WO2023143106A1 - 一种催化裂化过程的多目标优化方法及装置 - Google Patents

一种催化裂化过程的多目标优化方法及装置 Download PDF

Info

Publication number
WO2023143106A1
WO2023143106A1 PCT/CN2023/071999 CN2023071999W WO2023143106A1 WO 2023143106 A1 WO2023143106 A1 WO 2023143106A1 CN 2023071999 W CN2023071999 W CN 2023071999W WO 2023143106 A1 WO2023143106 A1 WO 2023143106A1
Authority
WO
WIPO (PCT)
Prior art keywords
individual
objective
optimization
objective optimization
population
Prior art date
Application number
PCT/CN2023/071999
Other languages
English (en)
French (fr)
Inventor
钟伟民
隆建
杜文莉
钱锋
杨明磊
Original Assignee
华东理工大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华东理工大学 filed Critical 华东理工大学
Publication of WO2023143106A1 publication Critical patent/WO2023143106A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • C—CHEMISTRY; METALLURGY
    • C10—PETROLEUM, GAS OR COKE INDUSTRIES; TECHNICAL GASES CONTAINING CARBON MONOXIDE; FUELS; LUBRICANTS; PEAT
    • C10G—CRACKING HYDROCARBON OILS; PRODUCTION OF LIQUID HYDROCARBON MIXTURES, e.g. BY DESTRUCTIVE HYDROGENATION, OLIGOMERISATION, POLYMERISATION; RECOVERY OF HYDROCARBON OILS FROM OIL-SHALE, OIL-SAND, OR GASES; REFINING MIXTURES MAINLY CONSISTING OF HYDROCARBONS; REFORMING OF NAPHTHA; MINERAL WAXES
    • C10G11/00—Catalytic cracking, in the absence of hydrogen, of hydrocarbon oils
    • C—CHEMISTRY; METALLURGY
    • C10—PETROLEUM, GAS OR COKE INDUSTRIES; TECHNICAL GASES CONTAINING CARBON MONOXIDE; FUELS; LUBRICANTS; PEAT
    • C10G—CRACKING HYDROCARBON OILS; PRODUCTION OF LIQUID HYDROCARBON MIXTURES, e.g. BY DESTRUCTIVE HYDROGENATION, OLIGOMERISATION, POLYMERISATION; RECOVERY OF HYDROCARBON OILS FROM OIL-SHALE, OIL-SAND, OR GASES; REFINING MIXTURES MAINLY CONSISTING OF HYDROCARBONS; REFORMING OF NAPHTHA; MINERAL WAXES
    • C10G2300/00—Aspects relating to hydrocarbon processing covered by groups C10G1/00 - C10G99/00
    • C10G2300/10—Feedstock materials
    • C10G2300/107—Atmospheric residues having a boiling point of at least about 538 °C
    • C—CHEMISTRY; METALLURGY
    • C10—PETROLEUM, GAS OR COKE INDUSTRIES; TECHNICAL GASES CONTAINING CARBON MONOXIDE; FUELS; LUBRICANTS; PEAT
    • C10G—CRACKING HYDROCARBON OILS; PRODUCTION OF LIQUID HYDROCARBON MIXTURES, e.g. BY DESTRUCTIVE HYDROGENATION, OLIGOMERISATION, POLYMERISATION; RECOVERY OF HYDROCARBON OILS FROM OIL-SHALE, OIL-SAND, OR GASES; REFINING MIXTURES MAINLY CONSISTING OF HYDROCARBONS; REFORMING OF NAPHTHA; MINERAL WAXES
    • C10G2300/00—Aspects relating to hydrocarbon processing covered by groups C10G1/00 - C10G99/00
    • C10G2300/10—Feedstock materials
    • C10G2300/1077—Vacuum residues
    • C—CHEMISTRY; METALLURGY
    • C10—PETROLEUM, GAS OR COKE INDUSTRIES; TECHNICAL GASES CONTAINING CARBON MONOXIDE; FUELS; LUBRICANTS; PEAT
    • C10G—CRACKING HYDROCARBON OILS; PRODUCTION OF LIQUID HYDROCARBON MIXTURES, e.g. BY DESTRUCTIVE HYDROGENATION, OLIGOMERISATION, POLYMERISATION; RECOVERY OF HYDROCARBON OILS FROM OIL-SHALE, OIL-SAND, OR GASES; REFINING MIXTURES MAINLY CONSISTING OF HYDROCARBONS; REFORMING OF NAPHTHA; MINERAL WAXES
    • C10G2400/00—Products obtained by processes covered by groups C10G9/00 - C10G69/14
    • C10G2400/02—Gasoline
    • C—CHEMISTRY; METALLURGY
    • C10—PETROLEUM, GAS OR COKE INDUSTRIES; TECHNICAL GASES CONTAINING CARBON MONOXIDE; FUELS; LUBRICANTS; PEAT
    • C10G—CRACKING HYDROCARBON OILS; PRODUCTION OF LIQUID HYDROCARBON MIXTURES, e.g. BY DESTRUCTIVE HYDROGENATION, OLIGOMERISATION, POLYMERISATION; RECOVERY OF HYDROCARBON OILS FROM OIL-SHALE, OIL-SAND, OR GASES; REFINING MIXTURES MAINLY CONSISTING OF HYDROCARBONS; REFORMING OF NAPHTHA; MINERAL WAXES
    • C10G2400/00—Products obtained by processes covered by groups C10G9/00 - C10G69/14
    • C10G2400/04—Diesel oil
    • C—CHEMISTRY; METALLURGY
    • C10—PETROLEUM, GAS OR COKE INDUSTRIES; TECHNICAL GASES CONTAINING CARBON MONOXIDE; FUELS; LUBRICANTS; PEAT
    • C10G—CRACKING HYDROCARBON OILS; PRODUCTION OF LIQUID HYDROCARBON MIXTURES, e.g. BY DESTRUCTIVE HYDROGENATION, OLIGOMERISATION, POLYMERISATION; RECOVERY OF HYDROCARBON OILS FROM OIL-SHALE, OIL-SAND, OR GASES; REFINING MIXTURES MAINLY CONSISTING OF HYDROCARBONS; REFORMING OF NAPHTHA; MINERAL WAXES
    • C10G2400/00—Products obtained by processes covered by groups C10G9/00 - C10G69/14
    • C10G2400/20—C2-C4 olefins
    • C—CHEMISTRY; METALLURGY
    • C10—PETROLEUM, GAS OR COKE INDUSTRIES; TECHNICAL GASES CONTAINING CARBON MONOXIDE; FUELS; LUBRICANTS; PEAT
    • C10G—CRACKING HYDROCARBON OILS; PRODUCTION OF LIQUID HYDROCARBON MIXTURES, e.g. BY DESTRUCTIVE HYDROGENATION, OLIGOMERISATION, POLYMERISATION; RECOVERY OF HYDROCARBON OILS FROM OIL-SHALE, OIL-SAND, OR GASES; REFINING MIXTURES MAINLY CONSISTING OF HYDROCARBONS; REFORMING OF NAPHTHA; MINERAL WAXES
    • C10G2400/00—Products obtained by processes covered by groups C10G9/00 - C10G69/14
    • C10G2400/26—Fuel gas
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]

Definitions

  • the invention belongs to the field of petrochemical technology, and in particular relates to a multi-objective optimization method for a catalytic cracking process, a multi-objective optimization device for a catalytic cracking process, and a computer-readable storage medium.
  • the oil refining industry is a pillar industry in my country, among which the catalytic cracking process is one of the important means of lightening heavy oil in the oil refining industry.
  • the catalytic cracking products such as gasoline and diesel obtained through the catalytic cracking (Fluid Catalytic Cracking, FCC) process are important fuel sources for transportation, and other products produced through the catalytic cracking process are also the main raw materials for the chemical industry.
  • FCC Fluid Catalytic Cracking
  • the present invention provides a method for multi-objective optimization of catalytic cracking process, a device for multi-objective optimization of catalytic cracking process, and a computer-readable storage medium capable of The regeneration operator improves the SPEA2 algorithm of the evolution process of the offspring, and the preset approx. Adjust the value of each process decision variable in the catalytic cracking process within the bundle range to determine the guiding values of multiple process decision variables corresponding to multiple optimization objectives, so as to meet the multi-objective optimization requirements of the catalytic cracking process and improve the catalytic cracking process. Multi-objective optimization efficiency, and improve the quality and stability of optimization results.
  • the multi-objective optimization method for the above-mentioned catalytic cracking process includes the following steps: determining multiple optimization objectives, multiple process decision variables corresponding to the multiple optimization objectives, and each The constraint range of the process decision variable; according to the plurality of optimization objectives and the plurality of process decision variables, determine the objective function; improve the SPEA2 algorithm of the offspring evolution process through the regeneration operator based on the path, within the constraint range Adjusting the value of each of the process decision variables to determine the operating data of the objective function on each of the process decision variables; according to the operation data of each of the process decision variables, determine the optimization target value of each of the optimization objectives; And according to the optimal optimization target value solution set, determine the corresponding operation data as guide values of multiple process decision variables corresponding to the multiple optimization targets.
  • the SPEA2 algorithm that improves the offspring evolution process through a path-based regeneration operator adjusts the value of each of the process decision variables within the constraint range to determine the
  • the step of calculating the fitness F(i) of each individual i in the population P t includes: determining the individual i according to the number S(i) of individuals dominated by individual i The original fitness value R(i), Wherein, the original fitness value R(i) represents individual i, and the sum of all individual numbers dominated by each individual (iv) that dominates individual i; calculate each individual i to the population P t and the reserve set The distances of all individuals in , and sorted in ascending order; select the kth individual as And calculate the corresponding distance value D(i), where, The N is the size of the population P t , the for the reserve set and determine the fitness F(i) of the individual i according to the original fitness value R(i) and the distance value D(i).
  • the pair of the reserve set The steps for environment selection include: the reserve set The number of individuals in and the size of the reserve set compare; if Then according to the sorting of fitness F(i), from the population P t and the reserve set Before selecting among all individuals i Individuals constitute the dominant solution set and are added to the reserve set and if The pruning operation is performed, and the individual i with the smallest distance to the adjacent individual is eliminated in each iteration.
  • the step of eliminating the individual i with the smallest distance from the adjacent individual in each iteration includes: eliminating the only individual i with the smallest distance from the adjacent individual in each iteration i.
  • the step of calculating the data set in the mating pool via the path-based regeneration operator comprises: determining the reserve set The center point Center g of the offspring population individuals, and determine the population P t and the reserve set The center point Center g-1 of the parent population individuals; according to the difference between the center points of the two generations of population individuals before and after, define the direction of the evolutionary path ep; according to the target survival rate of offspring individuals and the actual productivity p succ , determine the forward step length ⁇ of the evolutionary path ep; and take each parent individual as a starting point to generate corresponding offspring individuals within the rectangular range pointed to by the vector ⁇ *ep.
  • the step of recombining and mutating each individual i in the mating pool via the path-based regeneration operator further includes: After the generation of individuals, the gene sharing operation is performed between at least one excellent parent individual and each of the offspring individuals.
  • the The multi-objective optimization method before performing the gene sharing operation, also includes the following steps: obtaining the Pareto front layers or the fitness F(i) of each parent individual; and according to the Pareto front layers or the fitness F(i) , screening the at least one excellent parent individual.
  • the multi-objective optimization method further includes the following steps: evaluating each of the optimization target values according to the Pareto optimal solution set and the IGD index, so as to determine the optimal optimization target value solution set.
  • the plurality of optimization objectives at least include minimizing carbon dioxide emissions.
  • the multiple optimization objectives also include: maximizing economic benefits, minimizing sulfur dioxide emissions, minimizing total waste gas emissions, minimizing total waste liquid emissions, maximizing main At least one of product yield, maximizing primary product yield, minimizing input cost.
  • the plurality of process decision variables at least include: material flow rate, material state, material property, main fractionation tower state, main fractionation tower operation, absorption tower state, absorption tower operation, further At least one of absorber state, reabsorber operation, desorber state, desorber operation, stabilizer state, stabilizer operation.
  • the plurality of optimization objectives include maximizing economic benefits, minimizing carbon dioxide emissions, and minimizing sulfur dioxide emissions.
  • the objective function is as follows:
  • M indicates the economic benefit of the catalytic cracking process
  • P i indicates the price of product i
  • Y i indicates the flow rate of product i
  • i indicates the product, including acid gas, dry gas, ethylene, liquefied gas, propylene, gasoline, diesel, cycle At least one of oil, oil slurry, and coke
  • p j indicates the price of raw material j
  • y j indicates the flow rate of raw material j
  • j indicates raw materials, including tank farm residual oil, tank farm wax oil, and refined wax oil
  • P f Indicates the fixed cost per unit of processing, F indicates the processing load;
  • p c indicates the price of fresh catalyst, f c indicates the replenishment amount of fresh catalyst;
  • C CO2 indicates the emission of carbon dioxide;
  • C SO2 indicates the emission of sulfur dioxide.
  • the above-mentioned multi-objective optimization device provided by the second aspect of the present invention includes a memory and a processor.
  • the processor is connected to the memory and is configured to implement the above-mentioned multi-objective optimization method for catalytic cracking process provided by the first aspect of the present invention.
  • the above-mentioned computer-readable storage medium provided by the third aspect of the present invention has computer instructions stored thereon.
  • the above-mentioned multi-objective optimization method for catalytic cracking process provided by the first aspect of the present invention is implemented.
  • Fig. 1 shows a schematic flowchart of a multi-objective optimization method for a catalytic cracking process according to some embodiments of the present invention.
  • Fig. 2 shows a schematic diagram of a catalytic cracking unit provided according to some embodiments of the present invention.
  • Fig. 3 shows a schematic flowchart of adjusting process decision variables provided according to some embodiments of the present invention.
  • Fig. 4 shows a schematic flowchart of calculating data sets in a mating pool according to some embodiments of the present invention.
  • Fig. 5A shows a schematic diagram of an evolutionary path provided according to some embodiments of the present invention at an early stage of evolution.
  • Fig. 5B shows a schematic diagram of an evolutionary path provided according to some embodiments of the present invention at a later stage of evolution.
  • Fig. 6 shows a schematic diagram of generating offspring individuals according to some embodiments of the present invention.
  • Fig. 7 shows a schematic diagram of IGD indicators provided according to some embodiments of the present invention.
  • Fig. 8 shows a schematic diagram of a Pareto optimal solution set provided according to some embodiments of the present invention.
  • the existing catalytic cracking optimization technology in the field of petrochemical industry mainly focuses on the single-objective optimization operation, which cannot meet the actual needs of multi-objective optimization.
  • the phenomenon of ignoring the useful information generated in the process of individual evolution is common, so a large number of iterations are required, and there are problems such as long optimization time, poor quality of optimization results, and optimization problems. Defects that result in instability.
  • the present invention provides a method for multi-objective optimization of catalytic cracking process, a device for multi-objective optimization of catalytic cracking process, and a computer-readable storage medium capable of The regeneration operator improves the SPEA2 algorithm of the offspring evolution process, adjusts the value of each process decision variable in the catalytic cracking process within the preset constraint range, so as to determine the guidance value of multiple process decision variables corresponding to multiple optimization objectives, In order to meet the multi-objective optimization requirements of the catalytic cracking process, improve the multi-objective optimization efficiency of the catalytic cracking process, and improve the quality and stability of the optimization results.
  • the above-mentioned multi-objective optimization method for the catalytic cracking process provided by the first aspect of the present invention can be implemented by the above-mentioned multi-objective optimization device for the catalytic cracking process provided by the second aspect of the present invention.
  • the multi-objective optimization device is configured with a memory and a processor.
  • the storage includes, but is not limited to, the computer-readable storage medium provided by the third aspect of the present invention, on which computer instructions are stored.
  • the processor is connected to the memory and is configured to implement the above-mentioned multi-objective optimization method for catalytic cracking process provided by the first aspect of the present invention.
  • multi-objective optimization device The working principle of the above-mentioned multi-objective optimization device will be described below in conjunction with some embodiments of multi-objective optimization methods.
  • these examples of multi-objective optimization methods are only some non-limiting implementations provided by the present invention, and are intended to clearly demonstrate the main idea of the present invention and provide some specific solutions that are convenient for the public to implement. It is not intended to limit all functions or all working methods of the multi-objective optimization device.
  • the multi-objective optimization device is only a non-limiting embodiment provided by the present invention, and does not limit the subject of implementation of each step in these multi-objective optimization methods.
  • FIG. 1 shows a schematic flowchart of a multi-objective optimization method for a catalytic cracking process according to some embodiments of the present invention.
  • the multi-objective optimization device can first determine the mechanism model of the catalytic cracking process, and according to the mechanism model to determine multiple optimization objectives, corresponding to the multiple optimization Multiple process decision variables of the target, and the constraint range of each process decision variable.
  • the mechanism model can be pre-built offline by technicians for users to directly select and use in the process of multi-objective optimization, or it can be used by users in the process of multi-objective optimization based on the The process mechanism is directly established.
  • the catalytic cracking unit shown in Figure 2 is mainly composed of five parts: reaction-regeneration system, fractionation system, absorption stabilization system, product refining system and energy recovery. Heavy fractions such as crude oil, coker distillate oil, deasphalted oil, and vacuum gas oil are converted into high-quality light products such as dry gas, liquefied petroleum gas, stable gasoline, and light diesel oil.
  • the above-mentioned reaction regeneration system is generally composed of a riser reactor 11 and a catalyst regenerator 12, 13.
  • the reactor 11 is mainly used for reacting the passing raw material oil to produce the target product under the action of certain reaction temperature, pressure and catalyst. These complex products enter the fractionation system through the oil and gas pipeline in the form of gas at high temperature.
  • the surface of the catalyst is temporarily deactivated by the adhesion of the coke formed.
  • These waiting-life catalysts will enter the regenerators 12, 13, and the coke will be burned off with the oxygen in the air, so that the catalysts will be reactivated.
  • the heat released by the coke will be brought into the reactor 11 by the regenerated catalyst for use in the reaction, and the excess heat will be recycled by the equipment.
  • the above-mentioned fractionation system mainly includes a fractionation tower 14, a stripping tower 15, a raw oil buffer return tank, and a heat exchange system.
  • This system mainly separates the high-temperature oil and gas from the reactor 11 into rich gas, crude gasoline, sulfur-containing sewage, light gasoline, recycled oil and oil slurry according to the boiling point of each fraction.
  • the absorption stabilization system above mainly includes an absorption tower 16, a reabsorption tower 17, a desorption tower 18, a stabilization tower 18, a rich gas compressor and a corresponding heat exchange system.
  • the main task of this system is to separate the crude gasoline and rich gas separated by the oil-gas separator at the top of the fractionation tower 14, and use the difference in solubility of each component in the liquid to separate them into products such as dry gas, liquefied gas and stable gasoline. Control the indicators of each product to meet product requirements.
  • the above-mentioned product refining system is installed after the absorption and stabilization system, and is mainly used for refining operations such as desulfurization and sweetening of dry gas, liquefied gas and stabilized gasoline, so as to make them meet the relevant requirements of the environmental protection law for products.
  • the energy recovered by the above-mentioned energy recovery system is mainly used to maintain the heat balance of the reaction-regeneration system and the fractionation system.
  • the system usually consists of flue gas energy recovery unit, waste heat boiler, external heat extractor, fractionation tower bottom oil slurry steam generator and water supply system.
  • the recovered energy mainly includes the remaining heat in the regenerator, the pressure energy of the regeneration flue gas, the heat energy and the excess heat of the fractionation system.
  • the technicians can also correct the model parameters of the catalytic cracking mechanism model. Specifically, technicians can use the intelligent optimization algorithm to determine the optimal value of the model parameters that meet the accuracy requirements within the constraint range of the input parameters of the catalytic cracking mechanism model, and use the obtained optimized values of the model parameters to update the parameters of the catalytic cracking mechanism model. Model parameters, in order to improve the accuracy of the model, and make the output value obtained by the catalytic cracking mechanism model consistent with the production data, so that the output value obtained by the catalytic cracking mechanism model is more accurate.
  • the multi-objective optimization device can obtain multiple optimization objectives provided by the user, and determine multiple process decision variables corresponding to the multiple optimization objectives according to the mechanism model of the catalytic cracking process, and each process The bounds of the decision variable.
  • the plurality of optimization goals may include maximizing economic benefits, minimizing carbon dioxide emissions, minimizing sulfur dioxide emissions, minimizing total exhaust gas emissions, minimizing total waste liquid emissions, maximizing primary product yields , maximizing primary product yield, minimizing input cost at least one of.
  • the plurality of optimization objectives may at least include the optimization objective of minimizing carbon dioxide emissions.
  • M indicates the economic benefit of the catalytic cracking process
  • P i indicates the price of product i
  • Y i indicates the flow rate of product i
  • i indicates the product, including acid gas, dry gas, ethylene, liquefied gas, propylene, gasoline, diesel, cycle At least one of oil, oil slurry, and coke
  • p j indicates the price of raw material j
  • y j indicates the flow rate of raw material j
  • j indicates raw materials, including tank farm residual oil, tank farm wax oil, and refined wax oil
  • P f Indicates the fixed cost per unit of processing, F indicates the processing load;
  • p c indicates the price of fresh catalyst, f c indicates the replenishment of fresh catalyst;
  • C CO2 indicates the amount of carbon dioxide emissions;
  • C SO2 indicates the amount of sulfur dioxide emissions.
  • x 1 , x 2 , ..., x i are process decision variables, including but not limited to feed temperature, feed pressure , the outlet temperature of the first reaction zone, the reactor pressure, the mass flow rate of the feed, the activity index of the regenerated catalyst, and the oxygen index; the constraint range of the process decision variable is Among them, x i min is the lower limit value of the process decision variable, and xi max is the upper limit value of the process decision variable.
  • the multi-objective optimization device can improve the SPEA2 algorithm (Strength Pareto Evolutionary Algorithm) of the offspring evolution process through the path-based regeneration operator , adjust the value of each process decision variable within the constraint range to determine the operating data of the objective function on each process decision variable.
  • SPEA2 algorithm Strength Pareto Evolutionary Algorithm
  • Pareto (Pareto) analysis is a commonly used primary and secondary factor analysis method in project management. Its core idea is to distinguish the primary and secondary among the many factors that determine a thing, and to identify a few key factors that play a decisive role in things and the majority. Secondary factors that have less influence on things.
  • the above multi-objective optimization algorithm SPEA2PE based on the integrated path regeneration operator dominated by Pareto is an elite multi-objective evolutionary algorithm that enhances Pareto. It adopts the algorithm framework of the classic multi-objective optimization algorithm SPEA2 and can find multiple Pareto The optimal solution, and after the algorithm ends, the solution in the external set is used as an approximation to the Pareto optimal solution of the solving problem.
  • FIG. 3 shows a schematic flowchart of adjusting process decision variables according to some embodiments of the present invention.
  • the multi-objective optimization device can perform the following steps in sequence:
  • the multi-objective optimization device can first determine the number of individuals S(i) dominated by individual i, and then according to the individual The number of individuals S(i) dominated by i determines the original fitness value R(i) of individual i:
  • the original fitness value R(i) represents individual i, and the sum of all individual numbers dominated by each individual (iv) that dominates individual i.
  • the multi-objective optimization device can separately calculate each individual i to the population P t and the reserve set The distances of all individuals in , and sorted in ascending order. Afterwards, the multi-objective optimization device can select the kth individual as And calculate the corresponding distance value D(i):
  • N is the size of the population P t , for the reserve set
  • the size of add 2 to the denominator to ensure the distance value D(i) ⁇ (0, 1).
  • the multi-objective optimization device can determine the fitness F(i) of individual i according to the original fitness value R(i) and distance value D(i):
  • the multi-objective optimization device can first store the The number of individuals in and the size of the reserve set Compare. like Then the multi-objective optimization device can select from the population P t and the reserve set according to the order of the fitness F(i). Before selecting among all individuals i Individuals constitute the dominant solution set and are added to the reserve set Conversely, if Then the multi-objective optimization device can perform pruning operation, and eliminate the individual i with the smallest distance from the adjacent individual in each iteration, where the individual i satisfies:
  • the above formula indicates that all individuals and neighbors are sorted by distance, and the individual i with the smallest adjacent distance is selected to be deleted (k and i are not necessarily neighbors with the smallest distance to each other), thereby reducing the population Crowded conditions, and to ensure population diversity.
  • the multi-objective optimization device can determine that these individuals with equal distances have a greater correlation with the objective function P, and therefore retain these individuals first and consider deleting individuals with the second smallest value.
  • the individual i' with the adjacent distance is kept, while the individual i with the minimum distance value is kept.
  • the multi-objective optimization device can further consider the individual i" with the third smallest adjacent distance, and retain the individual i with the smallest distance value and the second smallest distance value individual i" with the distance value, and so on until the only individual i with the smallest distance is deleted.
  • FIG. 4 shows a schematic flowchart of calculating data sets in a mating pool according to some embodiments of the present invention.
  • Fig. 5A shows a schematic diagram of an evolutionary path provided according to some embodiments of the present invention at an early stage of evolution.
  • Fig. 5B shows a schematic diagram of an evolutionary path provided according to some embodiments of the present invention at a later stage of evolution.
  • Fig. 6 shows a schematic diagram of generating offspring individuals according to some embodiments of the present invention.
  • the multi-objective optimization device can first combine each offspring decision variable X g and parent decision
  • each individual participates in the calculation of the center point of the population.
  • the multi-objective optimization device can also be based on the target survival rate of offspring individuals and the actual productivity psucc, adaptively determine the forward step length ⁇ of the evolutionary path ep, namely
  • ⁇ ⁇ (1, ⁇ ) determines at what rate ⁇ increases or decreases.
  • the algorithm has basically completed the convergence, and the population is located near the optimal PS.
  • the length ⁇ of ep and the angle between ep and the optimal PS are both small, and ⁇ will decrease at a speed of ⁇ times, so that each individual can perform a local search in the PS neighborhood around itself,
  • the PE operator can use polynomial mutation to slightly increase the diversity of genes in the population.
  • the multi-objective optimization device can set a lower limit value ⁇ lb for ⁇ .
  • the multi-objective optimization device can randomly initialize an element value in Center g to ensure that the difference between Center g-1 and Center g is sufficient Obviously, that is to ensure that the element value in ep will not be too small, so as to prevent the population from being limited to a local optimum.
  • the multi-objective optimization device can use each parent individual as a starting point to generate corresponding offspring within the rectangular range pointed to by the vector ⁇ *ep individual.
  • the present invention can save the trouble of selecting a specific individual, so as to calculate ep more simply.
  • the ep calculated by all individuals is more robust and reliable, so that most offspring individuals can benefit from ep.
  • preliminary experiments show that most individuals show a relatively consistent evolutionary direction in the early and late stages of evolution, so the entire population can share this evolutionary direction, which can effectively improve the convergence speed of the operator and ensure the diversity of the population.
  • the multi-objective optimization device can also perform gene sharing operations between at least one excellent parent individual and each offspring individual (Gene- Sharing Operation). Specifically, the multi-objective optimization device may first obtain the Pareto front layers or fitness F(i) of each parent individual. If the number of Pareto front layers is equal to 1, the multi-objective optimization device can determine that it has a solution that is not dominated by any solution. In this way, the multi-objective optimization device can convert Pa The number of Reto frontier layers is equal to 1 or the parent individuals with better fitness F(i) are determined to be the carriers of excellent genes, so as to screen at least one excellent parent individual.
  • the multi-objective optimization device can refer to the simulated binary crossover (SBX) operator to perform a crossover operation on each variable with a probability of 0.5, so that the excellent genes in the parent individuals that are more correlated with the objective function are shared among the offspring .
  • SBX simulated binary crossover
  • the multi-objective optimization device can also preferably determine the process decision variables Whether the current operating value x i of is within the constraint range x i min ⁇ x i max of the process decision variable. If the current operating value x i is within the constraint range x i min ⁇ xi max of the process decision variable, the multi-objective optimization device may use the current operating value x i as the initial value of the process decision variable.
  • the multi-objective optimization device can execute the repair function in the algorithm (for example: re-execute the S1 step of the SPEA2 algorithm to obtain the initialization population P 0 , and the generated initial population P 0 is added to the existing population). Afterwards, the multi-objective optimization device can regenerate the operation data x i of each decision variable within the constraint range x i min ⁇ xi max of the process decision variable, as the initial value of each process decision variable.
  • the multi-objective optimization device can determine the optimization target value of each optimization objective according to the operation data of each process decision variable. Afterwards, the multi-objective optimization device can determine the corresponding operation data as guidance values of multiple process decision variables corresponding to the multiple optimization targets according to the optimal optimization target value solution set.
  • the multi-objective optimization device can firstly input various operating data into the mechanism model of the catalytic cracking process, so as to obtain the yield data of various products and gases. Afterwards, the multi-objective optimization device can input each yield data into each objective function, so as to obtain the optimization target value of each optimization objective.
  • the multi-objective optimization device can send the corresponding operation data into the catalytic cracking mechanism model through the interface of Matlab and process simulation software Aspen Hysys, so as to calculate the benefit values of each optimization target f 1 , f 2 , and f 3 respectively. Afterwards, the multi-objective optimization device can evaluate each optimization target value according to the Pareto optimal solution set and the IGD index as shown in FIG. 7 and FIG. 8 , so as to determine the optimal optimization target value solution set.
  • the present invention can realize the PE operator and the FCC mechanism model interface program through the programming means of Matlab software Data linkage for catalytic cracking mechanism models. Afterwards, the multi-objective optimization device can run the control program, automatically calculate the value of the objective function P under different decision variables by calling data, programs and algorithms, and continuously optimize the values of each process decision variable to make the objective function value To achieve better results, thereby obtaining the operation data of the objective function.
  • the improved SPEA2 algorithm will also automatically change different step size change strategies according to the specific situation as mentioned above, so as to perform incomplete polling calculations according to different step size change strategies within the constraint range, and continuously obtain the objective function Operate the data better until the objective function value is optimal.
  • each process decision variable includes but not limited to:
  • the multi-objective optimization device can send the corresponding operation data into the catalytic cracking mechanism model through the interface of Matlab and process simulation software Aspen Hysys, so as to calculate the benefit values of each optimization target f 1 , f 2 , and f 3 respectively.
  • Table 1 shows comparison data of optimization results provided according to some embodiments of the present invention.
  • the optimized operation data can provide guidance for the optimal operation of the catalytic cracking unit, and reduce the emission of carbon dioxide and sulfur dioxide while improving economic benefits.

Landscapes

  • Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Oil, Petroleum & Natural Gas (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Geometry (AREA)
  • Evolutionary Computation (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • General Chemical & Material Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Organic Chemistry (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Feedback Control In General (AREA)

Abstract

本发明提供了一种催化裂化过程的多目标优化方法及装置，以及一种计算机可读存储介质。所述多目标优化方法包括以下步骤：确定多个优化目标、对应于所述多个优化目标的多个工艺决策变量，以及各所述工艺决策变量的约束范围；根据所述多个优化目标及所述多个工艺决策变量，确定目标函数；经由基于路径的再生算子改进子代进化过程的SPEA2算法，在所述约束范围内调整各所述工艺决策变量的取值，以确定所述目标函数关于各所述工艺决策变量的操作数据；根据各所述操作数据，确定各所述优化目标的优化目标值；以及根据最优的优化目标值解集，确定对应于所述多个优化目标的多个工艺决策变量的指导值。

Description

一种催化裂化过程的多目标优化方法及装置 技术领域
本发明属于石油化工技术领域，尤其涉及一种催化裂化过程的多目标优化方法、一种催化裂化过程的多目标优化装置，以及一种计算机可读存储介质。
背景技术
炼油产业是我国的支柱型产业，其中催化裂化工艺是石油炼制工业中重油轻质化的重要手段之一。经过催化裂化(Fluid Catalytic Cracking，FCC)工艺得到的汽油、柴油等催化裂化产品是交通运输的重要燃料来源，而经过催化裂化工艺生产的其他产品也是化工产业提供了主要原料。针对炼油企业，如何充分发挥炼油装置能力、实现经济效益最大化和降低环境污染是关系到国计民生的重中之重，也是国家实现产业转型升级的有力保障。
在催化裂化工艺的实际应用中，需要综合考虑经济收益、环境保护等多个相互冲突的优化目标，即多目标优化。然而，石油化工领域现有的催化裂化优化技术主要集中在单一目标的优化操作，无法满足多目标优化的实际需求。即使信息科学技术等其他技术领域存在多目标优化的相关技术，也普遍存在忽视个体进化过程中产生的有用信息的现象，因而需要进行大量的迭代，并存在优化时间长、优化结果质量差、优化结果不稳定的缺陷。
为了克服现有技术存在的上述缺陷，本领域亟需一种催化裂化过程的多目标优化技术，一方面满足催化裂化工艺的多目标优化需求，另一方面提升催化裂化工艺的多目标优化效率，并提升优化结果的质量及稳定性。
发明内容
以下给出一个或多个方面的简要概述以提供对这些方面的基本理解。此概述不是所有构想到的方面的详尽综览，并且既非旨在指认出所有方面的关键性或决定性要素亦非试图界定任何或所有方面的范围。其唯一的目的是要以简化形式给出一个或多个方面的一些概念以为稍后给出的更加详细的描述之前序。
为了克服现有技术存在的上述缺陷，本发明提供了一种催化裂化过程的多目标优化方法、一种催化裂化过程的多目标优化装置，以及一种计算机可读存储介质，能够经由基于路径的再生算子改进子代进化过程的SPEA2算法，在预设的约 束范围内调整催化裂化过程的各工艺决策变量的取值，以确定对应于多个优化目标的多个工艺决策变量的指导值，从而满足催化裂化工艺的多目标优化需求，提升催化裂化工艺的多目标优化效率，并提升优化结果的质量及稳定性。
具体来说，本发明的第一方面提供的上述催化裂化过程的多目标优化方法，包括以下步骤：确定多个优化目标、对应于所述多个优化目标的多个工艺决策变量，以及各所述工艺决策变量的约束范围；根据所述多个优化目标及所述多个工艺决策变量，确定目标函数；经由基于路径的再生算子改进子代进化过程的SPEA2算法，在所述约束范围内调整各所述工艺决策变量的取值，以确定所述目标函数关于各所述工艺决策变量的操作数据；根据各所述工艺决策变量的操作数据，确定各所述优化目标的优化目标值；以及根据最优的优化目标值解集，将对应的各所述操作数据确定为对应于所述多个优化目标的多个工艺决策变量的指导值。
进一步地，在本发明的一些实施例中，所述经由基于路径的再生算子改进子代进化过程的SPEA2算法，在所述约束范围内调整各所述工艺决策变量的取值，以确定所述目标函数关于各所述工艺决策变量的操作数据的步骤包括：S1：初始化迭代次数变量t、种群Pt及储备集其中，初始化的种群P0是由所述多个工艺决策变量在t＝0时的集合构成，初始化的储备集为空集；S2：计算所述种群Pt中各个体i的适应度F(i)，其中，每个所述个体i对应一个所述工艺决策变量；S3：将所述种群Pt及所述储备集中的所有非支配解集复制到储备集并对所述储备集进行环境选择；S4：若迭代次数t未达到预设的迭代次数上限T，则对经过所述环境选择的储备集进行锦标赛选择，再将经过所述锦标赛选择的数据集放入交配池；S5：经由所述基于路径的再生算子，对所述交配池中的所述数据集进行计算，将计算结果存入所述储备集递增迭代次数，并返回步骤S2；以及S6：若迭代次数变量t达到所述迭代次数上限T，则输出所述储备集中非支配解所代表的工艺决策变量集A。
进一步地，在本发明的一些实施例中，所述计算所述种群Pt中各个体i的适应度F(i)的步骤包括：根据个体i支配的个体数S(i)，确定个体i的原始适应值R(i)， 其中，所述原始适应值R(i)表示个体i，以及支配个体i的每个个㈣所支配的所有个体数之和；计算每个个体i到所述种群Pt及所述储备集中所有个体的距离，并按照递增排序；选择第k个个体作为并计算对应的距离值D(i)，其中，所述N为所述种群Pt的大小，所述为所述储备集的大小；以及根据所述原始适应值R(i)及所述距离值D(i)，确定所述个体i的适应度F(i)。
进一步地，在本发明的一些实施例中，所述对所述储备集进行环境选择的步骤包括：将所述储备集中的个体数量与储备集大小进行比较；若则根据适应度F(i)的排序，从所述种群Pt及所述储备集的所有个体i中选择前个个体构成支配解集，并添加到所述储备集以及若则进行剪枝操作，在每次迭代中淘汰与相邻个体具有最小距离的个体i。
进一步地，在本发明的一些实施例中，所述在每次迭代中淘汰与相邻个体具有最小距离的个体i的步骤包括：在每次迭代中淘汰与相邻个体具有最小距离的唯一个体i。
进一步地，在本发明的一些实施例中，所述经由所述基于路径的再生算子，对所述交配池中的所述数据集进行计算的步骤包括：确定所述储备集的子代种群个体的中心点Centerg，并确定所述种群Pt及所述储备集的父代种群个体的中心点Centerg-1；根据前后两代种群个体中心点的差，定义进化路径ep的方向；根据子代个体的目标生存率及实际生产率psucc，确定所述进化路径ep的前进步长α；以及以每个父代个体为出发点，在向量α*ep所指向的矩形范围内分别产生对应的子代个体。
进一步地，在本发明的一些实施例中，所述经由所述基于路径的再生算子，对所述交配池中的各所述个体i进行重组和变异的步骤还包括：在产生所述子代个体之后，在至少一个优秀的父代个体及各所述子代个体之间进行基因共享操作。
进一步地，在本发明的一些实施例中，在进行所述基因共享操作之前，所述 多目标优化方法还包括以下步骤：获取各所述父代个体的帕雷托前沿层数或适应度F(i)；以及根据所述帕雷托前沿层数或所述适应度F(i)，筛选所述至少一个优秀的父代个体。
进一步地，在本发明的一些实施例中，在根据最优的优化目标值解集，将对应的各所述操作数据确定为对应于所述多个优化目标的多个工艺决策变量的指导值之前，所述多目标优化方法还包括以下步骤：根据帕雷托最优解集合以及IGD指标来评价各所述优化目标值，以确定所述最优的优化目标值解集。
进一步地，在本发明的一些实施例中，所述多个优化目标中至少包括最小化二氧化碳排放量。
进一步地，在本发明的一些实施例中，所述多个优化目标还包括：最大化经济效益、最小化二氧化硫排放量、最小化废气总排放量、最小化废液总排放量、最大化主要产品产量、最大化主要产品收率、最小化投入成本中的至少一者。
进一步地，在本发明的一些实施例中，所述多个工艺决策变量至少包括：物料流量、物料状态、物料性质、主分馏塔状态、主分馏塔操作、吸收塔状态、吸收塔操作、再吸收塔状态、再吸收塔操作、解吸塔状态、解吸塔操作、稳定塔状态、稳定塔操作中的至少一者。
进一步地，在本发明的一些实施例中，所述多个优化目标包括最大化经济效益、所述最小化二氧化碳的排放量及最小化二氧化硫的排放量。所述目标函数如下：
min imizeF(x)＝(1/f1，f2，f3)，
f2＝min CCO2，
f3＝min CSO2，
其中，M指示催化裂化过程的经济效益；Pi指示产品i的价格；Yi指示产品i的流量；i指示产品，包括酸性气、干气、乙烯、液化气、丙烯、汽油、柴油、循环油、油浆、焦炭中的至少一种；pj指示原料j的价格；yj指示原料j的流量；j指示原料，包括为罐区渣油、罐区蜡油、精制蜡油；Pf指示单位加工量的固定成本， F指示加工负荷；pc指示新鲜催化剂的价格，fc指示新鲜催化剂补充量；CCO2指示二氧化碳排放量；CSO2表示二氧化硫排放量。
此外，本发明的第二方面提供的上述多目标优化装置包括存储器及处理器。所述处理器连接所述存储器，并被配置用于实施本发明的第一方面提供的上述催化裂化过程的多目标优化方法。
此外，本发明的第三方面提供的上述计算机可读存储介质，其上存储有计算机指令。所述计算机指令被处理器执行时，实施本发明的第一方面提供的上述催化裂化过程的多目标优化方法。
附图说明
在结合以下附图阅读本公开的实施例的详细描述之后，能够更好地理解本发明的上述特征和优点。在附图中，各组件不一定是按比例绘制，并且具有类似的相关特性或特征的组件可能具有相同或相近的附图标记。
图1示出了根据本发明的一些实施例提供的催化裂化过程的多目标优化方法的流程示意图。
图2示出了根据本发明的一些实施例提供的催化裂化装置的示意图。
图3示出了根据本发明的一些实施例提供的调整工艺决策变量的流程示意图。
图4示出了根据本发明的一些实施例提供的计算交配池中数据集的流程示意图。
图5A示出了根据本发明的一些实施例提供的进化路径在进化前期的示意图。
图5B示出了根据本发明的一些实施例提供的进化路径在进化后期的示意图。
图6示出了根据本发明的一些实施例提供的产生子代个体的示意图。
图7示出了根据本发明的一些实施例提供的IGD指标的示意图。
图8示出了根据本发明的一些实施例提供的帕雷托最优解集的示意图。
具体实施方式
以下由特定的具体实施例说明本发明的实施方式，本领域技术人员可由本说明书所揭示的内容轻易地了解本发明的其他优点及功效。虽然本发明的描述将结合优选实施例一起介绍，但这并不代表此发明的特征仅限于该实施方式。恰恰相反，结合实施方式作发明介绍的目的是为了覆盖基于本发明的权利要求而有可能 延伸出的其它选择或改造。为了提供对本发明的深度了解，以下描述中将包含许多具体的细节。本发明也可以不使用这些细节实施。此外，为了避免混乱或模糊本发明的重点，有些具体细节将在描述中被省略。
如上所述，石油化工领域现有的催化裂化优化技术主要集中在单一目标的优化操作，无法满足多目标优化的实际需求。即使信息科学技术等其他技术领域存在多目标优化的相关技术，也普遍存在忽视个体进化过程中产生的有用信息的现象，因而需要进行大量的迭代，并存在优化时间长、优化结果质量差、优化结果不稳定的缺陷。
为了克服现有技术存在的上述缺陷，本发明提供了一种催化裂化过程的多目标优化方法、一种催化裂化过程的多目标优化装置，以及一种计算机可读存储介质，能够经由基于路径的再生算子改进子代进化过程的SPEA2算法，在预设的约束范围内调整催化裂化过程的各工艺决策变量的取值，以确定对应于多个优化目标的多个工艺决策变量的指导值，从而满足催化裂化工艺的多目标优化需求，提升催化裂化工艺的多目标优化效率，并提升优化结果的质量及稳定性。
在一些非限制性的实施例中，本发明的第一方面提供的上述催化裂化过程的多目标优化方法，可以由本发明的第二方面提供的上述催化裂化过程的多目标优化装置来实施。具体来说，该多目标优化装置中配置有存储器及处理器。该存储器包括但不限于本发明的第三方面提供的上述计算机可读存储介质，其上存储有计算机指令。该处理器连接该存储器，并被配置用于实施本发明的第一方面提供的上述催化裂化过程的多目标优化方法。
以下将结合一些多目标优化方法的实施例来描述上述多目标优化装置的工作原理。本领域的技术人员可以理解，这些多目标优化方法的实施例只是本发明提供的一些非限制性的实施方式，旨在清楚地展示本发明的主要构思，并提供一些便于公众实施的具体方案，而非用于限制该多目标优化装置的全部功能或全部工作方式。同样地，该多目标优化装置也只是本发明提供的一种非限制性的实施方式，不对这些多目标优化方法中各步骤的实施主体构成限制。
请首先参考图1，图1示出了根据本发明的一些实施例提供的催化裂化过程的多目标优化方法的流程示意图。
如图1所示，在多目标优化的过程中，多目标优化装置可以首先确定催化裂化过程的机理模型，并根据该机理模型来确定多个优化目标、对应于该多个优化 目标的多个工艺决策变量，以及各工艺决策变量的约束范围。
在一些实施例中，该机理模型可以由技术人员在线下预先构建，以供使用者在多目标优化的过程中直接选择使用，也可以由使用者在多目标优化的过程中基于催化裂化装置的过程机理来直接建立。
以图2所示的催化裂化装置为例，其主要由反应-再生系统、分馏系统、吸收稳定系统、产品精制系统和能量回收五部分组成，能够在加热和催化剂的作用下，将常压渣油、焦化馏分油、脱沥青油、减压蜡油等重质馏分转换成高品质的干气、液化石油气、稳定汽油和轻柴油等轻质产品。
具体来说，上述反应再生系统一般是由提升管反应器11和催化剂再生器12、13这两部分组成。反应器11主要用于在一定的反应温度、压力及催化剂的作用下，将通过的原料油反应生成目标产物。这些复杂的产物在高温下以气体的形式通过油气管线进入分馏系统。在反应过程中，催化剂表面由于被生成的焦炭附着而暂时失去活性。这些待生催化剂就要进入再生器12、13，用空气中的氧气烧掉焦炭，使催化剂再恢复活性。焦炭释放的热量将由再生的催化剂带入到反应器11中，以供反应使用，而多余的热量再由设备回收利用。
上述分馏系统主要包括分馏塔14、汽提塔15、原料油缓冲回炼罐以及换热系统等。该系统主要是把从反应器11来的高温油气根据各馏分沸点的不同分离成富气、粗汽油、含硫污水、轻汽油、回炼油和油浆等。
上述吸收稳定系统主要包括吸收塔16、再吸收塔17、解吸塔18和稳定塔18，还有富气压缩机以及相应的换热系统。该系统的主要任务是将分馏塔14顶油气分离器分离出来的粗汽油和富气，利用各个组分在液体中溶解度的不同把其分离成干气、液化气和稳定汽油等产品。控制各产品的指标使之达到产品要求。
上述产品精制系统设于吸收稳定系统之后，主要用于对干气、液化气和稳定汽油进行脱硫、脱硫醇等精制操作，使之达到环境保护法对产品的相关要求。
上述能量回收系统回收的能量主要用于维持反应-再生系统和分馏系统的热平衡。该系统通常由烟气能量回收机组、余热锅炉、外取热器、分馏塔底油浆蒸汽发生器和供水系统组成。回收的能量主要包括再生器内剩余的热量、再生烟气的压力能、热能及分馏系统的过剩热量。
技术人员可以基于催化裂化装置的上述架构及反应机理，进行原料油特征化、组分划分、反应网络划分、反应动力学模型建立等操作，使用耦合碳数分布、硫 和氮分布的烃类反应动力学反应系统构成催化裂化的全流程机理模型，其具体方案为本领域的现有技术，在此不做赘述。
进一步地，在一些实施例中，在建立催化裂化过程的机理模型之后，技术人员还可以对催化裂化机理模型的模型参数进行校正。具体来说，技术人员可以利用智能优化算法，在催化裂化机理模型输入参数的约束范围内确定满足精度要求的模型参数优化值，并使用得到的模型参数的优化值更新所述催化裂化机理模型的模型参数，以提高模型的精度，并使催化裂化机理模型得到的输出值与生产数据保持一致，从而使催化裂化机理模型得到的输出值更加精确。
在建立催化裂化过程的机理模型之后，多目标优化装置可以获取用户提供的多个优化目标，并根据催化裂化过程的机理模型确定该多个优化目标所对应的多个工艺决策变量，以及各工艺决策变量的约束范围。在一些实施例中，该多个优化目标可以包括最大化经济效益、最小化二氧化碳排放量、最小化二氧化硫排放量、最小化废气总排放量、最小化废液总排放量、最大化主要产品产量、最大化主要产品收率、最小化投入成本中的至少一者。尤其是在当前倡导“碳中和”的环保背景下，该多个优化目标中可以至少包括最小化二氧化碳排放量的优化目标。
以最大化经济效益、最小化二氧化碳的排放量及最小化二氧化硫的排放量的三个优化目标为例，其目标函数如下：
min imizeF(x)＝(1/f1，f2，f3)，
f2＝min CCO2，
f3＝min CSO2，
其中，M指示催化裂化过程的经济效益；Pi指示产品i的价格；Yi指示产品i的流量；i指示产品，包括酸性气、干气、乙烯、液化气、丙烯、汽油、柴油、循环油、油浆、焦炭中的至少一种；pj指示原料j的价格；yj指示原料j的流量；j指示原料，包括为罐区渣油、罐区蜡油、精制蜡油；Pf指示单位加工量的固定成本，F指示加工负荷；pc指示新鲜催化剂的价格，fc指示新鲜催化剂补充量；CCO2指示二氧化碳排放量；CSO2表示二氧化硫排放量。
进一步地，该目标函数可以被表达为：
P＝F(x1，x2，...，xi)
其中，P为目标函数；x1，x2，...，xi(i＝1，2，3，...，n)为工艺决策变量，包括但不限于进料温度、进料压力、第一反应区出口温度、反应器压力、进料质量流量、再生催化剂活性指数、氧气指数；工艺决策变量的约束范围为其中，xi min为工艺决策变量的下限值，xi max为工艺决策变量的上限值。
如图1所示，在根据多个优化目标及多个工艺决策变量确定目标函数P之后，多目标优化装置可以经由基于路径的再生算子改进子代进化过程的SPEA2算法(Strength Pareto Evolutionary Algorithm)，在约束范围内调整各工艺决策变量的取值，以确定目标函数关于各工艺决策变量的操作数据。
帕雷托(Pareto)分析法是项目管理中常用的主次因素分析法，其核心思想是在决定一事物的众多因素中分清主次，识别出少数但对事物起决定作用的关键因素和多数的但对事物影响较小的次要因素。
不失一般性，多目标优化问题可以表述如下：
minimize F(x)＝(f1(x)，...，fM(x))
subject to x∈Ω
其中，为D维决策空间；公式中各个目标fi(x)通常是相互冲突的；一个目标的改善通常会导致另一个目标变差；为将决策变量映射到M维目标空间的映射函数多目标优化试图在目标空间得到一组目标值，称为帕雷托前沿(Pareto Front，PF)；对应地，在决策空间得到一组解，称为帕雷托解集(Pareto Set，PS)。
上述基于帕雷托支配的集成路径再生算子的多目标优化算法SPEA2PE是增强Pareto的精英多目标进化算法，采用了经典多目标优化算法SPEA2的算法框架，能在单次运行中找到多个Pareto最优解，并在算法结束后以外部集合中的解作为对求解问题的Pareto最优解的近似。
请参考图3，图3示出了根据本发明的一些实施例提供的调整工艺决策变量的流程示意图。
如图3所示，在采用改进的SPEA2算法来调整工艺决策变量取值的过程中， 多目标优化装置可以依次执行以下步骤：
S1：初始化迭代次数变量t、种群Pt及储备集其中，初始化的种群P0是由多个工艺决策变量在t＝0时的集合构成，初始化的储备集为空集；
S2：计算种群Pt中各个体i的适应度F(i)，其中，每个个体i对应一个工艺决策变量；
S3：将种群Pt及储备集中的所有非支配解集复制到储备集并对储备集进行环境选择；
S4：若迭代次数t未达到预设的迭代次数上限T，则对经过环境选择的储备集进行锦标赛选择，再将经过锦标赛选择的数据集放入交配池；
S5：经由基于路径的再生算子，对交配池中的数据集进行计算，将计算结果存入所述储备集递增迭代次数，并返回步骤S2；以及
S6：若迭代次数变量t达到迭代次数上限T，则输出储备集中非支配解所代表的工艺决策变量集A。
具体来说，在计算种群Pt中各个体i的适应度F(i)(即步骤S2)的过程中，多目标优化装置可以首先确定个体i支配的个体数S(i)，再根据个体i支配的个体数S(i)确定个体i的原始适应值R(i)：

其中，原始适应值R(i)表示个体i，以及支配个体i的每个个㈣所支配的所有个体数之和。
之后，多目标优化装置可以分别计算每个个体i到种群Pt及储备集中所有个体的距离，并按照递增排序。再之后，多目标优化装置可以选择第k个个体作为并计算对应的距离值D(i)：
其中，N为种群Pt的大小，为所述储备集的大小，分母加2以保证距离值D(i)∈(0，1)。
再之后，多目标优化装置可以根据原始适应值R(i)及距离值D(i)，确定个体i的适应度F(i)：
F(i)＝R(i)+D(i)。
在对储备集进行环境选择(即步骤S3)的过程中，多目标优化装置可以首先将储备集中的个体数量与储备集大小进行比较。若则多目标优化装置可以根据适应度F(i)的排序，从种群Pt及所述储备集的所有个体i中选择前个个体构成支配解集，并添加到储备集反之，若则多目标优化装置可以进行剪枝操作，在每次迭代中淘汰与相邻个体具有最小距离的个体i，其中，个体i满足：

其中，代表个体i与第k个邻居的距离，上面公式表示将所有个体与邻居距离排序，并选择删除具有最小相邻距离的个体i(k和i不一定互为最小距离的邻居)，从而减少种群拥挤的情况，并保证种群多样性。进一步地，如果有多个个体都具有相同的距离最小值，则多目标优化装置可以判定这些距离相等的个体与目标函数P具有较大的相关度，因而优先保留这些个体而考虑删除具有次小相邻距离的个体i′，而保留最小距离值的个体i。同理，若如果有多个个体都具有相同的距离次小值，则多目标优化装置可以进一步考虑具有第三小相邻距离的个体i″，而保留最小距离值的个体i及具有次小距离值的个体i″，并以此类推，直到删除具有最小距离的唯一个体i。
请结合参考图4、图5A、图5B及图6。图4示出了根据本发明的一些实施例提供的计算交配池中数据集的流程示意图。图5A示出了根据本发明的一些实施例提供的进化路径在进化前期的示意图。图5B示出了根据本发明的一些实施例提供的进化路径在进化后期的示意图。图6示出了根据本发明的一些实施例提供的产生子代个体的示意图。
如图4所示，在经由基于路径的再生算子对交配池中的数据集进行计算(即步骤S5)的过程中，多目标优化装置可以首先将各子代决策变量Xg及父代决策变量Xg-1输入基于路径的再生算子(Path Evolution，PE)，并根据公式Centerg＝mean(x1，...，xi，...，xN)，确定储备集的子代种群个体的中心点Centerg，并确定种群Pt及储备集的父代种群个体的中心点Centerg-1，其中，Centerg是第g代种群个体的中心点，N是种群中的个体数量，D是决策变量的个数。这里，每一个个体都参与到了种群中心点的计算当中。
之后，多目标优化装置可以根据前后两代种群个体中心点的差，定义进化路径ep的方向，即ep＝Centerg-Centerg-1。
此外，多目标优化装置还可以根据子代个体的目标生存率及实际生产率psucc，自适应地确定进化路径ep的前进步长α，即
其中，是一个用户自定义的参数，代表子代个体的目标生存率，即子代个体进化到下一代的目标百分比；psucc是实际子代个体的生产率，其值等于进化到下一代的子代个体数量除以每代中的种群数(N)；参数ω∈(1，∞)决定了α以何种速度增大或减小。
如图5A所示，在进化过程的早期阶段，决策空间中大多数个体都远离最优PS，非支配个体相对容易产生。此时，ep的长度α、ep与最优PS之间的夹角均较大，且α将以ω倍的速度增大从而形成一个明确的进化路径，从而使种群快速向最优PS前进。进一步地，这将产生一个更大的α，从而形成一个更具潜力和更长的进化路径。上面这个过程是一个正反馈过程。通过增加α，该改进算法所需的计算资源较少，收敛速度显著更快。为了避免在某一代中朝着ep方向前进太远，从而错误地引导子代个体，多目标优化装置可以对α设置一个上限值αub。
如图5B所示，在进化过程的后期阶段，算法已基本完成收敛，种群位于最优PS的附近。此时，ep的长度α、ep与最优PS之间的夹角均较小，且α将以ω倍的速度减小，从而使每个个体在其自身周围的PS邻域内进行局部搜索，以增加种群的多样性，并获得沿最优PS均匀分布的一组解。在一些实施例中，PE算子可以采用多项式变异的方式来略微增加种群中基因的多样性。这个过程一般可以通过一个相对较小的α来实现，即每个个体在其原始位置有一个较小扰动来产生潜在的子代个体，或者大多数个体从PS的一端移动到另一端以实现更加均匀的分布。进一步地，为了确保每个基因都有一个可观的扰动，多目标优化装置可以对α设置一个下限值αlb。
进一步地，如果ep中是所有元素值都非常小，则产生的子代个体(Xtemp1(i，：))将非常靠近其父代个体(Xg(i，：))，这对进化过程将非常不利。因此当nep中的最大元素小于预设的最小归一化进化路径长度minC时，多目标优化装置可以随机初始化Centerg中的一个元素值，以保证Centerg-1和Centerg之间的差异足够明显，即保证ep中的元素值不会太小，从而防止种群限入局部最优。
如图6所示，在确定进化路径ep的方向及步长α之后，多目标优化装置可以以每个父代个体为出发点，在向量α*ep所指向的矩形范围内分别产生对应的子代个体。
通过采用前后两代种群中的所有个体来计算进化路径ep，本发明一方面可以省去挑选特定个体的麻烦，从而更加简单地计算ep。另一方面，与PSO算子相比，采用所有个体计算的ep更加鲁棒(Robust)、可靠，使得大多数子代个体都可以受益于ep。此外，初步实验表明，大多数个体在进化前期和后期都表现出了比较一致的进化方向，因此整个种群可以共享这一个进化方向，能够有效提高算子的收敛速度，并保证种群的多样性。
进一步地，为了在整个群体中共享一些优秀的个体基因，在产生子代个体之后，多目标优化装置还可以在至少一个优秀的父代个体及各子代个体之间进行基因共享操作(Gene-Sharing Operation)。具体来说，多目标优化装置可以首先获取各父代个体的帕雷托前沿层数或适应度F(i)。若帕雷托前沿层数等于1，则多目标优化装置可以判定其有不被任何解支配的解。如此，多目标优化装置可以将帕 雷托前沿层数等于1或适应度F(i)较好的父代个体，确定为优秀基因的携带者，从而筛选至少一个优秀的父代个体。之后，多目标优化装置可以参考模拟二进制交叉(SBX)算子，以0.5的概率对每个变量执行交叉操作，从而将父代个体中与目标函数相关性较大的优秀基因在子代中共享。通过只在子代个体和优秀的父代个体之间进行基因共享操作，本发明可以进一步提高算法的收敛性和种群的多样性。
进一步地，在一些实施例中，在利用改进的SPEA2算法在约束范围内调整工艺决策变量的取值，以得到目标函数P的操作数据之后，多目标优化装置还可以优选地判断各工艺决策变量的当前操作值xi是否在工艺决策变量的约束范围xi min～xi max内。若当前操作值xi在工艺决策变量的约束范围xi min～xi max内，则多目标优化装置可以将当前操作值xi作为该工艺决策变量的初始值。反之，若当前操作值xi超出工艺决策变量的约束范围xi min～xi max，则多目标优化装置可以执行算法中的修复函数(例如：重新执行SPEA2算法的S1步骤以得初始化种群P0，并将生成的初始种群P0加入到现有种群中)。之后，多目标优化装置可以在工艺决策变量的约束范围xi min～xi max内，重新生成各决策变量的操作数据xi，以作为各工艺决策变量的初始值。
如图1所示，在确定各工艺决策变量的操作数据之后，多目标优化装置可以根据各工艺决策变量的操作数据，确定各优化目标的优化目标值。之后，多目标优化装置可以根据最优的优化目标值解集，将对应的各操作数据确定为对应于该多个优化目标的多个工艺决策变量的指导值。
具体来说，多目标优化装置可以首先将各操作数据输入催化裂化过程的机理模型，以获取各产品及气体的收率数据。之后，多目标优化装置可以将各收率数据输入到各目标函数，以获取各优化目标的优化目标值。
例如，多目标优化装置可以通过Matlab和流程模拟软件Aspen Hysys的接口，将对应的操作数据送入催化裂化机理模型中，以分别计算出各优化目标f1、f2、f3的效益值。之后，多目标优化装置可以如图7及图8所示地根据帕雷托最优解集合以及IGD指标来评价各优化目标值，以确定最优的优化目标值解集。
通过采用改进的SPEA2算法在约束范围内调整各工艺决策变量的取值，本发明可以通过Matlab软件的编程手段和催化裂化机理模型接口程序，实现PE算子与 催化裂化机理模型的数据联动。之后，多目标优化装置可以运行控制程序，通过调用数据、程序和算法，在不同的决策变量下自动计算目标函数P的取值，并通过不断优化各工艺决策变量的取值，使得目标函数值达到更优，从而得到所述目标函数的操作数据。
进一步测，改进的SPEA2算法还会如上所述地根据具体情况，自动改变不同的步长变化策略，从而在约束范围内按照不同的步长变化策略进行不完全轮询计算，不断得到目标函数的更优的操作数据，直到使得目标函数值达到最优。
以图2所示的催化裂化装置为例，其工艺决策变量包括但不限于：进料温度x1、进料压力x2、第一反应区出口温度x3、反应器压力x4、加氢蜡油进料质量流量x5、罐区蜡油进料质量流量x6、罐区渣油进料质量流量x7再生催化剂活性指数x8、氧气指数x9。
对应地，各工艺决策变量的约束范围包括但不限于：
210≤x1≤220
850≤x2≤900
515≤x3≤530
210≤x4≤230
55≤x5≤61
2≤x6≤4
225≤x7≤240
105≤x8≤120
0≤x9≤45
多目标优化装置可以通过Matlab和流程模拟软件Aspen Hysys的接口，将对应的操作数据送入催化裂化机理模型中，以分别计算出各优化目标f1、f2、f3的效益值。请参考表1，表1示出了根据本发明的一些实施例提供的优化结果的对照数据。
表1

如图表1所示，若将各工艺决策变量进行优化前的操作数据代入目标函数P中进行计算，可以计算得出其经济效益为215100.0元/h，二氧化碳排放量为95.2t/h，二氧化硫排放量为0.125t/h。反之，若将各工艺决策变量进行优化后的操作数据代入目标函数P中进行计算，可以计算得出经济效益为226000元/h，二氧化碳排放量为71.7t/h，二氧化硫排放量为0.0818t/h。
可以理解的是，上述结果是从帕雷托最优解集中任意取出，并不代表本发明的最优效果。然而，通过优化前后的数据对比也可以看到全面、显著的优化效果。因此，优化后的操作数据可以为催化裂化装置的优化操作提供指导，在提高经济效益的同时，减小二氧化碳和二氧化硫的排放量。
尽管为使解释简单化将上述方法图示并描述为一系列动作，但是应理解并领会，这些方法不受动作的次序所限，因为根据一个或多个实施例，一些动作可按不同次序发生和/或与来自本文中图示和描述或本文中未图示和描述但本领域技术人员可以理解的其他动作并发地发生。
提供对本公开的先前描述是为使得本领域任何技术人员皆能够制作或使用本公开。对本公开的各种修改对本领域技术人员来说都将是显而易见的，且本文中所定义的普适原理可被应用到其他变体而不会脱离本公开的精神或范围。由此，本公开并非旨在被限定于本文中所描述的示例和设计，而是应被授予与本文中所公开的原理和新颖性特征相一致的最广范围。

Claims (15)

  1. 一种催化裂化过程的多目标优化方法，其特征在于，包括以下步骤：
    确定多个优化目标、对应于所述多个优化目标的多个工艺决策变量，以及各所述工艺决策变量的约束范围；
    根据所述多个优化目标及所述多个工艺决策变量，确定目标函数；
    经由基于路径的再生算子改进子代进化过程的SPEA2算法，在所述约束范围内调整各所述工艺决策变量的取值，以确定所述目标函数关于各所述工艺决策变量的操作数据；
    根据各所述工艺决策变量的操作数据，确定各所述优化目标的优化目标值；以及
    根据最优的优化目标值解集，将对应的各所述操作数据确定为对应于所述多个优化目标的多个工艺决策变量的指导值。
  2. 如权利要求1所述的多目标优化方法，其特征在于，所述经由基于路径的再生算子改进子代进化过程的SPEA2算法，在所述约束范围内调整各所述工艺决策变量的取值，以确定所述目标函数关于各所述工艺决策变量的操作数据的步骤包括：
    S1：初始化迭代次数变量t、种群Pt及储备集其中，初始化的种群P0是由所述多个工艺决策变量在t＝0时的集合构成，初始化的储备集为空集；
    S2：计算所述种群Pt中各个体i的适应度F(i)，其中，每个所述个体i对应一个所述工艺决策变量；
    S3：将所述种群Pt及所述储备集中的所有非支配解集复制到储备集并对所述储备集进行环境选择；
    S4：若迭代次数t未达到预设的迭代次数上限T，则对经过所述环境选择的储备集进行锦标赛选择，再将经过所述锦标赛选择的数据集放入交配池；
    S5：经由所述基于路径的再生算子，对所述交配池中的所述数据集进行计算，将计算结果存入所述储备集递增迭代次数，并返回步骤S2；以及
    S6：若迭代次数变量t达到所述迭代次数上限T，则输出所述储备集中非支配解所代表的工艺决策变量集A。
  3. 如权利要求2所述的多目标优化方法，其特征在于，所述计算所述种群Pt中各个体i的适应度F(i)的步骤包括：
    根据个体i支配的个体数S(i)，确定个体i的原始适应值R(i)，其中，所述原始适应值R(i)表示个体i，以及支配个体i的每个个体j所支配的所有个体数之和；
    计算每个个体i到所述种群Pt及所述储备集中所有个体的距离，并按照递增排序；
    选择第k个个体作为并计算对应的距离值D(i)，其中，所述N为所述种群Pt的大小，所述为所述储备集的大小；以及
    根据所述原始适应值R(i)及所述距离值D(i)，确定所述个体i的适应度F(i)。
  4. 如权利要求3所述的多目标优化方法，其特征在于，所述对所述储备集进行环境选择的步骤包括：
    将所述储备集中的个体数量与储备集大小进行比较；
    若则根据适应度F(i)的排序，从所述种群Pt及所述储备集的所有个体i中选择前个个体构成支配解集，并添加到所述储备集以及
    若则进行剪枝操作，在每次迭代中淘汰与相邻个体具有最小距离的个体i。
  5. 如权利要求4所述的多目标优化方法，其特征在于，所述在每次迭代中淘汰与相邻个体具有最小距离的个体i的步骤包括：
    在每次迭代中淘汰与相邻个体具有最小距离的唯一个体i。
  6. 如权利要求3所述的多目标优化方法，其特征在于，所述经由所述基于路径的再生算子，对所述交配池中的所述数据集进行计算的步骤包括：
    确定所述储备集的子代种群个体的中心点Centerg，并确定所述种群Pt及所述储备集的父代种群个体的中心点Centerg-1；
    根据前后两代种群个体中心点的差，定义进化路径ep的方向；
    根据子代个体的目标生存率及实际生产率psucc，确定所述进化路径ep的前进步长α；以及
    以每个父代个体为出发点，在向量α*ep所指向的矩形范围内分别产生对应的子代个体。
  7. 如权利要求6所述的多目标优化方法，其特征在于，所述经由所述基于路径的再生算子，对所述交配池中的各所述个体i进行重组和变异的步骤还包括：
    在产生所述子代个体之后，在至少一个优秀的父代个体及各所述子代个体之间进行基因共享操作。
  8. 如权利要求7所述的多目标优化方法，其特征在于，在进行所述基因共享操作之前，所述多目标优化方法还包括以下步骤：
    获取各所述父代个体的帕雷托前沿层数或适应度F(i)；以及
    根据所述帕雷托前沿层数或所述适应度F(i)，筛选所述至少一个优秀的父代个体。
  9. 如权利要求2所述的多目标优化方法，其特征在于，在根据最优的优化目标值解集，将对应的各所述操作数据确定为对应于所述多个优化目标的多个工艺决策变量的指导值之前，所述多目标优化方法还包括以下步骤：
    根据帕雷托最优解集合以及IGD指标来评价各所述优化目标值，以确定所述最优的优化目标值解集。
  10. 如权利要求1所述的多目标优化方法，其特征在于，所述多个优化目标中至少包括最小化二氧化碳排放量。
  11. 如权利要求10所述的多目标优化方法，其特征在于，所述多个优化目标还包括：最大化经济效益、最小化二氧化硫排放量、最小化废气总排放量、最小化废液总排放量、最大化主要产品产量、最大化主要产品收率、最小化投入成本中的至少一者。
  12. 如权利要求10所述的多目标优化方法，其特征在于，所述多个工艺决策变量至少包括：物料流量、物料状态、物料性质、主分馏塔状态、主分馏塔操作、吸收塔状态、吸收塔操作、再吸收塔状态、再吸收塔操作、解吸塔状态、解吸塔操作、稳定塔状态、稳定塔操作中的至少一者。
  13. 如权利要求10所述的多目标优化方法，其特征在于，所述多个优化目标包括最大化经济效益、所述最小化二氧化碳的排放量及最小化二氧化硫的排放量，所述目标函数如下：
    min imizeF(x)＝(1/f1，f2，f3)，

    f2＝min CCO2，
    f3＝min CSO2，
    其中，M指示催化裂化过程的经济效益；Pi指示产品i的价格；Yi指示产品i的流量；i指示产品，包括酸性气、干气、乙烯、液化气、丙烯、汽油、柴油、循环油、油浆、焦炭中的至少一种；pj指示原料j的价格；yj指示原料j的流量；j指示原料，包括为罐区渣油、罐区蜡油、精制蜡油；Pf指示单位加工量的固定成本，F指示加工负荷；pc指示新鲜催化剂的价格，fc指示新鲜催化剂补充量；CCO2指示二氧化碳排放量；CSO2表示二氧化硫排放量。
  14. 一种催化裂化过程的多目标优化装置，其特征在于，包括：
    存储器；以及
    处理器，所述处理器连接所述存储器，并被配置用于实施如权利要求1～13中任一项所述的催化裂化过程的多目标优化方法。
  15. 一种计算机可读存储介质，其上存储有计算机指令，其特征在于，所述计算机指令被处理器执行时，实施如权利要求1～13中任一项所述的催化裂化过程的多目标优化方法。
PCT/CN2023/071999 2022-01-25 2023-01-13 一种催化裂化过程的多目标优化方法及装置 WO2023143106A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210089249.9A CN114492029A (zh) 2022-01-25 2022-01-25 一种催化裂化过程的多目标优化方法及装置
CN202210089249.9 2022-01-25

Publications (1)

Publication Number Publication Date
WO2023143106A1 true WO2023143106A1 (zh) 2023-08-03

Family

ID=81474112

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/071999 WO2023143106A1 (zh) 2022-01-25 2023-01-13 一种催化裂化过程的多目标优化方法及装置

Country Status (2)

Country Link
CN (1) CN114492029A (zh)
WO (1) WO2023143106A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114492029A (zh) * 2022-01-25 2022-05-13 华东理工大学 一种催化裂化过程的多目标优化方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040033617A1 (en) * 2002-08-13 2004-02-19 Sonbul Yaser R. Topological near infrared analysis modeling of petroleum refinery products
EP3156124A1 (en) * 2015-10-16 2017-04-19 Reliance Industries Limited A catalyst, a process for preparation of the catalyst and application thereof
CN106971049A (zh) * 2017-04-17 2017-07-21 武汉理工大学 一种催化裂化分离系统的新型多目标优化方法
CN107609328A (zh) * 2017-08-30 2018-01-19 武汉理工大学 一种催化裂化装置模型的多目标优化方法
CN114492029A (zh) * 2022-01-25 2022-05-13 华东理工大学 一种催化裂化过程的多目标优化方法及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040033617A1 (en) * 2002-08-13 2004-02-19 Sonbul Yaser R. Topological near infrared analysis modeling of petroleum refinery products
EP3156124A1 (en) * 2015-10-16 2017-04-19 Reliance Industries Limited A catalyst, a process for preparation of the catalyst and application thereof
CN106971049A (zh) * 2017-04-17 2017-07-21 武汉理工大学 一种催化裂化分离系统的新型多目标优化方法
CN107609328A (zh) * 2017-08-30 2018-01-19 武汉理工大学 一种催化裂化装置模型的多目标优化方法
CN114492029A (zh) * 2022-01-25 2022-05-13 华东理工大学 一种催化裂化过程的多目标优化方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
"Master's Thesis", 1 May 2016, WUHAN UNIVERSITY OF TECHNOLOGY, CN, article ZHOU, JING: "Multi-Objective Optimization of Hydrocracking Fractionation Based on SPEA2 Algorithm", pages: 1 - 66, XP009547783 *

Also Published As

Publication number Publication date
CN114492029A (zh) 2022-05-13

Similar Documents

Publication Publication Date Title
WO2023143106A1 (zh) 一种催化裂化过程的多目标优化方法及装置
CN102520705B (zh) 一种炼化生产过程优化分析方法及系统
CN106971049A (zh) 一种催化裂化分离系统的新型多目标优化方法
CN107203687B (zh) 吸收塔脱硫过程多目标协同智能优化控制方法
CN112489733B (zh) 基于粒子群算法和神经网络的辛烷值损失预测方法
CN111461406B (zh) 一种基于混合建模的乙烯生产原料优化方法
CN104765347B (zh) 一种渣油延迟焦化过程中收率实时预测方法
CN112686296B (zh) 一种基于粒子群优化随机森林参数的辛烷损失值预测方法
Ullah et al. Optimization based comparative study of machine learning methods for the prediction of bio-oil produced from microalgae via pyrolysis
CN102252343A (zh) 一种多孔介质燃烧器燃烧优化的方法
Bi et al. Knowledge expression, numerical modeling and optimization application of ethylene thermal cracking: From the perspective of intelligent manufacturing
Han et al. A fuzzy constraint handling technique for decomposition-based constrained multi-and many-objective optimization
Cui et al. Production optimization and energy saving of complex chemical processes using novel competing evolutionary membrane algorithm: Emphasis on ethylene cracking
Tian et al. Automated selection of evolutionary multi-objective optimization algorithms
Zhang et al. Prediction of research octane number loss and sulfur content in gasoline refining using machine learning
Xu et al. A heuristic approach to design a cost-effective and low-CO2 emission synthesis in a heat exchanger network with crude oil distillation units
CN111914382A (zh) 一种基于代理模型的常减压装置的约束进化优化方法
Khaldi et al. Artificial intelligence perspectives: A systematic literature review on modeling, control, and optimization of fluid catalytic cracking
Leite et al. Multi-objective optimization of adiabatic styrene reactors using Generalized Differential Evolution 3 (GDE3)
Jung et al. Neural network models for atmospheric residue desulfurization using real plant data with novel training strategies
Chen et al. An improved differential evolution algorithm for operating optimization of a distillation unit
Wan et al. Multi-objective optimization strategy for industrial catalytic cracking units: Kinetic model and enhanced SPEA-2 algorithm with economic, CO2, and SO2 emission considerations
CN117313554B (zh) 炼焦生产多工段联合多目标优化方法、系统、设备及介质
CN108073784A (zh) 一种基于过程图的乙烯原料优化调度方案的求解方法
Bi et al. An interactive opposition matrix and regulatory gene fusion strategy for evolutionary multitasking

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23745989

Country of ref document: EP

Kind code of ref document: A1