CN110187635B - Method and apparatus for real-time optimization of continuous reformers - Google Patents
Method and apparatus for real-time optimization of continuous reformers Download PDFInfo
- Publication number
- CN110187635B CN110187635B CN201910285487.5A CN201910285487A CN110187635B CN 110187635 B CN110187635 B CN 110187635B CN 201910285487 A CN201910285487 A CN 201910285487A CN 110187635 B CN110187635 B CN 110187635B
- Authority
- CN
- China
- Prior art keywords
- real
- model
- time optimization
- continuous
- data
- 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
Links
Images
Classifications
-
- 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
- C10G35/00—Reforming naphtha
- C10G35/04—Catalytic reforming
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- 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
- C10G2300/1037—Hydrocarbon fractions
- C10G2300/1044—Heavy gasoline or naphtha having a boiling range of about 100 - 180 °C
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Oil, Petroleum & Natural Gas (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Organic Chemistry (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Evolutionary Computation (AREA)
- General Chemical & Material Sciences (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
- Production Of Liquid Hydrocarbon Mixture For Refining Petroleum (AREA)
Abstract
The embodiment provides a real-time optimization method and equipment for a continuous reforming device, which specifically comprise the following steps: establishing a real-time optimization model of the continuous reforming device, acquiring initial parameters for leading the real-time optimization model of the continuous reforming device, and performing effectiveness processing on the initial parameters to obtain data to be led in; transmitting data to be imported into a real-time optimization model of the continuous reforming device, solving a target optimization function in a range corresponding to constraint conditions, and determining optimal device operation parameters corresponding to the optimal solution of the target function; current control parameters of the continuous reformer are optimized based on the optimal plant control parameters. Due to the fact that the advantages of the mechanism model and the data experience model are fully exerted, the mixed model can reflect the real physical characteristics of the continuous reforming device process, model extrapolation capacity is improved, and real-time optimization solving efficiency and stability are improved.
Description
Technical Field
The invention belongs to the field of data optimization, and particularly relates to a real-time optimization method and equipment for a continuous reforming device.
Background
The comprehensive benefits of enterprises can be improved by applying the petrochemical industry process real-time optimization technology, and the method becomes a consensus of petrochemical industry and other process industry enterprises at home and abroad. By adopting the real-time optimization technology, the repetitive labor of the device operators is reduced, the operation value is determined by the technicians only according to the design parameters and experiences, the optimized operation cannot be achieved, the energy conservation and consumption reduction of enterprises can be greatly promoted, the reaction speed to the market is greatly increased, and considerable economic benefit is brought.
The process real-time optimization technology can monitor the running state of the process device at any time, and continuously adjust the working point on the premise of meeting all constraint conditions so as to overcome the influence of various factors on the process and ensure that the process device can obtain the best economic benefit all the time. However, most of the real-time optimization of the device depends on a strict steady-state mechanism model, and although the dynamic mechanism model based on complex material balance and energy balance equations can well represent the physical characteristics of the continuous reforming device, the real-time optimization often has difficulty in solving and even often cannot converge in the face of the actual on-site reforming feeding raw material property change and uncertain operation disturbance, and the on-line real-time calculation and stability performance are influenced. On the other hand, in order to avoid the problem, an empirical model of historical data is used in an attempt, so that the real-time performance and the stability are improved to a certain extent, but at the cost of sacrificing the model extrapolation capability, the actual application range is greatly narrowed.
Disclosure of Invention
In order to overcome the defects and shortcomings in the prior art, the invention provides the real-time optimization method and the real-time optimization equipment for the continuous reforming device, the advantages of a mechanism model and a data experience model are fully exerted, a mixed model can reflect the real physical characteristics of the process of the continuous reforming device, the model extrapolation capacity is improved, the complexity of the original model is reduced, and the real-time optimization solving efficiency and the stability are greatly improved.
In one aspect, the present embodiment provides a real-time optimization method for a continuous reformer, including:
establishing a real-time optimization model of a continuous reforming device, and establishing a target optimization function and constraint conditions in the process of solving the target optimization function in the real-time optimization model;
acquiring initial parameters for importing the real-time optimization model of the continuous reforming device from a near-infrared online analyzer and a database, performing effectiveness processing on the initial parameters, and performing operations including data setting and model parameter estimation on the processed data to obtain data to be imported;
transmitting the data to be imported into a real-time optimization model of the continuous reforming device, solving an objective optimization function in a range corresponding to the constraint condition, and determining an optimal device operation parameter corresponding to the optimal solution of the objective function;
and step four, optimizing the current control parameters of the continuous reforming device based on the optimal device control parameters.
Optionally, the establishing a real-time optimization model of the continuous reforming device includes:
establishing a simplified mechanism model of a continuous reforming device as shown in formula I
Wherein G (-) is a nonlinear system of equations, y is when j is 1,2,3,4 raw,j Respectively a straight-run naphtha raw material, a gasoline and diesel oil hydrogenated naphtha raw material, a cracking raffinate raw material, a hydrocracking heavy naphtha raw material, y burden Minimum and maximum reformer load, y oper,i The minimum value and the maximum value of the technological requirements of each operation condition are respectively, i is 1,2,3,4,5,6,7 and 8, and corresponds to the initial boiling point of reforming feed, the inlet temperature of a reforming reactor, the hydrogen-oil ratio, the water-chlorine ratio, the oxygen content of a heating furnace, the outlet temperature of the heating furnace, the reflux ratio of a fractionating tower, the reflux ratio of a rectifying tower, and the like,Operating conditions of the column pressure, y quality,m The minimum specification and the maximum specification are respectively the quality requirements of the generated oil product, m is 1,2 is respectively the product extracted from the bottom and the top of the tower, Ar _ yld is respectively the yield of the aromatic hydrocarbon, and the energy consumption of the EnergyCost device is reduced;
estimating the correlation coefficient in the simplified mechanism model by using a deep neural network model through the acquired steady-state data set of the historical process, and establishing a correlation mathematical model of the continuous reforming device as shown in a formula II
DNNf (·) is a deep learning network model used for estimating a reforming reaction rate constant by offline learning, and T is reaction temperature and has a unit of K; LHSV is liquid hourly space velocity, and the unit is 1/h; v c Is the catalyst loading in m 3 ；
The constant-pressure specific heat vector of the lumped component is expressed in kJ/kmol/K; Δ H j The reaction heat of the jth reaction is expressed in kJ/kmol; a is the flow direction cross section area of the reaction oil gas, and the unit is m 2 ；
A is 2 pi R H, wherein R is the radius of a reactor bed layer, and H is the height of the reactor;
the molar flow vector of the lumped components including hydrogen;
a matrix of reaction rate constants for each reforming reaction; r is i,j (j ═ 1 to m) i is the jth reaction rate of the lumped component.
Optionally, the establishing of the objective optimization function in the real-time optimization model and the constraint condition in the process of solving the objective optimization function include:
determining an optimization target for maximizing the aromatic hydrocarbon yield of a continuous reforming device and minimizing the energy consumption of the device on the premise of meeting the requirements of planned scheduling;
establishing an objective optimization function expression based on a confirmed optimization objective
OBJ＝min(w 1 /Ar_yld+w 2 ·EnergyCost)
Where OBJ is the minimization of the objective function, w 1 Ar-yld is the reciprocal of the maximum aromatics yield target calculation, w 2 The second term of EnergyCost is the minimum energy consumption target of the plant, Ar-yld is the yield of reformate aromatics, EnergyCost is the sum of the fuel gas consumption of the reforming reactor furnace and the fractionating tower furnace, w 1 、w 2 The weights are respectively two terms of the objective function;
the constraint conditions comprise raw material feeding amount constraint, device load constraint, operation condition constraint, reformate quality requirement constraint and model balance constraint.
Optionally, the performing validity processing on the initial parameter includes:
sequentially carrying out consistency detection, online preprocessing and process heuristic steady-state detection on the initial parameters;
the consistency detection comprises the steps of correcting the categories of all data in the initial parameters in a parameter matching mode;
the online pretreatment comprises the steps of filling missing values and filtering and denoising the corrected initial parameters;
the process heuristic steady-state detection comprises the operation of judging whether the reforming process corresponding to the initial parameters is in a steady state or not according to the difference between the processed initial parameters before and after denoising.
Optionally, the operation of determining whether the initial parameter is in a steady state by the difference between the processed initial parameter before and after denoising includes:
respectively calculating the high and low filtering values of each variable according to a formula;
in the formula (I), the compound is shown in the specification,
respectively represent the high and low filtered values at the current moment,
respectively representing the high and low filtered values at the last moment, f H 、f L Respectively representing high and low filter coefficients, x t Representing the raw data input at the current moment;
calculating whether the sum of the mean square deviations of the high and low filtering values is less than a threshold epsilon within a certain time interval t 1 ；
If the judgment condition of the formula IV is met, calculating whether the sum of the mean square deviations of the oldest high filtering value and the newest high filtering value is less than a threshold value epsilon within a certain time interval t or not as shown in the formula V 2 If the judgment condition of the formula five is met, judging that the reforming process corresponding to the initial parameter in the time interval t is in a steady state;
in the formula (I), the compound is shown in the specification,
representing the most recent high filtered value of the signal,
representing the oldest highly filtered value.
Optionally, the data tuning includes:
determining a setting condition based on the minimum square sum of the deviation of the setting value and the corresponding measured value on the basis of the original measured data by utilizing the material balance or energy balance and other relations in the production process of the continuous reforming device, and establishing a setting mathematical expression shown in a formula six on the basis of the setting condition
Wherein X represents a measured variable vector, U represents an unmeasured variable vector,
expressing a setting value vector, Q is a measured variable covariance matrix, G is an open constraint equation of a mixed model of the continuous reforming device, and the open constraint equation comprises a material balance equation, an energy balance equation, a chemical reflection rate equation, a chemical balance equation, a heat mass and momentum transfer equation, a connection equation, a physical property calculation equation and a molecular normalization equation of each device unit of the continuous reforming device model;
and setting the initial data based on four pairs of formulas.
Optionally, between the third step and the fourth step in the real-time optimization method, the method further includes:
and fifthly, performing steady state judgment and reliability judgment again on the obtained optimal device parameters, and optimizing the current control parameters of the continuous reforming device based on the judged optimal parameter values.
Optionally, the real-time optimization method further includes:
and setting a non-disturbance switching mechanism for realizing switching between a real-time optimization automatic operation mode and a conventional artificial set value operation mode.
Optionally, the undisturbed handover mechanism includes:
optimizing the operation of the loop in real time;
cutting off a real-time optimization loop;
and the undisturbed switching between the real-time optimization system and the conventional manual setting is realized.
In another aspect, the present embodiment also provides a real-time optimization apparatus for a continuous reformer, the real-time optimization apparatus including:
the model establishing unit is used for establishing a real-time optimization model of the continuous reforming device, and establishing a target optimization function and a constraint condition in the process of solving the target optimization function in the real-time optimization model;
the data sorting unit is used for acquiring initial parameters for importing the real-time optimization model of the continuous reforming device from the near-infrared online analyzer and the database, carrying out effectiveness processing on the initial parameters, and carrying out operations including data setting and model parameter estimation on the processed data to obtain data to be imported;
the parameter solving unit is used for transmitting the data to be imported into the real-time optimization model of the continuous reforming device, solving the objective optimization function in the range corresponding to the constraint condition and determining the optimal device operation parameters corresponding to the optimal solution of the objective function;
and the optimization realization unit is used for optimizing the current control parameters of the continuous reforming device based on the optimal device control parameters.
The technical scheme provided by the invention has the beneficial effects that:
due to the fact that the advantages of the mechanism model and the data experience model are fully exerted, the mixed model can reflect the real physical characteristics of the continuous reforming device process, model extrapolation capacity is improved, the complexity of the original model is reduced, and real-time optimization solving efficiency and stability are greatly improved. Meanwhile, the invention integrates the functions of modules such as data processing, steady-state detection, parameter dynamic correction, data setting and the like, effectively ensures the online real-time closed-loop operation of the continuous reforming device, and improves the operation and management level of the continuous reforming production process, thereby achieving the purposes of optimizing the operation of the device, eliminating the bottleneck, saving energy, reducing consumption and improving the economic benefit of the device.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on the drawings without creative efforts.
FIG. 1 is a schematic flow diagram of a method for real-time optimization of a continuous reformer as set forth herein;
fig. 2 is a schematic diagram of a real-time optimization apparatus for a continuous reformer according to the present application.
Detailed Description
To make the structure and advantages of the present invention clearer, the structure of the present invention will be further described with reference to the accompanying drawings.
Example one
In order to solve the problem of difficulty in solving caused by over-reliance on a strict steady-state mechanism model in the prior art, the present embodiment provides a real-time optimization method for a continuous reformer based on a combination of a simplified mechanism and a deep neural network model, specifically as shown in fig. 1, the real-time optimization method includes:
establishing a real-time optimization model of a continuous reforming device, and establishing a target optimization function and constraint conditions in the process of solving the target optimization function in the real-time optimization model;
acquiring initial parameters for importing the real-time optimization model of the continuous reforming device from a near-infrared online analyzer and a database, performing effectiveness processing on the initial parameters, and performing operations including data setting and model parameter estimation on the processed data to obtain data to be imported;
transmitting the data to be imported into a real-time optimization model of the continuous reforming device, solving an objective optimization function in a range corresponding to the constraint condition, and determining an optimal device operation parameter corresponding to the optimal solution of the objective function;
and step four, optimizing the current control parameters of the continuous reforming device based on the optimal device control parameters.
In the implementation, the real-time optimization method comprises the steps of synchronously acquiring reforming raw material property data output by a near-infrared online analyzer and reforming device operation data acquired by a real-time database in real time, performing data preprocessing such as denoising and missing value filling, process steady state detection and data setting, inputting the data into a continuous reforming device quality and energy consumption associated mathematical model established based on a simplified mechanism and deep learning neural network data mixed model, taking maximization of aromatic hydrocarbon yield and minimization of energy consumption as a comprehensive optimization target, simultaneously considering production raw materials, products and operation constraints such as raw material properties, device load, catalyst water-chlorine balance, heating furnace oxygen content, heating furnace outlet temperature, product quality and the like, and utilizing a nonlinear programming solver to online obtain optimal operation parameter settings such as a feeding initial boiling point, reactor inlet temperature, hydrogen-oil ratio, fractionating tower reflux ratio and the like in real time, and after steady state and reliability are judged again, the real-time optimization of the online closed loop of the continuous reforming device is realized, the maximum aromatic hydrocarbon yield and the minimum energy consumption are achieved, and the economic benefit of the device is improved.
The simplified mechanism and deep learning neural network data mixed model used here is that a simplified mechanism model of the continuous reforming device is firstly established, wherein the simplified mechanism model comprises a material balance equation, an energy balance equation, a chemical reflection rate equation, a chemical balance equation, a heat mass and momentum transfer equation, a connection equation, a physical property calculation equation, a molecular normalization equation and the like among device units, and then the deep neural network model is used for off-line learning and estimation of a correlation coefficient in the simplified mechanism model by using a historical process steady-state data set, so that a correlation mathematical model of the continuous reforming device is established for on-line real-time calculation.
The process for constructing the real-time optimization model of the continuous reforming device comprises the following steps:
11. establishing a simplified mechanism model of a continuous reforming device as shown in formula I
Wherein G (-) is a nonlinear system of equations, y is when j is 1,2,3,4 raw,j Respectively a straight-run naphtha raw material, a gasoline and diesel oil hydrogenated naphtha raw material, a cracking raffinate raw material, a hydrocracking heavy naphtha raw material, y burden Minimum and maximum reformer load, y oper,i The minimum value and the maximum value of the technological requirement are respectively set for each operation condition, i is 1,2,3,4,5,6,7 and 8 minRespectively corresponding to the initial boiling point of reforming feed, the inlet temperature of a reforming reactor, the hydrogen-oil ratio, the water-chlorine ratio, the oxygen content of a heating furnace, the outlet temperature of the heating furnace, the reflux ratio of a fractionating tower, the pressure operation condition of the fractionating tower, y quality,m The minimum specification and the maximum specification are respectively the quality requirements of the generated oil product, m is 1,2 is respectively the product extracted from the bottom and the top of the tower, Ar _ yld is respectively the yield of the aromatic hydrocarbon, and the energy consumption of the EnergyCost device is reduced;
12. estimating the correlation coefficient in the simplified mechanism model by using a deep neural network model through the acquired steady-state data set of the historical process, and establishing a correlation mathematical model of the continuous reforming device as shown in a formula II
DNNf (·) is a deep learning network model used for off-line learning and estimating a reforming reaction rate constant, T is reaction temperature and the unit is K; LHSV is liquid hourly space velocity, and the unit is 1/h; v c Is the catalyst loading in m 3 ；
The constant-pressure specific heat vector of the lumped component is expressed in kJ/kmol/K; Δ H j The reaction heat of the jth reaction is expressed in kJ/kmol; a is the flow direction cross section area of the reaction oil gas, and the unit is m 2 ；
A is 2 pi R H, wherein R is the radius of a reactor bed layer, and H is the height of the reactor;
the molar flow vector of the lumped components including hydrogen;
a matrix of reaction rate constants for each reforming reaction; r is i,j (j ═ 1 to m) i is the jth reaction rate of the lumped component.
After model parameter estimation, inputting the set current data set into a continuous reforming device correlation mathematical model established based on a simplified mechanism and a deep neural network data mixed model, taking maximization of aromatic hydrocarbon yield and minimization of energy consumption as a comprehensive optimization objective function, simultaneously considering production raw material, product and operation constraint conditions such as reforming feed raw material property, device load, catalyst water-chlorine balance, heating furnace oxygen content, heating furnace outlet temperature, product quality requirements and the like, and obtaining optimal reforming feed initial boiling point, reactor inlet temperature, hydrogen-oil ratio, fractionating tower reflux ratio and other operation parameter settings on line in real time by utilizing a nonlinear programming solver in a feasible region determined by the constraint conditions.
In order to realize the setting of the optimal operating parameters, an objective optimization function and a constraint condition in the process of solving the objective optimization function need to be established in a real-time optimization model, and the method comprises the following steps:
determining an optimization target for maximizing the aromatic hydrocarbon yield of a continuous reforming device and minimizing the energy consumption of the device on the premise of meeting the requirements of planned scheduling;
establishing an objective optimization function expression based on a confirmed optimization objective
OBJ＝min(w 1 /Ar_yld+w 2 ·EnergyCost)
Where OBJ is the minimization of the objective function, w 1 Ar-yld is the reciprocal of the maximum aromatics yield target calculation, w 2 The second term of EnergyCost is the minimum energy consumption target of the plant, Ar-yld is the yield of reformate aromatics, EnergyCost is the sum of the fuel gas consumption of the reforming reactor furnace and the fractionating tower furnace, w 1 、w 2 The weights are respectively two terms of the objective function;
the constraint conditions comprise raw material feeding amount constraint, device load constraint, operation condition constraint, reformate quality requirement constraint and model balance constraint.
After a real-time optimization model of the continuous reforming device and corresponding constraint conditions are established, the data input into the real-time optimization model are acquired from a near-infrared online analyzer, a real-time database and an Oracle database by a data integration unit through a data integration engine, and initial parameters required by the real-time optimization model of the continuous reforming device are stored in a data warehouse.
The initialization parameters comprise: properties of the reforming raw material and the reformed oil obtained by the near-infrared online analyzer, such as density, sulfur content, nitrogen content, initial boiling point, final boiling point, PONA value and the like; the operation conditions of equipment such as a reforming reactor, a fractionating tower and the like, such as reforming feeding load, reforming reaction inlet temperature, heating furnace outlet temperature, heating furnace oxygen content, fractionating tower reflux ratio, sensitive plate temperature, hydrogen-oil ratio, water-chlorine ratio and the like, acquired by a real-time database; and the tower plate efficiency coefficient, the reforming reaction kinetic coefficient, the pre-factor and the like of the fractionating tower are obtained by an Oracle database.
In order to ensure the validity of the initial parameters, the initial parameters need to be subjected to validity processing, and the specific processing steps include: and sequentially carrying out three processing modes of consistency detection, online preprocessing and heuristic steady-state detection on the initial parameters. The three types of processing modes specifically include:
1) the consistency detection comprises the steps of checking all data types in the initial parameters in a parameter matching mode;
2) the on-line preprocessing comprises the steps of filling missing values and filtering and denoising the corrected initial parameters;
3) the heuristic steady-state detection comprises the operation of judging whether the reforming process corresponding to the initial parameters is in a steady state or not according to the difference between the processed initial parameters before and after denoising.
Specifically, the processing step of the heuristic steady-state detection specifically includes:
respectively calculating the high and low filtering values of each variable according to a formula;
in the formula (I), the compound is shown in the specification,
respectively represent the high and low filtered values at the current moment,
respectively representing the high and low filtered values at the last moment, f H 、f L Respectively representing high and low filter coefficients, x t Representing the raw data input at the current moment;
calculating whether the sum of the mean square deviations of the high and low filtering values is less than a threshold epsilon within a certain time interval t 1 ；
If the judgment condition of the formula IV is met, calculating whether the sum of the mean square deviations of the oldest high filtering value and the newest high filtering value in a certain time interval t is less than a threshold value epsilon or not as shown in the formula V 2 If the judgment condition of the formula five is met, judging that the reforming process corresponding to the initial parameter in the time interval t is in a steady state;
in the formula (I), the compound is shown in the specification,
representing the most recent high filtered value of the signal,
representing the oldest highly filtered value.
After the effectiveness processing is performed, in order to ensure that the data input into the real-time optimization model can reflect the material balance or energy balance and other relations in the production process of the continuous reforming device, the data subjected to the effectiveness processing needs to be subjected to data setting and model parameter estimation.
Specifically, the data setting includes:
determining a setting condition based on the minimum square sum of the deviation of the setting value and the corresponding measured value on the basis of the original measured data by utilizing the material balance or energy balance and other relations in the production process of the continuous reforming device, and establishing a setting mathematical expression shown in a formula six on the basis of the setting condition
Wherein X represents a measured variable vector, U represents an unmeasured variable vector,
expressing a setting value vector, Q is a measured variable covariance matrix, G is an open constraint equation of a mixed model of the continuous reforming device, and the open constraint equation comprises a material balance equation, an energy balance equation, a chemical reflection rate equation, a chemical balance equation, a heat mass and momentum transfer equation, a connection equation, a physical property calculation equation and a molecular normalization equation of each device unit of the continuous reforming device model;
and setting the initial data based on six pairs of formulas.
And the step of model parameter estimation is to adjust the parameters (such as the efficiency coefficient of the tower plate of the fractionating tower, the kinetic coefficient of the reforming reaction and the pre-factor) of the mixed model of the continuous reforming device by using the data set after the data setting so as to ensure that the output of the model is consistent with the actual measured data on site or the deviation is minimum.
Optionally, between the third step and the fourth step in the real-time optimization method, the method further includes:
and fifthly, performing steady state judgment and reliability judgment again on the obtained optimal device parameters, and optimizing the current control parameters of the continuous reforming device based on the judged optimal parameter values.
After the optimal operation condition is set, after steady state and reliability judgment is carried out again, the optimization result is used as a set value of an Advanced Process Control (APC) controlled variable and automatically downloaded to an Advanced controller APC to be executed, and the APC system can ensure that the APC system can be smoothly and quickly drawn to the optimal value (set value), so that online closed-loop real-time optimization of the continuous reforming device is realized, the device can run to achieve the maximum aromatic hydrocarbon yield and the minimum energy consumption, and the economic benefit of device production is improved.
Optionally, the real-time optimization method further includes:
in order to ensure the system safety and prevent abnormal conditions such as communication interruption, upper computer crash and the like between the real-time optimization upper computer and the APC in the circulation online optimization process, a non-disturbance switching mechanism for realizing switching between a real-time optimization automatic operation mode and a conventional artificial set value operation mode is further arranged, so that the unification of safety and benefits is ensured.
A disturbance-free switching mechanism comprising commissioning of a real-time optimization loop; the method comprises the following steps of cutting off a real-time optimization loop and undisturbed switching between a real-time optimization system and conventional artificial setting. Specifically, the method comprises the following steps:
1) and (3) real-time optimization loop operation: when the switch position number of a corresponding loop is set to be in an ON state, the calculation mode of the loop is automatically changed to a real-time optimization mode.
2) And (3) real-time optimization loop cutting: (1) when a communication fault occurs, the switch of the optimizer is cut off, and the used optimization loop stops acting; (2) when a certain loop is cut off in an optimized mode, the optimized mode is automatically switched back to the original manual optimized setting mode; (3) when the optimized operation condition of a certain loop is not met, cutting off the loop optimization and giving an alarm for prompting;
3) undisturbed switching: when the operation is not optimized, the set value of the controlled variable automatically tracks the process measurement value of the bit number; the optimized output value of the operation variable automatically tracks the set value or the threshold value of the bit number, and undisturbed switching is realized when real-time optimized operation is carried out.
4) The switching of the conventional manual setting and the real-time optimization is the switching of the input source of the optimization loop. After the real-time optimization starting operation, if the switch of the optimization loop is opened, the set value provided by the real-time optimization system can be used for loop optimization, but the real-time optimization is not performed at the moment. If real-time optimization is to be adopted, the calculation mode of the optimization loop must be controlled tangentially and remotely, and the optimization loop only receives the set value given by the real-time optimization. If the calculation mode of the optimization loop is switched back to the manual setting mode, the optimization loop will select the conventional manual setting mode.
The embodiment provides a continuous reforming device real-time optimization method based on the combination of a simplified mechanism and a deep neural network model. Due to the fact that the advantages of the mechanism model and the data experience model are fully exerted, the mixed model can reflect the real physical characteristics of the continuous reforming device process, model extrapolation capacity is improved, the complexity of the original model is reduced, and real-time optimization solving efficiency and stability are greatly improved. Meanwhile, the invention integrates the functions of modules such as data processing, steady-state detection, parameter dynamic correction, data setting and the like, effectively ensures the online real-time closed-loop operation of the continuous reforming device, and improves the operation and management level of the continuous reforming production process, thereby achieving the purposes of optimizing the operation of the device, eliminating the bottleneck, saving energy, reducing consumption and improving the economic benefit of the device.
Example two
The present embodiment also proposes a real-time optimization apparatus 2 for a continuous reformer, as shown in fig. 2, comprising:
a model establishing unit 21, configured to establish a real-time optimization model of the continuous reforming apparatus, and establish an objective optimization function and constraint conditions in a process of solving the objective optimization function in the real-time optimization model;
the data sorting unit 22 is used for acquiring initial parameters for importing the real-time optimization model of the continuous reforming device from the near-infrared online analyzer and the database, performing effectiveness processing on the initial parameters, and performing operations including data setting and model parameter estimation on the processed data to obtain data to be imported;
the parameter solving unit 23 is configured to transmit data to be imported to the real-time optimization model of the continuous reforming device, solve the objective optimization function within a range corresponding to the constraint condition, and determine an optimal device operation parameter corresponding to the optimal solution of the objective function;
an optimization implementation unit 24 for optimizing the current control parameters of the continuous reformer based on the optimal plant control parameters.
In the implementation, the real-time optimization method comprises the steps of synchronously acquiring reforming raw material property data output by a near-infrared online analyzer and reforming device operation data acquired by a real-time database in real time, performing data preprocessing such as denoising and missing value filling, process steady state detection and data setting, inputting the data into a continuous reforming device quality and energy consumption associated mathematical model established based on a simplified mechanism and deep learning neural network data mixed model, taking maximization of aromatic hydrocarbon yield and minimization of energy consumption as a comprehensive optimization target, simultaneously considering production raw materials, products and operation constraints such as raw material properties, device load, catalyst water-chlorine balance, heating furnace oxygen content, heating furnace outlet temperature, product quality and the like, and utilizing a nonlinear programming solver to online obtain optimal operation parameter settings such as a feeding initial boiling point, reactor inlet temperature, hydrogen-oil ratio, fractionating tower reflux ratio and the like in real time, and after steady state and reliability are judged again, the real-time optimization of the online closed loop of the continuous reforming device is realized, the maximum aromatic hydrocarbon yield and the minimum energy consumption are achieved, and the economic benefit of the device is improved.
The simplified mechanism and deep learning neural network data mixed model used here is that a simplified mechanism model of the continuous reforming device is firstly established, wherein the simplified mechanism model comprises a material balance equation, an energy balance equation, a chemical reflection rate equation, a chemical balance equation, a heat mass and momentum transfer equation, a connection equation, a physical property calculation equation, a molecular normalization equation and the like among device units, and then the deep neural network model is used for off-line learning and estimation of a correlation coefficient in the simplified mechanism model by using a historical process steady-state data set, so that a correlation mathematical model of the continuous reforming device is established for on-line real-time calculation.
The process for constructing the real-time optimization model of the continuous reforming device comprises the following steps:
11. establishing a simplified mechanism model of a continuous reforming device as shown in formula I
Wherein G (-) is a nonlinear system of equations, y is when j is 1,2,3,4 raw,j Respectively a straight run naphtha feedstock, a gasoline and diesel hydrogenated naphtha feedstock, a cracked raffinate feedstock, a hydrocracked heavy naphtha feedstock, y burden Minimum and maximum reformer load, y oper,i The minimum value and the maximum value of the technological requirements of each operation condition are respectively represented by i ═ 1,2,3,4,5,6,7 and 8, which correspond to the initial boiling point of reforming feed, the inlet temperature of a reforming reactor, the hydrogen-oil ratio, the water-chlorine ratio, the oxygen content of a heating furnace, the outlet temperature of the heating furnace, the reflux ratio of a fractionating tower, the operation condition of the pressure of the fractionating tower and y quality,m The minimum specification and the maximum specification are respectively the quality requirements of the generated oil product, m is 1,2 is respectively the product extracted from the bottom and the top of the tower, Ar _ yld is respectively the yield of the aromatic hydrocarbon, and the energy consumption of the EnergyCost device is reduced;
12. estimating the correlation coefficient in the simplified mechanism model by using a deep neural network model through the acquired steady-state data set of the historical process, and establishing a correlation mathematical model of the continuous reforming device as shown in a formula II
DNNf (·) is a deep learning network model used for off-line learning and estimating a reforming reaction rate constant, T is reaction temperature and the unit is K; LHSV is liquid hourly space velocity, and the unit is 1/h; v c Is the catalyst loading in m 3 ；
The constant-pressure specific heat vector of the lumped component is expressed in kJ/kmol/K; Δ H j The reaction heat of the jth reaction is expressed in kJ/kmol; a is the flow direction cross section area of the reaction oil gas, and the unit is m 2 ；
A is 2 pi R H, wherein R is the radius of a reactor bed layer, and H is the height of the reactor;
the molar flow vector of the lumped components including hydrogen;
a matrix of reaction rate constants for each reforming reaction; r is i,j (j 1-m) i as lumped componentThe jth reaction rate of (c).
After model parameter estimation, inputting the set current data set into a continuous reforming device correlation mathematical model established based on a simplified mechanism and a deep neural network data mixed model, taking maximization of aromatic hydrocarbon yield and minimization of energy consumption as a comprehensive optimization objective function, simultaneously considering production raw material, product and operation constraint conditions such as reforming feed raw material property, device load, catalyst water-chlorine balance, heating furnace oxygen content, heating furnace outlet temperature, product quality requirements and the like, and obtaining optimal reforming feed initial boiling point, reactor inlet temperature, hydrogen-oil ratio, fractionating tower reflux ratio and other operation parameter settings on line in real time by utilizing a nonlinear programming solver in a feasible region determined by the constraint conditions.
The embodiment provides a continuous reforming device real-time optimization device based on the combination of a simplified mechanism and a deep neural network model. Due to the fact that the advantages of the mechanism model and the data experience model are fully exerted, the mixed model can reflect the real physical characteristics of the continuous reforming device process, model extrapolation capacity is improved, the complexity of the original model is reduced, and real-time optimization solving efficiency and stability are greatly improved. Meanwhile, the invention integrates the functions of modules such as data processing, steady-state detection, parameter dynamic correction, data setting and the like, effectively ensures the online real-time closed-loop operation of the continuous reforming device, and improves the operation and management level of the continuous reforming production process, thereby achieving the purposes of optimizing the operation of the device, eliminating the bottleneck, saving energy, reducing consumption and improving the economic benefit of the device.
The invention provides a real-time optimization system of a continuous reforming device, which is implemented on 150 ten thousand continuous reforming devices of a subsidiary company of middle-sea oil, and has obvious economic benefit.
On one hand, the pre-hydrogenation fractionating tower is optimized in real time, the initial boiling point (74 ℃ -78 ℃) of reforming feed is improved, and the conversion rate, the aromatic hydrocarbon yield and the tetraphenyl yield of the reforming product gasoline are obviously improved.
TABLE 1 comparison of multiple parameters before and after real-time optimization of a pre-hydrogenated fractionator
On the other hand, by carrying out real-time optimization operation on the reforming reactor, when the average reaction temperature WAIT at the inlet of the reforming reactor is increased to 522 ℃, the yield of the aromatic hydrocarbon is increased from 71.12% to 72.16%, and the yield of the aromatic hydrocarbon is increased by 1%.
TABLE 2 comparison of multiple parameters before and after real-time optimization of reforming reactors
The sequence numbers in the above embodiments are merely for description, and do not represent the sequence of the assembly or the use of the components.
The above description is only exemplary of the present invention and should not be taken as limiting the invention, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (5)
1. A real-time optimization method for a continuous reformer, the real-time optimization method comprising:
establishing a real-time optimization model of a continuous reforming device, and establishing a target optimization function and constraint conditions in the process of solving the target optimization function in the real-time optimization model;
acquiring initial parameters for importing the real-time optimization model of the continuous reforming device from a near-infrared online analyzer and a database, performing effectiveness processing on the initial parameters, and performing operations including data setting and model parameter estimation on the processed data to obtain data to be imported;
transmitting the data to be imported into a real-time optimization model of the continuous reforming device, solving an objective optimization function in a range corresponding to the constraint condition, and determining an optimal device operation parameter corresponding to the optimal solution of the objective function;
optimizing the current control parameters of the continuous reforming device based on the optimal device control parameters;
the establishing of the real-time optimization model of the continuous reforming device comprises the following steps:
establishing a simplified mechanism model of a continuous reforming device as shown in formula I
Wherein G (-) is a nonlinear system of equations, y is when j is 1,2,3,4 raw，j Respectively a straight-run naphtha raw material, a gasoline and diesel oil hydrogenated naphtha raw material, a cracking raffinate raw material, a hydrocracking heavy naphtha raw material, y burden In order to load the reformer,
for each operating condition, i is 1,2,3,4,5,6,7,8 respectively corresponding to the initial boiling point of reforming feed, the inlet temperature of reforming reactor, the hydrogen-oil ratio, the water-chlorine ratio, the oxygen content of heating furnace, the outlet temperature of heating furnace, the reflux ratio of fractionating tower, the operating condition of fractionating tower pressure, y quality，m In order to obtain the required quality value of the oil product, m is 1,2 respectively represents products extracted from the bottom and the top of the tower, Ar _ yld is the aromatic hydrocarbon yield of the reformed oil product, and EnergyCost is the sum of fuel gas consumption of a reforming reactor heating furnace and a fractionating tower heating furnace;
estimating the correlation coefficient in the simplified mechanism model by using a deep neural network model through the acquired steady-state data set of the historical process, and establishing a correlation mathematical model of the continuous reforming device as shown in a formula II
Wherein DNNf (. cndot.) is depthThe neural network model is used for off-line learning and estimating a reforming reaction rate constant, T is reaction temperature and the unit is K; LHSV is liquid hourly space velocity, and the unit is 1/h; v c Is the catalyst loading in m 3 ；
The constant-pressure specific heat vector of the lumped component is expressed in kJ/kmol/K; Δ H j’ The reaction heat of the j' th reaction is expressed in kJ/kmol; a is the flow direction cross section area of the reaction oil gas, and the unit is m 2 ；
A is 2 pi R H, wherein R is the radius of a reactor bed layer, and H is the height of the reactor;
the molar flow vector of the lumped components including hydrogen;
a matrix of reaction rate constants for each reforming reaction;
r j’ the reaction speed is jth 'of the lumped component, and j' is 1-70;
the establishing of the objective optimization function in the real-time optimization model and the constraint condition in the solving process of the objective optimization function comprises the following steps: determining an optimization target for maximizing the aromatic hydrocarbon yield of a continuous reforming device and minimizing the energy consumption of the device on the premise of meeting the scheduling requirement; establishing an objective optimization function expression based on a confirmed optimization objective
OBJ＝min(w 1 /Ar+_yld+w 2 ·EnergyCost)
Where OBJ is the minimization of the objective function, w 1 The second term w of the calculation of the maximum aromatics yield target for the plant, Ar _ yld 2 energyCost is the minimum energy consumption target of the device, Ar _ yld is the yield of aromatic hydrocarbons in the reformate product, energyCost is the sum of the fuel gas consumption of the reforming reactor heating furnace and the fractionating tower heating furnace, w 1 、w 2 The weights are respectively two terms of the objective function;
the constraint conditions comprise raw material feeding amount constraint, device load constraint, operation condition constraint, reformate quality requirement constraint and model balance constraint;
the validity processing of the initial parameters comprises:
sequentially carrying out consistency detection, online preprocessing and process heuristic steady-state detection on the initial parameters;
the consistency detection comprises the steps of correcting the categories of all data in the initial parameters in a parameter matching mode; the on-line preprocessing comprises the steps of filling missing values and filtering and denoising the corrected initial parameters;
the process heuristic steady-state detection comprises the steps of carrying out difference judgment on the processed initial parameters before and after denoising, and determining whether the reforming process corresponding to the initial parameters is in a steady-state operation;
the difference judgment before and after denoising is carried out on the processed initial parameters, and whether the reforming process corresponding to the initial parameters is in a steady-state operation or not is determined, including:
respectively calculating the high and low filtering values of each variable according to a formula;
in the formula (I), the compound is shown in the specification,
respectively represent the high and low filtered values at the current moment,
respectively representing the high and low filtered values at the last time, f H 、f L Respectively representing high and low filter coefficients, x t Representing the raw data input at the current moment;
calculating a certain time interval as shown in formula IV
Whether the sum of mean square deviations of high and low filtering values is less than a threshold value epsilon 1 ；
If the judgment condition of the formula IV is met, calculating whether the sum of the mean square deviations of the oldest high filtering value and the newest high filtering value is less than a threshold value epsilon within a certain time interval t or not as shown in the formula V 2 If the judgment condition of the formula five is met, judging that the reforming process corresponding to the initial parameter in the time interval t is in a steady state;
in the formula (I), the compound is shown in the specification,
representing the most recent high filtered value of the signal,
represents the oldest high filtered value;
the data setting comprises the following steps:
determining a setting condition based on the minimum square sum of the deviation of the setting value and the corresponding measured value on the basis of the original measured data by utilizing the material balance or energy balance relation in the production process of the continuous reforming device, and establishing a setting mathematical expression shown in a formula six on the basis of the setting condition
Wherein X represents a measured variable vector, U represents an unmeasured variable vector,
indicating a setting value vector, QFor measuring a variable covariance matrix, G is an open constraint equation of a continuous reforming device mixed model, and comprises a continuous reforming device model material balance equation, an energy balance equation, a chemical reflection rate equation, a chemical balance equation, a heat mass and momentum transfer equation, a connection equation among device units, a physical property calculation equation and a molecular normalization equation;
and setting the initial data based on a formula six.
2. The real-time optimization method for a continuous reformer according to claim 1, wherein between the third step and the fourth step in the real-time optimization method, further comprising:
and fifthly, performing steady state judgment and reliability judgment again on the obtained optimal device parameters, and optimizing the current control parameters of the continuous reforming device based on the judged optimal parameter values.
3. The real-time optimization method for a continuous reformer according to any one of claims 1 to 2, characterized in that it further comprises:
and setting a non-disturbance switching mechanism for realizing switching between a real-time optimization automatic operation mode and a conventional artificial set value operation mode.
4. The real-time optimization method for a continuous reformer according to claim 3, characterized in that said bumpless switching mechanism comprises:
optimizing the operation of the loop in real time;
cutting off a real-time optimization loop;
and a disturbance-free switching mechanism for realizing the switching between the real-time optimization automatic operation mode and the conventional artificial set value operation mode.
5. Real-time optimization apparatus for a continuous reformer, implementing the real-time optimization method of claim 1, characterized in that it comprises:
the model establishing unit is used for establishing a real-time optimization model of the continuous reforming device, and establishing a target optimization function and a constraint condition in the process of solving the target optimization function in the real-time optimization model;
the data sorting unit is used for acquiring initial parameters for importing the real-time optimization model of the continuous reforming device from the near-infrared online analyzer and the database, carrying out effectiveness processing on the initial parameters, and carrying out operations including data setting and model parameter estimation on the processed data to obtain data to be imported;
the parameter solving unit is used for transmitting the data to be imported into the real-time optimization model of the continuous reforming device, solving the objective optimization function in the range corresponding to the constraint condition and determining the optimal device operation parameters corresponding to the optimal solution of the objective function;
and the optimization realization unit is used for optimizing the current control parameters of the continuous reforming device based on the optimal device control parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910285487.5A CN110187635B (en) | 2019-04-10 | 2019-04-10 | Method and apparatus for real-time optimization of continuous reformers |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910285487.5A CN110187635B (en) | 2019-04-10 | 2019-04-10 | Method and apparatus for real-time optimization of continuous reformers |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110187635A CN110187635A (en) | 2019-08-30 |
| CN110187635B true CN110187635B (en) | 2022-08-02 |
Family
ID=67714071
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910285487.5A Active CN110187635B (en) | 2019-04-10 | 2019-04-10 | Method and apparatus for real-time optimization of continuous reformers |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110187635B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111242350A (en) * | 2019-12-31 | 2020-06-05 | 汉谷云智(武汉)科技有限公司 | Continuous reforming device fractionation and extraction optimization method, system and medium |
| CN111367181B (en) * | 2020-04-07 | 2022-06-07 | 北京工业大学 | Hybrid drive intelligent judgment control method for sewage treatment system |
| CN111899793B (en) * | 2020-06-12 | 2024-04-30 | 中国石油天然气股份有限公司 | Real-time optimization method, device and system for molecular-level device and storage medium |
| CN112147885B (en) * | 2020-08-21 | 2022-06-07 | 北京工业大学 | Dissolved oxygen concentration intelligent optimization control method based on execution-judgment structure |
| CN112489761A (en) * | 2020-11-19 | 2021-03-12 | 华中科技大学同济医学院附属协和医院 | Liquid and energy management system |
| CN112782979B (en) * | 2020-12-25 | 2022-05-03 | 杭州电子科技大学 | Real-time optimization control system and method for continuous catalytic reforming device |
| CN113110060B (en) * | 2021-04-29 | 2023-01-06 | 中海石油炼化有限责任公司 | Real-time optimization method of reforming device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000303076A (en) * | 1999-04-23 | 2000-10-31 | Idemitsu Kosan Co Ltd | Method for determining operation mode of gasoline production equipment and method for operating the equipment |
| CN102289579A (en) * | 2011-08-03 | 2011-12-21 | 浙江大学 | Modeling method for 38-lumping continuous reforming device reactor |
| CN107977736A (en) * | 2017-11-15 | 2018-05-01 | 浙江中控软件技术有限公司 | The optimization method and system of fuel gas system based on calorific value balance |
| CN108287474A (en) * | 2017-12-27 | 2018-07-17 | 上海交通大学 | Based on the probabilistic catalytic reforming reactor robust operation optimization method of raw material |
-
2019
- 2019-04-10 CN CN201910285487.5A patent/CN110187635B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000303076A (en) * | 1999-04-23 | 2000-10-31 | Idemitsu Kosan Co Ltd | Method for determining operation mode of gasoline production equipment and method for operating the equipment |
| CN102289579A (en) * | 2011-08-03 | 2011-12-21 | 浙江大学 | Modeling method for 38-lumping continuous reforming device reactor |
| CN107977736A (en) * | 2017-11-15 | 2018-05-01 | 浙江中控软件技术有限公司 | The optimization method and system of fuel gas system based on calorific value balance |
| CN108287474A (en) * | 2017-12-27 | 2018-07-17 | 上海交通大学 | Based on the probabilistic catalytic reforming reactor robust operation optimization method of raw material |
Non-Patent Citations (5)
| Title |
|---|
| On-line Production Quality Prediction for a Commercial Naphtha Catalytic Reforming Process;Hou Weifeng 等;《2013 Asian Network for Scientific Information》;20131231;第12卷(第18期);第4553-4560页全文 * |
| 催化重整装置的能量系统优化研究;汪佳;《中国优秀博硕士学位论文全文数据库(硕士) 工程科技Ⅰ辑》;20150615(第06期);第B017-34页全文 * |
| 基于Petro-SIM和遗传算法的连续重整优化研究;汪佳;《山东化工》;20150430;第44卷(第08期);第134-136页全文 * |
| 基于混合遗传算法的催化重整过程多目标优化;李鸿亮等;《化工学报》;20100229;第61卷(第02期);第432-438页全文 * |
| 基于统计模型的延迟焦化过程关键参数分析与优化;雷杨等;《石油学报(石油加工)》;20141231;第30卷(第06期);第1072-1079页全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110187635A (en) | 2019-08-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110187635B (en) | Method and apparatus for real-time optimization of continuous reformers | |
| US11993751B2 (en) | Predictive control systems and methods with fluid catalytic cracking volume gain optimization | |
| CN101305329B (en) | Dynamic constrained optimization of chemical manufacturing | |
| EP2408878B1 (en) | Automation and control of energy efficient fluid catalytic cracking processes for maximizing value added products | |
| CN104765346B (en) | A kind of oil refining process whole process modeling method | |
| US11720072B2 (en) | Plant state operating analysis and control | |
| US20220299952A1 (en) | Control system with optimization of neural network predictor | |
| CN109255461B (en) | Optimization method and optimization system of hydrogen resources | |
| CN104484714B (en) | A kind of real-time predicting method of catalytic reforming unit yield | |
| CN101957597B (en) | Real-time optimizer in continuous production process | |
| CN110020444B (en) | Optimization method and system of hydrogen resources | |
| CN100334047C (en) | Intelligent method for controlling cracking severity of cracking furnace in ethylene equipment | |
| CN104392098A (en) | Method for predicting yield of catalytically cracked gasoline | |
| US10614533B2 (en) | Methods for optimizing petrochemical facilities through stream lined transferal | |
| Van Wijk et al. | Advanced process control and on-line optimisation in shell refineries | |
| CN208477316U (en) | A kind of full cycle of operation cracking severity control system of ethane cracking furnace | |
| CN107450314B (en) | System and method for controlling full-operation period cracking depth of ethylene cracking furnace | |
| Gökçe | Model predictive controller design of hydrocracker reactors | |
| CA2807071A1 (en) | Computer apparatus and method for integration of process planning optimization and control | |
| CN112947343A (en) | Oil refining chemical plant production target tracking control method and device | |
| Macías‐Hernández et al. | Applications of evolving intelligent systems to oil and gas industry | |
| US20220266166A1 (en) | System and method for dynamic optimization of a crude distillation unit | |
| EP4334785A1 (en) | Predictive control systems and methods with fluid catalytic cracking volume gain optimization | |
| JP3016226B2 (en) | Apparatus and method for controlling properties of petroleum products | |
| WO2023235137A1 (en) | Control system with optimization of neural network predictor |
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 |