Dimensional control of turbine blades via RSM-based process parameter optimization in investment casting

Abstract

To address the dimensional accuracy challenges in investment casting of DD6 nickel-based superalloy hollow turbine blades, a multi-parameter collaborative optimization and deformation response prediction method based on response surface methodology was proposed. Using a Box-Behnken design, with pouring temperature, shell temperature, and withdrawal rate as key variables, deformation response data were obtained through numerical simulation, and a second-order model incorporating linear, interaction, and quadratic terms was established to characterize the nonlinear coupling effects of process parameters on dimensional deformation. The results indicate that withdrawal rate is the dominant factor influencing deformation, while shell temperature exhibits a pronounced “U”-shaped nonlinear trend. Significant interactions between process parameters are also observed. The constructed model demonstrates high predictive accuracy, with R2 of 0.978 and an RMSE of 0.0026 mm, and exhibits strong generalization capability, enabling the identification of optimal parameter combinations even beyond the simulated dataset. Compared with conventional orthogonal design methods, the maximum deformation of the optimized process was reduced from 0.2021 mm to 0.1905 mm, achieving an improvement of approximately 5.74%. This work provides a theoretical foundation and practical strategy for dimensional accuracy control and multi-parameter process optimization in the manufacturing of complex thin-walled castings.

1. Introduction

Investment casting is a critical manufacturing process for high-performance hollow turbine blades, in which process parameters exert a decisive influence on both the forming quality and dimensional accuracy of the castings [1, 2]. Blade dimensional accuracy is essential for reliable assembly and aerodynamic integrity, directly affecting energy-conversion efficiency and overall aeroengine performance [3]. With the increasing structural complexity and performance requirements of turbine blades, significant coupling effects have emerged among key process variables such as withdrawal rate, pouring temperature, and preheating temperature during casting [4, 5]. These interactions are recognized as major contributors to deformation and dimensional deviation in cast components [6-8]. Therefore, systematically identifying the dominant process parameters and their interactions has become a core challenge in improving the dimensional precision and consistency of turbine blades.

In recent years, extensive research has been conducted on solidification-based numerical analysis and process optimization in investment casting [4, 9-11]. These efforts have significantly contributed to shortening development cycles, reducing experimental costs, and improving product quality. Dini et al. [12] employed a single-factor control approach to investigate the effects of process parameters on residual stress and dimensional variation in castings. Ma et al. [13] studied the influence of temperature gradients and withdrawal rates on defect formation during Bridgman directional solidification. Miller et al. [14] determined optimal process conditions by analyzing the position and morphology of the solid-liquid interface, and found that a stabilized and planar interface is essential for reducing casting defects. Wang et al. [15] developed a two-dimensional response-surface model that elucidated the relationship between wax-injection parameters and dimensional accuracy, and reported that parameter interactions have a non-negligible effect on dimensional control. Galantucci et al. [16] used the finite element method to simulate temperature field evolution during directional solidification and examined the impact of filling temperature, withdrawal rate, and furnace geometry on the temperature distribution. For ring-on-ring castings, Wang et al. [17] established an RSM-based mapping between process variables and both casting diameter and ovality. Kavitha et al. [18] employed a Box-Behnken design (BBD) to study the effects of pouring temperature, insert temperature, insert thickness, and surface roughness on the bond strength of Al-SS bimetallic parts in lost-foam composite casting.