Mathematical model of a manufacturing process based on small-volume samples

Abstract

A method is proposed for obtaining an adequate mathematical model of a technological process (operation) in a multidimensional passive factorial experiment under conditions where the number of initial controlled production factors significantly exceeds the sample size for each factor. Analysis of the initial data table using the correlation cluster method allows the list of influencing factors to be reduced to a minimum without unnecessary duplication of information. The reduced table serves as the basis for an original method of statistical modeling.