Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw
In this thesis, enzymatic hydrolysis was utilized in the production of xylose from rice straw. The process model was developed by the modeling techniques using feed-forward artificial neural network (FANN) and optimized using both particle swarm optimization (PSO) and genetic algorithm (GA). The par...
Saved in:
Main Author: | |
---|---|
Format: | Thesis |
Language: | English |
Published: |
2015
|
Subjects: | |
Online Access: | http://eprints.usm.my/41069/1/NUR%E2%80%99ATIQAH_BINTI_NORHALIM_24_Pages.pdf http://eprints.usm.my/41069/ |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
my.usm.eprints.41069 |
---|---|
record_format |
eprints |
spelling |
my.usm.eprints.41069 http://eprints.usm.my/41069/ Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw Norhalim, Nur’atiqah TP155-156 Chemical engineering In this thesis, enzymatic hydrolysis was utilized in the production of xylose from rice straw. The process model was developed by the modeling techniques using feed-forward artificial neural network (FANN) and optimized using both particle swarm optimization (PSO) and genetic algorithm (GA). The parameters studied such as temperature, agitation speed and concentration of enzyme in the process were investigated in order to get an optimum yield of xylose during enzymatic hydrolysis process. Data collected from an experimental design using response surface methodology (RSM) were used to develop the FANN modeling. The data samples has been split into training, testing and validation data set before re-sampling with bootstrap re-sampling method. Then, the FANN model was used to predict the model performance with one hidden layer and the PSO and GA were used to predict the optimum conditions of the process. The number of nodes in the hidden layer obtained is six where the performance on the model is satisfactory with the architecture of FANN, 3-6-1. The correlation coefficient of training and testing set were indicated at 0.9970 and 0.9975 respectively though the correlation coefficient of validation obtained was 0.8501. The optimization of xylose production using the GA method obtained conditions of 50.3˚C, 154 rpm and 1.6944 g/l. The optimum xylose production was predicted as 0.1845 g/l at optimal condition obtained by using GA. Meanwhile with PSO, the optimum temperature observed was at 50 °C, 132 xviii rpm for optimum value of agitation speed and 1.6474 g/l optimum xylanase concentration respectively. The optimal yield of xylose predicted was 0.1845 g/l using PSO for the enzymatic hydrolysis process. The laboratory experiment was carried out to validate the prediction of optimization result. It is shown from the experiment that the concentration of xylose obtained by using prediction optimum parameters for both PSO and GA are 0.2331 g/l and 0.2398 g/l respectively. The average error for the prediction and experimental values for the optimization are 29.97% and 26.34% for GA and PSO respectively. Therefore, the enzymatic hydrolysis on the production of xylose has been enhanced by predicting the optimum conditions utilizing the developed model that fits the experimental data. 2015 Thesis NonPeerReviewed application/pdf en http://eprints.usm.my/41069/1/NUR%E2%80%99ATIQAH_BINTI_NORHALIM_24_Pages.pdf Norhalim, Nur’atiqah (2015) Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw. Masters thesis, Universiti Sains Malaysia. |
institution |
Universiti Sains Malaysia |
building |
Hamzah Sendut Library |
collection |
Institutional Repository |
continent |
Asia |
country |
Malaysia |
content_provider |
Universiti Sains Malaysia |
content_source |
USM Institutional Repository |
url_provider |
http://eprints.usm.my/ |
language |
English |
topic |
TP155-156 Chemical engineering |
spellingShingle |
TP155-156 Chemical engineering Norhalim, Nur’atiqah Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw |
description |
In this thesis, enzymatic hydrolysis was utilized in the production of xylose from rice straw. The process model was developed by the modeling techniques using feed-forward artificial neural network (FANN) and optimized using both particle swarm optimization (PSO) and genetic algorithm (GA). The parameters studied such as temperature, agitation speed and concentration of enzyme in the process were investigated in order to get an optimum yield of xylose during enzymatic hydrolysis process. Data collected from an experimental design using response surface methodology (RSM) were used to develop the FANN modeling. The data samples has been split into training, testing and validation data set before re-sampling with bootstrap re-sampling method. Then, the FANN model was used to predict the model performance with one hidden layer and the PSO and GA were used to predict the optimum conditions of the process. The number of nodes in the hidden layer obtained is six where the performance on the model is satisfactory with the architecture of FANN, 3-6-1. The correlation coefficient of training and testing set were indicated at 0.9970 and 0.9975 respectively though the correlation coefficient of validation obtained was 0.8501. The optimization of xylose production using the GA method obtained conditions of 50.3˚C, 154 rpm and 1.6944 g/l. The optimum xylose production was predicted as 0.1845 g/l at optimal condition obtained by using GA. Meanwhile with PSO, the optimum temperature observed was at 50 °C, 132
xviii
rpm for optimum value of agitation speed and 1.6474 g/l optimum xylanase concentration respectively. The optimal yield of xylose predicted was 0.1845 g/l using PSO for the enzymatic hydrolysis process. The laboratory experiment was carried out to validate the prediction of optimization result. It is shown from the experiment that the concentration of xylose obtained by using prediction optimum parameters for both PSO and GA are 0.2331 g/l and 0.2398 g/l respectively. The average error for the prediction and experimental values for the optimization are 29.97% and 26.34% for GA and PSO respectively. Therefore, the enzymatic hydrolysis on the production of xylose has been enhanced by predicting the optimum conditions utilizing the developed model that fits the experimental data. |
format |
Thesis |
author |
Norhalim, Nur’atiqah |
author_facet |
Norhalim, Nur’atiqah |
author_sort |
Norhalim, Nur’atiqah |
title |
Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw |
title_short |
Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw |
title_full |
Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw |
title_fullStr |
Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw |
title_full_unstemmed |
Neural Network Modeling And Optimization For Enzymatic Hydrolysis Of Xylose From Rice Straw |
title_sort |
neural network modeling and optimization for enzymatic hydrolysis of xylose from rice straw |
publishDate |
2015 |
url |
http://eprints.usm.my/41069/1/NUR%E2%80%99ATIQAH_BINTI_NORHALIM_24_Pages.pdf http://eprints.usm.my/41069/ |
_version_ |
1643710119958020096 |
score |
13.18916 |