This begins a series looking at bulk traffic flow when modelled with contemporary behaviour patterns. As with many aspects of transport design the underlying assumptions reference historic simplifications derived from empirical data hiding many subtle effects induced into the traffic that can be explored with the denser data available from simulation.
Real world data from several Australian examples were used to tune a driver model such that good agreement was achieved with those limited data sets and the resulting parameters produce recognisable traffic behaviour. Within each simulation run (represented by a single point on the plots) the drivers were in a diversity of vehicle shapes and each had varying targets for their speed and comfortable following distance. Here is the simple result for a divided 2 lane road comparing throughput for different speed limits.
For each speed limit the average speed gradually slows with increasing traffic density until they converge into a common limit as the flow becomes unstable and throughput drops. Already present in this simple example is the traffic organically separating into a slow and a fast flowing lane, and impatient drivers weaving through the traffic. The discontinuity as congestion increases further is as the flow collapses and becomes stop-start traffic with static waves of stopped cars forming on the roadway between smoothly flowing sections commonly seen on overloaded highways.
The units of cars per hour per lane is measuring the utility of the road and its ability to deliver traffic to a destination, more complex network analysis including routeing and intersections would be used to evaluate delay/travel times. Sadly this example is a little dry on insight but lays a confident basis for the following explorations.