Locomotive fuel usage analysis and insights for Pacific national
This analysis entails the findings of the fuel usage of two of Pacific National’s locomotives, 8339 and 8345, as they transport coal from the Sonoma Coal Mine to the Abbott Point Terminal. We found that in comparing Pacific National’s predicted fuel usage to the actual fuel usage of the trains, that there was a significant difference between the fuel consumption of some of the throttle settings (notches), these particularly being at the higher and lower end. Furthermore, we have also found that there was a difference in fuel usage between different crews on board, which we attribute mainly to notch skipping by the driver of the trip. However, in comparing the two physical locomotives themselves, we deemed that there was no discernible difference between their fuel consumption, as an analysis of the fuel use per notch between trains showed no outstanding differences
Is the throttle setting (notch) to estimate the fuel consumed accurate?
The predictions in lower notches are overestimated while in higher notches, the predicted fuel usage is understated. (Again only 4 trips, perhaps the trains were lucky and were not as stationary as normal, eg. waiting for passing trains, or unloading etc. which uses a bit of fuel in notch 0 and 1).
A significant portion of the Predicted Fuel Usage data, predicted 0 fuel usage for the trains when there was actually a small amount of fuel used. For example, the Predicted Fuel Usage showed 0 fuel usage, where the Actual Fuel Usage was 0.5. For LOCO 8345, this occurred 71 times out of the 207 data points (34%). For LOCO 8339, this occurred 476 times out of the 1051 data points (45%). This may be due to the devices not predicting fuel usage for quick notch changes.
Is there a difference between the two locomotives?
From the tables given above, we were then able to complete an Analysis of Variance (ANOVA) on the data set which compared the means of each separate notch of Locomotive 8345 and Locomotive 8339 to determine if there is any difference between each notch.
From the output of the ANOVA’s above you can see that the only notch that has variance with their means is Notch 1. All the other notches have no variance in their means of fuel consumption per minute which indicates that there is no difference between Locomotive 8345 and Locomotive 8339. The difference in notch 1 could be because the data for notch 0 is coded into the data for notch 1 and Locomotive 8339 is a longer journey and when the train is stopped at signals and loading and unloading it is using more fuel as it has had to do this function on more occasions then Locomotive 8345.
Is There Any Difference Between Drivers?
Is there a difference between the drivers when it comes to notch skipping?
Notch skipping is when a driver within one minute the notch is changed ± 3. For example at 20:51 the notch is set at 6, then at 20:52 the notch is at 3, this would be classified as notch skipping. Because it is very easy to accidentally skip a notch, anything above a ± 3 change would be considered a notch skip. This would also imply that the driver within one minute changed the notches every 20 seconds. Because of the sudden changes this could put more strain on the equipment.
From the ANOVA output we can say that with 95% confidence that there is a difference between the crew and notch skipping. From the confidence interval we can see that crews 1,2 & 3 are similar in intervals. However crew 4 & 5 have significantly higher intervals.
Does notch skipping use more fuel per minute?
From the ANOVA output we can say that with 95% confidence that there is a difference between the fuel used per minute and notch skipping. From the confidence interval yes (yes notch skipping occurs) has a higher mean than no.
From these findings we can conclude that notch skipping is not only bad on the train itself, but it also uses more fuel. This would mean that crews 4 & 5 are notch skipping more than the other crews, wearing down the train as well as consuming more fuel per minute.
Is there a difference between the position when it comes to fuel consumption?
From the ANOVA output we can say that with 95% confidence that there is a difference between the position and fuel consumption. From the confidence interval we can see that Pring to Sonoma and Sonoma to Abbotts Point use significantly more fuel per minute compared to the other positions.
As seen from the tables and ANOVA output, we can come to the conclusion that there is a very small difference between crews when it comes to fuel consumption. But with position and fuel used per minute, there is a significant difference. Pring to Sonoma and Sonoma to Abbotts Point are the fuel eaters. This could be because this is the longest part of the cycle. Even though ANOVA didn't show anything from crew vs fuel consumption, we can use the tables in the initial interpretation to see that crew 4 were the highest fuel consumers per minute in both Pring to Sonoma and Sonoma to Abbotts Point.
Questions for further research on crews.
Crew 4 & 5 were on the 2nd and 3rd cycle of loco 8339. Could it be that the crew is more inclined to skip notches because they may have been told they were running behind time and were under pressure to save time?
Could the time spent on a locomotive by a crew be a factor in notch skipping?
All crews have high notch skipping percentages from Sonoma to Abbott Point, Is this piece of track harder to operate for the crew?
What is the gender ratio of each crew? Do females skip fewer times than males?