Field tests were performed on three engines driving generators on a drilling rig during December 2009 and January 2010. Valid data points were extracted for the periods of December 17 to 22 2009 and January 1 to 6 of 2010. Selected data from Caterpillar’s engine control modules (ECM) were collected via an external data logging system every ten seconds. This allowed comparison of data from two engines in a common group during different time periods. The various engines are compared at the same time to consider identical ambient operating conditions. The same engine is compared over different times to consider similar engine condition data.
Fuel consumption reduction in the range of 1.7% to 2.5% is most often observed when the sample size is larger and the percent load deviation between the test and base engine is zero. Increasing the deviation of percent load up to 3% showed some data points with higher fuel consumption reduction (up to 11%) but a consistent trend is not evident. Many data points continued to show a 1.5% to 2.5% fuel consumption reduction even when the sample size was substantially increased by allowing the percent load deviation to increase up to 3%. G1 also showed a larger improvement than G3 when compared with G2 using the additive. In the case where G2 is compared G3 using the additive, the fuel consumption reduction is least. The variability is obviously very high when observing the data distribution. It is difficult to draw any conclusions with the current data set. Data quality issues, mostly imparted the ECM, suggest that the tests could be repeated with much better results and higher accuracy. Caterpillar specifies accuracy of their power estimates at +/- 3% and fuel flow estimates at +/-5%.
Clearly the reported fuel consumption reduction is less than the reported accuracy so the validity of the observations must be questioned. It is recommended that a follow-up field demonstration be performed with high accuracy direct measurement devices (i.e. flow meters, power measurement, etc.). This should increase the data accuracy and provide a more conclusive evaluation.