*3.2. Wear Rate Model*

Figure 3 illustrates the distribution of residuals and the comparison of model-predicted values to experimental results. From the figure, it can be seen that the residuals of 16 experimental runs were along a straight line, which indicated normal distribution and thus random error for lack of fit. The comparison data were approximate to the line *y* = *x*, showing the little divergence of model predictions from experimental results. Therefore, the model for wear rate was established with good prediction accuracy.

Figure 4 plots the 3D response surface and corresponding contour of wear rate with respect to laser power and particle ratio of TiC. From the figure, it is clear that the wear rate of the cladding layer decreased with an increase in laser power and weight fraction of TiC. Further investigation of phase composition, microstructure, and element distribution is required to understand the mechanism of how process parameters affect the wear rate. Hardness and wear resistance were positively correlated under the same friction situation.

**Figure 3.** (**a**) Distribution of residual wear rate, (**b**) Distribution of predicted and actual wear rate.

**Figure 4.** The response of wear rate to significant factors *LP* and *PR*. (**a**) 3D response surface; (**b**) contour plot.
