Modelling of Libyan crude oil using arti cial neural networks

178 Pages·2017·6.59 MB·English

Loughborough University
Institutional Repository
Modelling of Libyan crude
oil using artificial neural
networks
ThisitemwassubmittedtoLoughboroughUniversity’sInstitutionalRepository
by the/an author.
Additional Information:
• A Doctoral Thesis. Submitted in partial fulfilment of the requirements
for the award of Doctor of Philosophy of Loughborough University.
Metadata Record: https://dspace.lboro.ac.uk/2134/12117
Publisher: (cid:13)c AL Mahdi Al Hutmany
Please cite the published version.
This item was submitted to Loughborough University as a PhD thesis by the
author and is made available in the Institutional Repository
(https://dspace.lboro.ac.uk/) under the following Creative Commons Licence
conditions.

For the full text of this licence, please go to:
http://creativecommons.org/licenses/by-nc-nd/2.5/
Modelling of Libyan Crude Oil Using
Arti(cid:12)cial Neural Networks
by
Al Mahdi Al Hutmany
A doctoral thesis submitted in partial fulfilment of the requirements
for the award of the degree of Doctor of Philosophy (PhD)
Department of Chemical Engineering,
Loughborough University, Loughborough,
Leicestershire, UK, LE11 3TU
⃝c by AL Mahdi Al Hutmany, 2013
CERTIFICATE OF ORIGINALITY
This is to certify that I am responsible for the work submitted in this thesis,
that the original work is my own except as specified in acknowledgements
or in footnotes, and that neither the thesis nor the original work contained
therein has been submitted to this or any other institution for a degree.
................................ (Signed)
................................ (candidate)
Abstract
The preparation and analysis of input and model data was carried out. The
importance of the correct technique of data filtering was highlighted with
particular focus being emphasised on the removal of outliers in raw data.
An important process in the use of Artificial Neural Network (ANN) models
was identified as being the selection of the right input variables.A compari-
son between using factor analysis and statistical analysis in the selection of
inputs and it was observed that the former gave significantly better results.
The training and testing phase of Artificial Neural Network (ANN) model
developmentwasshowntobeanimportantstepinArtificialNeuralNetwork
(ANN) model development. If this phase was wrongly done then the ANN
model would not be accurate in its predictions.
OptimisationoftheANNmodelarchitecturewascarriedoutwiththeamount
of hidden layers, amount of neurons in the hidden layers, the transfer func-
tion used and the learning rate identified as key elements in obtaining an
Artificial Neural Network (ANN) architecture that gave fast and accurate
predictions.
Fresh water addition and demulsifier addition were identified as key param-
eters in the economic performance of the desalting process.
Due to a scarcity of water and the high cost of the demulsifier chemical it
was important to try and optimise these two input variables thus reducing
the cost of operations.
iii
Contents
Declaration iii
Abstract iv
Acknowledgments ix
List of acronyms x
List of symbols xii
List of (cid:12)gures xvi
List of tables xvii
1 INTRODUCTION 1
1.1 Research motivation 1
1.2 Objectives of the Study 2
1.3 Structuring the Thesis 3
2 LITERATURE REVIEW 4
2.1 Introduction 4
2.2 Crude oil components 4
2.3 Stability of emulsions 6
2.4 Water in crude oil emulsion stabilisation 8
2.5 Destabilisation of crude oil emulsions 9
2.6 Crude oil desalting 12
iv
v
Contents
2.7 Artificial Neural Networks 15
2.8 Closure 19
3 ARTIFICIAL NEURAL NETWORKS 20
3.1 Introduction 20
3.2 The neuron 21
3.3 Neural Network models 23
3.3.1 The perceptron 23
3.3.2 Layered networks 24
3.3.3 Fully connected network 26
3.3.4 Neural Network activation function 26
3.3.5 Linear function 27
3.3.6 Sigmoidal function 27
3.3.7 Threshold function 28
3.3.8 Network structure 29
3.3.9 Single input single output 30
3.3.10 Single input multiple output 30
3.3.11 Multiple input single output 31
3.3.12 Multiple input multiple output 33
3.3.13 Network training 34
3.3.14 Learning 35
3.3.14.1 Supervised learning 35
3.3.14.2 Unsupervised learning 35
3.3.15 Artificial Neural Network applications 36
3.3.16 Merits and demerits of Artificial Neural Networks 37
3.3.17 Building steps of an Artificial Neural Network 39
3.3.17.1 Normalisation 40
3.3.17.2 Hidden layer selection 41
3.3.18 MATLAB simulations 42
vi
Contents
3.3.19 Closure 44
4 METHODOLOGY 45
4.1 Dehydration and desalting 45
4.1.1 Chemical treatment using demulsifiers 46
4.1.2 The use of gravity and residence time 47
4.1.3 Heating 47
4.1.4 Electric treatment 48
4.2 Desalting process 48
4.2.1 Desalting process description 49
4.2.2 Instrumentation 52
4.2.2.1 Mixing valve 53
4.2.2.2 Interface level controller 53
4.2.2.3 Effluent draw-off valve 53
4.2.3 Instrument installation 53
4.3 Factors affecting desalter perfomance 54
4.3.1 Crude oil feed rate 55
4.3.2 Demulsifier dosage 55
4.3.3 Crude oil temperature 55
4.3.4 Fresh water addition 55
4.4 Data acquisition 56
4.4.1 Equipment and materials 56
4.4.2 Investigated variables 56
4.4.3 Testing methods 57
4.4.3.1 Salt in crude oil testing method 60
4.4.3.2 Water in crude oil testing method 61
4.5 Normalising and filtering data 63
4.6 Closure 72
5 MODELLING OF LIBYAN CRUDE OIL DESALTER US-
vii
Contents
ING AN ARTIFICIAL NEURAL NETWORK 73
5.1 Desalting 73
5.2 Preparation and analysis of input and output model data 75
5.3 Selection of best input variables 91
5.3.1 Factor analysis 91
5.3.2 Input data selection with the use of statistics 94
5.4 Neural network based model for the predicton of salt removal
efficiency 102
5.4.1 Division of data to obtain training data sets and test-
ing data sets 102
5.4.2 Development of the neural network model 104
5.4.3 Comparisonsofstatisticalmodelpredictionswithneu-
ral network model 118
5.5 Optimisation of demulsifier injection and fresh water addition 121
5.6 Closure 127
6 CONCLUSION AND FUTURE WORK 128
6.1 Conclusion 128
6.2 Future work 130
Appendix 150
References 150
Acknowledgements
I would firstly like to thank Allah who has given me the health and strength
to be successful in my research and persevere throughout this critical stage
of my life.
IamgreatlyindebtedtoProfessorVahidNassehiandProfessorVictorStarov
for their supervision, guidance and encouragement, which helped to moti-
vate me and ultimately achieve all that I have in this study. Their continued
support also provided me with the drive that also allowed me to complete
this thesis.
Further, I would like to also thank the Department of chemical engineering
for giving me the opportunity to carry out my work.
I would like to especially express my appreciation to the Libyan government
for sponsoring me during the course of my research.
I wish to thank the Arabian Gulf oil company for providing me free oil field
data, which was used in my study and the overall advice they provided to
me. I am glad to thank my family and friends for their encouragement dur-
ing this period of my PhD research.
FinallyIwouldliketopersonallythankmyofficecolleaguesfromthebottom
of my heart for all the support and encouragement they provided me during
this research period.
M. ALHUTMANY
March, 2013