Empirical Study on CNG Refilling Behaviors of Different Timescale Based on Boxplot Method

Abstract

The number of CNG secondary filling stations is limited in cities because of their security requirement. However, the number of CNG vehicles is increasing. It is important to improve the efficiency of CNG secondary filling stations, as critical infrastructures in CNG industry, to meet the increasing demand. As a first step to find possible improvement methods, the periodical CNG refueling behaviors are empirically studied in this article. By boxplot method, the intraday, day-of-the-week and monthly CNG refueling behaviors are studied. Based on analysis of more 250 thousand records of a CNG secondary filling station in Tianjin, the hourly, daily, and monthly volume, initial pressure, final pressure and the time spent in each refueling behavior are presented. From the central tendency and outliers analysis, it is found that one possible way to improve efficiency is to consider factors in intraday and monthly timescale. Another way to improve efficiency is to reduce outliers, which are independent of time.