Contd....

4. The calibration curve is frequently written in the form

where is the value of when and are calibration constants that must be determined from the calibration data using a least squares procedure. The value of is usually between 0.45 and 0.5. When is not expected to be zero anywhere in the flow domain it is convenient to set and treat and as curve-fitting parameters. The generalized least squares procedure for fitting the statistically smoothed data through a function with parameters is:

This gives equations for the unknowns . Depending on the form chosen for the - simultaneous equations can be linear or non-linear. In the latter problem iterations will be required to solve for the 's.