Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/62472
Title: Zeolite structure prediction with artificial neural networks
Authors: Kevin Tangkas
Keywords: DRNTU::Engineering::Materials::Functional materials
Issue Date: 2015
Abstract: With zeolites consumption exceeding 3 million tons and hundreds of new zeolites structures are being synthesize, zeolites are an important part of the world in science and industry. Furthermore, there are still millions of hypothetical zeolites structures that are still not able to synthesized and waiting to be explored. Characterizations of zeolites still take a long time and involve complicated procedures. Hence, it is imperative to speed up the process of characterizing zeolites structure in order to advance zeolites science and technologies. With advancement of computational science, machine learning method will be explored here in order to expedite the characterization of zeolites. Artificial Neural Network will be utilized to build a prediction model that will predict the Framework Density of zeolites. This prediction model will only use simple inputs that can be easily obtained through chemical analysis of zeolites. The prediction models built produced promising results with relatively small error. The best model in this project was built with simple input of the Al/Si ratio and the type of element that present in the zeolite with Radial Basis Function Network algorithm in machine learning software called Waikato Environment for Knowledge Analysis (WEKA). The model’s Mean Absolute Error is just 0.3119 with Root Mean Squared Error of 0.5029.
URI: http://hdl.handle.net/10356/62472
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:MSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
62472 FYP+Report+(KEVIN).pdf
  Restricted Access
1.98 MBAdobe PDFView/Open

Page view(s) 20

333
checked on Oct 20, 2020

Download(s) 20

16
checked on Oct 20, 2020

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.