A numerical method was proposed to predict transient temperature distribution in friction welding of two similar materials of steel. In this method, heat input was estimated using the data of thrust pressure and rotation speed which could be easily measured under a commercial friction welder. The estimated heat input was given as a boundary condition of heat flux at the weld interface, and a finite element method was used for solving an unsteady heat conduction problem of two-dimensional axis symmetry. Transient temperature distributions during whole friction welding process were calculated using variable thermal properties of the specimen materials of mild steel and stainless steel respectively. Calculated results were compared with experimental results, and the relationship between calculated temperature distribution and measured hardness distribution was investigated in the vicinity of weld interface.


Friction welding is a solid-state welding process for joining two similar or dissimilar materials. It is used widely throughout many manufacturing processes where high production rates are required. Friction welding conditions for the process are generally selected on the basis of past experience or study report. However, establishment of a method to decide the friction welding conditions in response to the phenomena which occur during friction welding has been needed because optimum welding conditions depend on each welding machine. Accordingly, many researchers have investigated the relationship between mechanical work, namely heat input, during friction welding and joint performance. Shinoda et al. (1993) revealed that the mechanical properties were correlated with the heat input, and they concluded that the optimum welding condition was obtained with large heat input. Sawai et al. (1999; 2001; 2002) have demonstrated that the mechanical work in the upsetting stage affects tensile strength of the weld joints.

This content is only available via PDF.
You can access this article if you purchase or spend a download.