MELT ATOMIZATION - MULTIPLE-REGRESSION ANALYSIS

The goal of energetic melt atomization is to produce fine metal powder and is achieved by generating atomization conditions that maximize th e surface are of a given volume of melt. To this end, it is useful to identify and determine processing variables which have a significant i mpact on atomization. We performed multiple regression analysis on exp erimental data obtained with a rapidly spinning cup atomizer. This mod e of atomization is of particular interest as a way of producing very fine powders relatively easily and economically for use is application s such as metal injection molding. The analysis took into consideratio n planned and unplanned variations in all the processing variables adn identified several main and multi-factor interaction effects. It was found that variables that maximize atomization energy and minimize par ticle-particle collisions produced the finest particle size. For examp le, morphological evidence and Stokesian analysis showed that increasi ng the sample size increased the probability of particle-melt encounte rs; this resulted in the formation of encapsulated particles which wer e larger than the individual particles and which shifted the size dist ribution to coarser sizes. Applying the principles derived from the mu ltiple regression analysis, we have designed an improved larger atomiz er which is described briefly.