Title: A new neural network-based control scheme for fault detection and fault diagnosis in fuzzy multivariate multinomial data
Authors: Mohammad Reza Maleki; Seyed Meysam Mousavi; Amirhossein Amiri
Addresses: Industrial Engineering Department, Faculty of Engineering, Shahed University, Tehran, Iran ' Industrial Engineering Department, Faculty of Engineering, Shahed University, Tehran, Iran ' Industrial Engineering Department, Faculty of Engineering, Shahed University, Tehran, Iran
Abstract: In some multivariate statistical control applications, the data of the process cannot be precise and defined linguistically in practice. Using multivariate control charts in such situations with non-precise data leads to misleading results. In this paper, a new neural network-based monitoring scheme is presented by considering fuzzy multivariate multinomial data. The proposed approach is also able to identify the attribute(s) that cause an out-of-control signal. An application example is provided to evaluate the performance of the proposed approach in detecting different shifts as well as diagnosing the out-of-control attribute quality characteristic(s). The results of applying the proposed approach in both fault detection and the fault diagnosis are satisfactory.
Keywords: linguistic form; fuzzy set theory; multi-attribute processes; multilayer perceptron; MLP neural networks; fault detection; fault diagnosis; multivariate control charts; SPC; statistical process control; fuzzy multinomial data; fuzzy logic.
International Journal of Applied Decision Sciences, 2015 Vol.8 No.2, pp.127 - 144
Received: 06 Aug 2014
Accepted: 25 Jan 2015
Published online: 27 May 2015 *