CN113035287A - Prediction method for steam thermal cracking process - Google Patents
Prediction method for steam thermal cracking process Download PDFInfo
- Publication number
- CN113035287A CN113035287A CN202110254597.2A CN202110254597A CN113035287A CN 113035287 A CN113035287 A CN 113035287A CN 202110254597 A CN202110254597 A CN 202110254597A CN 113035287 A CN113035287 A CN 113035287A
- Authority
- CN
- China
- Prior art keywords
- basis function
- data
- regression model
- thermal cracking
- steam thermal
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/10—Analysis or design of chemical reactions, syntheses or processes
Landscapes
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Crystallography & Structural Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Feedback Control In General (AREA)
Abstract
The invention belongs to the technical field of steam thermal cracking, and relates to a prediction method of a steam thermal cracking process. The method comprises the steps of generating a basis function set related to industrial data by obtaining a plurality of groups of corresponding production data of a plurality of variables in the steam thermal cracking process, and establishing a corresponding multivariate self-adaptive spline regression model. And inputting the independent variable in the steam thermal cracking process acquired in real time into the multivariate self-adaptive spline regression model, and outputting to obtain a dependent variable predicted value of the steam thermal cracking process so as to realize prediction of the steam thermal cracking process. As the multivariate self-adaptive spline regression model has the capability of self-adaptive screening of industrial variables, the process of establishing the model has interpretability. The modeling and predicting process is on an industrial data set, has the capability of efficiently, quickly and accurately modeling a small amount of data, can control and optimize the steam thermal cracking process, improves the production operation level, ensures the safe and stable operation of an industrial device, and realizes the quality improvement and efficiency improvement of the process.
Description
Technical Field
The invention belongs to the technical field of steam thermal cracking, relates to a prediction method of a steam thermal cracking process, and particularly relates to a prediction method of a steam thermal cracking process based on multivariate self-adaptive spline regression.
Background
Steam thermal cracking is an important secondary petroleum processing process, in which various distillate oils are decomposed into products such as olefins and aromatics by the heat provided by the combustion of fuel gas and superheated steam. The ethylene yield is regarded as one of important marks for measuring the development level of the national petrochemical industry, the total ethylene yield reaches 2052 ten thousand tons in 2019, about 92 percent of ethylene products come from the steam thermal cracking process, and simultaneously, a large number of ethylene plants are newly built and expanded. The operating level of the steam thermal cracking process will determine the final product distribution. The process is accurately modeled, simulated and predicted, and the method plays an important role in improving the quality and the efficiency of the process.
In the prior art, a mechanism model is mainly used for simulating a steam thermal cracking process, but the mechanism model has long modeling period and high requirement on the quality of industrial data, and under the conditions that the demand of the model is large and the industrial data is difficult to collect, the mechanism model is slow to popularize and difficult to apply. However, machine learning models using methods such as neural networks have a problem of poor extensibility and interpretability. Therefore, a high-efficiency, accurate and high-interpretability product yield prediction model is developed, and the method has important practical significance for the steam thermal cracking process.
Disclosure of Invention
The invention aims to provide a prediction method for a steam thermal cracking process, aiming at the defects of low modeling efficiency and poor interpretability of a machine learning model of a steam thermal cracking mechanism model in the prior art, and the prediction method is used for establishing a product yield prediction model with an explicit expression according to the characteristics of an industrial data set so as to facilitate the continuous development of a subsequent control and optimization model and improve the device benefit.
The invention provides a prediction method of a steam thermal cracking process, which comprises the following steps:
acquiring historical data in the steam thermal cracking process, generating a basis function set related to the historical data, and establishing a corresponding multivariate self-adaptive spline regression model; and acquiring production operation conditions in the steam thermal cracking process in real time as independent variables X, inputting the independent variables X into a multivariate self-adaptive spline regression model, outputting to obtain a dependent variable predicted value of the steam thermal cracking process, and realizing prediction of the steam thermal cracking process.
The invention provides a prediction method for a steam thermal cracking process, which has the advantages that:
the prediction method for the steam thermal cracking process generates a basis function set related to historical data samples by acquiring a plurality of groups of production historical data corresponding to a plurality of variables in the steam thermal cracking process, and establishes a corresponding multivariate self-adaptive spline regression model. And predicting the independent variable in the steam thermal cracking process collected in real time by using the multivariate self-adaptive spline regression model to obtain the type and the content of the thermal cracking product in the steam thermal cracking process. The multivariate self-adaptive spline regression model related in the method has the self-adaptive screening capacity to the industrial variables, so the establishment process of the model has interpretability. Because the modeling and predicting process is based on industrial data, the method has the capability of efficiently, quickly and accurately modeling a small amount of data, can control and optimize the steam thermal cracking process, improves the steam thermal cracking production operation level, ensures the safe and stable operation of industrial devices, and realizes the quality improvement and efficiency improvement of the process.
Drawings
FIG. 1 is a block flow diagram of a predictive method of steam thermal cracking in accordance with the present invention.
FIG. 2 is a schematic view of a steam thermal cracking reactor according to the present invention.
FIG. 3 is a generalized cross validation function GCV trend graph in an embodiment of the present invention.
Detailed Description
The invention provides a prediction method of a steam thermal cracking process, which comprises the following steps:
acquiring historical data in the steam thermal cracking process, generating a basis function set related to the historical data, and establishing a corresponding multivariate self-adaptive spline regression model; and acquiring production operation conditions in the steam thermal cracking process in real time as independent variables X, inputting the independent variables X into a multivariate self-adaptive spline regression model, outputting to obtain a dependent variable predicted value of the steam thermal cracking process, and predicting the steam thermal cracking process.
The flow chart of the prediction method of the steam thermal cracking process of the invention is shown in figure 1, and specifically comprises the following steps:
(1) setting the type i of the steam thermal cracking product and the content w of the steam thermal cracking productiIs a dependent variable Y, where wiSetting physical properties of the steam pyrolysis feedstock (including the group composition PIONA of the feedstock, the density D of the feedstock, the enmerdian distillation range ASTM D86 of the feedstock) and production operating conditions including the feed amount F to the reactor, the outlet temperature COT of the reactor, the outlet pressure COP of the reactor, the cross-section temperature CIT of the reactor, the cross-section pressure CIP of the reactor as independent variables X for the content of the ith product, as shown in fig. 2;
(2) acquiring independent variable X and dependent variable Y in the steam thermal cracking process from a historical record, wherein the independent variable X has P types, the dependent variable Y has I types, and each type has n groups of data, and preprocessing the P + I types and n groups of data;
specifically, the p variables comprise physical properties of raw materials, operating parameters of the reactor, and types and contents of cracking products, and the n groups of data are subjected to data cleaning, namely, if a missing variable exists in one group of data, the group of data is deleted, and if the variable in the data deviates from the overall data distribution, the group of data is deleted.
(3) Respectively establishing a plurality of basis functions g (X) related to the independent variable X according to the n groups of data in the independent variable X in the step (2), and recording a set of the plurality of basis functions g (X) as C1；
For each independent variable XjAll n groups of data of the independent variable are used as nodes to establish a basis function, and a basis function g (X) corresponding to the datap) The specific expression is as follows:
where p is the type of the independent variable X, t is the independent variable value in n groups of data, and the basis function g (X)p) Appears in pairs with respect to the value t, to
And
it is shown that,
after a basis function is taken from all data in the independent variable X, a basis function set C is obtained1,
Wherein n is the number of data, P is the kind of independent variable X, and P is 1, 2 … … P;
(4) adopting a multivariate self-adaptive spline regression model and utilizing the basis function set C in the step (3)1Training the multivariate self-adaptive spline regression model to obtain the multivariate self-adaptive spline regression model for predicting the dependent variable Y, and the method comprises the following steps of:
(4-1) setting the raising power number k of a multivariate self-adaptive spline regression model;
(4-2) set of basis functions C according to step (3)1A basis function of
And
and (3) as independent variables, carrying out forward stepwise regression on the dependent variable Y to obtain a regression model based on a first-order basis function, wherein the process is as follows:
(4-2-1) establishing each basis function g (X)p) First regression equation with dependent variable Y, i.e. Y ═ betatpg(Xp)+β0Wherein g (X)p) A basis function, β, for the value t of the argument X on the class ptpIs a basis function g (X)p) Is determined by a least squares calculation formula as betatp＝(g(Xp)Tg(Xp))-1g(Xp)TY, where superscript T is the matrix transpose, β0Constant terms of the first one-dimensional regression equation; respectively calculating each first regression equationF statistic of the equation
Where n is the number of data in the data set and m is the equation Y ═ betatpgt(Xp)+β0The dimension of the medium independent variable X, the SSR is the residual square sum, and the calculation formula of the SSR is
Wherein
Is an estimate of the dependent variable Y,
is the average value of Y, SSE is the regression average value, and the calculation formula of SSE is
The superscript T is matrix transposition, the largest F statistic is selected from the F statistics of the first one-time regression equations and is recorded as
k is the kth regression equation corresponding to the largest F statistic; let the significance critical value of the F statistic be alpha, and the maximum value of the F statistic
Make a judgment if
Stopping screening, and proceeding to step (4-3), if so
Then will be connected with
The corresponding basis functions are put into a set of basis functions, which are recorded as
Carrying out the step (4-2-2) wherein FαObtaining an F distribution value corresponding to the critical value alpha by inquiring an F distribution value table, wherein n is the number of data in the data set;
(4-2-2) separately combining the sets C1Each basis function and set of basis functions remaining in
The step (4-2-1) is carried out, and the maximum value in the plurality of F statistics is recorded as
To pair
Make a judgment if
Stopping screening, and proceeding to step (4-3), if so
The basis function in the second linear regression equation with the largest F statistic is put into the basis function set, and the basis function set is recorded as
Wherein FαPerforming step (4-2-3) for the F distribution value corresponding to the critical value alpha and n for the number of data in the data set;
(4-2-3) setting a threshold value of the F statistic of the first Primary regression equation
Repeating the step (4-2-2) until
Wherein q is a set of basis functions C1The number of the basis functions in the data set is n, and a forward regression model based on the first-order basis function is obtained
Where X is all data for the argument, M is the total number of selected basis functions in the model,
is a basis function in the model, betamIs the regression coefficient of the basis function, beta0Is a constant term in a forward regression model of the first order basis functions; and will be
Is recorded as
The set of basis functions has q functions;
(4-3) carrying out backward deletion process on the forward regression model of the first-order basis function in the step (4-2) to obtain a backward regression model based on the first-order basis function, wherein the specific process is as follows:
(4-3-1) calculating the value of the generalized cross validation function GCV of the forward regression model of the first-order basis function in the step (4-2), and marking the value as GCVqThe calculation formula is
Wherein wizThe content of the steam thermal cracking product type i in the data z; f. ofiz(Xp) The predicted value of the forward regression model of the first-order basis function on the data z is shown, and n is the number of data in the data set; g is the number of effective parameters, and G is 2q + 1;
(4-3-2) arbitrarily deleting one basis function from the q basis functions of the forward regression model of the first basis functions in the step (4-2) to obtain q backward regression models of the first basis functions containing q-1 basis functions, and respectively calculating the values of the generalized cross validation functions GCV of the q models, wherein the minimum value is GCVq-1For the GCVq-1Make a judgment ifGCVq-1<GCVqThen, the regression model is judged to be more optimal, and the corresponding set of basis functions is recorded as
The corresponding first order basis function is backward regression modeled as
(4-3-3) repeating the step (4-3-2) until the value of GCV is no longer reduced, i.e., GCVq-1<GCVqThe set of basis functions at this time is denoted as h1(X) obtaining a multivariate adaptive spline regression model of the first order basis function as f1(X)；
(4-4) set h of Functions to the step (4-3-3)1(X) performing an ascending iteration, i.e. set h of basis functions1All items in (X) and set C of step (3)1All basis functions g (X) in (1)p) Multiplying to obtain a basis function set C2Set the basis functions to C2The basis function in (4) is used as a new basis function, and the step (4-2) and the step (4-3) are repeated to obtain a basis function set h2(X) obtaining a multivariate adaptive spline regression model f of the quadratic base function2(X)；
(4-5) setting a threshold value of a generalized cross validation function GCV, repeating the power raising process in the step (4-4), and if the corresponding GCV of the multivariate self-adaptive spline regression model reaches the threshold value or the power raising times of the multivariate self-adaptive spline regression model reaches the power raising times set in the step (4-1), confirming that the multivariate self-adaptive spline regression model is the multivariate self-adaptive spline regression model for predicting the dependent variable Y;
(5) and (4) acquiring an independent variable X in the steam thermal cracking process in real time, inputting the independent variable X into the multivariate self-adaptive spline regression model in the step (4), and outputting to obtain a dependent variable predicted value of the steam thermal cracking process so as to realize prediction of the steam thermal cracking process.
In one embodiment of the present invention, the dependent variable is selected as ethylene in the cracked gas, the content is the dependent variable, and the finally established multivariate adaptive regression model is:
Y＝31.5290+0.0066×max(0,819.86-COT)-0.0487×max(0,CIT-599.60)
+0.0214×max(0,169.78-CIP)-0.0388×max(0,F-37496.87)
+0.0093×max(0,CIT-599.60)×max(0,18.861-NAP_N)
+0.0138×max(0,819.86-COT)×max(0,34.735-NAP_I)
the calculations show that the ethylene content Y in the cracked gas is closely related to the reactor outlet temperature (COT), the reactor cross-section temperature (CIT), the reactor cross-section pressure (CIP), the total reactor feed (F), the naphthene content in the naphtha feed (NAP _ N), and the isoparaffin content in the naphtha feed (NAP _ I). In the multivariate adaptive spline regression process, the change trend of the generalized cross validation function GCV is shown in FIG. 3 as the number of basis functions increases.
In the embodiment, a plurality of groups of corresponding production data of a plurality of variables in the steam thermal cracking process are obtained, a basis function set related to an industrial data sample is generated, and a corresponding multivariate self-adaptive spline regression model is established. The model building process is interpretable due to the capability of the model to carry out adaptive screening on the industrial variables. The modeling and predicting process is on an industrial data set, has the capability of efficiently, quickly and accurately modeling a small amount of data, can control and optimize the steam thermal cracking process, improves the production operation level, ensures the safe and stable operation of an industrial device, and realizes the quality improvement and efficiency improvement of the process.
The foregoing is merely a preferred embodiment of the invention, which is intended to be illustrative and not limiting. It will be understood by those skilled in the art that various changes, modifications and equivalents may be made therein without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (2)
1. A method for predicting a thermal steam cracking process, comprising:
acquiring historical data in the steam thermal cracking process, generating a basis function set related to the historical data, and establishing a corresponding multivariate self-adaptive spline regression model; and acquiring production operation conditions in the steam thermal cracking process in real time as independent variables X, inputting the independent variables X into a multivariate self-adaptive spline regression model, outputting to obtain a dependent variable predicted value of the steam thermal cracking process, and realizing prediction of the steam thermal cracking process.
2. A method of predicting a thermal steam cracking process, the method comprising the steps of:
(1) setting the type i of the steam thermal cracking product and the content w of the steam thermal cracking productiIs a dependent variable Y, where wiSetting the physical property and production operation condition of the steam thermal cracking raw material as independent variables X for the content of the ith product;
(2) acquiring independent variable X and dependent variable Y in the steam thermal cracking process from a historical record, wherein the independent variable X has P types, the dependent variable Y has I types, and each type has n groups of data, and preprocessing the P + I types and n groups of data;
(3) respectively establishing a plurality of basis functions g (X) related to the independent variable X according to the n groups of data in the independent variable X in the step (2), and recording a set of the plurality of basis functions g (X) as C1；
For each independent variable XjAll n groups of data of the independent variable are used as nodes to establish a basis function, and a basis function g (X) corresponding to the datap) The specific expression is as follows:
where p is the type of the independent variable X, t is the independent variable value in n groups of data, and the basis function g (X)p) Appears in pairs with respect to the value t, to
And
it is shown that,
after a basis function is taken from all data in the independent variable X, a basis function set C is obtained1,
Wherein n is the number of data, P is the kind of independent variable X, and P is 1, 2 … … P;
(4) adopting a multivariate self-adaptive spline regression model and utilizing the basis function set C in the step (3)1Training the multivariate self-adaptive spline regression model to obtain the multivariate self-adaptive spline regression model for predicting the dependent variable Y, and the method comprises the following steps of:
(4-1) setting the raising power number k of a multivariate self-adaptive spline regression model;
(4-2) set of basis functions C according to step (3)1A basis function of
And
and (3) as independent variables, carrying out forward stepwise regression on the dependent variable Y to obtain a regression model based on a first-order basis function, wherein the process is as follows:
(4-2-1) establishing each basis function g (X)p) First regression equation with dependent variable Y, i.e. Y ═ betatpg(Xp)+β0Wherein g (X)p) A basis function, β, for the value t of the argument X on the class ptpIs a basis function g (X)p) Is determined by a least squares calculation formula as betatp＝(g(Xp)Tg(Xp))-1g(Xp)TY, where superscript T is the matrix transpose, β0Constant terms of the first one-dimensional regression equation; respectively calculating the F statistic of each first one-time regression equation, wherein the calculation formula is
Where n is the number of data in the data set and m is the equation Y ═ betatpgt(Xp)+β0The dimension of the medium independent variable X, the SSR is the residual square sum, and the calculation formula of the SSR is
Wherein
Is an estimate of the dependent variable Y,
is the average value of Y, SSE is the regression average value, and the calculation formula of SSE is
The superscript T is matrix transposition, the largest F statistic is selected from the F statistics of the first one-time regression equations and is recorded as
k is the kth regression equation corresponding to the largest F statistic; let the significance critical value of the F statistic be alpha, and the maximum value of the F statistic
Make a judgment if
Stopping screening, and proceeding to step (4-3), if so
Then will be connected with
Putting corresponding basis functions into a set of basis functionsIn, the basis function set is recorded as
Carrying out the step (4-2-2) wherein FαObtaining an F distribution value corresponding to the critical value alpha by inquiring an F distribution value table, wherein n is the number of data in the data set;
(4-2-2) separately combining the sets C1Each basis function and set of basis functions remaining in
The step (4-2-1) is carried out, and the maximum value in the plurality of F statistics is recorded as
To pair
Make a judgment if
Stopping screening, and proceeding to step (4-3), if so
The basis function in the second linear regression equation with the largest F statistic is put into the basis function set, and the basis function set is recorded as
Wherein FαPerforming step (4-2-3) for the F distribution value corresponding to the critical value alpha and n for the number of data in the data set;
(4-2-3) setting a threshold value of the F statistic of the first Primary regression equation
Repeating the step (4-2-2) until
Wherein q is a set of basis functions C1The number of the basis functions in the data set is n, and a forward regression model based on the first-order basis function is obtained
Where X is all data for the argument, M is the total number of selected basis functions in the model,
is a basis function in the model, betamIs the regression coefficient of the basis function, beta0Is a constant term in a forward regression model of the first order basis functions; and will be
Is recorded as
The set of basis functions has q functions;
(4-3) carrying out backward deletion process on the forward regression model of the first-order basis function in the step (4-2) to obtain a backward regression model based on the first-order basis function, wherein the specific process is as follows:
(4-3-1) calculating the value of the generalized cross validation function GCV of the forward regression model of the first-order basis function in the step (4-2), and marking the value as GCVqThe calculation formula is
Wherein wizThe content of the steam thermal cracking product type i in the data z; f. ofiz(Xp) The predicted value of the forward regression model of the first-order basis function on the data z is shown, and n is the number of data in the data set; g is the number of effective parameters, and G is 2q + 1;
(4-3-2) arbitrarily deleting one of the q basis functions of the forward regression model of the first basis function in the step (4-2)The basis functions are used for obtaining q backward regression models containing the linear basis functions of q-1 basis functions, the values of the generalized cross validation functions GCV of the q models are respectively calculated, and the minimum value is recorded as GCVq-1For the GCVq-1Make a judgment if GCVq-1<GCVqThen, the regression model is judged to be more optimal, and the corresponding set of basis functions is recorded as
The corresponding first order basis function is backward regression modeled as
(4-3-3) repeating the step (4-3-2) until the value of GCV is no longer reduced, i.e., GCVq-1<GCVqThe set of basis functions at this time is denoted as h1(X) obtaining a multivariate adaptive spline regression model of the first order basis function as f1(X)；
(4-4) set h of Functions to the step (4-3-3)1(X) performing an ascending iteration, i.e. set h of basis functions1All items in (X) and set C of step (3)1All basis functions g (X) in (1)p) Multiplying to obtain a basis function set C2Set the basis functions to C2The basis function in (4) is used as a new basis function, and the step (4-2) and the step (4-3) are repeated to obtain a basis function set h2(X) obtaining a multivariate adaptive spline regression model f of the quadratic base function2(X)；
(4-5) setting a threshold value of a generalized cross validation function GCV, repeating the power raising process in the step (4-4), and if the corresponding GCV of the multivariate self-adaptive spline regression model reaches the threshold value or the power raising times of the multivariate self-adaptive spline regression model reaches the power raising times set in the step (4-1), confirming that the multivariate self-adaptive spline regression model is the multivariate self-adaptive spline regression model for predicting the dependent variable Y;
(5) and (4) acquiring an independent variable X in the steam thermal cracking process in real time, inputting the independent variable X into the multivariate self-adaptive spline regression model in the step (4), and outputting to obtain a dependent variable predicted value of the steam thermal cracking process so as to realize prediction of the steam thermal cracking process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110254597.2A CN113035287B (en) | 2021-03-09 | 2021-03-09 | Prediction method for steam thermal cracking process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110254597.2A CN113035287B (en) | 2021-03-09 | 2021-03-09 | Prediction method for steam thermal cracking process |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113035287A true CN113035287A (en) | 2021-06-25 |
| CN113035287B CN113035287B (en) | 2022-11-04 |
Family
ID=76467166
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110254597.2A Active CN113035287B (en) | 2021-03-09 | 2021-03-09 | Prediction method for steam thermal cracking process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113035287B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113779865A (en) * | 2021-08-09 | 2021-12-10 | 清华大学 | Modeling method and device for steam thermal cracking process, electronic equipment and readable storage medium thereof |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101414158A (en) * | 2008-11-19 | 2009-04-22 | 华东理工大学 | Method for optimizing cracking reaction operating condition of ethylene cracking furnace |
| WO2010016110A1 (en) * | 2008-08-05 | 2010-02-11 | 富士通株式会社 | Method of creating estimation model by means of multiple regression analysis, creation system and creation program |
| US20100070441A1 (en) * | 2007-03-27 | 2010-03-18 | Fujitsu Limited | Method, apparatus, and program for generating prediction model based on multiple regression analysis |
| CN101727609A (en) * | 2008-10-31 | 2010-06-09 | 中国石油化工股份有限公司 | Pyrolyzate yield forecasting method based on support vector machine |
| US20110173144A1 (en) * | 2010-01-14 | 2011-07-14 | Shan Jerry Z | System and method for constructing forecast models |
-
2021
- 2021-03-09 CN CN202110254597.2A patent/CN113035287B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100070441A1 (en) * | 2007-03-27 | 2010-03-18 | Fujitsu Limited | Method, apparatus, and program for generating prediction model based on multiple regression analysis |
| WO2010016110A1 (en) * | 2008-08-05 | 2010-02-11 | 富士通株式会社 | Method of creating estimation model by means of multiple regression analysis, creation system and creation program |
| CN101727609A (en) * | 2008-10-31 | 2010-06-09 | 中国石油化工股份有限公司 | Pyrolyzate yield forecasting method based on support vector machine |
| CN101414158A (en) * | 2008-11-19 | 2009-04-22 | 华东理工大学 | Method for optimizing cracking reaction operating condition of ethylene cracking furnace |
| US20110173144A1 (en) * | 2010-01-14 | 2011-07-14 | Shan Jerry Z | System and method for constructing forecast models |
Non-Patent Citations (3)
| Title |
|---|
| HEDAN ZHANG 等: "SSD Thermal Throttling Prediction using Improved Fast Prediction Model", 《2019 18TH IEEE INTERSOCIETY CONFERENCE ON THERMAL AND THERMOMECHANICAL PHENOMENA IN ELECTRONIC SYSTEMS (ITHERM)》 * |
| 张利军: "裂解反应模型的建立与应用研究", 《中国优秀博士学位论文全文数据库 信息科技辑》 * |
| 鲍长庚等: "基于多元自适应回归样条的光伏并网系统日输出功率预测", 《电力系统保护与控制》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113779865A (en) * | 2021-08-09 | 2021-12-10 | 清华大学 | Modeling method and device for steam thermal cracking process, electronic equipment and readable storage medium thereof |
| CN113779865B (en) * | 2021-08-09 | 2024-04-02 | 清华大学 | Modeling method and device for steam thermal cracking process, electronic equipment and readable storage medium thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113035287B (en) | 2022-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | A hybrid deep learning and mechanistic kinetics model for the prediction of fluid catalytic cracking performance | |
| CN108469805B (en) | Distributed dynamic process monitoring method based on dynamic optimal selection | |
| Michalopoulos et al. | Modelling of an industrial fluid catalytic cracking unit using neural networks | |
| CN104361153B (en) | A kind of method for predicting RFCC settler coking amount | |
| CN101414158A (en) | Method for optimizing cracking reaction operating condition of ethylene cracking furnace | |
| CN113035287B (en) | Prediction method for steam thermal cracking process | |
| CN112131785B (en) | Screening and optimizing method of methanation nickel-based catalyst formula based on ANN-NSGA-II | |
| CN103605325A (en) | Complete period dynamic optimization method for industrial ethylene cracking furnace and based on surrogate model | |
| Bhuyan et al. | Comparative assessment of artificial neural network and response surface methodology for evaluation of the predictive capability on bio-oil yield of Tithonia diversifolia pyrolysis | |
| CN105740960A (en) | Optimization method of industrial hydrocracking reaction condition | |
| CN113420500B (en) | Intelligent atmospheric and vacuum system | |
| CN110766234B (en) | Cement cooling process grate down pressure prediction method based on information fusion | |
| Vasseghian et al. | Artificial Intelligent modeling and optimizing of an industrial hydrocracker plant | |
| Rosli et al. | Review of neural network modelling of cracking process | |
| CN114492988A (en) | Method and device for predicting product yield in catalytic cracking process | |
| CN114783541A (en) | Method, system and readable storage medium for predicting catalytic cracking product yield | |
| WO2018042337A1 (en) | Olefin plant optimization using steady-state models | |
| Nian et al. | Strategy of changing cracking furnace feedstock based on improved group search optimization | |
| Moghadassi et al. | Prediction of research octane number in catalytic naphtha reforming unit of Shazand Oil Refinery | |
| CN113420498B (en) | AI modeling method of atmospheric and vacuum distillation unit | |
| CN112800672A (en) | Boiler fouling coefficient evaluation method, system, medium and electronic equipment | |
| Monteiro et al. | Application of artificial neural networks for identification of catalysts used in thermogravimetry lignocellulosic biomass | |
| Pradhan et al. | Machine learning application in slow pyrolysis of biomass to predict biochar yield and quality | |
| Lv et al. | Developing a novel method for estimating the speed of sound in biodiesel known as grey wolf optimizer support vector machine algorithm | |
| Chohan et al. | An integrated approach of artificial neural networks and polynomial chaos expansion for prediction and analysis of yield and environmental impact of oil shale retorting process under uncertainty |
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 |