• Login
    View Item 
    •   ResearchSpace Home
    • College of Agriculture, Engineering and Science
    • School Mathematics, Statistics and Computer Science
    • Applied Mathematics
    • Doctoral Degrees (Applied Mathematics)
    • View Item
    •   ResearchSpace Home
    • College of Agriculture, Engineering and Science
    • School Mathematics, Statistics and Computer Science
    • Applied Mathematics
    • Doctoral Degrees (Applied Mathematics)
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    On paired decoupled quasi-linearization methods for solving nonlinear systems of differential equations that model boundary layer fluid flow problems.

    Thumbnail
    View/Open
    Otegbeye_Olumuyiwa_2018.pdf (7.679Mb)
    Date
    2018
    Author
    Otegbeye, Olumuyiwa.
    Metadata
    Show full item record
    Abstract
    Two numerical methods, namely the spectral quasilinearization method (SQLM) and the spectral local linearization method (SLLM), have been found to be highly efficient methods for solving boundary layer flow problems that are modeled using systems of differential equations. Conclusions have been drawn that the SLLM gives highly accurate results but requires more iterations than the SQLM to converge to a consistent solution. This leads to the problem of figuring out how to improve on the rate of convergence of the SLLM while maintaining its high accuracy. The objective of this thesis is to introduce a method that makes use of quasilinearization in pairs of equations to decouple large systems of differential equations. This numerical method, hereinafter called the paired quasilinearization method (PQLM) seeks to break down a large coupled nonlinear system of differential equations into smaller linearized pairs of equations. We describe the numerical algorithm for general systems of both ordinary and partial differential equations. We also describe the implementation of spectral methods to our respective numerical algorithms. We use MATHEMATICA to carry out the numerical analysis of the PQLM throughout the thesis and MATLAB for investigating the influence of various parameters on the flow profiles in Chapters 4, 5 and 6. We begin the thesis by defining the various terminologies, processes and methods that are applied throughout the course of the study. We apply the proposed paired methods to systems of ordinary and partial differential equations that model boundary layer flow problems. A comparative study is carried out on the different possible combinations made for each example in order to determine the most suitable pairing needed to generate the most accurate solutions. We test convergence speed using the infinity norm of solution error. We also test their accuracies by using the infinity norm of the residual errors. We also compare our method to the SLLM to investigate if we have successfully improved the convergence of the SLLM while maintaining its accuracy level. Influence of various parameters on fluid flow is also investigated and the results obtained show that the paired quasilinearization method (PQLM) is an efficient and accurate method for solving boundary layer flow problems. It is also observed that a small number of grid-points are needed to produce convergent numerical solutions using the PQLM when compared to methods like the finite difference method, finite element method and finite volume method, among others. The key finding is that the PQLM improves on the rate of convergence of the SLLM in general. It is also discovered that the pairings with the most nonlinearities give the best rate of convergence and accuracy.
    URI
    https://researchspace.ukzn.ac.za/handle/10413/17131
    Collections
    • Doctoral Degrees (Applied Mathematics)

    DSpace software copyright © 2002-2013  Duraspace
    Contact Us | Send Feedback
    Theme by 
    @mire NV
     

     

    Browse

    All of ResearchSpaceCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsAdvisorsTypeThis CollectionBy Issue DateAuthorsTitlesSubjectsAdvisorsType

    My Account

    LoginRegister

    DSpace software copyright © 2002-2013  Duraspace
    Contact Us | Send Feedback
    Theme by 
    @mire NV