CN110728031B - Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling - Google Patents

Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling Download PDF

Info

Publication number
CN110728031B
CN110728031B CN201910890104.7A CN201910890104A CN110728031B CN 110728031 B CN110728031 B CN 110728031B CN 201910890104 A CN201910890104 A CN 201910890104A CN 110728031 B CN110728031 B CN 110728031B
Authority
CN
China
Prior art keywords
model
temperature
phase change
algorithm
optimization
Prior art date
Legal status (The legal status 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 status listed.)
Active
Application number
CN201910890104.7A
Other languages
Chinese (zh)
Other versions
CN110728031A (en
Inventor
王俊
朱群雄
贺彦林
王钰
叶玮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Chemical Technology
Original Assignee
Beijing University of Chemical Technology
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 Beijing University of Chemical Technology filed Critical Beijing University of Chemical Technology
Priority to CN201910890104.7A priority Critical patent/CN110728031B/en
Publication of CN110728031A publication Critical patent/CN110728031A/en
Application granted granted Critical
Publication of CN110728031B publication Critical patent/CN110728031B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P20/00—Technologies relating to chemical industry
    • Y02P20/10—Process efficiency

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Physiology (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses a multi-objective optimization method for balancing complex petrochemical process productivity based on ANN modeling, which is characterized in that an MINLP model is established on the basis of ANNs models, the MINLP model comprises a logical variable compressor position and a heat exchanger position, an optimal position is determined through a GA (genetic algorithm), and then multi-objective integrated optimization is carried out at the optimal position through an NSGA-II (non-subsampled genetic algorithm-II) algorithm to obtain a pareto front edge and a key decision variable front edge. The invention can greatly reduce the operation complexity of the petrochemical industry and has important guiding significance for energy conservation and emission reduction of factories.

Description

Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling
Technical Field
The invention belongs to the technical field of process simulation, and particularly relates to a multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling.
Technical Field
The ethylene industry is the dominant position in the petrochemical industry, about three quarters of the petrochemicals being related to ethylene products. In 2018, the yield of the Chinese ethylene is about 2330 ten thousand tons/year, and compared with the last year, the yield of the Chinese ethylene is increased by 9.6 percent and accounts for 13.8 percent of the global ethylene yield. However, the energy consumption is as high as more than 50% of the total cost, which is increased by 3.5% compared with the last year. Therefore, balancing energy consumption and production efficiency in the petrochemical industry plays a crucial role in improving energy efficiency.
At present, with the high-speed development of the computer level, modeling simulation for the petrochemical industry is mature. A number of rigorous models have been proposed. However, given the complexity of industrial processes, some assumptions are inevitable in these models. Therefore, the current optimization model can not accurately and truly reflect the actual industrial situation. On the other hand, since a strict model has a large number of differential operations, the calculation amount is large and convergence is difficult.
Disclosure of Invention
The invention provides a complex petrochemical process statistical modeling optimization method based on Artificial Neural Networks (ANNs), aiming at the problems that a complex petrochemical industry model, particularly an ethylene production industry model, is too complex, and important attributes of the process cannot be completely understood, multi-system multi-objective optimization is difficult to perform, and the like.
The technical scheme is as follows:
a multi-objective optimization method for balancing complex petrochemical process productivity based on ANN modeling is characterized in that an MINLP model is established on the basis of an ANNs model, the MINLP model comprises a logical variable compressor position and a heat exchanger position, an optimal position is determined through a GA algorithm, then multi-objective integrated optimization is carried out at the optimal position through an NSGA-II algorithm, and a pareto front edge and a key decision variable front edge are obtained.
Preferably, the method for establishing the ANNs model of the petrochemical process in the method comprises the following steps:
1): establishing a strict model;
2): obtaining uniformly distributed random input samples by utilizing a Latin hypercube sampling method:
xm,i＝xm,l+(xm,u-xm,l)·PLHS
wherein xm,u，xm,lRespectively, the upper and lower limits of the variable, PLHSIs the probability distribution of Latin hypercube sampling;
3): compiling an automatic step, taking the samples obtained in the second step as the input of the strict model in sequence, automatically executing the strict model and collecting corresponding output;
4): classifying ANNs models into two categories according to whether the output is converged or not, wherein the feasibility model ANN1 is used for judging whether the input data is feasible or not; the rectification model ANN2 was used to calculate all the outputs related to the basic optimization;
5): and (5) training and verifying the ANN model.
Preferably, in the method, the method for establishing the heat capacity model considering the influence of temperature and phase change comprises the following steps:
I) heat capacity of the stream, which is not affected by temperature and whose value is only related to the composition of the stream, during the heat exchange, the following function is established to calculate the enthalpy of the stream:
Figure BDA0002208434280000021
Figure BDA0002208434280000022
wherein subscript i represents the number of streams; a is0,a1,a2… are coefficients of linear terms; b0,b1,b2… are coefficients of the mixture term; c is a constant term; m is mass fraction;
II) a stream with partial phase change, the heat capacity of the stream will change with temperature change in the heat exchange temperature range, in order to reflect the effect of the phase change on enthalpy, on the basis of group I), a gas phase fraction is introduced, the degree of phase change at different temperatures will be different, and therefore the value of the gas phase fraction will also be different, a model is established to predict the gas fraction at different temperatures:
Figure BDA0002208434280000023
VFRAC＝A0T5+A1T4+A2T3+A3T2+A4T+A5
Figure BDA0002208434280000024
wherein the parameters a, b, c, d and e are coefficients of a linear term, a mixed term, a quadratic term, a cubic term and a constant term, respectively; vFRACThe gas phase fraction is shown, A is a regression coefficient of temperature and gas phase fraction, and the proved fifth-order polynomial regression effect is the best;
III) a stream with a complete phase change, exchanging only latent heat, with a temperature difference of almost zero, considering that the heat capacity does not vary with temperature during the heat exchange, and the enthalpy can be assumed as a linear relationship with temperature, calculated by the following formula:
Figure BDA0002208434280000031
wherein H3,OUT、H3,IN、T3,OUTAnd T3,INRespectively representing the inlet and outlet enthalpy values and the inlet and outlet temperatures of the heat exchanger under the condition III; the coefficients of the above equations can all be obtained by using least squares regression, with regression data from a validated distillation model ANN2 or rigorous model.
Preferably, in the method, the optimization framework based on the hybrid algorithm is solved:
the method comprises the following steps of using an ANN2 distillation model for calculating yields of ethylene and propylene and integrally optimizing other variables, calculating the minimum energy consumption by using a heat recovery model, firstly initializing, generating samples containing 1000 groups of operating conditions by using a feasible neural network, and calculating the optimal combination of the 1000 groups of samples by using an NSGA-II algorithm, wherein the sample with the pareto front value of 1; profit P is then calculated from diene production and energy consumption and maximized using GA algorithm.
The technical effects are as follows:
the invention provides a Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling, wherein a Multi-objective mixed integer nonlinear programming model (MOMINLP) is obtained on the basis of the modeling, and the model is solved through a hybrid algorithm. Pareto Frontier (PF) and corresponding Key Decision Variables (KDV) are derived that simultaneously maximize diene production and minimize energy consumption. The method can greatly reduce the operation complexity of the petrochemical industry and has important guiding significance for energy conservation and emission reduction of factories.
Drawings
Fig. 1 is a diagram of an artificial neural network.
Fig. 2 is a graph of the training results of ANN 1.
FIG. 3 is a graph of the results of ANN2 training
FIG. 4 is an optimization framework.
FIG. 5 is a flow diagram of the cryogenic separation of ethylene.
FIG. 6 is a full process modeling and optimization sketch.
In FIG. 7, (a) is a multi-objective optimization optimal leading edge; (b-k) are leading edges of manipulated variables.
In FIG. 8, (a) is a multi-objective optimization optimal leading edge; (b, c) is the leading edge of the key operating variable.
Detailed Description
The data-driven ANNs method is applied to modeling optimization of the large petrochemical industry, so that the problems can be well solved. Robustness and accuracy are essential for an integrated model for multiple system multiple targets in a petrochemical process. And the model built needs to provide enough information to represent the synergy between the entire separation process and other systems. When selected regression data and model parameters are appropriate, ANNs models can predict results that are the same as or even more accurate than rigorous distillation models. In addition, because the ANNs model is built based on engineering data, strict mechanisms and complex process limitations do not need to be considered, and therefore the ANNs model is simpler and more robust. Based on these advantages, the distillation model of the artificial neural network is more suitable as an optimization model to simulate complex petrochemical processes than a strictly simplified model.
The ANNs principle of operation stems from biological neurons that correlate input data with output data through a series of nodes. Each layer in the ans is composed of one or more neurons, each neuron having a corresponding transfer function. By adjusting the number of neural layers and meridians, the ANN can represent any linear or non-linear function with any desired precision.
FIG. 1 shows a block diagram of a typical single hidden layer feedforward neural network. The artificial neural network modeling process can be divided into the following steps:
1. data acquisition and processing, 2, model structure determination, and 3, model training and verification.
Equations (1) and (2) are algebraic expressions of the network graph, respectively.
a＝f(W1x+b1) (1)
y＝g(W2x+b2) (2)
Wherein: x represents the input, a represents the hidden layer output, and y represents the prediction output; w represents a weight between layers, b represents a threshold; superscripts 1 and 2 represent the hidden layer and the output layer, respectively; the transfer function f is a Sigmoid function and g is a linear function.
In order to make those skilled in the art better understand the technical solution of the present invention, the following describes in detail a multi-objective optimization method for balancing the production capacity of a complex petrochemical process based on ANN modeling, which is provided by the present invention, with reference to the following embodiments. The following examples are intended to illustrate the invention only and are not intended to limit the scope of the invention.
In order to better balance the energy consumption and yield of the ethylene industry in the petrochemical process, the invention provides a modeling, analyzing and optimizing method aiming at the low-temperature separation process of complex ethylene based on an artificial neural network. On the basis of ANNs models, an MINLP model is established, a Genetic Algorithm (GA) and an NSGA-2 algorithm are combined to optimize the model, and the obtained pareto front edge and the key decision variable front edge can greatly reduce the operation complexity of the petrochemical industry and have important guiding significance for energy conservation and emission reduction of factories.
1. Petrochemical process ANNs model establishing method
The first step is as follows: and (5) establishing a strict model.
The second step is that: and (2) obtaining uniformly distributed random input samples by utilizing a Latin Hypercube Sampling (LHS) method. The principle of the method is given in equation (3).
xm,i＝xm,l+(xm,u-xm,l)·PLHS (3)
Wherein xm,u，xm,lRespectively, the upper and lower limits of the variable, PLHSIs the probability distribution of latin hypercube sampling.
The third step: and compiling an automatic step, taking the samples obtained in the second step as the input of the strict model in sequence, automatically executing the strict model and collecting corresponding output.
The fourth step: ANNs models are classified into two categories depending on whether the output converges or not. The feasibility model ANN1 is used for judging whether the input data is feasible or not; the rectification model ANN2 was used to calculate all the outputs related to the basic optimization.
The fifth step: and (5) training and verifying the ANN model.
The training results for ANN1 and ANN2 are given in fig. 2 and 3, respectively. The accuracy of the two models is high, and the models can be used for later optimization.
2. Heat capacity model establishing method considering temperature and phase change influence
Thermal integration design methods typically implement simplifying assumptions, such as assuming that the heat capacity of all process streams is constant. These assumptions are only valid when there is no phase change, the temperature difference between the stream inlet and outlet is small, or the temperature has little effect on the heat capacity. However, in the complex petrochemical industry, the temperature range is large, the heat capacity of most streams is greatly affected by temperature, and other properties of the stream (e.g., density, viscosity, etc.) also vary with temperature. Therefore, an accurate model is required to calculate the heat capacity of the flow.
Process streams participating in thermal integration can generally be divided into three cases: I) no phase-changed stream during heat exchange, II) a stream with partial phase change, III) a stream with complete phase change.
For group (I), the heat capacity is not affected by the temperature when there is no phase change during the heat exchange, and its value is only related to the composition of the stream. To calculate the heat capacity of the flow, the following function is established to calculate the enthalpy of the flow:
Figure BDA0002208434280000061
Figure BDA0002208434280000062
where the subscript i represents the stream number. a is0,a1,a2… are coefficients of linear terms. b0,b1,b2… are coefficients of the mixture term. c is a constant term. M is a mass fraction.
For group (II), when a partial phase change occurs in the stream, the heat capacity of the stream will vary with temperature over the heat exchange temperature range. To reflect the effect of phase change on enthalpy, a gas phase fraction was introduced based on group (I). Since the degree of phase change will be different at different temperatures, the value of the gas phase fraction will also be different. Therefore, a model must also be built to predict the gas fraction at different temperatures.
Figure BDA0002208434280000063
VFRAC＝A0T5+A1T4+A2T3+A3T2+A4T+A5 (7)
Figure BDA0002208434280000064
The parameters a, b, c, d and e in the formula (6) are coefficients of a linear term, a mixed term, a quadratic term, a cubic term and a constant term, respectively. VFRACIs the gas phase fraction, the value of which can be calculated by equation (7), a is the regression coefficient of temperature and gas phase fraction, and the best regression effect of the fifth order polynomial is verified.
For group (III), only latent heat is exchanged in this case, so the temperature difference is almost zero. Therefore, it is considered that the heat capacity does not change with temperature during the heat exchange. And the enthalpy can be assumed to be a linear relationship of temperature, which can be calculated by equation (9):
Figure BDA0002208434280000065
wherein H3,OUT、H3,IN、T3,OUTAnd T3,INRespectively representing the inlet and outlet enthalpy values and the inlet and outlet temperatures of the heat exchanger under the condition III. The coefficients of equations (4), (6), and (7) can all be obtained by regression using the least square method. These regression data were from a validated distillation model ANN2 or rigorous model.
3. Optimization framework based on hybrid algorithm
Among the established models for ANNs, the ANN2 distillation model was used to calculate yields of ethylene and propylene and to integrate optimization of other variables. The heat recovery model calculates the minimum energy consumption. Since the model built contains the logical variables compressor position and heat exchanger position, the model is a typical MINLP model, which is very difficult to solve. The invention combines the GA algorithm and the NSGA-2 algorithm to carry out optimization solution on the model. Firstly, determining an optimal position through GA, and then performing multi-objective integrated optimization at the optimal position through NSGA-II algorithm. FIG. 4 shows the detailed optimization steps, j and g represent the maximum number of iterations for GA and NSGA-II, respectively. i represents the position of the compressor and the heat exchanger. For each iteration in the genetic algorithm, i is updated once. The Flag is used to determine whether the genetic algorithm is finished. And when the position i of the compressor reaches the optimal position or the position j reaches the maximum iteration number, ending the genetic algorithm, wherein the Flag is 1, otherwise, continuously optimizing the position of the compressor, and the Flag is 0. When Flag is 1, then start with the optimal position ioptAnd performing multi-objective optimization. The optimization process is divided into four parts, first, initialization, and the feasible neural network is used to generate a sample containing 1000 sets of operating conditions. The second part uses a multi-objective algorithm (NSGA-II) to find the optimal combination of 1000 sets of samples (samples with a pareto front value of 1). Followed by utilization of diene yieldAnd calculating profit P by energy consumption, maximizing profit P by using GA, and returning to the optimal structure i.
Example 1
The ANNs model was established for the ethylene separation process for the ethylene plant and all modeling data was derived from the rigorous model Aspen Plus (V8.4). Figure 5 shows a flow diagram of an ethylene cryogenic separation system. Based on the ANNs model, the ethylene yield and the energy consumption are optimized in a multi-objective mode through a mixing algorithm. Equations (10), (11) are the optimized objective functions:
Figure BDA0002208434280000071
Figure BDA0002208434280000072
wherein MIN J1Denotes minimum energy consumption, MAX J2Indicating maximum ethylene production. CuAnd CWRespectively the unit price of the shaft work of the utility and the compressor. Specific data are given in table 1.
TABLE 1 Utility data and unit price
Figure BDA0002208434280000073
Figure BDA0002208434280000081
There are 16 operating variables that need to be optimized, with compressor and heat exchanger positions being logical variables and the other 15 being continuous variables. Table 2 shows the operating ranges for 15 continuous variables. Figure 6 gives a structural sketch of the entire modeling and optimization process.
TABLE 2 operating variables and their optimization Range
Figure BDA0002208434280000082
The optimization is carried out according to the optimization method provided by the invention, and a good optimization result is obtained. Table 3 gives the optimum operating conditions for each flash tank. FIG. 7 shows the optimal leading edge for multi-objective optimization.
TABLE 3 optimal operating conditions for the flash tank
Figure BDA0002208434280000091
It can be readily seen from fig. 7 that only the bottom draw of the DE column and the reflux ratio of the DE column were varied with ethylene production, while the other operating variables were constant throughout the optimization process. Thus, it is reasonable to believe that the ethylene production is only affected by the bottom draw of the DE column and the reflux ratio of the DE column, which are key operating variables. To verify this conclusion, a control experiment can be performed on the basis of fig. 7. In the control group, only the reflux ratio of the deethanizer and the bottom yield are taken as operation variables, and other variables are taken as fixed values. Figure 8 shows the results of the optimization of the control group. As can be seen from fig. 8, the optimization result when only two key operation variables are considered is almost completely consistent with the optimization result of fig. 7, but the operation difficulty is greatly reduced. By adjusting the values of the two key decision variables, the plant's ethylene production can be increased by 73164 kilograms per year, corresponding to a revenue increase of $ 95.1 ten thousand per year.
The present invention is not limited to the above-described examples, and various changes can be made without departing from the spirit and scope of the present invention within the knowledge of those skilled in the art.

Claims (1)

1. A multi-objective optimization method for balancing complex petrochemical process capacity based on ANN modeling is characterized in that an MINLP model is established on the basis of an ANNs model, the MINLP model comprises a logical variable compressor position and a heat exchanger position, an optimal position is determined through a GA algorithm, then multi-objective integrated optimization is carried out at the optimal position through an NSGA-II algorithm, and a pareto front edge and a key decision variable front edge are obtained;
the method for establishing the ANNs model in the petrochemical process comprises the following steps:
1): establishing a strict model;
2): obtaining uniformly distributed random input samples by utilizing a Latin hypercube sampling method:
xm,i＝xm,l+(xm,u-xm,l)·PLHS
wherein xm,u，xm,lRespectively, the upper and lower limits of the variable, PLHSIs the probability distribution of Latin hypercube sampling;
3): compiling an automatic step, taking the samples obtained in the second step as the input of the strict model in sequence, automatically executing the strict model and collecting corresponding output;
4): classifying ANNs models into two categories according to whether the output is converged or not, wherein the feasibility model ANN1 is used for judging whether the input data is feasible or not; the rectification model ANN2 was used to calculate all the outputs related to the basic optimization;
5): training and verifying an ANN model;
the method for establishing the heat capacity model considering the temperature and the phase change influence comprises the following steps:
I) heat capacity of the stream, which is not affected by temperature and whose value is only related to the composition of the stream, during the heat exchange, the following function is established to calculate the enthalpy of the stream:
Figure FDA0003052647280000011
Figure FDA0003052647280000012
wherein subscript i represents the number of streams; a is0,a1,a2… are coefficients of linear terms; b0,b1,b2… are coefficients of the mixture term; c. CIs a constant term; m is mass fraction; h1Is not subjected to phase change at temperature T1Lower corresponding enthalpy value, CP1The specific pressure heat capacity is the specific pressure heat capacity when the phase change does not occur, and is a constant when the phase change does not occur;
II) a stream with partial phase change, the heat capacity of the stream will change with temperature change in the heat exchange temperature range, in order to reflect the effect of the phase change on enthalpy, on the basis of group I), a gas phase fraction is introduced, the degree of phase change at different temperatures will be different, and therefore the value of the gas phase fraction will also be different, a model is established to predict the gas fraction at different temperatures:
Figure FDA0003052647280000021
VFRAC＝A0T5+A1T4+A2T3+A3T2+A4T+A5
Figure FDA0003052647280000022
wherein the parameters a, b, c, d and e are coefficients of a linear term, a mixed term, a quadratic term, a cubic term and a constant term respectively; vFRACThe method is characterized in that the gas phase fraction is under the temperature T when partial phase change occurs, A is a regression coefficient of the temperature and the gas phase fraction, and the fifth-order polynomial regression effect is proved to be the best; h2Is the temperature T at which the partial phase change occurs2Corresponding enthalpy value, which is no longer a linear function of temperature, CP2Is the temperature T at which the partial phase change occurs2Specific pressure heat capacity at a value of H2For T2Partial differentiation is obtained;
III) a stream with a complete phase change, exchanging only latent heat, with a temperature difference of almost zero, considering that the heat capacity does not vary with temperature during the heat exchange, and the enthalpy can be assumed as a linear relationship with temperature, calculated by the following formula:
Figure FDA0003052647280000023
wherein H3,OUT、H3,IN、T3,OUTAnd T3,INRespectively representing the inlet and outlet enthalpy values and inlet and outlet temperatures, CP, of the heat exchanger under the condition III3 *The specific pressure heat capacity is assumed on the basis of the previous formula, coefficients of the formula are obtained by using least square regression, and regression data are from a verified distillation model ANN2 or a strict model;
solving an optimization framework based on a hybrid algorithm: the method comprises the following steps of using an ANN2 distillation model for calculating yields of ethylene and propylene and integrally optimizing other variables, calculating the minimum energy consumption by using a heat recovery model, firstly initializing, generating samples containing 1000 groups of operating conditions by using a feasible neural network, and calculating the optimal combination of the 1000 groups of samples by using an NSGA-II algorithm, wherein the sample with the pareto front value of 1; profit P is then calculated from diene production and energy consumption and maximized using GA algorithm.
CN201910890104.7A 2019-09-20 2019-09-20 Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling Active CN110728031B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910890104.7A CN110728031B (en) 2019-09-20 2019-09-20 Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910890104.7A CN110728031B (en) 2019-09-20 2019-09-20 Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling

Publications (2)

Publication Number Publication Date
CN110728031A CN110728031A (en) 2020-01-24
CN110728031B true CN110728031B (en) 2021-06-22

Family

ID=69218178

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910890104.7A Active CN110728031B (en) 2019-09-20 2019-09-20 Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling

Country Status (1)

Country Link
CN (1) CN110728031B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220317644A1 (en) * 2020-02-28 2022-10-06 Boe Technology Group Co., Ltd. Production programming system and method based on nonlinear program model, and computer-readable storage medium
CN114547823B (en) * 2022-03-16 2024-06-04 厦门大学 Heat exchange network optimization design method comprising phase-change flow strands

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1308055A1 (en) * 2000-06-30 2003-05-07 BRITISH TELECOMMUNICATIONS public limited company A memetic method for multiobjective optimisation
CN102339040A (en) * 2010-07-15 2012-02-01 清华大学 Optimization and control method of rectification column
CN103087751A (en) * 2011-10-28 2013-05-08 中国石油化工股份有限公司 Industrial cracking furnace value maximization model construction method
CN103488085A (en) * 2013-09-22 2014-01-01 上海交通大学 Multi-objective optimization control method of methyl alcohol four-tower rectification system
CN107807518A (en) * 2017-09-05 2018-03-16 昆明理工大学 A kind of Multiobjective Optimal Operation method of Chemical Manufacture raw material still-process
CN109242088A (en) * 2018-07-23 2019-01-18 大冶特殊钢股份有限公司 Heat treatment method and device based on GA-ANN artificial nerve network model
CN109919485A (en) * 2019-03-06 2019-06-21 桂林理工大学 A kind of Multipurpose Optimal Method of the Water Resources Allocation based on NSGA-II

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1308055A1 (en) * 2000-06-30 2003-05-07 BRITISH TELECOMMUNICATIONS public limited company A memetic method for multiobjective optimisation
CN102339040A (en) * 2010-07-15 2012-02-01 清华大学 Optimization and control method of rectification column
CN103087751A (en) * 2011-10-28 2013-05-08 中国石油化工股份有限公司 Industrial cracking furnace value maximization model construction method
CN103488085A (en) * 2013-09-22 2014-01-01 上海交通大学 Multi-objective optimization control method of methyl alcohol four-tower rectification system
CN107807518A (en) * 2017-09-05 2018-03-16 昆明理工大学 A kind of Multiobjective Optimal Operation method of Chemical Manufacture raw material still-process
CN109242088A (en) * 2018-07-23 2019-01-18 大冶特殊钢股份有限公司 Heat treatment method and device based on GA-ANN artificial nerve network model
CN109919485A (en) * 2019-03-06 2019-06-21 桂林理工大学 A kind of Multipurpose Optimal Method of the Water Resources Allocation based on NSGA-II

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
MINLP模型及其在石化行业中的应用——Ⅱ．反应网络与反应精馏***MINLP模型的建立;解新安 等;《炼油技术与工程》;20031130;第33卷(第11期);第50-54页 *
Multiobjective Stochastic Optimization of Dividing-wall Distillation Columns Using a Surrogate Model Based on Neural Networks;C.Gutiérrez-Antonio等;《Chemical & Biochemical Engineering Quarterly》;20151231;第29卷(第4期);第491页第2段，第493页第2-3段，第498页第2段 *
人工神经网络在化工过程中的应用进展;宋泓阳 等;《化工进展》;20161231;第35卷(第12期);第3755-3762页 *

Also Published As

Publication number Publication date
CN110728031A (en) 2020-01-24

Similar Documents

Publication Publication Date Title
CN102693451B (en) Method for predicting ammonia process flue gas desulphurization efficiency based on multiple parameters
Yu et al. A combined genetic algorithm/simulated annealing algorithm for large scale system energy integration
CN109084415A (en) Central air-conditioning operating parameter optimization method based on artificial neural network and genetic algorithms
CN106845012B (en) A kind of blast furnace gas system model subordinating degree function based on multiple target Density Clustering determines method
CN108846526A (en) A kind of CO2 emissions prediction technique
CN110728031B (en) Multi-objective optimization method for balancing complex petrochemical process production energy based on ANN modeling
CN109872012A (en) Based on the determination method for thermal power plant's operation multiple-objection optimization that operating condition divides
CN110046408A (en) Bounded-but-unknown uncertainty optimization method based on BP neural network differential and interval analysis
CN109670625A (en) NOx emission concentration prediction method based on Unscented kalman filtering least square method supporting vector machine
CN114239400A (en) Multi-working-condition process self-adaptive soft measurement modeling method based on local double-weighted probability hidden variable regression model
CN112836429A (en) Multi-objective optimization coal blending method based on coal quality prediction
CN105868867A (en) Method and system for optimized operation of heating boiler cluster
CN111461406A (en) Ethylene production raw material optimization method based on mixed modeling
CN110207094A (en) IQGA-SVR boiler heating surface fouling characteristics discrimination method based on principal component analysis
CN111931436A (en) Burner nozzle air quantity prediction method based on numerical simulation and neural network
CN111832839A (en) Energy consumption prediction method based on sufficient incremental learning
CN105808945B (en) A kind of hybrid intelligent boiler efficiency burning optimization method
CN109919401B (en) Multi-dimensional energy efficiency analysis method of multi-energy complementary system
CN108446452B (en) A kind of mixed-flow pump impeller Robust Optimal Design
Mengistu et al. Aerodynamic shape optimization of turbine blades using a design-parameter-based shape representation
Liu et al. Mixed-flow assembly line balancing with uncertain assembly times in remanufacturing
CN109101683B (en) Model updating method for pyrolysis kettle of coal quality-based utilization and clean pretreatment system
CN102682336A (en) Method for optimizing design on number of regulating stage nozzles of steam turbine based on improved genetic algorithm
Ye et al. Comprehensive evaluation of cleaner production in thermal power plants based on an improved least squares support vector machine model
CN110033118A (en) Elastomeric network modeling and the blower multiobjective optimization control method based on genetic algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant