EXPERIMENTAL AND SIMULATION STUDY OF DIRECTIONAL SOLIDIFICATION PROCESS FOR INDUSTRIAL GAS TURBINE BLADES PREPARED BY LIQUID METAL COOLING

Abstract

Advanced aero and power generation industry needs high-performance gas turbine. As key parts of gas turbine directionally solidified (DS) columnar grain and single crystal (SX) blades operate in heavy stress and high temperature conditions. The continuous demand for increasing turbine inlet temperature and aggressive environment has pushed alloy designers to develop DS and SX Ni-based blade alloys that contain high amount of alloying elements. DS process of blades using such alloys has become a challenging task. The small DS and SX blades are usually produced by high rate solidification (HRS) process. However, the growth of large DS and SX blades requires directional solidification with a sustained thermal gradient along the DS direction. By increasing the thermal gradient, the dendrites are refined, which results in a mechanically-superior DS and SX with reduced defects. One method to achieve consistent and higher thermal gradients is the utilization of the liquid metal cooling (LMC) process. In this method, heat extraction from the outer surface of the mold during DS relies on heat conduction rather than radiation in the conventional HRS process. The optimization of the LMC process is difficult and costly by experimental methods, especially for the complexly shaped industry gas turbine (JOT) blades because of the complicated process parameters associated with the technique. Numerical simulation is an efficient method to solve this problem. In this work, directionally solidified industry gas turbine hollow blades were prepared by high gradient LMC process. Liquid Sn was used as cooling medium. The temperature fields, macrostructures, primary dendrite arm spacing (PDAS) at various withdrawal rates during LMC process have been calculated with ProCAST software. The impact of withdrawal rate on formation of stray grains and freckles was predicted. The calculated results and the experimental observations agreed well. The solidification rates and cooling rates were found to increase with the increase of withdrawal rate. The axial thermal gradient was high and stable during the LMC process. It was found that stray grains would not block the growth of original grains at optimized withdrawal rate. No freckles were observed in the industry gas turbine hollow blades prepared by LMC technique due to the high cooling rate. Though the mean diameters of columnar grains in LMC blades were almost identical to that observed in HRS blades, the PDAS were more than 50% refined in LMC blades than those in HRS blades.