An Effective Soft-Sensor Method Based on Belief-Rule-Base and Differential Evolution for Tipping Paper Permeability Measurement. (1st November 2018)

• Record Type: Journal Article
• Authors: Hu, Rong; Zhang, Qinli; Qian, Bin; Chang, Leilei; Zhou, Zhijie
• Other Names: Yang Zhile Academic Editor.
• Is Part Of: Complexity. Volume 2018(2018)
• Publication Date: 2018-11-01
• Subjects: Chaotic behavior in systems -- Periodicals; Complexity (Philosophy) -- Periodicals; 003
• DOI: 10.1155/2018/4378701
• Languages: English
• ISSNs: 1076-2787

Abstract

Abstract : The current paper presents a soft-sensor method based on belief-rule-base (BRB) system for solving the problem of tipping paper permeability measurement in the tobacco industry. Firstly, BRB is utilized to establish a model between the feature variables in the tipping paper image and the corresponding paper permeability obtained by the traditional measuring device. Unlike the traditional case of BRB, this paper adds the output attribute as the optimization parameters. In this way, the feasible solution space can be enlarged to obtain an effective BRB model. Second, in order to find the reasonable parameters of BRB in a complex nonconvex solution space, an enhanced differential evolutionary (DE) algorithm is developed to train BRB, which not only embeds a simplex method to stress the balance between the global and local search but also designs a perturbation operation and an adaptively selected mutation strategy to maintain the diversity of search direction. The test results and comparisons based on the data collected from a cigarette factory in China show that the presented method is effective and robust.