A vehicle testing programme for calibration and validation of an evaporative emissions model Giorgos Mellios a , Zissis Samaras a, * , Giorgio Martini b , Urbano Manfredi b , Steve McArragher c , Ken Rose c a Laboratory of Applied Thermodynamics, Aristotle University, 54124 Thessaloniki, Greece b Transport and Air Quality Unit, Institute for Environment and Sustainability, Joint Research Centre, Ispra, Italy c CONCAWE, Brussels, Belgium article info Article history: Received 4 November 2007 Received in revised form 14 March 2009 Accepted 17 March 2009 Available online 5 April 2009 Keywords: Evaporation losses Carbon canister Hydrocarbon emissions abstract A vehicle testing programme has been designed in order to calibrate and validate an empirical evapora- tive emissions model developed in previous work. To this aim, a large number of ‘‘targeted” tests have been performed on four vehicles covering a wide range of the model input parameters such as fuel vol- atility, ambient temperature, fuel tank and carbon canister size, fuel system materials. The fair agreement between modelled and measured values demonstrates that ‘‘bottom-up” modelling work and ‘‘top- down” vehicle testing may be combined to predict evaporative emissions on a vehicle level. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction Breathing losses through the tank vent and fuel permeation are in general the most important sources of evaporative emissions in a vehicle. Breathing losses are due to evaporation of gasoline in the tank during driving and parking as a result of normal diurnal tem- perature variation. In current vehicles vapour emissions are con- trolled by means of an activated carbon canister connected to the fuel tank. Various studies [1,2] indicate that liquid fuel seepage and permeation through plastic and rubber components of the fuel and vapour control system contribute significantly to the total evaporative emissions. An empirical model for the estimation of evaporative emissions from canister-equipped gasoline vehicles has been developed in previous work [3]. The model uses various fuel, vehicle and test- ing-related input parameters to calculate vapour generation in the fuel tank, canister breakthrough emissions and emissions due to fuel permeation and/or leakage. Experimental data from vehi- cles tests conducted in the framework of the EVAP Programme [4], using the current regulatory test procedure, were used to cal- ibrate and validate the model. However, only the standard temper- ature profile envisaged by this procedure was used, not allowing thus for a complete validation of the model outside this tempera- ture range. Therefore, the Laboratory of Applied Thermodynamics (LAT), the oil companies association (CONCAWE) and the Joint Research Centre of the European Commission (JRC) jointly carried out a test programme specifically designed to investigate the influence of various input parameters on evaporative emissions from modern passenger cars and eventually to calibrate and validate the empir- ical model. Special attention was given to the temperature profiles and the vehicle design characteristics (e.g., canister and fuel tank size) in order to cover a wide range of the model parameters. The present paper describes the main results of this investigation. From the validated model, a calculation methodology has been developed based on appropriate emission factors [6]. The applica- tion of the methodology, coupled with relevant activity data, e.g., parking duration, enables the calculation of evaporative emissions from a vehicle fleet in any spatial and temporal resolution [7]. 2. Experimental 2.1. Test vehicles In total four modern vehicles were tested, of which three were Euro 4 and one Euro 3. All vehicles complied with the 2 g/test emission standard of the current European SHED (Sealed Housing for Evaporative Determination) test procedure [5]. The selected vehicles had engine sizes ranging from 1.2 to 1.8 l and were all equipped with a multi-point injection (MPI) fuel system. Three vehicles had plastic fuel tanks and one had a metal one. Further- more, one vehicle had a pressurised fuel tank. Data on the selected test vehicles are listed in Table 1. For the application of the empirical model, data on vehicles’ fuel and vapour control systems are required. The data needed 0016-2361/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2009.03.018 * Corresponding author. Tel.: +30 2310996014; fax: +30 2310996019. E-mail address: zisis@auth.gr (Z. Samaras). Fuel 88 (2009) 1504–1512 Contents lists available at ScienceDirect Fuel journal homepage: www.elsevier.com/locate/fuel