Scalable speed-flow models for Singapore open expressways

To meet transportation demands and ease traffic congestion on local expressways, the Land Transport Authority (LTA) plans to enlarge existing highways and construct new expressways for Singapore. An important guide when planning and designing expressways is the use of a relevant and up-to-date set of traffic stream models. These speed-flow models will be useful for measuring the level of service, the effectiveness of a road segment and estimating expressway characteristics. However, existing models suggested in the Highway Capacity Manual (HCM) and other models developed overseas reflect traffic conditions vastly different from Singapore. Current models developed for local expressways also appear to differ from those defined in HCM (2010). Furthermore, none of them considered the effects of speed limits and number of lanes on expressway characteristics. As such, this study aims to model a worthy and up-to-date set of speed-flow curves for the 80 and 90 km/h, 3-lane and 4-lane expressways. Additionally, the study seeks to investigate the effects of speed limit on expressway characteristics. Expressway data was gathered using video graphic technology and speeds were collected using the speed gun method. Data was collected between June 2013 and January 2014 on weekdays, (Tuesdays, Wednesdays and Thursdays) for the peak (1700-1930) and non-peak (1400-1700) periods. Data extraction was conducted meticulously in the Transport and Geospatial Lab (TGS) over the research period. The modelling proceeded with Passenger Car Units (PCU) and the Passenger Car Equivalents (PCE) was obtained from an earlier study. A regression analysis was conducted on the 3 established models (Drake, Drew and Greenshields) to determine the best fit for the modelling of local expressway data. Results of the analysis indicated that the Generalised Drake model provided the best fit for the data points, with the highest R2 value of 0.979 (on average). Finally, the calibration of the speed-flow curves was done by statistical fitting of data points onto the Drakes’ model. Resultant curves for the 3-lane 80 km/h, 3-lane 90 km/h, 4-lane 80 km/h and 4-lane 90 km/h expressway models yielded capacities of 1912 pcu/h/l, 2281 pcu/h/l, 2010 pcu/h/l and 2112 pcu/h/l respectively. Correspondingly, free-flow speeds of 109.9 km/h, 87.6 km/h, 90.7 km/h and 69.8 km/h were also observed from the models. However, due to the lack of data points at the free-flow regime, the 109.9 km/h and 90.7 km/h free-flow speeds are extrapolated values for the 80 km/h 3-lane and 4-lane expressways. Comparison between the current 3-lane 90 km/h model and the model obtained in the CIS study in 2008 showed that both capacity and free-flow speed have decreased. Lastly, investigation into effects of speed limit on model characteristics revealed that both capacity and optimum speed increased when speed limit is increased. Overall, models obtained in this study showed similarities in shape and characteristics as compared to those developed in the past. The consistency of the models advocated their usefulness to LTA and engineers in predicting relevant characteristics of the 80 and 90 km/h, 3-lane and 4-lane expressways in Singapore. However, PCE values for each expressway site should be developed uniquely for each particular model, under different levels of service and operational speeds. The idea of slow lane modelling should also be explored in the future, so as to account for the effects of the different speed limits imposed on the goods vehicles and buses. Lastly, due to the tendency for motorcycles to travel in between lanes and their free-flow nature, separate motorcycle lanes should be modelled to account for their effects on capacity and flow. Bachelor of Engineering (Civil