Self-adaptive soft measurement method and system based on vine copula quantile regression

A technique of quantile regression and soft measurement, applied in the direction of adaptive control, general control system, control/regulation system, etc., can solve the problem of large model error

Active Publication Date: 2020-10-16
EAST CHINA UNIV OF SCI & TECH
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, when there is heteroscedasticity in the process data, the model error is large

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Self-adaptive soft measurement method and system based on vine copula quantile regression
  • Self-adaptive soft measurement method and system based on vine copula quantile regression
  • Self-adaptive soft measurement method and system based on vine copula quantile regression

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0134] The present invention discloses an adaptive soft sensor method based on vine copula quantile regression, and the specific steps are as follows:

[0135] [Step S1]: According to actual industrial production conditions and expert knowledge, obtain data under normal operating conditions, and select appropriate auxiliary variables for the soft sensor model;

[0136] [Step S2]: Normalize the training data to obtain data that can be used for copula modeling. Normalize the data see formula (15)

[0137]

[0138] Among them, X i is the variable before transformation, X i ’ is the variable after zero-mean standardization, mean(X i ) is the variable X i mean, sd(X i ) is the variable X i The standard deviation of , m is the dimension of the vector X.

[0139] [Step S3]: Establish a quantile regression model using D-vine copula.

[0140] Step 3.1: Equation (16) constructs the analytical model of D-vine copula:

[0141]

[0142] (F(X j |X j+1 ,...,X j+i-1 ), F(X ...

Embodiment 2

[0184] The description of the following examples will help to understand the present invention, but does not limit the content of the present invention. see figure 2 , this embodiment realizes the prediction of numerical examples, and the data of this implementation example is generated according to formula (25).

[0185] x i ~N(0,1)i=1,2

[0186]

[0187] x 4 =sin(x 2 +1)+x 2 x 3

[0188]

[0189] And add error e~N(0,0.01) on each dimension data. The prediction target is y. A total of 500 sets of data were selected, 400 sets were used to train the copula model, and 100 sets were used for testing.

[0190] (1) According to prior information, four auxiliary variables are selected to be x in formula (24) 1 ,x 2 ,x 3 ,x 4 , the key variable is y. (2) Data preprocessing: normalize the zero-mean value of the training samples, and obtain the transformed data as [V, U 1 ,U 2 ,U 3 ,U 4 ]. where V represents the key variable after normalization, U 1 ,U 2 ,U ...

Embodiment 3

[0203] see Figure 4 , the present embodiment realizes the prediction (PER) of the degree of ethylene cracking in the ethylene cracking process. The data of this implementation example comes from the SRT-III model ethylene cracking furnace, and the prediction target is the cracking rate of ethylene. Ratio) to indicate that 500 sets of data under normal working conditions were selected, 400 sets were used to train the copula model, and 100 sets were used for testing.

[0204] (1) According to the prior information, four auxiliary variables are selected: the average outlet temperature of the cracking furnace x 1 , the density of pyrolysis feedstock x 2 , total feed x 3 and steam hydrocarbon ratio x 4 . The key variable y is the lysis depth indicator PER.

[0205] (2) Data preprocessing: normalize the zero mean value of the training samples to obtain the transformed data [V, U 1 ,U 2 ,U 3 ,U 4 ].

[0206] (3) Using training samples [V, U 1 ,U 2 ,U 3 ,U 4 ] Build the...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a self-adaptive soft measurement method and system based on vine copula quantile regression, and the method comprises the following steps: obtaining data under a normal operation condition, and selecting a proper auxiliary variable; normalizing training data; establishing a quantile regression model by utilizing the D-vine copula; calculating an average prediction interval difference between a conditional quantile function and the training sample according to the training sample; collecting and standarizing auxiliary variables of a to-be-predicted sample online; determining the conditional quantile function; calculating a function value of each conditional quantile according to a set quantile; performing inverse normalization on the conditional quantile function value to obtain a median and a confidence interval of a final prediction value; and calculating a confidence interval difference of the prediction value, and comparing whether the confidence interval difference of the prediction value exceeds an average prediction interval difference of the training sample.

Description

technical field [0001] The invention belongs to the technical field of process control, in particular to an adaptive soft sensor method based on vine copula quantile regression; meanwhile, the invention also relates to an adaptive soft sensor system based on vine copula quantile regression. Background technique [0002] As the production scale of modern industry gradually increases, the production process becomes more complex. In order to ensure the normal and orderly operation of the factory's production process, workers can monitor the system status in time, which is very important for the measurement and control of product quality. Limited by technology and technology, it is difficult for them to be detected online in real time by hardware sensors. In the actual production process, manual offline analysis is usually used, that is, samples are taken every few hours and sent to the laboratory for manual analysis. However, the method of manual offline analysis is costly and...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G05B13/04
CPCG05B13/042Y02P90/02
Inventor 李绍军倪佳能周洋刘漫丹贾琼田一彤李雪梅王世豪
Owner EAST CHINA UNIV OF SCI & TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap