Population selection method for optimizing natural gas liquefaction process parameters and electronic equipment

Abstract

The invention discloses a population selection method for optimizing natural gas liquefaction process parameters and electronic equipment. According to the method, during population screening, through preliminary search of a large sample phase space, on the premise that calculation conditions are allowed, all samples in the phase space are covered as many as possible, then the number of the samples in the phase space is reduced through a proper grouping rule, a new population is formed, and meanwhile the difference of all variables in the population is reserved as much as possible; then, new population numbers generated by the grouping rules are optimized, and a globally optimal solution is searched; according to the technical scheme, the points in the phase space can be covered as many as possible, the uniqueness of each point in the population, especially the uniqueness of the strong correlation variable of the target function, is stored, and the purpose of greatly reducing the calculation amount is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of natural gas liquefaction, and more specifically, relates to a population selection method and electronic equipment for optimizing natural gas liquefaction process parameters. Background Art

[0002] Natural gas liquefaction is the core technology of energy companies, and achieving low energy consumption and low cost is the key goal of natural gas liquefaction technology. Liquefied natural gas production is affected by ambient temperature, gas source pressure and processing volume, and process parameters such as mixed refrigerant formula and compressor operating parameters need to be adjusted. The optimization of process parameters of such complex systems can be reduced to an optimization problem with multiple degrees of freedom and multiple constraints. Conventional operating condition analysis methods or gradient-based classical optimization methods are often helpless for systems with such multiple degrees of freedom (for example, more than 10 degrees of freedom). The development of modern artificial intelligence optimization methods provides an effective means for optimizing natural gas liquefaction process parameters.

[0003] Conventional algorithms have great limitations when optimizing complex systems with multiple variables and multiple constraints, and they tend to converge to local optimal solutions. Genetic algorithms (GAs) originate from computer simulation studies of biological systems. The algorithm uses digital strings to analogize the dyed individuals in organisms in accordance with natural evolutionary methods such as natural selection, mutation, and crossover in the biological world. It simulates the basic evolutionary process of organisms through genetic operators such as selection, crossover, and mutation. It uses fitness functions to represent the quality of the problem solutions implied by chromosomes. Through the continuous "renewal" of the population, the average fitness of the population is improved, and the evolutionary direction of the population is guided by the fitness function. On this basis, the problem solutions represented by the optimal individuals are close to the global optimal solution of the problem. Different from classical optimization algorithms based on gradient information, genetic algorithms are intelligent. After determining the coding scheme, fitness function, and genetic operators, they use the information obtained during the evolution process to organize the search by themselves. This self-organizing and adaptive feature gives it the ability to automatically discover environmental characteristics and laws according to environmental changes, while traditional algorithms need to fully describe the problem in advance during the design process, and explain the measures that the algorithm should take for different problems.

[0004] Genetic algorithm is a group operation, and an initial population consisting of several initial solutions must be prepared for genetic operation. The quality of the initial population has an important impact on the quality of the genetic algorithm calculation results and the efficiency of the algorithm. For multivariable systems, the larger the initial population, the more monomers are contained in the phase space, and the greater the possibility of finding the global optimal value, but it also causes the disadvantage of excessive calculation.

[0005] Since the optimization of the natural gas liquefaction process requires simultaneous iteration of genetic algorithms and complex process simulation programs, the amount of calculation is huge. How to reduce the amount of calculation while seeking the global optimal solution is an important issue in lowering the application difficulty threshold. Summary of the invention

[0006] In view of this, an embodiment of the present invention provides a population selection method for optimizing natural gas liquefaction process parameters, which can generate a suitable initial population for a genetic algorithm and significantly reduce the amount of calculation while seeking a global optimal solution.

[0007] According to one aspect of the present invention, a population selection method for optimizing natural gas liquefaction process parameters is provided, characterized in that the method comprises:

[0008] Step 1, determining the optimization objective function according to the process objectives of natural gas liquefaction;

[0009] Step 2, taking the process parameters to be optimized as the independent variables of the genetic algorithm, and setting the constraint relationship between the independent variables;

[0010] Step 3: Set the upper and lower bounds of each independent variable, and generate multiple monomers Z based on the set upper and lower bounds. 0,i , forming the original population G0, i = 1, 2, ... I, I represents the number of monomers in the original population G0;

[0011] Step 4: Perform the first round of genetic algorithm iterations to optimize the objective function based on the original population G0. The number of iterations does not exceed the preset number of iterations, and obtain multiple monomers Z after iteration. i , i = 1, 2, ... I, constituting the population G;

[0012] Step 5: Set multiple grouping rules. For each grouping rule, group multiple monomers Z in population G according to the grouping rule. i Divide into M groups, and extract one monomer from each group to enter the initial population G corresponding to the grouping rule 0,n , n = 1…N, N represents the number of grouping rules;

[0013] Step 6: Based on the initial population G under each grouping rule 0,n Perform genetic algorithm iterations of the optimization objective function respectively to obtain the optimal value of the optimization objective function under each grouping rule, and the monomer corresponding to the optimal value;

[0014] Step 7: Select a global optimal value from the optimal values ​​of the optimization objective function under each grouping rule, and determine the value of the independent variable of the monomer corresponding to the global optimal value as the value of the process parameter to be optimized.

[0015] Optionally, the process target is the energy consumption of the natural gas liquefaction unit, the operating cost of the natural gas liquefaction unit, or the capital expenditure of the natural gas liquefaction unit.

[0016] Optionally, the process parameters to be optimized include some or all of the following: flow rates of different components in the mixed refrigerant, inlet and outlet pressures of each section of the mixed refrigerant compressor, and outlet temperatures of each section of the cold box.

[0017] Optionally, in step 3, multiple monomers Z are generated based on the set upper and lower boundaries. 0,i include:

[0018] For each independent variable, randomly generate individual monomers Z between the upper and lower boundaries of the independent variable o,i The value of the independent variable in .

[0019] Optionally, in step 4, when any one of the following conditions is met, the first round of genetic algorithm iteration ends:

[0020] The number of iterations reaches the preset number of iterations;

[0021] Although the number of iterations has not reached the preset number of iterations, the population after this iteration meets the preset termination condition.

[0022] Optionally, the preset termination condition is:

[0023] The number of monomers satisfying the constraint condition in the population after this iteration exceeds a preset threshold.

[0024] Optionally, in step 5, the grouping rules include part or all of the following:

[0025] Calculate the Z of each monomer in the population G i The distance between the lower boundary and the Z of each monomer is calculated according to the distance. i Sorting, dividing the sorted monomers into M groups in order of I / M monomers per group;

[0026] Calculate the Z of each monomer in the population G i The distance between the upper boundary and the Z of each monomer is calculated according to the distance. i Sorting, dividing the sorted monomers into M groups in order of I / M monomers per group;

[0027] Calculate the Z of each monomer in the population G i The variance with the lower and upper boundaries is calculated according to the variance of each monomer Z i Sorting, and dividing the sorted monomers into M groups in order of I / M monomers per group.

[0028] Optionally, in step 5, one monomer is drawn from each group into the initial population G in the following manner: 0,n:

[0029] When the proportion of monomers in the group that meet the constraints is less than the preset proportion, the probability of extracting monomers is adjusted so that the ratio of the probability of extracting monomers that meet the constraints to the probability of extracting monomers that do not meet the constraints is greater than the preset probability ratio, and then randomly extract monomers from the group to enter the initial population G 0,n ;

[0030] When the proportion of monomers in the group that meet the constraints is greater than or equal to the preset proportion, monomers are directly randomly drawn from the group to enter the initial population G 0,n .

[0031] Optionally, the preset proportion is e / (1+e), where e is a natural constant; and the preset probability ratio is e.

[0032] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising:

[0033] A memory storing executable instructions;

[0034] A processor runs the executable instructions in the memory to implement the population selection method for optimizing natural gas liquefaction process parameters as described above.

[0035] According to the above-mentioned embodiment of the present invention, a large-scale original population G0 can be generated to cover as many points in the phase space as possible; then, a finite number of genetic algorithm iterations are performed on the original population. Since the number of iterations is limited, a small amount of calculation can be guaranteed even though the population is large; then, the population G updated after a finite number of iterations is grouped according to different grouping rules, and monomers are extracted from each group to form an initial population G0,n under different grouping rules. At this time, the scale of each initial population is much smaller than the original population G0, and the uniqueness of each point in the population is fully retained; then, a genetic algorithm iteration is performed based on each initial population. Since the scale of the initial population is significantly smaller than the original population, the amount of calculation can be significantly reduced; then, the optimal values ​​obtained based on different initial populations are comprehensively considered to find the global optimal value, and the value of the independent variable of the monomer corresponding to the global optimal value is determined as the optimized natural gas liquefaction process parameter.

[0036] According to the present invention, while fully covering the points in the phase space, the uniqueness of different populations is retained as much as possible, and the amount of calculation is significantly reduced, thereby improving the calculation efficiency.

[0037] Other features and advantages of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0039] Figure 1 A flow chart of a population selection method for optimizing natural gas liquefaction process parameters according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0040] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0041] Figure 1 The diagram shows a population selection method for optimizing natural gas liquefaction process parameters according to an embodiment of the present invention. The method includes steps 1-7.

[0042] Step 1: Determine the optimization objective function according to the process objectives of natural gas liquefaction.

[0043] In some examples, the energy consumption of the natural gas liquefaction plant can be used as a process target; in other examples, the operating cost (also known as OPEX) of the natural gas liquefaction plant can be used as a process target; in other examples, the capital expenditure (also known as CAPEX) of the natural gas liquefaction plant can also be used as a process target.

[0044] Step 2: Use the process parameters to be optimized as the independent variables of the genetic algorithm, and set the constraint relationship between the independent variables.

[0045] There are many process parameters related to the process objectives, such as the outlet temperature of the water cooler or air cooler, the flow rate of different components in the mixed refrigerant (such as the percentage molar flow rate and total flow rate of single components such as nitrogen, methane, C2, C3, C4, and C5), the inlet and outlet pressures of each section of the mixed refrigerant compressor, the outlet temperature of each section of the cold box, and other factors affecting liquefaction.

[0046] According to experience, process parameters that are strongly correlated with process objectives can be selected from numerous process parameters as process parameters to be optimized in the present invention. In a preferred embodiment according to the present invention, some or all of the following process parameters can be selected as independent variables of monomers: flow rates of different components in the mixed refrigerant (such as nitrogen, methane, carbon 2, carbon 3, carbon 4, carbon 5, etc.), inlet and outlet pressures of each section of the mixed refrigerant compressor, and outlet temperatures of each section of the cold box. After in-depth research, the inventors found that these process parameters are strongly correlated with the process objectives. Using them as independent variables of population monomers can reduce the amount of calculation and improve calculation efficiency while ensuring the optimization effect.

[0047] The constraints between independent variables can be set based on experience and device requirements / indicators.

[0048] Step 3: Set the upper and lower bounds of each independent variable, and generate multiple monomers Z based on the set upper and lower bounds. 0,i , constituting the original population G0, i = 1, 2,…I, I represents the number of monomers in the original population G0.

[0049] Assume that the number of independent variables contained in each monomer is L, then any independent variable is represented by J l , l=1,2,…,L, can be used for any independent variable J l Set its upper boundary J l,lb and the lower boundary J l,hb , and then generate each monomer Z based on the upper and lower boundaries 0,i The independent variable J in l , so that its value is at the corresponding upper boundary J l,lb and the lower boundary J l,hb between.

[0050] In one embodiment, for monomer Z 0,i Each independent variable J l , can be found at the upper boundary J of the independent variable l,lb and the lower boundary J l,hb Each monomer Z is randomly generated o,i The random generation method can ensure that the population randomly covers points in the phase space, reducing deviations and limitations introduced by human intervention.

[0051] Let the upper bounds of all independent variables form a vector J lb ={J 1,lb ,J 2,lb ,…J L,lb}, the lower boundary forms the vector J hb ={J 1,hb ,J 2,hb ,…J L,hb}, can be found in J lb ≤Z 0,i ≤J hb Under the condition of 0,i .

[0052] Step 4: Perform the first round of genetic algorithm iterations to optimize the objective function based on the original population G0. The number of iterations does not exceed the preset number of iterations, and obtain multiple monomers Z after iteration. i , i=1,2,…I, constituting the population G.

[0053] According to the technical solution of the present invention, in step 4, a limited number of first round genetic iterative algorithms can be performed on the original population G0, which not only ensures that the population covers as much phase space as possible, but also limits the computational complexity of the genetic algorithm iteration.

[0054] In some implementations, the first round of genetic algorithm iterations may end when any one of the following conditions is met:

[0055] The number of iterations reaches the preset number of iterations;

[0056] Although the number of iterations has not reached the preset number of iterations, the population after this iteration meets the preset termination condition.

[0057] For example, if the preset number of iterations is 5, then when the number of iterations reaches 5, or although the number of iterations has not reached 5, but the updated population after this iteration - which can be the 1st, 2nd, 3rd, or 4th iteration - has met specific conditions, the first round of genetic algorithm iterations is terminated.

[0058] In some embodiments, the preset condition is: the number of monomers satisfying the constraint condition in the population updated after this iteration exceeds a preset threshold. Those skilled in the art can design the preset threshold as needed. In one example, the preset threshold can be set to 3 times the number of independent variables L.

[0059] Step 5: Set multiple grouping rules. For each grouping rule, group multiple monomers Z in population G according to the grouping rule. i Divide into M groups, and extract one monomer from each group to enter the initial population G corresponding to the grouping rule 0,n , n=1…N, N represents the number of grouping rules.

[0060] In some implementations, the grouping rules set include some or all of the following:

[0061] Grouping rule 1: Calculate the Z of each monomer in population G i The distance between the lower boundary and the Z of each monomer is calculated according to the distance. i Sorting, dividing the sorted monomers into M groups in order of I / M monomers per group;

[0062] Grouping rule 2: Calculate the Z of each monomer in population G i The distance between the upper boundary and the Z of each monomer is calculated according to the distance. i Sorting, dividing the sorted monomers into M groups in order of I / M monomers per group;

[0063] Grouping rule 3: Calculate the Z of each monomer in population G i The variance with the lower and upper boundaries is calculated according to the variance of each monomer Z iSorting, and dividing the sorted monomers into M groups in order of I / M monomers per group.

[0064] The above grouping rule 1 is used as an example. Each monomer Z in population G can be calculated based on the following formula i Distance S from the lower boundary l,i , i=1,2,…I:

[0065] S l,i =∑ l＝1,…,L |J l -J l,lb |,

[0066] Among them, |J l -J l,lb | indicates monomer Z i The independent variable J l The value of the independent variable is related to the lower boundary J l,lb The absolute value of the interval.

[0067] For all monomers Z i According to the corresponding distance S l,i Sort from large to small or from small to large, and sort the sorted monomers Z i Sequentially, the monomers are divided into M groups according to the number of I / M. For example, if I is 1000 and M is 50, then for the sorted monomers, the first 20 are grouped into one group, the next 20 are grouped into one group, and so on.

[0068] Take grouping rule 2 as an example. Each monomer Z can be calculated based on the following formula i Distance S from the upper boundary h,i , i=1,2,…I:

[0069] S h,i =∑ l＝1,…,L |J l -J l,hb |,

[0070] Among them, |J l -J l,hb | indicates monomer Z i The independent variable J l The value of the independent variable is related to the upper boundary J l,hb The absolute value of the interval.

[0071] For all monomers Z i According to the corresponding distance S h,i Sort from large to small or from small to large, and sort the sorted monomers Z i Sequentially, the monomers are divided into M groups according to the number of I / M. For example, if I is 1000 and M is 50, then for the sorted monomers, the first 20 are grouped into one group, the next 20 are grouped into one group, and so on.

[0072] Take grouping rule 3 as an example. Each monomer Z can be calculated based on the following formula i The variance S with the lower and upper boundaries D,i , i=1,2,…I:

[0073]

[0074] For all monomers Z i According to the corresponding distance S D,i Sort from large to small or from small to large, and sort the sorted monomers Z i Sequentially, the monomers are divided into M groups according to the number of I / M. For example, if I is 1000 and M is 50, then for the sorted monomers, the first 20 are grouped into one group, the next 20 are grouped into one group, and so on.

[0075] In some implementations, the grouping rules set may include two of the above three grouping rules, that is, N=2, for example 1 and 2, 1 and 3, 2 and 3, which is equivalent to grouping the population G twice and obtaining two grouping results; in some implementations, all three grouping rules 1, 2, and 3 may be included, that is, N=3, which is equivalent to grouping the population G three times and obtaining three grouping results.

[0076] Take N=3, i.e., the embodiment including all three grouping rules 1, 2, and 3 as an example. Using the above assumptions, let I be 1000, M be 50, and the 1000 monomers Z in the population G after the first round of genetic algorithm iteration can be grouped according to grouping rule 1. i Divide into 50 groups, each group includes 20 monomers, and extract one monomer from each group into the initial population G 0,1 , G 0,1 The number of monomers in the population G is 50; according to grouping rule 2, the 1000 monomers Z in the population G after the first round of genetic algorithm iteration can be i Divide into 50 groups again, each group includes 20 monomers, and draw one monomer from each group into the initial population G 0,2 , G 0,2 The number of monomers in the population G is 50; according to grouping rule 3, the 1000 monomers Z in the population G after the first round of genetic algorithm iteration can be i Divide into 50 groups, each group includes 20 monomers, and extract one monomer from each group into the initial population G 0,3 , G 0,3 The number of monomers in is 50.

[0077] According to this step, one monomer is extracted from each group to form the initial population, which can not only retain the uniqueness of the population as much as possible (monomers in each range are extracted into the initial population), but also make the population cover the entire phase space as much as possible, and at the same time significantly reduce the population size. For example, in the above example, the number of monomers in each initial population is 50, which is significantly less than the number of monomers in the original population of 1000, thereby significantly reducing the amount of computational complexity of subsequent genetic algorithm iterations.

[0078] In some embodiments, the initial population G can be further improved by setting appropriate extraction rules. 0,n For example, one monomer can be drawn from each group into the initial population G in the following way 0,n :

[0079] When the proportion of monomers in the group that meet the constraints is less than the preset proportion, the probability of extracting monomers is adjusted so that the ratio of the probability of extracting monomers that meet the constraints to the probability of extracting monomers that do not meet the constraints is greater than the preset probability ratio, and then randomly extract monomers from the group to enter the initial population G 0,n ;

[0080] When the proportion of monomers in the group that meet the constraints is greater than or equal to the preset proportion, monomers are directly randomly drawn from the group to enter the initial population G 0,n .

[0081] In one embodiment, the preset proportion is e / (1+e), wherein e is a natural constant, and 1 / (1+e) can be approximately equal to 26.89%; the preset probability ratio is e.

[0082] The above extraction method will be described below with examples.

[0083] For a certain group, you can first confirm whether each monomer in the group meets the constraints. If the quotient of the number of monomers in the group that meet the constraints divided by the total number of monomers in the group is less than 26.89%, the extraction probability of the monomers in the group is adjusted, that is, the probability of the monomers that meet the constraints being drawn is increased, rather than all monomers being drawn with equal chances. By adjusting the extraction probability, the following formula is ensured:

[0084]

[0085] Then, based on the adjusted extraction probability, a single cell is randomly drawn from the group and enters the initial population G. 0,n .

[0086] If the quotient of the number of monomers in the group that meet the constraints divided by the total number of monomers in the group is greater than or equal to 26.89%, then a monomer is randomly drawn from the group and enters the initial population G. 0,n .

[0087] Step 6: Based on the initial population G under each grouping rule 0,n The genetic algorithm iteration of the optimization objective function is performed respectively to obtain the optimal value of the optimization objective function under each grouping rule and the monomer corresponding to the optimal value.

[0088] The grouping rules include all three grouping rules mentioned above. 0,1 , G 0,2 , G 0,3 Perform genetic algorithm iteration to optimize the objective function. 0,1 , G 0,2 , G 0,3 The scale of is much smaller than that of the original population G0, so compared with the genetic algorithm iteration based on the original population G0, the amount of calculation according to the present invention is significantly reduced. 0,1 Perform genetic algorithm iteration to obtain the optimal value D1 of the optimization objective function and the monomer Z1 corresponding to the optimal value; based on the population G 0,2 Perform genetic algorithm iteration to obtain the optimal value D2 of the optimization objective function and the monomer Z2 corresponding to the optimal value; based on the population G 0,3 The genetic algorithm is iterated to obtain the optimal value D3 of the optimization objective function and the monomer Z3 corresponding to the optimal value.

[0089] Step 7: Select a global optimal value from the optimal values ​​of the optimization objective function under each grouping rule, and determine the value of the independent variable of the monomer corresponding to the global optimal value as the value of the process parameter to be optimized.

[0090] Continuing to use the above example as an example for explanation, assuming that the optimal values ​​D1, D2, and D3 of the optimization objective function are obtained based on the above three grouping rules, then in step 7, the value that best meets the expectations is selected from D1, D2, and D3 as the global optimal value. Those skilled in the art can select the global optimal value from D1, D2, and D3 based on common sense and experience. For example, if the process target of the optimization objective function is the energy consumption of the natural gas liquefaction device, the minimum value among D1, D2, and D3 can be selected as the global optimal value. Assuming that the value of D1 is the smallest among D1, D2, and D3, the value of D1 is determined to be the global optimal value, and the value of the independent variable in the corresponding monomer Z1 is determined to be the optimized natural gas liquefaction process parameter.

[0091] According to the above embodiments of the present invention, there are at least the following beneficial effects:

[0092] (1) Cover as many points in phase space as possible;

[0093] (2) The uniqueness of each point in the population is preserved as much as possible;

[0094] (3) retaining as much uniqueness of strongly correlated variables in the population as possible;

[0095] (4) The amount of calculation is significantly reduced.

[0096] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

[0097] An electronic device according to an embodiment of the present invention includes a memory and a processor.

[0098] The memory is used to store non-temporary computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0099] The processor may be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. In one embodiment of the present invention, the processor is used to run the computer-readable instructions stored in the memory.

[0100] Those skilled in the art should be able to understand that in order to solve the technical problem of how to obtain a good user experience, the present embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the protection scope of the present invention.

[0101] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0102] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A population selection method for optimizing natural gas liquefaction process parameters, characterized in that: The method comprises: Step 1, determining the optimization objective function according to the process objectives of natural gas liquefaction; Step 2, taking the process parameters to be optimized as the independent variables of the genetic algorithm, and setting the constraint relationship between the independent variables; Step 3: Set the upper and lower bounds of each independent variable, and generate multiple monomers Z based on the set upper and lower bounds. 0,i , forming the original population G0, i = 1, 2, ... I, I represents the number of monomers in the original population G0; Step 4: Perform the first round of genetic algorithm iterations to optimize the objective function based on the original population G0. The number of iterations does not exceed the preset number of iterations, and obtain multiple monomers Z after iteration. i , i = 1, 2, ... I, constituting the population G; Step 5: Set multiple grouping rules. For each grouping rule, group multiple monomers Z in population G according to the grouping rule. i Divide into M groups, and extract one monomer from each group to enter the initial population G corresponding to the grouping rule 0,n , n = 1…N, N represents the number of grouping rules; Step 6: Based on the initial population G under each grouping rule 0,n Perform genetic algorithm iterations of the optimization objective function respectively to obtain the optimal value of the optimization objective function under each grouping rule, and the monomer corresponding to the optimal value; Step 7: Select a global optimal value from the optimal values ​​of the optimization objective function under each grouping rule, and determine the value of the independent variable of the monomer corresponding to the global optimal value as the value of the process parameter to be optimized.

2. The method according to claim 1, characterized in that: The process target is the energy consumption of the natural gas liquefaction unit, the operating cost of the natural gas liquefaction unit or the capital expenditure of the natural gas liquefaction unit.

3. The method according to claim 1, characterized in that The process parameters to be optimized include part or all of the following: flow rates of different components in the mixed refrigerant, inlet and outlet pressures of each section of the mixed refrigerant compressor, and outlet temperatures of each section of the cold box.

4. The method according to claim 1, characterized in that In step 3, multiple monomers Z are generated based on the upper and lower boundaries set. 0,i include: For each independent variable, randomly generate individual monomers Z between the upper and lower boundaries of the independent variable 0,i The value of the independent variable in .

5. The method according to claim 1, characterized in that In step 4, when any of the following conditions is met, the first round of genetic algorithm iteration ends: The number of iterations reaches the preset number of iterations; Although the number of iterations has not reached the preset number of iterations, the population after this iteration meets the preset termination condition.

6. The method according to claim 5, characterized in that The preset termination condition is: The number of monomers satisfying the constraint condition in the population after this iteration exceeds a preset threshold.

7. The method according to claim 1, characterized in that In step 5, the grouping rules include some or all of the following: Calculate the Z of each monomer in the population G i The distance between the lower boundary and the Z of each monomer is calculated according to the distance. i Sorting, dividing the sorted monomers into M groups in order of I / M monomers per group; Calculate the Z of each monomer in the population G i The distance between the upper boundary and the Z of each monomer is calculated according to the distance. i Sorting, dividing the sorted monomers into M groups in order of I / M monomers per group; Calculate the Z of each monomer in the population G i The variance with the lower and upper boundaries is calculated according to the variance of each monomer Z i Sorting, and dividing the sorted monomers into M groups in order of I / M monomers per group.

8. The method according to claim 1, characterized in that In step 5, one monomer is drawn from each group into the initial population G in the following manner: 0,n : When the proportion of monomers in the group that meet the constraints is less than the preset proportion, the probability of extracting monomers is adjusted so that the ratio of the probability of extracting monomers that meet the constraints to the probability of extracting monomers that do not meet the constraints is greater than the preset probability ratio, and then randomly extract monomers from the group to enter the initial population G 0,n ; When the proportion of monomers in the group that meet the constraints is greater than or equal to the preset proportion, monomers are directly randomly drawn from the group to enter the initial population G 0,n .

9. The method according to claim 8, characterized in that: The preset ratio is e / (1+e), where e is a natural constant; The preset probability ratio is e.

10. An electronic device, characterized in that: The electronic device comprises: A memory storing executable instructions; A processor, wherein the processor runs the executable instructions in the memory to implement the population selection method for optimizing natural gas liquefaction process parameters according to any one of claims 1 to 9.

Images