International Scientific Journals
of Scientific Technical Union of Mechanical Engineering "Industry 4.0"

Using the method of binary logistic regression, the dependence of dichotomous variables on one or more, quantitative or qualitative independent variables is studied. In order to solve problems with binary dependent variables, binary choice model estimation methodologies, such as logistic regression (logit) and probit regression (probit), can be implemented. The logistic model is based on the logistic distribution, while the probit model is based on the normal distribution. In this paper the binary logistic regression methodology is applied for obtaining defect-free welds by electron beam welding process, based on experimental data.

Experimental results for the quality of electron beam melting and refining of pyro-refined copper samples are analysed by an empirical model estimation approach. Investigation of the influence of the process parameters – melting power and time of refining on the removal efficiency of specific impurities, overall removal efficiency and material losses is performed.