Tolerance analysis plays an important role in design and manufacturing stages for reducing manufacturing cost by improving producibility. In most production processes encountered in practice, a process mean may shift or drift in the long run although process is in control. This study discusses the feasibility of three most common inflation factors(Bender, Gilson and Six Sigma) as a correction to Root Sum of Squares(RSS) method to compensate heuristically for a shift of process mean and nonnormal component distributions from simulation experiments and proposes the guidelines for choosing the inflation factor.