Nonlinear Interaction of Friction and Interpolation Errors in Unsteady Flow Analyses

Masashi Shimada, A. Vardy

    Research output: Contribution to journalArticlepeer-review

    6 Citations (Scopus)

    Abstract

    Explicit analytical assessments of numerical errors in solutions of decaying standing waves in pipes are obtained using analyses based on the fixed-grid method of characteristics and are used as a quantitative basis for comparing alternative grid types, sizes, and shapes. It is shown that the influences of friction and interpolation combine in a nonlinear manner, but that the combination has the form of a product so that the two effects may be considered independently to a certain extent. The analysis is presented for finite pipes, thus creating both upper and lower limits for the range of possible standing wave frequencies. It is found that frequencies in the upper half of the range are strongly damped. The performances of diamond and rectangular grids are compared, but there is no clear winner because, for equivalent computer resources, each has both advantages and disadvantages. The overall purposes of the paper are to move toward methods of (1) relieving untrained software users from tasks related to grid selection and (2) assisting third parties in assessing the validity of numerical results undertaken by others. DOI: 10.1061/(ASCE)HY.1943-7900.0000685. (C) 2013 American Society of Civil Engineers.

    Original languageEnglish
    Pages (from-to)397-409
    Number of pages13
    JournalJournal of Hydraulic Engineering
    Volume139
    Issue number4
    DOIs
    Publication statusPublished - Apr 2013

    Keywords

    • Time-line interpolation
    • PIPELINE TRANSIENTS
    • LEAK DETECTION
    • Friction
    • Method of characteristics
    • Pipe flows
    • Water hammer
    • Diamond grids
    • MODELS
    • Numerical errors
    • Rectangular grids
    • TIME-LINE INTERPOLATIONS

    Fingerprint

    Dive into the research topics of 'Nonlinear Interaction of Friction and Interpolation Errors in Unsteady Flow Analyses'. Together they form a unique fingerprint.

    Cite this