Integration of statistical process control methods with engineering modeling tools and digital counterparts to increase the maturity of quality management systems in mechanical engineering

Abstract

The article examines the integration of statistical process control (SPC) methods with engineering modeling tools and digital twin technologies to improve the maturity of quality management systems in high-tech mechanical engineering. The purpose of the study is to substantiate a methodology and a set of tools that enable the transformation of a traditional reactive quality assurance model based on sampling inspection into a proactive predictive control loop for precision machining of gas turbine engine components. The empirical basis consists of data from monitoring key process parameters over 18 months of production operation: more than 2,5 million measurements of geometric, thermal, and vibration characteristics recorded by an IoT sensor network integrated into the SPID system. The methodology is based on combining classical SPC tools (Cp, Cpk indices, X̄/R control charts, regression and variance analysis) with high-detail engineering modeling of thermoelastic deformations using the finite element method as part of a digital twin of the equipment and processed components, as well as adapting the CMMI maturity model to the requirements of ISO 9001 and AS9100 standards. It is shown that the implementation of the integrated system leads to a reduction of the standard deviation of critical dimensional parameters by more than 60%, an increase in Cp and Cpk indices to levels above 2,7, effective compensation of thermal deformations (92-93%), and a multiple (more than 50-fold) reduction in the average time to detect deviations. Additionally, there was a reduction in internal defect costs from 3,84% to 0,52% of turnover, a decrease in the share of false alarms in the control system, and an increase in the comprehensive maturity index of the QMS from 2,15 to 4,35 points, with the share of processes featuring automatic correction increasing to 85%. A high correlation (r≈0.92) was established between the level of digitalization and product accuracy characteristics, confirming the synergistic effect of combining stochastic and deterministic models within a unified digital quality control loop.