Please use this identifier to cite or link to this item:
Title: Parametric Optimization of Slurry Erosion in Pipeline Materials Using Fuzzy Logic and Artificial Neural Networks
Authors: Mishra, Zubin
Supervisor: Kumar, Satish
Mohapatra, S. K.
Issue Date: 7-Aug-2015
Abstract: Wear is most common problems encountered in industries like thermal power plants. More than 51% of India’s energy is met through large coal reserves. Coal combustion in thermal power plant generates nearly 20-25 % bottom ash. Bottom ash generated by thermal power plant is usually transported to ash ponds via mild steel pipelines. Many researchers have studied the effect of various operating parameters on the slurry erosion behavior of Mild steel using different environmental conditions. The present work deals with the parametric study of erosion wear on coated and uncoated ferrous material in solid-liquid mixture. Erosion wear of Al2O3 + 13% TiO2 coated Mild steel, SS202 and grey cast iron is investigated using slurry pot tester. Atmospheric Plasma Spraying is used to deposits ceramic coatings on the base material. Erosion wear at three different speeds 700rpm, 1100rpm and 1400rpm is evaluated at 25% and 45% slurry concentrations. The effect of two different particle sizes of 75m and 250-350m has also been investigated. The erosion mechanism shown by coated specimen was brittle in nature with grain pull out, grain boundary dislodgment and brittle fracture as major mechanism. Fuzzy logic and Artificial Neural Networks (ANN) has been implemented for predicting the erosion wear. ANN model is based on the database obtained from the experiments and involves training, testing and prediction protocols. The results predicted by ANN were more accurate than Fuzzy logic with average percentage error less than 5 %.
Description: ME, MED
Appears in Collections:Masters Theses@MED

Files in This Item:
File Description SizeFormat 
3520.pdf1.72 MBAdobe PDFThumbnail

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.