An on-line controlmodel for road tunnelwith longitudinal ventilation was proposed based
on neural networks and by combining traffic mode,l aerodynamics model and pollution mode.l
Simulationwas conductedwith themode,l where the traffic volume changed from 1 000 to1 400 veh/h
and the benchmark emission ratewas from 0. 008 to 0. 010m3/(veh·km). The results show that the
proposed system increases or decreases the number of fans in realtime following the changes in the
traffic volume and the benchmark emission rate, so that the volumetric fraction ofCO, as a control
targe,t is controlled justunder the standard limit(150×10-6).