Title: Vibration signal responses classification in AA 6063 aluminium alloy friction stir welded joint using optimal neural network

Authors: Kalathi Madhivanan; M. SenthilKumar; Rajagopal Ramesh

Addresses: Department of Mechanical Engineering, Anna University, Tamil Nadu, Chennai – 600-025, India ' SKR Engineering College, Poonamallee, Chennai – 123, Tamilnadu, India ' Department of Mechanical Engineering, Sri Venkateshwara College of Engineering, Sriperumbudur, Kaancheepuram District, India

Abstract: Friction-stir-welding (FSW) is a firm stipulation dual procedure and its possessions are reliant on the welding procedure confines. Investigational results are executed by an advance speed revolving apparatus negotiating AA 6063 aluminium alloy 4 mm width laminate substance by FSW apparatus which is formed by HSS M2 substance. The DWT is utilised to decay the vibration signal in different welded dual. Following the signal decay procedure, the signals are indicated to categorisation procedure. Arithmetical and chronological limits of decay vibration signals by wavelet transform have been employed as the input of the FFBN. For improving the classification performance optimises the network structure hidden the layer and hidden neuron optimisation techniques are used. The optimal hidden layer and neuron attained in OGA technique to organise the vibration signals. This experimental and simulation analysis proves that vibration signal analysis method could be used to concern in process condition monitoring in friction stir welding process.

Keywords: wavelet analyses; friction stir welding; FSW; process monitoring; weld defects; neural network; classification process.

DOI: 10.1504/IJBIDM.2017.086985

International Journal of Business Intelligence and Data Mining, 2017 Vol.12 No.4, pp.358 - 382

Received: 28 Sep 2016
Accepted: 19 Jan 2017
Published online: 03 Oct 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article