Emissions produced by internal combustion engines during laboratory testing have been shown to not fully represent real world applications. Raw emissions measurements have aided the study of vehicle emissions resulting from in-use applications. In-use emissions measurements may be cumbersome; being that direct measurement of exhaust flow rate will often require that the vehicle exhaust system be modified. This paper investigated the feasibility of substituting exhaust air-to-fuel ratio and ECU fuel flow rates for the inferred measurement of exhaust flow rates. The air-to-fuel ratio for a diesel application was solved from the measured raw emissions of CO2 , O2 , and NOx. A 2002 Ford F-650, powered by a 2002 Cummins ISB diesel engine, was fitted with an Annubar™ exhaust flow rate measurement system (averaging pitot tube) in order to directly measure continuous exhaust flow rates. Concurrently, exhaust flow rates were estimated from air-to-fuel ratio, while “theoretical” exhaust flow rates were derived from engine speed, intake air density, and assumed volumetric efficiency. These surrogate measurements of exhaust flow rates were then compared with the direct measurements of flow rates obtained by the Annubar™ system. Fuel consumption estimates based on the air-to-fuel ratio derived exhaust flow rate and CO2 , theoretical exhaust flow rate and CO2 , and measured exhaust flow rates and CO2 were then compared with reported ECU fueling rates over the entire test and during NTE events. An error analysis was performed on the air-to-fuel ratio exhaust flow rate equation to quantify uncertainty resulting from the measurements and assumed parameters. The results showed that the air-to-fuel ratio derived exhaust flow rates compared well with the measured exhaust flow rates and the theoretical exhaust flow rates with an R2 value of 0.982 and 0.986, respectively. During highly transient events and motoring conditions, the air-to-fuel ratio derived exhaust flow rates were inaccurate due to analyzer response and zero fueling conditions, respectively. However, during steadier operation, the air-to-fuel ratio derived exhaust flow rate compared to the measured exhaust flow rate and the theoretical exhaust flow rate within 3.5% and 1.9%, respectively. Overall measurement uncertainty was most affected by the CO2 analyzer at high AFRs. The resulting fuel consumptions from AFR derived exhaust flow rate, theoretical exhaust flow rate, and MEMS exhaust flow rate compared to within 2% of each other. The ECU fuel consumption was 5–7% higher than the MEMS, AFR derived, and theoretical.Copyright © 2007 by ASME