<p>This study analyzes the results of more than 100 CFD simulations of smoke spreading in a double-deck tunnel with a branch duct exhaust system. A single fire source is positioned in the lower deck of the tunnel, flush with the floor and in the center between the side walls. The CFD study provides detailed flow information which could not be measured in the experiments reported in earlier work (Wang et al. in Fire Saf J 146:104144, 2024, <a href="https://doi.org/10.1016/j.fi​resaf.2024.104144">https://doi.org/10.1016/j.fi​resaf.2024.104144</a>), hence the CFD study provides very useful complementary information, which allows for better understanding of the observed phenomena. The following parameters are varied: fire heat release rate (HRR); smoke extraction flow rate; and the number of vents, as well as their location, opened in the lower deck (and leading to the branch ducts). The impact of leakage is discussed as well. The results reveal that: (i) the smoke spreading distance (SSD) increases with HRR as the ceiling jet strengthens, but once smoke/heat escapes through a tunnel portal, the effective HRR inside the tunnel drops and the opposite-side SSD shortens; (ii) leakage significantly reduces flow velocities and enlarges SSD, quantifying the sensitivity of smoke control effectiveness to sealing quality; and (iii) with symmetric openings around the fire and one-sided extraction, the ratio of downstream SSD to the total SSD is approximately proportional to the ratio of upstream to total air inflow volume flow rate, providing a predictive indicator for vent performance. Applicability bounds and long-tunnel usage are stated to guide design and operations.</p>