CN109636050B - MDS-based prediction model for petrochemical industry production capacity by RBF - Google Patents

MDS-based prediction model for petrochemical industry production capacity by RBF Download PDF

Info

Publication number
CN109636050B
CN109636050B CN201811552864.9A CN201811552864A CN109636050B CN 109636050 B CN109636050 B CN 109636050B CN 201811552864 A CN201811552864 A CN 201811552864A CN 109636050 B CN109636050 B CN 109636050B
Authority
CN
China
Prior art keywords
ethylene
matrix
neural network
basis function
distance
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201811552864.9A
Other languages
Chinese (zh)
Other versions
CN109636050A (en
Inventor
韩永明
武昊
耿志强
朱群雄
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Chemical Technology
Original Assignee
Beijing University of Chemical Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing University of Chemical Technology filed Critical Beijing University of Chemical Technology
Priority to CN201811552864.9A priority Critical patent/CN109636050B/en
Publication of CN109636050A publication Critical patent/CN109636050A/en
Application granted granted Critical
Publication of CN109636050B publication Critical patent/CN109636050B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/04—Manufacturing
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P80/00—Climate change mitigation technologies for sector-wide applications
    • Y02P80/10—Efficient use of energy, e.g. using compressed air or pressurized fluid as energy carrier
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Abstract

The invention discloses a prediction model of MDS-based RBF for petrochemical industry production capacity, which comprises the following steps: obtaining ethylene data; performing dimensionality reduction processing on the ethylene data by using a multidimensional dimension analysis algorithm so that the distance of the ethylene data in a low-dimensional space is the same as that of the ethylene data in a high-dimensional space; obtaining corresponding categories in the low-dimensional space as a training set and a test set of the radial basis function neural network; forming a radial basis function neural network prediction model according to the training set and the test set; and predicting the ethylene yield according to the basis function neural network prediction model. According to the technical scheme provided by the invention, the prediction precision of the ethylene production energy efficiency is improved, so that the effective prediction of the petrochemical industry energy efficiency is realized, the inaccuracy of the traditional neural network model in the petrochemical industry energy efficiency prediction is solved, the energy efficiency of the complex petrochemical industry is improved, and the purposes of energy conservation and emission reduction are realized.

Description

MDS-based prediction model for petrochemical industry production capacity by RBF
Technical Field
The invention relates to the technical field of ethylene production prediction, in particular to a prediction model of MDS-based RBF (radial basis function) for petrochemical industry production capacity.
Background
At present, the petrochemical industry is the industry with the largest energy consumption in China. As the 'grains in the petrochemical industry', ethylene is a basic organic chemical raw material for synthesizing and manufacturing materials, synthetic fibers and other products, and is widely applied to various fields of life, national defense, science and technology and the like. The scale, yield and technical level of the ethylene production process represents a national petrochemical industry level. The petrochemical ethylene production and average fuel consumption in china in 2015 were 11005.2 kilotons/year and 559.06 kilograms/ton of ethylene, respectively. The ethylene production and average fuel consumption of the petroleum and gas group in China were 5032 kilotons/year and 594 kg/ton of ethylene, respectively. Therefore, it is of great significance to improve the energy efficiency in the ethylene production process in the petrochemical industry in terms of production level and production efficiency.
The prior art provides various solutions for energy efficiency prediction in the petrochemical industry, which include prediction by using neural networks such as BP and RBF, but prediction results are inaccurate and incomplete.
Disclosure of Invention
In order to solve the limitations and defects of the prior art, the invention provides a prediction model of an RBF (radial basis function) based on MDS (MDS-based data processing) for petrochemical industry production capacity, which comprises the following steps:
obtaining ethylene data;
performing dimensionality reduction processing on the ethylene data by using a multidimensional dimension analysis algorithm so that the distance of the ethylene data in a low-dimensional space is the same as that of the ethylene data in a high-dimensional space;
obtaining corresponding categories in the low-dimensional space as a training set and a test set of the radial basis function neural network;
forming a radial basis function neural network prediction model according to the training set and the test set;
and predicting the ethylene yield according to the basis function neural network prediction model.
Optionally, the step of performing dimension reduction processing on the ethylene data by using a multidimensional scale analysis algorithm includes:
obtaining a distance matrix D, wherein the distance matrix D is a real symmetric matrix with all 0 diagonal lines;
obtaining the following according to the distance matrix D:
d 2 rs ＝(x r -x s ) T (x r -x s ) (1)
wherein x is r And x s For m dimensional spatial samples (x) 1 ，x 2 …x n ) Element of (a), d rs For the elements of the distance matrix D, x is represented r And x s The distance therebetween;
obtained according to equation (1):
d 2 rs ＝x r T x r +x s T x s -2x r T x s (2)
obtaining an inner product matrix B:
b rs ＝x r T x s (3)
wherein, b rs Is an element of the inner product matrix B;
obtained according to equation (2):
Figure BDA0001911090630000021
Figure BDA0001911090630000022
Figure BDA0001911090630000023
obtained from equations (3) to (6):
b rs ＝a rs -a r1 -a s1 +a 11 (7)
wherein:
Figure BDA0001911090630000024
Figure BDA0001911090630000025
Figure BDA0001911090630000026
Figure BDA0001911090630000027
obtaining the inner product matrix B according to the distance matrix D:
Figure BDA0001911090630000031
wherein the content of the first and second substances,
Figure BDA0001911090630000032
i is a column vector with elements all being 1;
acquiring k maximum eigenvalues of the inner product matrix B and corresponding eigenvectors;
the coordinate matrix in k-dimensional space is obtained as:
Figure BDA0001911090630000033
wherein, V = diag (λ) 1 ，λ 2 ，…λ k ) V is a diagonal matrix composed of k maximum eigenvalues,
Figure BDA0001911090630000034
Λ is an n × k matrix of k orthonormal eigenvectors.
Optionally, the step of forming a radial basis function neural network prediction model according to the training set and the test set includes:
obtaining a radial basis function of the radial basis function neural network as follows:
Figure BDA0001911090630000035
wherein R is i (x) I =1,2, \8230; m, which is the output of the ith node of the hidden layer; sigma i The variance of the ith node of the hidden layer;
the vector distance between the weight vector of neuron j (j =1,2, \8230;, R) and the input vector is:
dist＝x p -ω j (15)
wherein, ω is j Is the weight vector, X, of the neuron j (j =1,2, \ 8230;, R) p Is the p-th input vector;
the input to obtain the neuron j (j =1,2, \8230;, R) from the vector distance is:
Figure BDA0001911090630000036
wherein bj is the deviation;
the output of neuron j (j =1,2, \8230;, R) is obtained from the input of neuron j (j =1,2, \8230;, R):
Figure BDA0001911090630000041
the invention has the following beneficial effects:
the model for predicting the production capacity of the MDS-based RBF in the petrochemical industry comprises the following steps: obtaining ethylene data; performing dimensionality reduction processing on the ethylene data by using a multidimensional dimension analysis algorithm so that the distance of the ethylene data in a low-dimensional space is the same as that of the ethylene data in a high-dimensional space; obtaining corresponding categories in the low-dimensional space as a training set and a test set of the radial basis function neural network; forming a radial basis function neural network prediction model according to the training set and the test set; and predicting the ethylene yield according to the basis function neural network prediction model. According to the technical scheme provided by the invention, the prediction precision of the ethylene production energy efficiency is improved, so that the effective prediction of the petrochemical industry energy efficiency is realized, the inaccuracy of the traditional neural network model in the petrochemical industry energy efficiency prediction is solved, the energy efficiency of the complex petrochemical industry is improved, and the purposes of energy conservation and emission reduction are realized.
Drawings
FIG. 1 is a flow chart of ethylene production according to a first embodiment of the present invention.
FIG. 2 is a flowchart of an MDS-RBF prediction model according to an embodiment of the present invention.
Fig. 3 is a schematic diagram illustrating comparison between the actual value and the predicted value of the ethylene yield according to an embodiment of the present invention.
FIG. 4 is a diagram illustrating a comparison between an MDS-RBF prediction model and other prediction models according to an embodiment of the present invention.
Wherein the reference numerals are: 01. a raw material port; 02. a cracking furnace; 03. a quench oil tower; 04. an acid gas; 05. a gas measurer; 06. a quench water tower; 07. a desiccant; 08. a depropanizer; 09. a debutanizer column; 10. acetylene; 11. hydrogen gas; 12. a cold box; 13. a fuel; 14. a demethanizer column; 15. a deethanizer; 16. mixing the raw materials; 17. rectifying ethylene; 18. an ethylene rectifying tower.
Detailed Description
In order to enable those skilled in the art to better understand the technical solution of the present invention, the prediction model of the MDS-based RBF for petrochemical industry production capacity provided by the present invention is described in detail below with reference to the accompanying drawings.
Example one
In order to better predict the energy efficiency of the petrochemical industry, the embodiment provides a model for predicting the production capacity of the petrochemical industry by using an RBF (radial basis function) based on MDS (multidimensional scaling). In this embodiment, the categories of the raw materials can be automatically obtained in the low-dimensional space, the distance between the raw data in the high-dimensional space is maintained, and then the suitable categories are analyzed and selected in the low-dimensional space as the training set and the testing set of the RBF to predict the ethylene yield.
FIG. 1 is a flow chart of ethylene production according to a first embodiment of the present invention. As shown in fig. 1, the present embodiment uses a multidimensional scaling (MDS) dimension reduction method to reduce raw data from a high-dimensional space to a low-dimensional space and enable the raw data to maintain a distance in the high-dimensional space, so that an appropriate class can be analyzed and selected in the low-dimensional space as a training set and a testing set of a Radial Basis Function (RBF) neural network to predict ethylene production. Referring to fig. 1, 01 represents a raw material port; 02 represents a cracking furnace; 03 represents a quench oil column; 04 represents an acid gas; 05 represents a gasometer; 06 represents a quench water tower; 07 represents a desiccant; 08 represents a depropanizer; 09 represents a debutanizer column; 10 represents acetylene; 11 represents hydrogen; 12 represents a cold box; 13 represents fuel; 14 represents a demethanizer; 15 represents a deethanizer column; 16 represents a mixed raw material; 17 represents the rectification of ethylene; 18 represents an ethylene rectification column.
FIG. 2 is a flowchart of an MDS-RBF prediction model according to an embodiment of the present invention. As shown in fig. 2, firstly, a data set to be processed is selected, ethylene data in the complex petrochemical industry is selected, and then, dimension reduction is performed on the data by using a dimension reduction algorithm of multidimensional scaling (MDS), so that a distance of the data in a low-dimensional space can be consistent with a distance of the data in a high-dimensional space. The essence of the MDS algorithm is to find a matrix B in a low-dimensional space n×k So that it can maintain a high-dimensional matrix A n×m The association between data points.
In this embodiment, a distance matrix D is obtained, where the distance matrix D is a real symmetric matrix whose diagonals are all 0, and the distance matrix D is obtained according to:
d 2 rs ＝(x r -x s ) T (x r -x s ) (1)
wherein x is r And x s For m dimensional spatial samples (x) 1 ，x 2 …x n ) Element of (a), d rs For the elements of the distance matrix D, x is represented r And x s The distance between them.
The present embodiment is obtained according to formula (1):
d 2 rs ＝x r T x r +x s T x s -2x r T x s (2)
this example obtains the inner product matrix B:
b rs ＝x r T x s (3)
wherein, b rs Is an element of the inner product matrix B.
The present embodiment is obtained according to equation (2):
Figure BDA0001911090630000061
Figure BDA0001911090630000062
Figure BDA0001911090630000063
the present embodiment is obtained from equations (3) to (6):
b rs ＝a rs -a r1 -a s1 +a 11 (7)
wherein:
Figure BDA0001911090630000064
Figure BDA0001911090630000065
Figure BDA0001911090630000066
Figure BDA0001911090630000067
in this embodiment, the inner product matrix B is obtained according to the distance matrix D:
Figure BDA0001911090630000068
wherein the content of the first and second substances,
Figure BDA0001911090630000069
i is a column vector with elements all being 1.
In this embodiment, k maximum eigenvalues and corresponding eigenvectors of the inner product matrix B are obtained, and the coordinate matrix in the k-dimensional space is obtained as follows:
Figure BDA0001911090630000071
wherein, V = diag (λ) 1 ，λ 2 ，…λ k ) V is a diagonal matrix composed of k maximum eigenvalues,
Figure BDA0001911090630000072
Λ is an n × k matrix of k orthonormal eigenvectors.
In this embodiment, the clustered result is used as a training set and a test set of a Radial Basis Function (RBF) for prediction. The RBF is a simple and effective feedforward neural network learning algorithm, and consists of three parts, namely an input layer, a hidden layer and an output layer. The input layer comprises a large number of sensors to be connected with the external environment, the hidden layer uses a radial basis function as an excitation function, therefore, the selection of a proper radial basis function is the key for realizing network prediction, and the embodiment selects a Gaussian function as the radial basis function of the RBF.
In this embodiment, the radial basis function of the radial basis function neural network is selected as:
Figure BDA0001911090630000073
wherein R is i (x) I =1,2, \8230; m, which is the output of the ith node of the hidden layer; sigma i Is the variance of the ith node of the hidden layer.
The present embodiment obtains the vector distance between the weight vector of neuron j (j =1,2, \8230;, R) and the input vector as:
dist＝x p -ω j (15)
wherein, ω is j Is the weight vector, X, of the neuron j (j =1,2, \ 8230;, R) p Is the p-th input vector.
The present embodiment obtains the input of neuron j (j =1,2, \8230;, R) from the vector distance as:
Figure BDA0001911090630000074
where bj is the offset.
The present embodiment obtains the output of the neuron j (j =1,2, \8230;, R) from the input of the neuron j (j =1,2, \8230;, R) as:
Figure BDA0001911090630000081
in the training process of the RBF neural network, the determination of the number of the hidden layer neurons is a key problem, when a plurality of input vectors exist, the training is started from 0 neurons, the neurons are automatically added to the network by checking output errors, the input vector corresponding to the maximum error generated by the network is used as a weight vector for each cycle, a new hidden layer neuron is generated, then the error of the new network is checked, and the process is repeated until the error requirement or the maximum number of the hidden layer neurons is reached. Therefore, the RBF neural network has the advantages of self-adaptive determination of the structure, independence of output and initial weight and the like.
Fig. 3 is a schematic diagram illustrating comparison between actual and predicted values of ethylene yield according to an embodiment of the present invention, and fig. 4 is a schematic diagram illustrating comparison between an MDS-RBF prediction model and other prediction models according to an embodiment of the present invention. As shown in fig. 3 and 4, the prediction of ethylene yield in the production of the complex petrochemical industry can be realized by training the RBF neural network, and the effectiveness and accuracy of the prediction model are proved by comparing and analyzing the prediction result of the model with the prediction results of other neural networks. To validate the effectiveness of the MDS-RBF predictive model, it is first necessary to perform a test with a standard data set. This example selects two classical data sets in the UCI, and the detailed description is shown in table 1.
TABLE 1MDS-RBF clustering UCI dataset samples
Figure BDA0001911090630000082
In this embodiment, original data of Energy and Cortex-Nuclear are first subjected to dimension reduction by using MDS, and analyzed and selected in a low-dimensional space to be used as a test set and a training set of RBFs for prediction, and finally, prediction results are compared with other neural networks. The results of the comparison are shown in tables 2 and 3, respectively.
TABLE 2 comparative analysis of different predictive models for Energy
Figure BDA0001911090630000083
As can be seen from Table 2, the error of the MDS-RBF prediction model for Energy is minimal compared to other prediction models. In addition, compared with other prediction models, the prediction accuracy of the MDS-RBF prediction model is respectively improved by 6.2%, 60% and 62%.
TABLE 3 comparative analysis of different predictive models of Cortex-Nuclear
Figure BDA0001911090630000091
As can be seen from Table 3, the error of the MDS-RBF prediction model on Cortex-Nuclear is minimal compared to other prediction models. In addition, compared with other prediction models, the prediction accuracy of the MDS-RBF prediction model is improved by 1.8%, 3.8% and 3.8% respectively.
The embodiment verifies the effectiveness and accuracy of the MDS-RBF prediction model on ethylene yield prediction in complex petrochemical industry production. Meanwhile, the model is applied to modeling analysis of complex petrochemical production data, the yield of energy in the complex petrochemical industry can be accurately predicted, and the energy condition of the complex petrochemical industry can be objectively analyzed. Meanwhile, experimental results prove that the model can reasonably distribute the investment of raw materials by guiding a production department so as to improve the production efficiency and save energy.
The model for predicting the production capacity of the MDS-based RBF in the petrochemical industry provided by the embodiment comprises the following steps: obtaining ethylene data; performing dimensionality reduction processing on the ethylene data by using a multidimensional dimension analysis algorithm so that the distance of the ethylene data in a low-dimensional space is the same as that of the ethylene data in a high-dimensional space; obtaining corresponding categories in the low-dimensional space as a training set and a test set of the radial basis function neural network; forming a radial basis function neural network prediction model according to the training set and the test set; and predicting the ethylene yield according to the basis function neural network prediction model. According to the technical scheme provided by the embodiment, the prediction precision of the ethylene production energy efficiency is improved, so that the effective prediction of the petrochemical industry energy efficiency is realized, the inaccuracy of the traditional neural network model in the petrochemical industry energy efficiency prediction is solved, the energy efficiency of the complex petrochemical industry is improved, and the purposes of energy conservation and emission reduction are realized.
It will be understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principles of the present invention, and the present invention is not limited thereto. It will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the spirit and substance of the invention, and these modifications and improvements are also considered to be within the scope of the invention.

Claims (2)

1. A model for predicting petrochemical industry production capacity by an MDS-based RBF, comprising:
obtaining ethylene data;
performing dimensionality reduction processing on the ethylene data by using a multidimensional dimension analysis algorithm so that the distance of the ethylene data in a low-dimensional space is the same as that of the ethylene data in a high-dimensional space;
obtaining corresponding categories in the low-dimensional space as a training set and a test set of the radial basis function neural network;
forming a radial basis function neural network prediction model according to the training set and the test set;
predicting the ethylene yield according to the basis function neural network prediction model;
the step of performing dimension reduction processing on the ethylene data by using a multi-dimensional dimension analysis algorithm comprises the following steps:
obtaining a distance matrix D, wherein the distance matrix D is a real symmetric matrix with all 0 diagonal lines;
obtaining the following according to the distance matrix D:
d 2 rs ＝(x r -x s ) T (x r -x s ) (1)
wherein x is r And x s For m dimensional spatial samples (x) 1 ，x 2 …x n ) Element of (a), d rs For the elements of the distance matrix D, x is represented r And x s The distance between them;
obtained according to equation (1):
d 2 rs ＝x r T x r +x s T x s -2x r T x s (2)
obtaining an inner product matrix B:
b rs ＝x r T x s (3)
wherein, b rs Is an element of the inner product matrix B;
obtained according to equation (2):
Figure FDA0003945160650000021
Figure FDA0003945160650000022
Figure FDA0003945160650000023
obtained from formula (3) to formula (6):
b rs ＝a rs -a r1 -a s1 +a 11 (7)
wherein:
Figure FDA0003945160650000024
Figure FDA0003945160650000025
Figure FDA0003945160650000026
Figure FDA0003945160650000027
obtaining the inner product matrix B according to the distance matrix D:
Figure FDA0003945160650000028
wherein the content of the first and second substances,
Figure FDA0003945160650000029
i is a column vector with elements all being 1;
acquiring k maximum eigenvalues of the inner product matrix B and corresponding eigenvectors;
the coordinate matrix in k-dimensional space is obtained as:
Figure FDA00039451606500000210
wherein, V = diag (λ) 1 ，λ 2 ，…λ k ) V is a diagonal matrix composed of k maximum eigenvalues,
Figure FDA00039451606500000211
Λ is an n × k matrix of k orthonormal eigenvectors.
2. The model for predicting petrochemical industry production capacity by MDS based RBF according to claim 1, wherein the step of forming a radial basis function neural network prediction model from the training set and the test set comprises:
obtaining a radial basis function of the radial basis function neural network as follows:
Figure FDA0003945160650000031
wherein R is i (x) I =1,2, \8230; m, which is the output of the ith node of the hidden layer; sigma i The variance of the ith node of the hidden layer;
the vector distance between the weight vector of the neuron j (j =1,2, \8230;, R) and the input vector is obtained as:
dist＝x p -ω j (15)
wherein, ω is j Is the weight vector, X, of the neuron j (j =1,2, \ 8230;, R) p Is the p-th input vector;
the input to obtain the neuron j (j =1,2, \8230;, R) from the vector distance is:
Figure FDA0003945160650000032
wherein bj is the deviation;
the output of neuron j (j =1,2, \8230;, R) is obtained from the input of neuron j (j =1,2, \8230;, R):
Figure FDA0003945160650000033
/>
CN201811552864.9A 2018-12-19 2018-12-19 MDS-based prediction model for petrochemical industry production capacity by RBF Active CN109636050B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811552864.9A CN109636050B (en) 2018-12-19 2018-12-19 MDS-based prediction model for petrochemical industry production capacity by RBF

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811552864.9A CN109636050B (en) 2018-12-19 2018-12-19 MDS-based prediction model for petrochemical industry production capacity by RBF

Publications (2)

Publication Number Publication Date
CN109636050A CN109636050A (en) 2019-04-16
CN109636050B true CN109636050B (en) 2023-04-07

Family

ID=66075329

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811552864.9A Active CN109636050B (en) 2018-12-19 2018-12-19 MDS-based prediction model for petrochemical industry production capacity by RBF

Country Status (1)

Country Link
CN (1) CN109636050B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111027799A (en) * 2019-10-24 2020-04-17 山西云时代技术有限公司 National enterprise productivity analysis system

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010017300A1 (en) * 2008-08-06 2010-02-11 Halliburton Energy Services, Inc. Systems and methods employing cooperative optimization-based dimensionality reduction
CN108376302A (en) * 2018-05-23 2018-08-07 贵州大学 The prediction model of ELM complexity petrochemical industry production capacity based on AP clusters

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010017300A1 (en) * 2008-08-06 2010-02-11 Halliburton Energy Services, Inc. Systems and methods employing cooperative optimization-based dimensionality reduction
CN108376302A (en) * 2018-05-23 2018-08-07 贵州大学 The prediction model of ELM complexity petrochemical industry production capacity based on AP clusters

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
PCA-RBF网络预测模型在水质信息预测中的应用;周艳柳等;《测绘与空间地理信息》;20170925(第09期);第1段至108页右栏最后1段，图1 *
基于主成分分析―RBF神经网络模型的备件预测研究;关子明等;《物流科技》;20090410(第04期);全文 *
基于径向基神经网络的太阳能电池缺陷检测;沈凌云等;《发光学报》;20150115(第01期);全文 *
基于监督等距映射和支持向量回归的料位软测量;郭磊等;《仪表技术与传感器》;20160515(第05期);全文 *
基于相似性保持和特征变换的高维数据聚类改进算法;王家耀等;《测绘学报》;20110615(第03期);全文 *
改进的双模型结构RBF神经网络及其应用;李全善等;《化工学报》;20110815(第08期);全文 *

Also Published As

Publication number Publication date
CN109636050A (en) 2019-04-16

Similar Documents

Publication Publication Date Title
CN111126575B (en) Gas sensor array mixed gas detection method and device based on machine learning
CN107403196B (en) Method for predicting concentration of butane at bottom of debutanizer by instant learning modeling based on spectral clustering analysis
Li et al. On the norm of dominant difference for many-objective particle swarm optimization
CN104376434B (en) The integrated evaluating method of cigarette multiple spot processing quality
CN110046377B (en) Selective integration instant learning soft measurement modeling method based on heterogeneous similarity
CN109658988A (en) A kind of Hydrobon catalyst performance prediction method
CN109636050B (en) MDS-based prediction model for petrochemical industry production capacity by RBF
CN102621953A (en) Automatic online quality monitoring and prediction model updating method for rubber hardness
CN111861002A (en) Building cold and hot load prediction method based on data-driven Gaussian learning technology
CN114118292B (en) Fault classification method based on linear discriminant neighborhood preserving embedding
Peng et al. An Effective Fault Diagnosis Approach Based On Gentle AdaBoost and AdaBoost. MH
CN113191614A (en) Polypropylene product quality real-time monitoring method based on slow characteristic regression analysis
CN113379002B (en) Industrial soft measurement method based on depth partial least square model
Li et al. Industrial Process Soft Sensing Based on Random Forest-Optimized Mallow's Cp Variable Selection
CN113191616B (en) Polypropylene product quality abnormity detection method based on double-layer correlation characteristic analysis
CN111292811B (en) Product prediction method and system for aromatic disproportionation production link
Zhang et al. Optimization of process conditions for catalytic preparation of C4 olefins from ethanol based on Stochastic Forest algorithm
Yang et al. Prediction effect of ADMET properties based on machine learning
Wu et al. Real-time prediction of styrene production volume based on machine learning algorithms
Meng et al. Research on the Importance of Feature Screening Algorithm Based on the GBDT-XGBOOST Algorithm
Deng et al. A fault diagnosis method in industrial processes with integrated feature space and optimized random forest
Chu et al. MP 2-Method: A Visual Analysis Method for Material Performance Prediction Supporting Model Selection
Bi et al. Support vector machine qualitative correction algorithm for multicomponent gases based on whale optimization algorithm
CN117612617A (en) Method for constructing initial thermal decomposition temperature prediction model of energetic compound
Li et al. Study on Influencing Factors of Ethanol Coupling to Prepare C4 Olefins

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