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

The article is dedicated to an approach optimizing a task of statistical modeling of the mechanical properties of products in real production metallurgy design. The approach is designed for the benefit of producers-metallurgists aimed at providing panels of Ti – alloys of a specific set of eventual industrial properties. This is accomplished by a procedure of composition optimizing based on existing certificates of brands Ti – alloys. The article presents an approach using mathematical models of optimization problem following the implementation of approach the classical methodology capable of decision-making in the production practice.

The numerical experiment has proved the ability to improve the quality of titanium alloys. Mathematical models suitable for forecasting and optimization have been derived. The approach of Taguchi applied has lead to a desired result, to separate variables Xi for the examined parameters that do not influence significantly on the final result. With this limit, the numerical optimization for maximum search has been conducted with each chemical composition. That allows improving it.

An approach to the study of the relationship between the extreme values of the mechanical properties of titanium alloys . The approach is based on the approximated dependence of the mechanical properties of the chemical composition of the alloy of the ANN model . Approximating dependencies are included in the genetic optimization model based on optimization algorithm NSGA II, which was built by Pareto – fronts , modeling the relationship between the studied mechanical parameters.

The possibilities of regression analysis to model the properties and the related to it multicriteria optimization of the composition of titanium alloys are introduced in the presented research. This is done after performing characteristics of established applications of various titanium alloys. To predict elongation in this paper there have been obtained nonlinear regression models that allow to estimate the strength properties and the plasticity depending on the equivalents for aluminum and molybdenum for titanium alloys.