A type-2 fuzzy expert system based on a hybrid inference method for steel industry

In this paper, a novel type-2 fuzzy expert system for prediction the amount of reagents in desulfurization process of a steel industry in Canada is developed. In this model, the new interval type-2 fuzzy c-regression clustering algorithm for structure identification phase of Takagi–Sugeno (T–S) syst...

Full description

Saved in:
Bibliographic Details
Published in:International journal of advanced manufacturing technology Vol. 71; no. 5-8; pp. 857 - 885
Main Authors: Fazel Zarandi, M. H., Gamasaee, R., Turksen, I. B.
Format: Journal Article
Language:English
Published: London Springer London 01.03.2014
Springer Nature B.V
Subjects:
Adaptive systems
Algorithms
CAE) and Design
Canonical forms
Clustering
Computer-Aided Engineering (CAD
Engineering
Error reduction
Expert systems
Fuzzy systems
Hybrid systems
Industrial and Production Engineering
Industrial development
Inference
Iron and steel industry
Mechanical Engineering
Media Management
Neural networks
Original Article
Probabilistic models
Reagents
Statistical analysis
Statistical tests
Steel industry
Interval type-2 fuzzy hybrid system
Fuzzy disjunctive normal forms
IT2F c-regression clustering
Steel industry
Gaussian mixture model
Fuzzy conjunctive normal forms
ISSN:0268-3768, 1433-3015
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, a novel type-2 fuzzy expert system for prediction the amount of reagents in desulfurization process of a steel industry in Canada is developed. In this model, the new interval type-2 fuzzy c-regression clustering algorithm for structure identification phase of Takagi–Sugeno (T–S) systems is presented. Gaussian Mixture Model is used to generate partition matrix in clustering algorithm. Then, an interval type-2 hybrid fuzzy system, which is the combination of Mamdani and Sugeno method, is proposed. The new hybrid inference system uses fuzzy disjunctive normal forms and fuzzy conjunctive normal forms for aggregation of antecedents. A statistical test, which uses least square method, is implemented in order to select variables. In order to validate our method, we develop three system modeling techniques and compare the results with our proposed interval type-2 fuzzy hybrid expert system. These techniques are multiple regression, type-1 fuzzy expert system, and interval type-2 fuzzy TSK expert system. For tuning parameters of the system, adaptive-network-based fuzzy inference system is used. Finally, neural network is utilized in order to reduce error of the system. The results show that our proposed method has less error and high accuracy.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0268-3768
1433-3015
DOI:10.1007/s00170-013-5372-4