Phd thesis artificial neural networks


Description:

The authors suggested that SHRT was able to reduce the energy consumption of the column significantly.

Highlight your research

An energy saving of about Efficient operation of the distillation unit cannot be over emphasised. A slight improvement in thermal efficiency of the unit can make a large difference in profitability. Energy generation leads to release of greenhouse gas GNG especially in developing countries that are just coming up industrially leading to environmental pollution and thus defeating the concept of sustainable development Dincer and Rosen, , Dincer, Improved efficiency of chemical processes is one of the ways of reducing the GNG.

Also, improved efficiency of distillation column will result in better yield of. It will also reduce the consumption of energy and thereby lengthen the depleting energy reserve. Although distillation process has been extensively studied by researchers, the issue of energy efficiency still remains an unresolved.

Determination of minimum energy requirements for a distillation process is useful in the design stage. This is because a high energy requirement in distillation column culminates in a high number of stages and increases the complexity and subsequently the cost of the column. An initial understanding of the minimum energy for a separation process gives room for the understanding of its cost implication as there is a direct relationship between the minimum energy and the diameter and length of the column.

How can this minimum energy requirement be actualised in the operation of the column especially for the special case of multicomponent non ideal system? There is a need to develop a model that will evolve methods of column improvement which will incorporate the principle of second law analysis in determining points of inefficiency in the column, bring about a control mechanism of the unit to achieve the optimum energy efficiency and develop optimization procedures aimed at minimising the inefficiencies without compromising the qualities of the products.

This research work is focused on novel methods of increasing the efficiency of distillation units either in the simple binary system or the complex crude distillation system of the refinery by the application of second law principle using the tools of advanced control and optimisation. This increased efficiency can then translate to reduction in the energy requirements of the column. The main aim of this research work is to develop energy efficient operating techniques for distillation columns. In this thesis, the contribution is on developing methods to bring about the application of second law analysis to improve the energy efficiency of various forms of distillation columns.

Past studies in examining the second law analysis of distillation columns has been majorly focused on quantifying its energy and exergy efficiency and pinpointing the area of the column with most exergy loss.

Artificial neural networks and their application to sequence recognition

This study goes a step further to use exergy analysis not only as an analytical tool but also a retrofitting and a design tool in improving the exergy efficiency of the distillation unit. In selection of energy efficient control structures for distillation columns, relative exergy array REA has been used in some past work.

However, detailed analysis of the control structure in the steady and dynamic state is lacking. This study presents methods of selecting energy efficient control structures in the steady state and validating the methods in the dynamic state. This study also presents an optimisation framework based on exergy analysis to improve the energy efficiency of distillation column. Bootstrap aggregated neural network BANN is presented in this study for enhanced model accuracy.

This is in lieu of artificial neural network ANN which has been commonly used in past studies. The presented techniques were applied on a wide range of columns from the simple binary system to more complex crude distillation systems with remarkable results. Usually, to improve the process unit, design and process engineers often use advance control and optimisation in lieu of designing new processes because of its cost implications.

This study introduces a new way of controlling and optimising the column using the second law of thermodynamics. In Chapter 2, the knowledge gap on energy efficiency of chemical processes, distillation systems and application of thermodynamics analysis is highlighted. The theory on which the research is based on and a general review of literatures related to the research is presented. Chapter 3 is focused on modelling and the choice of the models used in this research. Rigorous models and dynamic models of distillation column are presented as well as method of solution of the models.

Artificial neural network model are also discussed extensively. The developed method incorporates energy efficiency in control structure selection.

Tyler Santander PhD Thesis - 2017

Detailed analysis of the control structures in the steady and dynamic states is reported. Application of the methods to binary distillation systems was highlighted. In Chapter 5, a neural network based strategy for the modelling and optimisation of distillation columns incorporating the second law of thermodynamics was discussed. Neural network models for exergy efficiency and product compositions are developed from simulated process operation data and are used to maximise exergy efficiency while satisfying product quality constraints.

Applications to binary and multicomponent systems demonstrate the effectiveness of the method. Bootstrap aggregated neural network BANN is introduced. Multiobjective optimisation of the column based on BANN was developed. Application to a complex system of crude distillation unit comprising of atmospheric distillation unit and vacuum distillation unit was discussed. A further study on effect of preflash unit preflash drum and preflash column on the crude distillation unit was also made.

Chapter 7 is focused on highlighting the conclusions from the research and recommendations for further work. Distillation is the separation of a mixture into its different constituents due to boiling point difference. The process could be batch or continuous. In batch distillation the feed is fed to the distillation tower, the products are removed one at a time until the distillation process is ended. Continuous distillation on the other hand separates large quantities of mixture such as those found in the chemical and petrochemical industries.

The feed are fed continuously to a distillation tower and the products or fractions are withdrawn continuously at the same time from the different stages of the tower. A distillation process that separates only two distinct products is said to be binary distillation while multicomponent distillation has at least three products from the separation process.

Crude oil distillation is a multicomponent, continuous distillation process. Crude oil is made up of a mixture of different hydrocarbons and sometimes referred to as fractions. The separation of these fractions based on boiling point differences is referred to as fractionation. The distillation process does not produce a product of a distinct boiling point rather it produces fraction based on a boiling point range.

Each fraction can be further separated into a number of components. For example the naphtha product from the top of the Atmospheric distillation ADU column of the refinery can be separated to liquefied petroleum gas LPG such as propane, butane etc. A distillation tower is a series of distillation process with tray such as sieve and bubble gas stacked on each other or with a series of packings such as Raschig ring and Berl saddle.

The feed is introduced to the column at its flash point where a portion of the feed vaporises. The portion above the flash zone is the rectifying section and the lower part is the stripping section. At the onset of separation process within the distillation tower, the openings on each tray allows the vapour to travel up the column and as the vapour comes in contact with the liquid, the heavier part of the vapour are condensed back to the liquid state and the lighter part move up the column.

The tray is built with weirs that allow liquid to be retained; liquid that overflows the weirs flows into the downcomer and moves down the column. The heavier fractions that may still be retained in the upper part of the column are condensed by the external reflux to give a higher purity of product while the lighter fraction in the bottom of the column are vaporised by the reboiler.

Distillate Reflux. Reboiler Reboiler Bottoms energy product. Processes such as distillation bring phases of different compositions into contact. When the phases are not in equilibrium, mass transfer between the phases alters the composition. The rate of change of compositions in different phases depend largely on the deviation from equilibrium hence the equilibrium temperature, phase composition and pressure is of great importance. The tendency for given chemical species to co-exist in the liquid and vapour phase is defined as its equilibrium ratio.

Equilibrium ratio is the ratio of vapour to liquid phase of mole fraction of species. The constant for the expression of this ratio is K. For non ideality of the liquid phase, a more rigorous VLE data is required in terms of activity coefficient and it is given as.

Share your Details to get free

Pure component pressure will be modelled from Antoine equation. Usually enthalpy of the vapour and liquid stream should be calculated as functions of temperature, pressure and composition of each stream. However, because liquid are incompressible and if low to moderate pressure system is assumed, then the enthalpy is calculated as function of temperature and composition based on a linear fit of heat capacity with temperature.

For multicomponent systems, mixing rule applies. The molar average of the pure component enthalpy is the vapour enthalpy while for liquid enthalpy, non idealities is accounted for by heat of mixing. Method for accurate prediction of equilibrium constant is of great interest in the correct modelling of a crude distillation unit Tarighaleslami et al. There are many approaches to obtaining the K value. This could be from experimental measurements, use of correlations empirical such as nomographs and use of Equation of state EOS.

Methods based on EOS and empirical correlations are explored further. Correlations that could be applicable for the predictions of K value for crude mixtures at low pressure and high pressure include but not limited to the following. Almehaideb further modified Whitson and Torp to give Almehaideb et al. Almehaideb et al.

For low pressure of under psia, Standing correlation and Wilson correlation were considered in the work of Almehaideb and were found to compare relatively well with the extracted K values from the new work. To determine K value, some Hydrocarbon properties need to be determined. These properties include critical temperature, critical pressure, acentric factor etc. Several methods of estimating parameters for the evaluation of K value are available in the literatures.

Deep neural network models for image classification and regression

The VLE data are of paramount importance in the modelling of distillation unit because it has great impact on the accuracy of the model. Thermodynamics is a concept derived from the two words thermo and dynamics, which means heat in motion. It was limited to heat engines of which the steam engine was an example but now is applicable in determining heat and work effects associated with processes in terms of obtainable maximum work from a process or minimum work required for processes.

It is also applicable in describing relationships among variables for a process in equilibrium. The history of thermodynamics could be dated back to the 18th century when James Watts in built the first steam engine and thus starting the. In , Sadi Carnot claimed that the efficiency of a heat engine depends on the temperature only, the so-called Carnot factor Smith et al. Thermodynamics developed out of a desire to increase the efficiency of the steam engine. Carnot introduced two ideas in his analysis of heat engines.

The first was the idea of a cycle; a process which occurs through several stages but which leads back to the same conditions of temperature, pressure and volume. The second was that the cycle could be reversible if equilibrium is maintained in the system at all times. In , William J.

phd thesis artificial neural networks Phd thesis artificial neural networks
phd thesis artificial neural networks Phd thesis artificial neural networks
phd thesis artificial neural networks Phd thesis artificial neural networks
phd thesis artificial neural networks Phd thesis artificial neural networks
phd thesis artificial neural networks Phd thesis artificial neural networks
phd thesis artificial neural networks Phd thesis artificial neural networks
phd thesis artificial neural networks Phd thesis artificial neural networks
phd thesis artificial neural networks Phd thesis artificial neural networks
Phd thesis artificial neural networks

Related phd thesis artificial neural networks



Copyright 2019 - All Right Reserved