On the distribution of individual daily driving distances
研究分析了三种分布函数对个人每日行驶里程的拟合效果，发现威布尔分布最适合多数车辆，并能可靠预测插电式电动汽车的适用性。
Plug-in electric vehicles (PEV) can reduce greenhouse gas emissions. However, the utility of PEVs, as well as reduction of emissions is highly dependent on daily vehicle kilometres travelled (VKT). Further, the daily VKT by individual passenger cars vary strongly between days. A common method to analyse individual daily VKT is to fit distribution functions and to further analyse these fits. However, several distributions for individual daily VKT have been discussed in the literature without conclusive decision on the best distribution. Here we analyse three two-parameter distribution functions for the variation in daily VKT with four sets of travel data covering a total of 190,000 driving days and 9.5 million VKT. Specifically, we look at overall performance of the distributions on the data using four goodness of fit measures, as well as the consequence of choosing one distribution over the others for two common PEV applications: the days requiring adaptation for battery electric vehicles and the utility factor for plug-in hybrid electric vehicles. We find the Weibull distribution to fit most vehicles well but not all and at the same time yielding good predictions for PEV related attributes. Furthermore, the choice of distribution impacts PEV usage factors. Here, the Weibull distribution yields reliable estimates for electric vehicle applications whereas the log-normal distribution yields more conservative estimates for PEV usage factors. Our results help to guide the choice of distribution for a specific research question utilising driving data and provide a methodological advancement in the application of distribution functions to longitudinal driving data.