimization system comprises: The model building module is used to build a process simulation model of the mixed working fluid refrigeration system, determine the constraints and the range of values ​​for the distribution ratio of each component of the mixed working fluid, as well as the range and step size of the high pressure and the low pressure. An initialization module is used to generate an initial population within the range of values. The initial population consists of a combination of mechanism seed individuals, derived individuals generated based on local perturbations of the mechanism seed individuals, and globally randomly sampled individuals. Each individual in the initial population corresponds to a set of optimized variable values, which at least include the distribution ratio of each group of the mixed working fluid, as well as the high pressure and low pressure. The iterative optimization module is used to iteratively optimize the population using a population iterative optimization algorithm to generate offspring individuals; wherein, the current population at the time of the first iteration is the initial population, and the offspring individuals are generated through at least one of differential mutation, crossover, position update, and perturbation search; The evaluation module is used to input the optimization variable values ​​of the offspring individuals into the process simulation model for evaluation, and to obtain the performance indicators and constraint violation information of each offspring individual. The fitness determination module is used to classify and determine the fitness of each offspring individual based on the constraint violation information: if all preset constraints are met, the performance index of the individual is used as the fitness; if the heat exchange temperature difference related constraint is violated and is within the preset near-feasible repair range, the component ratio of the individual is corrected based on the partial molar enthalpy difference of each component of the mixed working fluid in the critical heat exchange temperature zone; if the compressor exhaust temperature constraint or intake phase constraint is violated, the high pressure, low pressure or component ratio is corrected based on the thermodynamic causal correspondence; otherwise, a penalty is imposed on its performance index according to the degree of constraint violation, and the result after penalty is used as the fitness. The selection module is used to select individuals based on their fitness to form a new generation of the population. The output module is used to repeatedly trigger the iterative optimization module, evaluation module, fitness determination module and selection module until the preset termination condition is met, and output the optimal mixed working fluid ratio, optimal high pressure, optimal low pressure and corresponding performance indicators.

10. A computer device, characterized in that, The computer device includes a processor and a memory coupled to the processor, wherein the memory stores program instructions for implementing the method for optimizing a mixed refrigerant refrigeration system according to any one of claims 1-8; the processor is used to execute the program instructions stored in the memory to optimize the mixed refrigerant refrigeration system.

11. A computer-readable storage medium, characterized in that, The system stores processor-executable program instructions for performing the method for optimizing a mixed refrigerant refrigeration system according to any one of claims 1-8.

Images