CN117815860A - Catalytic regeneration flue gas SO based on time delay estimation 2 Prediction method - Google Patents
Catalytic regeneration flue gas SO based on time delay estimation 2 Prediction method Download PDFInfo
- Publication number
- CN117815860A CN117815860A CN202211208008.8A CN202211208008A CN117815860A CN 117815860 A CN117815860 A CN 117815860A CN 202211208008 A CN202211208008 A CN 202211208008A CN 117815860 A CN117815860 A CN 117815860A
- Authority
- CN
- China
- Prior art keywords
- variable
- time delay
- flue gas
- solution
- prediction
- 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.)
- Pending
Links
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 title claims abstract description 81
- 239000003546 flue gas Substances 0.000 title claims abstract description 81
- 230000008929 regeneration Effects 0.000 title claims abstract description 77
- 238000011069 regeneration method Methods 0.000 title claims abstract description 77
- 230000003197 catalytic effect Effects 0.000 title claims abstract description 70
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000012216 screening Methods 0.000 claims abstract description 22
- 238000004523 catalytic cracking Methods 0.000 claims abstract description 17
- 230000006870 function Effects 0.000 claims description 49
- 239000002245 particle Substances 0.000 claims description 38
- 238000004422 calculation algorithm Methods 0.000 claims description 28
- 230000008569 process Effects 0.000 claims description 15
- 238000010238 partial least squares regression Methods 0.000 claims description 12
- 229910052717 sulfur Inorganic materials 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000006477 desulfuration reaction Methods 0.000 claims description 9
- 230000023556 desulfurization Effects 0.000 claims description 9
- 238000009826 distribution Methods 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 9
- 238000003860 storage Methods 0.000 claims description 7
- 238000010219 correlation analysis Methods 0.000 claims description 6
- 230000001133 acceleration Effects 0.000 claims description 3
- 238000009827 uniform distribution Methods 0.000 claims description 3
- 238000004989 laser desorption mass spectroscopy Methods 0.000 abstract description 5
- TXKMVPPZCYKFAC-UHFFFAOYSA-N disulfur monoxide Inorganic materials O=S=S TXKMVPPZCYKFAC-UHFFFAOYSA-N 0.000 description 52
- XTQHKBHJIVJGKJ-UHFFFAOYSA-N sulfur monoxide Chemical compound S=O XTQHKBHJIVJGKJ-UHFFFAOYSA-N 0.000 description 52
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 18
- 238000006243 chemical reaction Methods 0.000 description 14
- 239000003054 catalyst Substances 0.000 description 13
- 239000003921 oil Substances 0.000 description 11
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 10
- 239000001301 oxygen Substances 0.000 description 10
- 229910052760 oxygen Inorganic materials 0.000 description 10
- NINIDFKCEFEMDL-UHFFFAOYSA-N Sulfur Chemical compound [S] NINIDFKCEFEMDL-UHFFFAOYSA-N 0.000 description 9
- 229910052757 nitrogen Inorganic materials 0.000 description 9
- 239000011593 sulfur Substances 0.000 description 9
- 239000004215 Carbon black (E152) Substances 0.000 description 6
- RAHZWNYVWXNFOC-UHFFFAOYSA-N Sulphur dioxide Chemical compound O=S=O RAHZWNYVWXNFOC-UHFFFAOYSA-N 0.000 description 6
- 238000001514 detection method Methods 0.000 description 6
- 229930195733 hydrocarbon Natural products 0.000 description 6
- 150000002430 hydrocarbons Chemical class 0.000 description 6
- 239000007788 liquid Substances 0.000 description 6
- 230000003044 adaptive effect Effects 0.000 description 5
- 239000003795 chemical substances by application Substances 0.000 description 5
- 238000004590 computer program Methods 0.000 description 5
- 230000008878 coupling Effects 0.000 description 5
- 238000010168 coupling process Methods 0.000 description 5
- 238000005859 coupling reaction Methods 0.000 description 5
- 238000005984 hydrogenation reaction Methods 0.000 description 5
- 239000002994 raw material Substances 0.000 description 5
- 239000002002 slurry Substances 0.000 description 5
- 239000000779 smoke Substances 0.000 description 5
- 238000005406 washing Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- MWUXSHHQAYIFBG-UHFFFAOYSA-N nitrogen oxide Inorganic materials O=[N] MWUXSHHQAYIFBG-UHFFFAOYSA-N 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000000889 atomisation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000003009 desulfurizing effect Effects 0.000 description 2
- 239000000428 dust Substances 0.000 description 2
- 239000003344 environmental pollutant Substances 0.000 description 2
- 239000007789 gas Substances 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 231100000719 pollutant Toxicity 0.000 description 2
- 238000004064 recycling Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 238000003915 air pollution Methods 0.000 description 1
- 239000001273 butane Substances 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 239000000571 coke Substances 0.000 description 1
- 239000010779 crude oil Substances 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 239000002283 diesel fuel Substances 0.000 description 1
- 238000004231 fluid catalytic cracking Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- IJDNQMDRQITEOD-UHFFFAOYSA-N n-butane Chemical compound CCCC IJDNQMDRQITEOD-UHFFFAOYSA-N 0.000 description 1
- OFBQJSOFQDEBGM-UHFFFAOYSA-N n-pentane Natural products CCCCC OFBQJSOFQDEBGM-UHFFFAOYSA-N 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000001172 regenerating effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/46—Removing components of defined structure
- B01D53/48—Sulfur compounds
- B01D53/50—Sulfur oxides
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Chemical & Material Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Probability & Statistics with Applications (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Operations Research (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Environmental & Geological Engineering (AREA)
- Artificial Intelligence (AREA)
- Analytical Chemistry (AREA)
- General Chemical & Material Sciences (AREA)
- Oil, Petroleum & Natural Gas (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Exhaust Gas Treatment By Means Of Catalyst (AREA)
Abstract
The invention provides a catalytic regeneration flue gas SO based on time delay estimation 2 A prediction method, comprising: first-stage screening data during operation of catalytic cracking unit to obtain SO 2 The related variable is used as an auxiliary variable, and the range of the time delay variable in the auxiliary variable is determined; based on the auxiliary variable, building SO 2 A multi-objective delay estimation model of the auxiliary variable; optimizing a multi-target time delay estimation model, and realizing secondary screening of auxiliary variables to obtain an optimal time delay variable of the auxiliary variables; taking the optimal time delay variable as the SO of catalytic regeneration flue gas 2 The input of a prediction model corrects the time sequence characteristics of each variable and improves the SO of catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a). The invention solves the problem of SO in the regenerated flue gas of the catalytic cracking device 2 Related variables of (1) are respectively obtained by different acquisition modesThe problem of significant time lag exists between the real-time data collected by each sensor from the DCS system and the discrete data from the LIMS system.
Description
Technical Field
The invention relates to the technical field of regenerated flue gas of catalytic cracking devices, in particular to a catalytic regenerated flue gas SO based on time delay estimation 2 A prediction method.
Background
Catalytic cracking regeneration flue gas is the largest source of gas emissions in refinery production units. With the continuous increase of crude oil deterioration, the regenerated flue gas contains a large amount of sulfur dioxide (SO 2 ) Nitrogen oxides, have become the major source of air pollution in refineries. As the pollutant concentration of the flue gas at the outlet of the catalytic cracking regenerator has larger fluctuation range along with the adjustment of working conditions, no adjustment mechanism is established among the inlet concentration, the operation parameters and the outlet concentration at present, the operation parameters of the catalytic cracking desulfurization device mostly adopt empirical values, the addition amount of the desulfurizing agent does not establish the adjustment mechanism, and the method does not exist according to SO (sulfur oxide) 2 The inlet concentration value is dynamically adjusted. Therefore, when the inlet concentration fluctuates, the outlet flue gas is easy to have the problems of no stable standard, waste of desulfurizing agent, higher dangerous waste production and the like. Regenerated flue gas SO of catalytic cracking device 2 The related variables of the system are respectively acquired through different acquisition modes, obvious time delay exists between the real-time data acquired by each sensor and the discrete data of the LIMS system, and the characteristics of the process cannot be completely explained by directly modeling by utilizing the acquired data. The delay time of the data of different detection positions and different detection methods is different, SO that the regenerated flue gas SO is excavated 2 The delay characteristics of the related variables of (a) are key links of modeling.
However, the current state of the art is directed to catalytic regeneration of flue gas SO 2 The time delay characteristic research of the related variable is not mature, and the time delay variable is not accurately identified and the time delay time is not determined.
Aiming at the problems in the prior art, the invention provides a delay estimation-based methodCatalytic regeneration flue gas SO 2 A prediction method.
Disclosure of Invention
The method aims to solve the problem that in the prior art, obvious time delay exists between real-time data and discrete data acquired by each sensor. The invention provides a catalytic regeneration flue gas SO based on time delay estimation 2 A method of prediction, the method comprising the steps of:
first-stage screening data during operation of catalytic cracking unit to obtain SO 2 The related variable is used as an auxiliary variable, and the range of a time delay variable in the auxiliary variable is determined;
based on the auxiliary variable, building SO 2 A multi-objective delay estimation model of the auxiliary variable;
optimizing the multi-target time delay estimation model to realize secondary screening of the auxiliary variable to obtain an optimal time delay variable of the auxiliary variable;
taking the optimal time delay variable as catalytic regeneration flue gas SO 2 The input of a prediction model corrects the time sequence characteristics of each variable and improves the SO of catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
According to one embodiment of the invention, the catalytic cracker run time data comprises: reactor data, regenerator data, desulfurization and denitrification facility data, product distribution data, and abatement facility feedstock data.
According to one embodiment of the invention, the SO is obtained by first-stage screening of data during operation of the catalytic cracker by a sliding window strategy 2 And determining the range of delay variable in the auxiliary variable and the maximum delay time by utilizing a sliding window strategy.
According to one embodiment of the present invention, the multi-objective time delay estimation model includes two objective functions, wherein a first objective function is a sum of partial least squares regression values of all other auxiliary variables except the time delay variable after the time delay result is included, and a second objective function is catalytic regeneration of flue gas SO after the time delay variable is included 2 Prediction accuracy of the prediction model.
According to one embodiment of the invention, SO 2 Multi-objective delay estimation model of auxiliary variables:
minimize F(x i,(k) (t))=(S PLS (x i,(k) (t)),Δf(x i,(k) (t))) T ,
first objective function:
second objective function:
catalytic regeneration flue gas SO 2 Prediction model:
wherein: x is x i,(k) (t) is the variable value of the variable x after the ith time delay at the t moment, and the objective function of the multi-objective time delay estimation model is F (x) i,(k) (t)),S PLS (x i,(k) (t)) is the sum of partial least squares regression values of all other relevant variables except the delay variable after the delay result is taken in, S m (x i,(k) (t)) is the result of partial least squares regression calculation of variable x after the ith time delay at time t, and SO based on regression kernel function is introduced 2 Predictive model, f (x) i,(k) (t)) is the time instant SO 2 Expression of predictive function of model, f (x i,(k) (t)) * Is the true value of the variable x at time t, Δf (x i,(k) (t)) is the time instant SO 2 Model prediction accuracy expression, c r (t) is the center of the (r) th kernel function at time t, b r (t) is the width of the (r) th kernel function at time t, W r And (t) is the connection weight of the r-th kernel function at the moment t.
According to one embodiment of the invention, a multi-target particle swarm algorithm based on an adaptive global optimal solution is established to optimize the multi-target time delay estimation model, and the method is implementedNow the secondary screening of the auxiliary variable, the identification of the auxiliary variable relative to SO 2 And determining the time delay time difference of the time delay variable and extracting the optimal time delay variable.
According to the embodiment of the invention, two convergence indexes of particles are defined based on a multi-target particle swarm algorithm of a self-adaptive global optimal solution, namely convergence strength and dominance, so as to realize self-adaptive selection of the global optimal solution and enhance convergence performance of an algorithm optimizing process.
According to one embodiment of the invention, in a multi-objective particle swarm algorithm based on an adaptive global optimal solution:
velocity v i The (k+1) update formula is:
v i (k+1)=ωv i (k)+c 1 r 1 (p i (k)-a i (k))+c 2 r 2 (gBest(k)-a i (k)),
wherein p is i (k) The method is a historical optimal position of particles in the kth iteration, gBest (k) is an optimal position established through a population, and a global optimal solution gBest can be found through the whole particle population; ω is inertial weight for controlling the effect of previous speed on current speed; c 1 And c 2 Is the acceleration constant, r 1 And r 2 Is in [0,1 ]]Random values of the uniform distribution, position a of the particles i The (k+1) update formula is:
a i (k+1)=a i (k)+v i (k+1),
the initial position of each particle is randomly generated, the distribution of the particles in the evolution space can be uneven, the convergence strength of an optimized solution obtained by defining a function is determined by the number of solutions dominated by the optimized solution and the convergence distance of the current iteration, the optimal global optimal solution is obtained by solving two objective functions of the dominance and the convergence distance, and the expression of the global optimal solution in the kth iteration is as follows:
gBest(k+1)=argmaxG(a i (k)),
minimize G(a i (k))=(Z(a i (k)),Q(a i (k))) T ,
wherein: q (a) i (k) Is the ith solution a) i (k) Is of the convergence degree, Z (a) i (k) Is the ith solution a) i (k) I=1, 2,3 … I, I being the number of optimal solutions within the optimal solution set.
According to one embodiment of the invention, the convergence strength is calculated by the following formula:
solution a i (k)∈A k Is the dominance Z (a) i (k) A) is the solution a at the kth iteration i (k) Solution b capable of being dominant in optimal solution set j (k)∈B k-1 The expression of dominance is:
Z(a i (k))=|{b j (k)|b j (k)∈B k-1 ∧a i (k)>b j (k)∧a i (k)∈A k }
solution a i (k)∈A k With all solutions b governed by it j (k)∈B k-1 The convergence distance is expressed as:
where j=1, 2,3 … J, J is the number of solutions governed by the optimal solution, a k Is the non-dominant solution set obtained by the kth iteration in the optimal solution set, B k-1 Is the set of solutions governed by the solution generated in the k-1 th iteration in the optimal solution set, the dominance Z (a i (k) Is solution a) i (k) Can dominate B k-1 The number of solutions in (a), when Z (a) i (k) When zero), the solution is non-dominant, Q (a) i (k) Is solution a) i (k) Convergence distance of b j (k) Is solved a i (k) The dominant ith solution.
According to one embodiment of the invention, the optimal delay variation and SO are performed 2 Concentration correlation analysis is carried out to obtain the catalytic regeneration flue gas SO 2 And predicting a new input sequence of the model, correcting the model parameters through a rapid descent algorithm, and evaluating the performance of the model through root mean square error and prediction precision.
According to another aspect of the invention there is also provided a storage medium containing a series of instructions for performing the method steps as described in any one of the above.
According to another aspect of the present invention, there is also provided a catalytic regeneration flue gas SO based on time delay estimation 2 A predictive device for performing the method of any one of the preceding claims, the device comprising:
an auxiliary variable module for performing primary screening on data during operation of the catalytic cracking unit to obtain SO 2 The related variable is used as an auxiliary variable, and the range of a time delay variable in the auxiliary variable is determined;
an estimation model module for establishing SO based on the auxiliary variable 2 A multi-objective delay estimation model of the auxiliary variable;
the optimizing module is used for optimizing the multi-target time delay estimation model, realizing the secondary screening of the auxiliary variable and obtaining the optimal time delay variable of the auxiliary variable;
an accuracy module for taking the optimal time delay variable as catalytic regeneration flue gas SO 2 The input of a prediction model corrects the time sequence characteristics of each variable and improves the SO of catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
The invention provides a catalytic regeneration flue gas SO based on time delay estimation 2 Compared with the prior art, the prediction method has the following beneficial characteristics:
1. solves the problem of SO in the regenerated flue gas of the catalytic cracking device 2 The related variables of the system are respectively acquired through different acquisition modes, and obvious time delay exists between the real-time data acquired by each sensor and the discrete data acquired by the LIMS system, so that the characteristics of the process can not be completely explained by directly modeling by utilizing the acquired data.
2. Establishment of SO 2 The multi-target time delay estimation model of the auxiliary variable identifies the time delay variable and determines the time delay time, and the time delay estimated variable is used as the catalytic regeneration flue gas SO 2 The input of the prediction model corrects the time sequence characteristics of each variable and improves the SO of the catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The accompanying drawings are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention, without limitation to the invention. In the drawings:
FIG. 1 shows a catalytic regeneration flue gas SO based on time delay estimation according to an embodiment of the invention 2 A prediction method flow chart;
FIG. 2 shows a time delay profile of scrubber inlet flue gas temperature according to one embodiment of the present invention.
In the drawings, like parts are designated with like reference numerals. In addition, the drawings are not drawn to scale.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the following embodiments of the present invention will be described in further detail with reference to the accompanying drawings.
At present, in order to realize catalytic regeneration of flue gas SO 2 Is stable and reaches the standard, reduces the load of desulfurization and denitrification facilities, zhou Ziyang and the like (the industrial application of the sulfur transfer agent RFS09 in RFCC flue gas desulfurization) improves SO by reasonably adding the sulfur transfer agent 2 The industrial operation results show that the method reduces the outlet SO 2 Without reducing the yield of the main product and the quality of the balancing agent. Against influencing SO 2 In (2) under the assumption that the catalyst coke is completely burned, yatee et al (Fluid Catalytic Cracking Unit Emissions and Their Impact) are based on SO 2 And the material balance of the particles are used for exploring the emission rule of pollutants, and experimental exploration shows that the feed rate and the sulfur content in the feed can influence SO 2 Discharge amount. None of the above methods have been examinedThe timing characteristics of the variables cannot be corrected in consideration of the delay problem of the relevant input variables.
The time delay characteristics of the variables reflect important dynamic causal relationships of the process, and the accuracy of time sequence matching directly influences the calculation result of the soft measurement model. At present, part of scholars have developed research on time delay characteristics of input variables, the prior art (time delay joint estimation based on complex correlation coefficients and application thereof, li Haijun, china test, 2019, 8 months) utilizes a complex correlation coefficient method to calculate time delay characteristics of a plurality of auxiliary variables simultaneously, ruan Hongmei and the like (dynamic soft measurement method based on joint mutual information) use the joint mutual information to determine process time delay parameters, so that the problem of predicting butane concentration at the bottom of a debutanizer is solved, and the calculation complexity is higher when the neighbor mutual information is used for analyzing the correlation. Bear, etc. (adaptive soft measurement modeling method for time-lapse selective integration LTDGPR model) selectively integrates gaussian process regression model.
In the prior art (an online soft measurement modeling method with process variable time lag estimation, li Yanjun, information and control, and 2016, 6 months) uses a time lag parameter set obtained under an offline condition to reconstruct soft measurement modeling data, and adopts a time difference-Gaussian process regression model to conduct online prediction on a current moment dominant variable value. The prior art (CN 107273633 a) solves the multi-process variable time-lag estimation problem by establishing an optimization model of the multi-process variable time-lag estimation problem and adopting an improved adaptive particle swarm algorithm.
However, the current state of the art is directed to catalytic regeneration of flue gas SO 2 The time delay characteristic research of the related variable is not mature, and the time delay variable is not accurately identified and the time delay time is not determined.
In view of the state of the art, the present invention establishes SO 2 The multi-target time delay estimation model of the auxiliary variable identifies the time delay variable and determines the time delay time, and the time delay estimated variable is used as the catalytic regeneration flue gas SO 2 The input of the prediction model corrects the time sequence characteristics of each variable and improves the SO of the catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
FIG. 1 shows a delay-based system according to an embodiment of the inventionEstimated catalytic regeneration flue gas SO 2 A flow chart of a prediction method.
As shown in FIG. 1, in step S1, data during the operation of a catalytic cracker is first screened to obtain SO 2 The related variable is used as an auxiliary variable, and the range of the delay variable in the auxiliary variable is determined. Specifically, the data during the operation of the catalytic cracker comprises: reactor data, regenerator data, desulfurization and denitrification facility data, product distribution data, and abatement facility feedstock data.
In one embodiment, the SO is obtained by first screening data during operation of the catalytic cracker via a sliding window strategy 2 Sample set V (v= [ V1, V2,) VM, sample set variable types M]) And using a sliding window strategy to determine the range of delay variable in the auxiliary variable and the maximum delay time T as the auxiliary variable.
As shown in fig. 1, in step S2, the SO is established based on the auxiliary variable 2 A multi-objective delay estimation model of the auxiliary variable. Specifically, the multi-objective delay estimation model comprises two objective functions, wherein a first objective function is the sum of partial least squares regression values of all other auxiliary variables except the delay variable after the delay variable is included, and a second objective function is the catalytic regeneration of flue gas SO after the delay variable is included 2 Prediction accuracy of the prediction model.
In one embodiment, the SO 2 Multi-objective delay estimation model of auxiliary variables:
min imize F(x i,(k) (t))=(S PLS (x i,(k) (t)),Δf(x i,(k) (t))) T ,
first objective function:
second objective function:
catalytic regeneration flue gas SO 2 Prediction model:
wherein: x is x i,(k) (t) is the variable value of the variable x after the ith time delay at the t moment, and the objective function of the multi-objective time delay estimation model is F (x) i,(k) (t)),S PLS (x i,(k) (t)) is the sum of partial least squares regression values of all other relevant variables except the delay variable after the delay result is taken in, S m (x i,(k) (t)) is the result of partial least squares regression calculation of variable x after the ith time delay at time t, and SO based on regression kernel function is introduced 2 Predictive model, f (x) i,(k) (t)) is the time instant SO 2 Expression of predictive function of model, f (x i,(k) (t)) * Is the true value of the variable x at time t, Δf (x i,(k) (t)) is the time instant SO 2 Model prediction accuracy expression, c r (t) is the center of the (r) th kernel function at time t, b r (t) is the width of the (r) th kernel function at time t, W r And (t) is the connection weight of the r-th kernel function at the moment t.
As shown in fig. 1, in step S3, a multi-objective delay estimation model is optimized, and secondary screening of auxiliary variables is implemented, so as to obtain an optimal delay variable of the auxiliary variables. Specifically, a multi-target particle swarm algorithm based on a self-adaptive global optimal solution is established to optimize a multi-target time delay estimation model, secondary screening of auxiliary variables is realized, and the auxiliary variables are identified relative to SO (SO) 2 And determining the time delay time difference of the time delay variable and extracting the optimal time delay variable.
In one embodiment, two convergence indexes of particles are defined based on a multi-target particle swarm algorithm of a self-adaptive global optimal solution, namely convergence strength and dominance, so as to realize self-adaptive selection of the global optimal solution and enhance convergence performance of an algorithm optimizing process.
In one embodiment, in a multi-objective particle swarm algorithm based on an adaptive global optimal solution:
velocity v i The (k+1) update formula is:
v i (k+1)=ωv i (k)+c 1 r 1 (p i (k)-a i (k))+c 2 r 2 (gBest(k)-a i (k)),
wherein p is i (k) The method is a historical optimal position of particles in the kth iteration, gBest (k) is an optimal position established through a population, and a global optimal solution gBest can be found through the whole particle population; ω is inertial weight for controlling the effect of previous speed on current speed; c 1 And c 2 Is the acceleration constant, r 1 And r 2 Is in [0,1 ]]Random values of the uniform distribution, position a of the particles i The (k+1) update formula is:
a i (k+1)=a i (k)+v i (k+1),
the initial position of each particle is randomly generated, the distribution of the particles in the evolution space can be uneven, the convergence strength of an optimized solution obtained by defining a function is determined by the number of solutions dominated by the optimized solution and the convergence distance of the current iteration, the optimal global optimal solution is obtained by solving two objective functions of the dominance and the convergence distance, and the expression of the global optimal solution in the kth iteration is as follows:
gBest(k+1)=argmaxG(a i (k)),
min imize G(a i (k))=(Z(a i (k)),Q(a i (k))) T ,
wherein: q (a) i (k) Is the ith solution a) i (k) Is of the convergence degree, Z (a) i (k) Is the ith solution a) i (k) I=1, 2,3 … I, I being the number of optimal solutions within the optimal solution set.
In one embodiment, the convergence strength is calculated by the following formula:
solution a i (k)∈A k Is the dominance Z (a) i (k) A) is the solution a at the kth iteration i (k) Solution b capable of being dominant in optimal solution set j (k)∈B k-1 The expression of dominance is:
Z(a i (k))=|{b j (k)|b j (k)∈B k-1 ∧a i (k)>b j (k)∧a i (k)∈A k }
solution a i (k)∈A k With all solutions b governed by it j (k)∈B k-1 The convergence distance is expressed as:
where j=1, 2,3 … J, J is the number of solutions governed by the optimal solution, a k Is the non-dominant solution set obtained by the kth iteration in the optimal solution set, B k-1 Is the set of solutions governed by the solution generated in the k-1 th iteration in the optimal solution set, the dominance Z (a i (k) Is solution a) i (k) Can dominate B k-1 The number of solutions in (a), when Z (a) i (k) When zero), the solution is non-dominant, Q (a) i (k) Is solution a) i (k) Convergence distance of b j (k) Is solved a i (k) The dominant ith solution.
As shown in fig. 1, in step S4, the optimal delay variable is used as the catalytic regeneration flue gas SO 2 The input of a prediction model corrects the time sequence characteristics of each variable and improves the SO of catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
In one embodiment, the optimal delay variation and SO are performed 2 Concentration correlation analysis to obtain catalytic regenerated flue gas SO 2 And predicting a new input sequence of the model, correcting the model parameters through a rapid descent algorithm, and evaluating the performance of the model through root mean square error and prediction precision.
The invention aims at regenerating flue gas SO of a catalytic cracking device 2 The related variables of the system are respectively acquired through different acquisition modes, obvious time delay exists between the real-time data acquired by each sensor and the discrete data of the LIMS system, and the characteristics of the process cannot be completely explained by directly modeling by utilizing the acquired data. The invention is mainly applied to the regeneration of flue gas SO of a catalytic cracking device 2 Is to identify a delay variable and determine a delay time, and delayThe estimated variable is used as the SO of catalytic regeneration flue gas 2 The input of the prediction model can accurately correct the time sequence characteristics of each variable, SO as to improve the SO of the catalytic regeneration flue gas 2 Plays a key guiding role in the prediction precision of the model (C).
The invention provides a catalytic regeneration flue gas SO based on time delay estimation 2 The prediction method can be matched with a computer readable storage medium, the storage medium is stored with a computer program, and the computer program is executed to run a catalytic regeneration flue gas SO based on time delay estimation 2 A prediction method. The computer program is capable of executing computer instructions, which include computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc.
The computer readable storage medium may include: any entity or device capable of carrying computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth.
It should be noted that the content of the computer readable storage medium may be appropriately increased or decreased according to the requirements of the jurisdiction's legislation and the patent practice, for example, in some jurisdictions, the computer readable storage medium does not include an electrical carrier signal and a telecommunication signal according to the legislation and the patent practice.
In addition, the invention also provides a catalytic regeneration flue gas SO based on time delay estimation 2 A prediction device, comprising: the system comprises an auxiliary variable module, an estimation model module, an optimization module and an accuracy module. Wherein, the auxiliary variable module is used for carrying out primary screening on data during the operation of the catalytic cracking device to obtain SO 2 The related variable is used as an auxiliary variable, and the range of the delay variable in the auxiliary variable is determined. The estimation model module is used for establishing SO based on the auxiliary variable 2 A multi-objective delay estimation model of the auxiliary variable. The optimization module is used for optimizing the multi-target time delay estimation model to realize auxiliary changeAnd (5) carrying out secondary screening on the quantity to obtain the optimal time delay variable of the auxiliary variable. The precision module is used for taking the optimal time delay variable as the SO of the catalytic regeneration flue gas 2 The input of a prediction model corrects the time sequence characteristics of each variable and improves the SO of catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
FIG. 2 shows a time delay profile of scrubber inlet flue gas temperature according to one embodiment of the present invention.
In one embodiment, reactor data, regenerator data, desulfurization and denitrification facility data, product distribution data, abatement facility feedstock data, etc. during operation of the catalytic cracker are obtained by on-line instrumentation or laboratory analysis, which contains the following 47 parameters, each of 1000, consisting essentially of: scrubber inlet oxygen concentration, regenerator ejector outlet oxygen concentration, regenerator outlet SO 2 Concentration, catalyst activity value, regeneration pressure, catalyst inventory, gasoline yield, liquid hydrocarbon yield, raw material sulfur, nitrogen, fresh feed quantity, reaction pressure, reaction temperature, feed preheating temperature, reactor regeneration inclined tube temperature, diesel oil yield raw material nitrogen content, regenerator outlet smoke dust concentration, regenerator dense phase inventory, regeneration main air quantity, riser slurry feed quantity, riser upper temperature, regenerator outlet smoke gas temperature, denitration reaction temperature, hydrogenation wax oil flow from a hydrogenation device, total feed quantity, regenerator bottom dense phase temperature, regenerator dilute phase section pressure and the like. Because the data set has the problem of large data dimension difference among variables and large data acquisition frequency difference, the exploration of the time delay characteristics of the variables in the data set plays a key role in the establishment of a model.
First-stage screening by utilizing sliding window strategy to obtain SO 2 The related variable sample set V of (1) specifically comprises the inlet oxygen concentration of a washing tower, the outlet oxygen concentration of a regenerator, the outlet sulfur dioxide concentration of the regenerator, the catalyst activity value, the regeneration pressure, the catalyst inventory, the gasoline yield, the dense phase inventory of the regenerator, the raw material sulfur, nitrogen, fresh feed, the reaction pressure, the total feed, the feed preheating temperature, the liquid hydrocarbon yield, the outlet flue gas temperature of the regenerator, the denitration reaction temperature, the dense phase temperature of the bottom of the regenerator and the sampleThe number of the concentrated variables is 18; and simultaneously, determining the range of the time delay variable in the auxiliary variable by utilizing a sliding window strategy, wherein the maximum time delay time is 2 hours.
Establishment of SO 2 The multi-objective time delay estimation model of the auxiliary variable, wherein a first objective function is the sum of partial least squares regression values of all other auxiliary variables except the time delay variable after the time delay result is included, and a second objective function is the catalytic regeneration of flue gas SO after the time delay variable is included 2 Prediction accuracy of the prediction model.
The multi-target particle swarm algorithm based on the self-adaptive global optimal solution is provided, two convergence indexes of particles are defined, namely convergence strength and dominance, so that the self-adaptive selection of the global optimal solution is realized, and the convergence performance of the algorithm optimizing process is enhanced.
The time delay estimation variable is obtained after calculation through a multi-target particle swarm algorithm based on a self-adaptive global optimal solution, and the time delay variable comprises two variables, namely the denitration reaction temperature and the washing tower inlet flue gas temperature, obtained through calculation through a time delay estimation model. The time delay step number of the denitration reaction temperature is 16 steps, the interval time of detection samples is 3 minutes, and the total delay time is 48 minutes; the delay steps of the temperature of the flue gas at the inlet of the washing tower are 15 steps, the interval time of the detection samples is 3 minutes, and the total delay time is 45 minutes. Performing correlation analysis on the variables subjected to time lag estimation to obtain SO 2 Modeling new parameter sequences with building SO based on regression kernel functions 2 Prediction model, by applying fast descent algorithm to SO 2 And (5) evaluating the performance of the prediction model by the root mean square error and the prediction precision of the parameters of the prediction model.
In one embodiment, reactor data, regenerator data, desulfurization and denitrification facility data, product distribution data, abatement facility feedstock data, etc. during operation of the catalytic cracker are obtained by on-line instrumentation or laboratory analysis, which contains the following 47 parameters, each of 1000, consisting essentially of: the riser inlet A.B nozzle raw oil line, the concentration of oxygen at the top outlet of the regenerator, the concentration of sulfur dioxide at the outlet of the regenerator, the catalyst activity value, the regeneration pressure, the catalyst inventory, the gasoline yield, the liquid hydrocarbon yield, the raw sulfur, nitrogen, fresh feed quantity, the reaction pressure, the reaction temperature, the feed preheating temperature, the reactor regeneration inclined tube temperature, the content of raw nitrogen of the diesel yield, the concentration of smoke at the outlet of the regenerator, the dense phase inventory of the regenerator, the regeneration main air quantity, the feed quantity of slurry oil of the riser, the upper temperature of the riser, the temperature of flue gas at the outlet of the regenerator, the denitration reaction temperature, the flow of hydrogenated wax oil from a hydrogenation device, the total feed quantity, the dense phase temperature at the bottom of the regenerator, the dilute phase section pressure of the regenerator and the like. Because the data set has the problem of large data dimension difference among variables and large data acquisition frequency difference, the exploration of the time delay characteristics of the variables in the data set plays a key role in the establishment of a model.
First-stage screening by utilizing sliding window strategy to obtain SO 2 The related variable sample set V of (1) specifically comprises the inlet oxygen concentration of a washing tower, the outlet oxygen concentration of a regenerator and the outlet SO of the regenerator 2 Concentration, catalyst activity value, regeneration pressure, catalyst reserve, riser feed A.B nozzle raw oil line, liquid hydrocarbon yield, raw sulfur, nitrogen, fresh feed quantity, reaction pressure, total feed quantity, feed preheating temperature, regenerator dense phase reserve, regenerator bottom dense phase temperature, and the variety of sample concentrated variables are 16; meanwhile, the range of the time delay variable in the auxiliary variable is determined by utilizing a sliding window strategy, and the maximum time delay time is 1.7 hours.
Establishment of SO 2 The multi-objective time delay estimation model of the auxiliary variable, wherein a first objective function is the sum of partial least squares regression values of all other auxiliary variables except the time delay variable after the time delay result is included, and a second objective function is the catalytic regeneration of flue gas SO after the time delay variable is included 2 Prediction accuracy of the prediction model.
The multi-target particle swarm algorithm based on the self-adaptive global optimal solution is provided, two convergence indexes of particles are defined, namely convergence strength and dominance, so that the self-adaptive selection of the global optimal solution is realized, and the convergence performance of the algorithm optimizing process is enhanced.
Obtaining a time delay estimation variable through calculation of a multi-target particle swarm algorithm based on a self-adaptive global optimal solution, and obtaining the time delay variation through calculation of a time delay estimation modelThe amount is that the lifting pipe enters the A.B nozzle raw oil line. The time delay step number of the lifting pipe feeding the A.B nozzle raw material oil line is 36 steps, the detection sample interval time is 3 minutes, and the total delay time is 108 minutes. Performing correlation analysis on the variables subjected to time lag estimation to obtain SO 2 Modeling new parameter sequences with building SO based on regression kernel functions 2 Prediction model, by applying fast descent algorithm to SO 2 And (5) evaluating the performance of the prediction model by the root mean square error and the prediction precision of the parameters of the prediction model.
In one embodiment, reactor data, regenerator data, desulfurization and denitrification facility data, product distribution data, abatement facility feedstock data, etc. during operation of the catalytic cracker are obtained by on-line instrumentation or laboratory analysis, which contains the following 42 parameters, each of 1500, consisting essentially of: riser slurry oil recycling atomization steam line, regenerator top outlet oxygen concentration and regenerator outlet SO 2 Concentration, catalyst activity value, regeneration pressure, catalyst reserve, gasoline yield, liquid hydrocarbon yield, regeneration slide valve position, raw sulfur, nitrogen, fresh feed quantity, reaction pressure, reaction temperature, feed preheating temperature, reactor regeneration inclined tube temperature, diesel yield raw nitrogen content, regenerator outlet smoke dust concentration, regenerator dense phase reserve, regeneration main air quantity, riser slurry oil feed quantity, riser upper temperature, regenerator outlet smoke temperature, denitration reaction temperature, hydrogenation wax oil flow from a hydrogenation device, total feed quantity, regenerator bottom dense phase temperature, regenerator dilute phase section pressure and the like. Because the data set has the problem of large data dimension difference among variables and large data acquisition frequency difference, the exploration of the time delay characteristics of the variables in the data set plays a key role in the establishment of a model.
First-stage screening by utilizing sliding window strategy to obtain SO 2 The related variable sample set V of (1) specifically comprises the inlet oxygen concentration of a washing tower, a valve position of a regeneration slide valve, the outlet oxygen concentration of a top of a regenerator and the outlet SO of the regenerator 2 Concentration, catalyst activity value, regeneration pressure, catalyst inventory, riser slurry oil recycling atomization steam line, liquid hydrocarbon yield, raw material sulfur, nitrogen, fresh feeding amount and reactionThe pressure, the total feeding quantity, the feeding preheating temperature and the dense phase temperature at the bottom of the regenerator, and the number of types of variables in the sample set is 16; meanwhile, the range of the time delay variable in the auxiliary variable is determined by utilizing a sliding window strategy, and the maximum time delay time is 1.5 hours.
Establishment of SO 2 The multi-objective time delay estimation model of the auxiliary variable, wherein a first objective function is the sum of partial least squares regression values of all other auxiliary variables except the time delay variable after the time delay result is included, and a second objective function is the catalytic regeneration of flue gas SO after the time delay variable is included 2 Prediction accuracy of the prediction model.
The multi-target particle swarm algorithm based on the self-adaptive global optimal solution is provided, two convergence indexes of particles are defined, namely convergence strength and dominance, so that the self-adaptive selection of the global optimal solution is realized, and the convergence performance of the algorithm optimizing process is enhanced.
And obtaining a time delay variable which is a valve position of the regeneration slide valve through calculation by a multi-target particle swarm algorithm based on a self-adaptive global optimal solution. The number of time delay steps of the valve position of the regenerated slide valve is 23 steps, the interval time of detection samples is 3 minutes, and the total delay time is 69 minutes. Performing correlation analysis on the variables subjected to time lag estimation to obtain SO 2 Modeling new parameter sequences with building SO based on regression kernel functions 2 Prediction model, by applying fast descent algorithm to SO 2 And (5) evaluating the performance of the prediction model by the root mean square error and the prediction precision of the parameters of the prediction model.
In summary, the invention provides a catalytic regeneration flue gas SO based on time delay estimation 2 Compared with the prior art, the prediction method has the following beneficial characteristics:
1. solves the problem of SO in the regenerated flue gas of the catalytic cracking device 2 The related variables of the system are respectively acquired through different acquisition modes, and obvious time delay exists between the real-time data acquired by each sensor and the discrete data acquired by the LIMS system, so that the characteristics of the process can not be completely explained by directly modeling by utilizing the acquired data.
2. Establishment of SO 2 Multi-target time delay estimation model of auxiliary variable, and identifying time delay variable anddetermining time delay time, and taking the variable after time delay estimation as catalytic regeneration flue gas SO 2 The input of the prediction model corrects the time sequence characteristics of each variable and improves the SO of the catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
It is to be understood that the disclosed embodiments are not limited to the specific structures, process steps, or materials disclosed herein, but are intended to extend to equivalents of these features as would be understood by one of ordinary skill in the relevant arts. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting.
In the description of the present invention, unless otherwise indicated, the meaning of "a plurality" is two or more; the terms "upper," "lower," "left," "right," "inner," "outer," "front," "rear," "head," "tail," and the like are used as an orientation or positional relationship based on that shown in the drawings, merely to facilitate description of the invention and to simplify the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore should not be construed as limiting the invention. Furthermore, the terms "first," "second," "third," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
In the description of the present invention, it should be noted that, unless explicitly specified and limited otherwise, the terms "connected," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
Certain terminology is used throughout this application to refer to particular system components. As one skilled in the art will recognize, identical components may generally be indicated by different names, and thus this document does not intend to distinguish between components that differ only in name, but not function. In this document, the terms "include", "include" and "have" are used in an open form and are therefore to be construed as meaning "including but not limited to …". Furthermore, the terms "substantially," "substantially," or "approximately" as may be used herein relate to the tolerances accepted by the industry for the respective terms. The term "coupled," as may be used herein, includes direct coupling and indirect coupling via another component, element, circuit, or module where, for indirect coupling, the intervening component, element, circuit, or module does not alter the information of the signal but may adjust its current level, voltage level, and/or power level. Inferred coupling (e.g., where one element is coupled to another element by inference) includes direct and indirect coupling between two elements in the same manner as "coupled".
Reference in the specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. Thus, the appearances of the phrase "one embodiment" or "an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment.
The embodiments of the invention have been presented for purposes of illustration and description, and are not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments with various modifications as are suited to the particular use contemplated.
Although the embodiments of the present invention are disclosed above, the embodiments are only used for the convenience of understanding the present invention, and are not intended to limit the present invention. Any person skilled in the art can make any modification and variation in form and detail without departing from the spirit and scope of the present disclosure, but the scope of the present disclosure is still subject to the scope of the appended claims.
Claims (12)
1. Catalytic regeneration flue gas SO based on time delay estimation 2 A method of prediction, the method comprising the steps of:
first-stage screening data during operation of catalytic cracking unit to obtain SO 2 The related variable is used as an auxiliary variable, and the range of a time delay variable in the auxiliary variable is determined;
based on the auxiliary variable, building SO 2 A multi-objective delay estimation model of the auxiliary variable;
optimizing the multi-target time delay estimation model to realize secondary screening of the auxiliary variable to obtain an optimal time delay variable of the auxiliary variable;
taking the optimal time delay variable as catalytic regeneration flue gas SO 2 The input of a prediction model corrects the time sequence characteristics of each variable and improves the SO of catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
2. A catalytic regeneration flue gas SO based on time delay estimation as defined in claim 1 2 A predictive method, wherein the catalytic cracker run-time data comprises: reactor data, regenerator data, desulfurization and denitrification facility data, product distribution data, and abatement facility feedstock data.
3. A catalytic regeneration flue gas SO based on time delay estimation as claimed in claim 1 or 2 2 The prediction method is characterized in that the data during the operation of the catalytic cracking unit is subjected to primary screening through a sliding window strategy to obtain SO 2 And determining the range of delay variable in the auxiliary variable and the maximum delay time by utilizing a sliding window strategy.
4. A catalytic regeneration flue gas SO based on a time delay estimation according to any one of claims 1 to 3 2 The prediction method is characterized in thatThe multi-objective time delay estimation model comprises two objective functions, wherein a first objective function is the sum of partial least squares regression values of all other auxiliary variables except the time delay variable after the time delay result is included, and a second objective function is the catalytic regeneration of flue gas SO after the time delay variable is included 2 Prediction accuracy of the prediction model.
5. A catalytic regeneration flue gas SO based on time delay estimation as defined in claim 4 2 A prediction method is characterized in that 2 Multi-objective delay estimation model of auxiliary variables:
minimize F(x i,(k) (t))=(S PLS (x i,(k) (t)),Δf(x i,(k) (t))) T ,
first objective function:
second objective function:
catalytic regeneration flue gas SO 2 Prediction model:
wherein: x is x i,(k) (t) is the variable value of the variable x after the ith time delay at the t moment, and the objective function of the multi-objective time delay estimation model is F (x) i,(k) (t)),S PLS (x i,(k) (t)) is the sum of partial least squares regression values of all other relevant variables except the delay variable after the delay result is taken in, S m (x i,(k) (t)) is the result of partial least squares regression calculation of variable x after the ith time delay at time t, and SO based on regression kernel function is introduced 2 Predictive model, f (x) i,(k) (t)) is the time instant SO 2 Prediction function expression of modelF (x) i,(k) (t)) * Is the true value of the variable x at time t, Δf (x i,(k) (t)) is the time instant SO 2 Model prediction accuracy expression, c r (t) is the center of the (r) th kernel function at time t, b r (t) is the width of the (r) th kernel function at time t, W r And (t) is the connection weight of the r-th kernel function at the moment t.
6. A catalytic regeneration flue gas SO based on a time delay estimation as claimed in any one of claims 1 to 5 2 The prediction method is characterized by establishing a multi-target particle swarm algorithm based on a self-adaptive global optimal solution to optimize the multi-target time delay estimation model, realizing the secondary screening of the auxiliary variable, and identifying the auxiliary variable relative to SO (SO) 2 And determining the time delay time difference of the time delay variable and extracting the optimal time delay variable.
7. A catalytic regeneration flue gas SO based on time delay estimation as defined in claim 6 2 The prediction method is characterized in that two convergence indexes of particles are defined based on a multi-target particle swarm algorithm of a self-adaptive global optimal solution, the convergence indexes are respectively convergence strength and dominance, the self-adaptive selection of the global optimal solution is realized, and the convergence performance of an algorithm optimizing process is enhanced.
8. A catalytic regeneration flue gas SO based on time delay estimation as claimed in claim 6 or 7 2 The prediction method is characterized in that in a multi-target particle swarm algorithm based on a self-adaptive global optimal solution:
velocity v i The (k+1) update formula is:
v i (k+1)=ωv i (k)+c 1 r 1 (p i (k)-a i (k))+c 2 r 2 (gBest(k)-a i (k)),
wherein p is i (k) The method is a historical optimal position of particles in the kth iteration, gBest (k) is an optimal position established through a population, and a global optimal solution gBest can be found through the whole particle population; omega is the inertial weight used for control toThe effect of the previous speed on the current speed; c 1 And c 2 Is the acceleration constant, r 1 And r 2 Is in [0,1 ]]Random values of the uniform distribution, position a of the particles i The (k+1) update formula is:
a i (k+1)=a i (k)+v i (k+1),
the initial position of each particle is randomly generated, the distribution of the particles in the evolution space can be uneven, the convergence strength of an optimized solution obtained by defining a function is determined by the number of solutions dominated by the optimized solution and the convergence distance of the current iteration, the optimal global optimal solution is obtained by solving two objective functions of the dominance and the convergence distance, and the expression of the global optimal solution in the kth iteration is as follows:
gBest(k+1)=arg max G(a i (k)),
min imize G(a i (k))=(Z(a i (k)),Q(a i (k))) T ,
wherein: q (a) i (k) Is the ith solution a) i (k) Is of the convergence degree, Z (a) i (k) Is the ith solution a) i (k) I=1, 2,3 … I, I being the number of optimal solutions within the optimal solution set.
9. A catalytic regeneration flue gas SO based on time delay estimation as defined in claim 8 2 The prediction method is characterized in that the convergence strength is calculated by the following formula:
solution a i (k)∈A k Is the dominance Z (a) i (k) A) is the solution a at the kth iteration i (k) Solution b capable of being dominant in optimal solution set j (k)∈B k-1 The expression of dominance is:
Z(a i (k))=|{b j (k)|b j (k)∈B k-1 ∧a i (k)>b j (k)∧a i (k)∈A k }|
solution a i (k)∈A k With all solutions b governed by it j (k)∈B k-1 The convergence distance is expressed as:
where j=1, 2,3 … J, J is the number of solutions governed by the optimal solution, a k Is the non-dominant solution set obtained by the kth iteration in the optimal solution set, B k-1 Is the set of solutions governed by the solution generated in the k-1 th iteration in the optimal solution set, the dominance Z (a i (k) Is solution a) i (k) Can dominate B k-1 The number of solutions in (a), when Z (a) i (k) When zero), the solution is non-dominant, Q (a) i (k) Is solution a) i (k) Convergence distance of b j (k) Is solved a i (k) The dominant ith solution.
10. A catalytic regeneration flue gas SO based on a time delay estimation as claimed in any one of claims 1 to 9 2 The prediction method is characterized in that the optimal time delay variable and SO are carried out 2 Concentration correlation analysis is carried out to obtain the catalytic regeneration flue gas SO 2 And predicting a new input sequence of the model, correcting the model parameters through a rapid descent algorithm, and evaluating the performance of the model through root mean square error and prediction precision.
11. A storage medium containing a series of instructions for performing the method steps of any one of claims 1-10.
12. Catalytic regeneration flue gas SO based on time delay estimation 2 A prediction apparatus, characterized in that it performs the method according to any one of claims 1-10, said apparatus comprising:
an auxiliary variable module for performing primary screening on data during operation of the catalytic cracking unit to obtain SO 2 The related variable is used as an auxiliary variable, and the range of a time delay variable in the auxiliary variable is determined;
an estimation model module for establishing SO based on the auxiliary variable 2 A multi-objective delay estimation model of the auxiliary variable;
the optimizing module is used for optimizing the multi-target time delay estimation model, realizing the secondary screening of the auxiliary variable and obtaining the optimal time delay variable of the auxiliary variable;
an accuracy module for taking the optimal time delay variable as catalytic regeneration flue gas SO 2 The input of a prediction model corrects the time sequence characteristics of each variable and improves the SO of catalytic regeneration flue gas 2 Is used for the prediction accuracy of (a).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211208008.8A CN117815860A (en) | 2022-09-30 | 2022-09-30 | Catalytic regeneration flue gas SO based on time delay estimation 2 Prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211208008.8A CN117815860A (en) | 2022-09-30 | 2022-09-30 | Catalytic regeneration flue gas SO based on time delay estimation 2 Prediction method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117815860A true CN117815860A (en) | 2024-04-05 |
Family
ID=90523497
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211208008.8A Pending CN117815860A (en) | 2022-09-30 | 2022-09-30 | Catalytic regeneration flue gas SO based on time delay estimation 2 Prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117815860A (en) |
-
2022
- 2022-09-30 CN CN202211208008.8A patent/CN117815860A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111804146B (en) | Intelligent ammonia injection control method and intelligent ammonia injection control device | |
CA2118885C (en) | Process control system | |
CN110807577A (en) | Pollution emission prediction method and device | |
CN113724800A (en) | Catalytic cracking unit simulation prediction method based on molecular level mechanism model and big data technology | |
CN101067828A (en) | Coke oven fire path temperature integrated moulding and soft measuring method | |
CN104789256A (en) | Real-time yield predicting method for catalytic cracking device | |
CN112651166B (en) | Denitration system inlet nitrogen oxide concentration prediction method and device and denitration system | |
CN114225662A (en) | Flue gas desulfurization and denitrification optimization control method based on hysteresis model | |
CN116688754A (en) | Ship flue gas desulfurization automatic control system and method thereof | |
CN104765347A (en) | Yield real-time prediction method in residual oil delayed coking process | |
CN112364562A (en) | Cooperative control method and system for flue gas environmental protection island | |
CN117815860A (en) | Catalytic regeneration flue gas SO based on time delay estimation 2 Prediction method | |
CN110931085B (en) | Continuous reforming catalyst coking carbon content online prediction and optimization method | |
WO2007009322A1 (en) | Real-time operating optimized method of multi-input and multi-output continuous manufacture procedure | |
JP7301876B2 (en) | Driving state estimation system, learning device, estimation device, generation method of state estimator, and estimation method | |
US10899972B1 (en) | Fluid catalytic cracking of petroleum oil with reduced emissions | |
McFarlane et al. | Empirical strategies for open‐loop on‐line optimization | |
CN101859103A (en) | On-line calculation and self-adaptation nonlinear prediction control method of catalytic cracking reaction depth | |
CN115201408A (en) | Method for predicting concentration of sulfur dioxide at desulfurization outlet under all working conditions | |
CN116312869A (en) | Method, device and system for predicting nitrogen oxides in catalytic cracking regenerated flue gas | |
CN111650894A (en) | Bayesian network complex industrial process soft measurement method based on hidden variables | |
TWI829479B (en) | Information processing programs, information processing devices, and information processing methods | |
CN117990881A (en) | Method and system for predicting total sulfur content of gas for desulfurization system | |
Zhang et al. | Enhanced Performance Assessment of Subspace Model‐Based Predictive Controller with Parameters Tuning | |
CN116343932A (en) | Method, device and system for predicting micro-reaction activity of catalyst of catalytic cracking device |
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 |