Critical evaluation of on-engine fuel consumption measurement

Abstract

To enable continued development and facilitate the adoption of new internal combustion engine technologies, the accuracy and repeatability of measurement methods used for verification need to be improved upon. A variety of methods are available for the measurement of fuel consumption based on the volumetric or gravimetric principle or by equating carbon in the fuel to carbon in the exhaust flow. Measurements of fuel consumption from five different experimental campaigns with varying engine set-ups are presented, highlighting discrepancies between gravimetric fuel balance and exhaust feed gas carbon balance. Differences were larger for cold-start tests and if all correction factors are neglected offsets can reach 7 per cent. The carbon balance and gravimetric methods have been considered independently to identify sources of inaccuracy and improvements have been suggested in the form of correction factors. The carbon balance estimate is dependent on a number of separate measurements, all taken at different conditions. To account for these, two correction factors were compared, the first proposed by the British Standards Institution and the second derived from the experimental conditions of each of the measurements. The gravimetric measurement was affected by changes in fuel temperature within the fuelling circuit and a corrective method was proposed based on the change in fuel density. When correction factors were applied to each of the measurement techniques for the five experimental campaigns, discrepancies were less than 1 per cent, which, in most cases, was small enough for there to be no statistically significant difference between measurements. In addition, in some cases the scatter of results was reduced, contributing to improved test-to-test repeatability. The improved performance when using correction factors was explained by including known disturbances, such as fuel temperature and ambient humidity, as inputs to the measurement system.