Sampling approaches for road vehicle fuel consumption monitoring
EU Regulations introduced in 2019 for light- and heavy- duty vehicles contain provisions requiring the European Commission to set up a mechanism to monitor the real-world representativeness of the fuel consumption determined during the type-approval tests. This study proposes a sampling based approach to collect these data. Two probability-sampling methods (simple random sampling and stratified sampling) and one non-probability sampling method (quota sampling) are discussed. We use data from three user-based datasets (IFPEN, Travelcard and Spritmonitor) and the 2018 European Environment Agency CO2 monitoring dataset. All three user-based datasets provide fairly good representations of their respective countries’ sub-fleets and to a lesser extent the whole fleet. The standard deviation of the fuel consumption gap was consistently found to be approximately 20%. For a population of 15 million vehicles, using simple random sampling, and the standard deviation of the fuel consumption set at 20%, a sample of fewer than 3000 vehicles is required for estimating the average gap with a confidence level of 99% and sampling error less than 1%. Multivariate stratification with three stratification variables (vehicle manufacturer, fuel type and engine rated power) was the optimal combination, reducing the sample size by around 28% compared to simple random sample. Requiring strata specific estimators resulted to an increase of the sample size, as the number of stratification variables increased. Non-sampling errors, such as inaccuracy of On-Board Fuel and/or energy Consumption Monitor (OBFCM) device measurements, are expected to lead to an increase of the required sample size by at least 20%. Samples using quota sampling were taken and had a sampling error less than 3.5%.
KTISTAKIS Markos;
PAVLOVIC Jelica;
FONTARAS Georgios;
2021-08-11
Publications Office of the European Union
JRC122063
978-92-76-23986-4 (online),
1831-9424 (online),
EUR 30420 EN,
OP KJ-NA-30420-EN-N (online),
https://publications.jrc.ec.europa.eu/repository/handle/JRC122063,
10.2760/39369 (online),
Additional supporting files
| File name | Description | File type | |
|---|
