Table 6 and
Figures 6-9 show correlation coefficients with their corresponding p-values and scatter
plots, respectively. Surprisingly, there was no statistically significant correlation between payloads for the experimental period and fuel/cycle as indicated by the p-value of 0.1801 (greater than α = 0.05). This was contrary to expectation and hence the correlation between payload for the entire available data set (May 3 to July2) and fuel/cycle was also analyzed. This yielded a statistically significant correlation (p-value of 0.0000). Modeling fuel/cycle per ton is desirable so that the model can be extended to different truck payloads. In fact, it is expected that fuel consumption should correlate to amount of material carried since more work is done. Hence, correlations between cycle time components in Table 6 and fuel/cycle/ton was tested and statistically significant positive correlation was found. Based on this, the regression model in Equation (3) was formulated. In this model, ti is cycle time in minutes for component i. Subscripts es, et, l, ls, and lt mean empty stopped, empty travel, loading, loaded stopped, and loaded travel.