Scalable speed-flow models for Singapore open expressways

• Main Author: Seet, Dai Qiang
• Other Authors: Wong Yiik Diew
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Civil engineering::Transportation
• Online Access: http://hdl.handle.net/10356/60006
• Institution: Nanyang Technological University

Travel demand in Singapore is expected to increase significantly by 2030 due to its huge projected population growth. To tackle this issue, one approach is the construction and expansion of new and existing roadways. In the planning and design of new roadways, speed-flow models are crucial in the determination of road capacities. However, existing speed-flow models have been developed according to the Highway Capacity Manual (HCM) which is calibrated for US roadways and may not be suitable for Singapore roadways. Although some local studies have been done, they were conducted many years ago and might not represent the recent traffic in Singapore accurately. The main focus of this study is to develop four speed-flow models for Singapore roadways, specifically for 3-lane, 4-lane, 80km/h speed limit and 90km/h speed limit open expressways under dry weather, clear visibility, no construction activity, good pavement and accident-free condition. This study also aims to investigate the effect of the number of lanes on the speed-flow models developed. Traffic speed and traffic volume data were collected from PIE, AYE, CTE and KJE using laser speed gun and video-recording technology. Traffic composition analysis and data points connectivity analysis were first done to better understand the characteristics of the traffic data collected. It was noted that KJE Chua Chu Kang survey site has significantly more goods vehicles and motorcycles as compared to the other survey sites. The data points were observed to be contiguous with no obvious ‘breakpoint’ between non-congested and congested regimes indicating the suitability of single-regime models for the traffic data collected in this study. The collected data were used to do multi-linear regression analysis to three well-known single-regime traffic stream models namely the Greenshields, Drew and Drake to develop speed-density models for the entire roadway. The Drake model was determined to be the best fit for the traffic data collected in this study for all four models developed. PCE values from Fan et al. (1997) were used to convert the vehicles into equal passenger car units. It is concluded that no definite trend on the capacity and free-flow speed can be determined in this study by investigating on the number of lanes alone. A more updated set of PCE values is suggested to be used to better represent the different types of vehicles in the traffic composition. The safety of the road due to the speed limit and number of lanes can also be incorporated into the speed-flow model development. Lastly, it is recommended that speed-flow models under bad weather conditions can be developed to determine the road capacities under degraded operating conditions.