Functional approxmations to experimental data

Bera, Rajendra K (1975) Functional approxmations to experimental data. Technical Report. National Aeronautical Laboratory, Bangalore, India.

Full text available as:
[img] PDF
Restricted to Archive staff only

Download (5Mb)

    Abstract

    In engineering sciences it is often necessary to approximate a function to a given set of experimental data either for computational ease or to evaluate the emperical constants in a semi-emperical formulation. In either case 9 one is basically interested in getting the closest fit to13; the provided data by selecting a function with one. or more13; unknown constants. A commonly used method is to choose a13; sequence of functions - usually polynomials - with unknow13; constants which is approximated in a linear least squares13; sense. This results in a set of linear simultaneous equations in the unknown constants. The drawbacks of the method are that not all types of experimental data have a good fit with few terms, and with larger number of terms the resulting coefficient matrix often becomes quickly ill-conditioned. One may cite. the straight polynomials as an example.

    Item Type: Proj.Doc/Technical Report (Technical Report)
    Uncontrolled Keywords: Gauss-Markov theorem
    Subjects: AERONAUTICS > Aerodynamics
    Division/Department: Aerodynamics
    Depositing User: Poornima Narayana
    Date Deposited: 31 Aug 2006
    Last Modified: 24 May 2010 09:49
    URI: http://nal-ir.nal.res.in/id/eprint/2557

    Actions (login required)

    View Item