On the Median Volume Diameter Approximation for Droplet Collision Efficiency

View More View Less
  • 1 Norwegian Hydrotechnical Laboratory, Trondherin, Norway
  • | 2 Division of Meteorology, The University of Alberta, Edmonton, Canada
  • | 3 Laboratory of Structural Engineering Technical Research Centre of Finland, Espoo, Finland
© Get Permissions
Full access

Abstract

In this note, we examine a shortcut for calculating the overall collision efficiency of a droplet spectrum, known as the “median volume diameter” (mvd) approximation. By calculating the overall collision efficiency of a circular cylinder for a variety of natural droplet spectra, first precisely using a spectrum weighting approach, and then as approximated using the mvd, as well as several other representative droplet sizes, we show by comparison that the mvd approximation is a good one, with an average absolute error of about 0.02. While trying to give some mathematical justification for why the mvd approximation works, we show that it can be derived from a single-point numerical integration formula, and that extension of this formula to 2, 3 or 4 points should give correspondingly better approximations. Detailed comparisons confirm that use of the 2-point formula reduces the average error by one-half, while the 3- and 4-point formulae can reduce it even more, depending on the type of spectrum.

Abstract

In this note, we examine a shortcut for calculating the overall collision efficiency of a droplet spectrum, known as the “median volume diameter” (mvd) approximation. By calculating the overall collision efficiency of a circular cylinder for a variety of natural droplet spectra, first precisely using a spectrum weighting approach, and then as approximated using the mvd, as well as several other representative droplet sizes, we show by comparison that the mvd approximation is a good one, with an average absolute error of about 0.02. While trying to give some mathematical justification for why the mvd approximation works, we show that it can be derived from a single-point numerical integration formula, and that extension of this formula to 2, 3 or 4 points should give correspondingly better approximations. Detailed comparisons confirm that use of the 2-point formula reduces the average error by one-half, while the 3- and 4-point formulae can reduce it even more, depending on the type of spectrum.

Save