Title: Thermal stress deformation prediction for rotary air-preheater rotor using deep learning approach

Authors: Jing Xin; Rong Yu; Ding Liu; Youmin Zhang

Addresses: Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing, Xi'an University of Technology, Shaanxi 710048, China; Shaanxi Key Laboratory of Integrated and Intelligent Navigation, China ' Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing, Xi'an University of Technology, Shaanxi 710048, China ' Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing, Xi'an University of Technology, Shaanxi 710048, China ' Department of Mechanical, Industrial and Aerospace Engineering, Concordia University, Quebec, Canada

Abstract: Failures often occur in the seal clearance measuring sensor due to the harsh operating conditions of the rotary air-preheater in power plant boilers. Therefore, it is necessary to predict the rotor deformation to eliminate the effects of failures on the gap control system. An air-preheater rotor thermal stress deformation prediction method is proposed in this paper based on deep learning. Firstly, a stacked auto-encoder (SAE) is constructed and trained to learn the feature information which is hidden within the input data (the temperature of flue gas side inlet, air side outlet, flue gas side outlet, air side inlet); then, an Elman neural network is constructed and trained using the output of the encoder part of the well trained stacked auto-encoder to predict rotor deformation. Simulation and experimental results show that the proposed SAE-Elman prediction method can obtain the effective feature representation and has better prediction precision compared with other traditional prediction methods.

Keywords: deep learning; stacked auto-encoder; SAE; Yadvendra SAE-Elman; rotary air-preheater; thermal stress deformation prediction.

DOI: 10.1504/IJMIC.2019.099824

International Journal of Modelling, Identification and Control, 2019 Vol.31 No.4, pp.293 - 302

Received: 29 Jul 2017
Accepted: 14 May 2018

Published online: 23 May 2019 *

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