Simulation of duplicate tooling consumption at aircraft manufacturing enterprises

Abstract

To simulate the consumption of duplicate technological tooling (hereinafter referred to as TT), it is proposed to use the apparatus of multivariate regression analysis, which provides the main patterns formalization of consumption formation value in the form of linear regression dependencies. An important difference of the proposed approach to the development of regression models is an increased degree of detail due to the use of the number of detail operations performed using specific TT sizes as independent variables, which makes it possible to eliminate the calculation error associated with the presence of passing positions in the planning period (batches of parts and assembly units ). As a result, a quantitative assessment of the constructed dependencies based on the calculation of multiple correlation coefficients, the Fisher dispersion ratio, the Theil’s coefficient of irrelevance allowed us to conclude that they are adequate to the conditions of consumption TT at the aviation enterprise. To improve the adequacy of displaying real processes of TT consumption, the developed models are equipped with a parametric adaptation mechanism. The use of the steepest descent method makes it possible to adjust the regression parameters based on incoming data, taking into account the main trends in the development of the simulated process and providing for smoothing its random fluctuations.