Stream of Variation Modeling Method for Turbine Blade in the Investment Casting Process

Investment casting turbine blade is an important component which affect the performance of aero-engines with their complex structural design, cumbersome manufacturing processes and frequent state changes. But the investment casting turbine blade final product has low accuracy and a high rate of waste, which seriously affects the production and development capabilities of Chinese aero-engines. To solve this problem, this study proposes a stream of variation model construction method for the entire investment casting process of turbine blade. Based on the state space theory, this method constructs the stream of variation models for the four states in the investment casting process and the dimensional variation prediction of each link which are influenced by the Multi-source and heterogeneous variation is realized. The feasibility of the stream of variation model in practical applications was verified through ICT (Industrial Computed Tomography) simulation experiments using approximated blade constructed by the real blade cross-sectional lines. Finally, the heuristic algorithm was added to the stream of variation model to optimize the parameter search process to improve the applicability of the model. This method implements the real-time prediction of variations at each stage and provides an important tool for quality control of investment casting turbine blade.