Powertrain modeling for realtime simulation
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesisAlternative title
Modellering av drivlinemodell för realtidssimulering (Swedish)
The goal of this thesis was to develop a powertrain model of a vehicle and parametrize itusing non-invasive sensors. The non-invasive sensors available were chassis dynamometer,the pedal robot and the vehicle’s on-board diagnostics which was accessed using a scantool. Non-invasive sensors were used so that the vehicle to model can easily be changed. Aparametrization methodology to parametrize the model for a new vehicle was also developedto facilitate the change of vehicle. The powertrain model is for cars with a combustion engineand a manual gearbox.
The engine model consist of two static maps, a pedal map and an engine map. The pedalmap is created using the fact that a constant pedal position and engine speed gives a constantthrottle position. The engine map is created in similar manner using that a constant throttleposition and engine speed gives a constant engine torque. The engine model also uses a firstorder lag element to model the time delay from a change in pedal position to a change inwheel torque. The driveline model is a rigid driveline model that assumes that the clutch,driveshaft and propeller shaft are stiff.
The developed parametrization methodology contains information on how to estimate theparameters of the model which are gear ratios, engine and driveline inertias, engine anddriveline losses, engine and pedal maps and the time constant for the time delay.
The powertrain model was validated component wise, as standalone and integrated intothe vehicle model against data gathered with the help of the chassis dynamometer. For thestandalone and integrated validation the gathered data were for different driving cases, suchas up and down gear-shifting, engine braking and skipping gears. The standalone validationshowed that the model performed well for the presented driving cases and the results hadgood data fit for 3rd gear and higher. However not for 1st and 2nd gear due to problemsin the pedal map. The pedal map was constructed on the assumption that the same pedalposition for all gears gives the same throttle position, which was not always the case. Thiscaused problems in some areas of the engine and pedal maps however in the validation ofthe maps it was shown that the maps for the most part gave good results.
Place, publisher, year, edition, pages
2014. , 74 p.
Engine, driveline, model, realtime, simulation
IdentifiersURN: urn:nbn:se:liu:diva-112126ISRN: LiTH-ISY-EX–14/4808–SEOAI: oai:DiVA.org:liu-112126DiVA: diva2:763494
Statens väg- & transportforskningsinstitut
Subject / course
Myklebust, AndreasAndersson, AndersKharrazi, Sogol