Loughborough University
Leicestershire, UK
LE11 3TU
+44 (0)1509 263171
Loughborough University

Loughborough University Institutional Repository

Please use this identifier to cite or link to this item: https://dspace.lboro.ac.uk/2134/17959

Title: The energy consumption mechanisms of a power-split hybrid electric vehicle in real-world driving
Authors: Lintern, Matthew A.
Keywords: Real-world driving
On-road driving
Hybrid electric vehicle
Fuel consumption
Energy consumption
Chassis dynamometer
Drive cycles
Drive cycle development
Battery testing
Issue Date: 2015
Publisher: © Matthew A. Lintern
Abstract: With increasing costs of fossil fuels and intensified environmental awareness, low carbon vehicles, including hybrid electric vehicles (HEVs), are becoming more popular for car buyers due to their lower running costs. HEVs are sensitive to the driving conditions under which they are used however, and real-world driving can be very different to the legislative test cycles. On the road there are higher speeds, faster accelerations and more changes in speed, plus additional factors that are not taken into account in laboratory tests, all leading to poorer fuel economy. Future trends in the automotive industry are predicted to include a large focus on increased hybridisation of passenger cars in the coming years, so this is an important current research area. The aims of this project were to determine the energy consumption of a HEV in real-world driving, and investigate the differences in this compared to other standard drive cycles, and also compared to testing in laboratory conditions. A second generation Toyota Prius equipped with a GPS (Global Positioning System) data logging system collected driving data while in use by Loughborough University Security over a period of 9 months. The journey data was used for the development of a drive cycle, the Loughborough University Urban Drive Cycle 2 (LUUDC2), representing urban driving around the university campus and local town roads. It will also have a likeness to other similar driving routines. Vehicle testing was carried out on a chassis dynamometer on the real-world LUUDC2 and other existing drive cycles for comparison, including ECE-15, UDDS (Urban Dynamometer Driving Schedule) and Artemis Urban. Comparisons were made between real-world driving test results and chassis dynamometer real-world cycle test results. Comparison was also made with a pure electric vehicle (EV) that was tested in a similar way. To verify the test results and investigate the energy consumption inside the system, a Prius model in Autonomie vehicle simulation software was used. There were two main areas of results outcomes; the first of which was higher fuel consumption on the LUUDC2 compared to other cycles due to cycle effects, with the former having greater accelerations and a more transient speed profile. In a drive cycle acceleration effect study, for the cycle with 80% higher average acceleration than the other the difference in fuel consumption was about 32%, of which around half of this was discovered to be as a result of an increased average acceleration and deceleration rate. Compared to the standard ECE-15 urban drive cycle, fuel consumption was 20% higher on the LUUDC2. The second main area of outcomes is the factors that give greater energy consumption in real-world driving compared to in a laboratory and in simulations being determined and quantified. There was found to be a significant difference in fuel consumption for the HEV of over a third between on-road real-world driving and chassis dynamometer testing on the developed real-world cycle. Contributors to the difference were identified and explored further to quantify their impact. Firstly, validation of the drive cycle accuracy by statistical comparison to the original dataset using acceleration magnitude distributions highlighted that the cycle could be better matched. Chassis dynamometer testing of a new refined cycle showed that this had a significant impact, contributing approximately 16% of the difference to the real-world driving, bringing this gap down to 21%. This showed how important accurate cycle production from the data set is to give a representative and meaningful output. Road gradient was