Optimisation and validation of Plume Chasing for robust and automated NOx and particle vehicle emission measurements
Christina Schmidt, David C. Carslaw, Naomi J. Farren, N. René Gijlswijk, Markus Knoll, N.E. Ligterink et al. · 2025 · 2 citationsRead the paper
High-emitting vehicles comprise a small proportion ( < 20 % ) of the vehicle fleet, yet are responsible for the majority ( > 50 % ) of vehicle emissions. Plume Chasing is a reliable, high-precision measurement technique that derives emissions without interfering with the vehicle being tested. Its characteristics make it well suited for high emitter identification. In this study, the influence of several Plume Chasing measurement and data processing methods on the results of derived on-road NO x and particle fuel-specific emission factors are investigated. A range of vehicles, representative of a common vehicle fleet, were tested under different driving conditions on a test track. The derived results were evaluated against on-board SEMS (Smart Emission Measurement System) emission measurements. We found that one of the best performing Plume Chasing data processing methods is based on the use of a rolling minimum for background determination. The average absolute deviation of the determined NO x / CO 2 emission ratios from the reference was − 0 . 2 ( 46 ) p p m / % for the heavy duty vehicle and 0 . 3 ( 29 ) p p m / % for the light duty vehicles tested. The methods were easy to automate and suitable for high emitter detection and quantification of emissions from two-wheeled vehicles. Inaccurate emission factors derived from Plume Chasing measurements occurred only in situations when emissions were significantly influenced by a strong plume from vehicles driving directly ahead of the vehicle of interest.
What the full text shows
Working prototype- The rolling minimum (RolMin) data processing method for NOx/CO2 emission ratios showed high accuracy compared to on-board SEMS reference. — Average absolute deviation of -0.2(46) ppm/% for HDV and 0.3(29) ppm/% for LDV (Test track driving)
- Plume Chasing demonstrated high correlation with on-board SEMS measurements for various vehicle types. — R2 between 0.88 and 0.98 (NOx measurements)
- Linear and robust regression methods can significantly underestimate emissions if the pollutant-to-CO2 ratio changes during measurement. — Underestimations of 50% (LinReg) and 75% (RobReg) (Situations with engine or SCR efficiency changes)
- Using an additional roof-mounted background inlet (BGinlet) reduces the influence of preceding high-emitting HDVs. — Influence reduced from ~35% to ~24% (LDVs driving behind high-emitting HDVs)
- Plume Chasing successfully quantified emissions for two-wheeled vehicles. — Median NOx emissions between 0.6–2.1 ppm/% (approx. 56–110 mg/km) (Motorcycle and scooter tests)
- Setup
- Diverse vehicle fleet (HDV Ford Truck, LDV VW Caddy/Transporter, motorcycle, scooter) tested on a test track under non-convoy and convoy driving conditions.
- How
- The study compared various Plume Chasing data processing methods (rolling minimum, linear/robust regression, and roof-inlet background) against on-board SEMS, Mini-PEMS, and NanoMet3 reference measurements. A measurement vehicle followed target vehicles at a 2-second gap, sampling the exhaust plume for NOx, CO2, and particle numbers (PN).
- Limitations
- Inaccurate emission factors occur when a strong plume from a vehicle directly ahead interferes with the target vehicle. · Particle number (PN) measurements may be overestimated depending on the metric and interfering sources. · Roof-mounted background inlets are costly with relatively small improvements in data quality.
Summarized by Gemma 4 from the open-access full text · 8 Oct 2026. Read it