An integrated real-time learning industrial process soft measurement modeling method based on multi-objective optimization

A multi-objective optimization and industrial process technology, applied in the field of integrated real-time learning soft sensor modeling based on multi-objective optimization, to achieve the effect of dealing with nonlinearity and improving computational efficiency

Active Publication Date: 2019-06-04
KUNMING UNIV OF SCI & TECH
View PDF6 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is how to improve the quality of the historical sample database in

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
  • An integrated real-time learning industrial process soft measurement modeling method based on multi-objective optimization
  • An integrated real-time learning industrial process soft measurement modeling method based on multi-objective optimization
  • An integrated real-time learning industrial process soft measurement modeling method based on multi-objective optimization

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0028]Embodiment 1: as figure 1 As shown, an integrated real-time learning industrial process soft-sensing modeling method based on multi-objective optimization includes the following steps. The first step is to collect industrial process data through a distributed control system or offline detection method, and construct a soft-sensing modeling method for soft-sensing modeling database. Through the mechanism analysis of the industrial process, the auxiliary variables related to the predictor variables are determined.

[0029] Step 2: Normalize the samples in the database and divide them into training sets (D train ∈R J×Q ), validation set (D validate ∈R K×Q ) and the test set (D test ∈R T×Q ). The training set is used for model training, the validation set is used for model parameter optimization, and the test set is used for model performance evaluation.

[0030] The third step: use the multi-objective evolutionary optimization method (NSGA-II algorithm) to optimize ...

Embodiment 2

[0048] Example 2: The following will be further described in conjunction with the industrial process of the debutanizer. The debutanizer is a part of the device for desulfurization and naphtha separation in the industrial refining process, and its goal is to minimize the concentration of butane at the bottom of the tower. However, it is still difficult to realize real-time online detection of butane concentration at present. The on-line prediction of butane concentration by soft sensing method can effectively improve the desulfurization efficiency of debutanizer. According to the mechanism analysis, the x 1 Tower top temperature; x 2 pressure at the top of the tower; x 3 Return flow at the top of the tower; x 4 Outflow of top product; x 5 The sixth layer tray temperature; x 6 tower low temp 1;x 7 The seven monitoring variables of tower bottom temperature 2 are used as auxiliary variables to construct the soft sensor model, and the output variable is butane concentration....

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 discloses an integrated real-time learning industrial process soft measurement modeling method based on multi-objective optimization, and belongs to the field of industrial process softmeasurement modeling. The invention aims to solve the common problems of redundancy and process nonlinearity of industrial data. An evolutionary multi-objective optimization method is adopted to optimize input variables and model structures in a historical sample database, variables irrelevant to quality variables or weakly relevant to the quality variables are removed, the sample quality of the database is improved, and meanwhile, the relation between the model complexity and the prediction precision is effectively balanced. Besides, a part of samples similar to the query samples are selectedfrom the optimized historical sample library to construct a local extreme learning machine model, and a selective integration strategy is adopted to integrate the Pareto optimal solution obtained through multi-objective optimization, so that the nonlinear problem of the industrial process can be effectively solved. By optimizing the modeling data structure and the extreme learning machine model structure, the prediction precision and the calculation efficiency of industrial process soft measurement modeling are improved.

Description

technical field [0001] The invention belongs to the field of industrial process soft sensor modeling, in particular to an integrated real-time learning soft sensor modeling method based on multi-objective optimization. Background technique [0002] In the modern industrial production process, in order to ensure that the product quality meets the increasingly stringent production indicators, it is necessary to conduct real-time online detection of some key process variables or quality variables, so as to realize real-time control of the production process. However, the actual industrial production process is sometimes in harsh environments such as high temperature, high pressure, and strong corrosiveness, so the requirements for sensor performance are extremely strict, and high-performance sensors often have problems such as high cost and difficult maintenance. In addition, Long offline analysis time is another important factor restricting the real-time control of the product...

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): G06F17/50G06N20/00
Inventor 金怀平潘贝
Owner KUNMING 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