It certainly helps reduce turbulent airflow as air passes between the fins.
(And improves cooling along the way).
Funny you should mention that.
Can't go wrong removing casting flash that seems to block airflow. However,
turbulent flow
is generally considered to be highly desirable in the case of a heat exchanger. The goal is to thin
the boundary layer (an insulator) and thus increase heat transfer, allowing the use of less mass to
carry away the same quantity of heat.
Consider the design of a cooling system for an airplane with a wide speed range...you know, total
performance, where the designer tries to makes intelligent trade-offs so no part of the operating
range is crippled. One might select a large, external diffusion, low Vi/Vo inlet, and couple it with
variable exit area.
In low velocity, high power flight like WOT climb, the slow inlet wouldn't be as sensitive to high AOA,
and thus maintain good upper cowl static pressure. With the outlet wide open, the lower cowl pressure
would be low, meaning a large pressure delta exists across the cooling fins. That increases velocity
between the fins, pushing the Reynolds number toward turbulent flow and thus increasing heat transfer.
Lots of mass, high heat transfer...good recipe for cooling at WOT. Sure, high mass, turbulent flow,
and slow exit velocity drives up cooling drag, but in this flight regime drag isn't a great concern.
Cooling demand is proportional to power. In cruise, we might need only 65% as much cooling as we needed
at WOT down low. So, we close the outlet, which does three things. First, the mass flow is reduced; air
that can't enter the inlet streamlines itself around the nose, hopefully without separation. Second, lower
cowl pressure is increased, slowing flow through the fins, which transitions toward laminar; heat transfer is
reduced, as is momentum loss due to drag. Third, exit velocity is increased; no surprise, given smaller area
and increased lower cowl pressure.
Alas, such a cowl design might suffer some external inlet separation in level flight at top speed. It might be
kinda blunt-nosed. Maybe it would only exceed VNE by 3 or 4 knots in level flight, autopilot on, despite
having enough motor (and cooling) to do a steady 1800 FPM two-up on the sort of 80F afternoon we had
here yesterday. Such is the nature of compromise. It spends most of its time at cruise power, and climbs
to some altitude for every flight; those seemed like the best conditions to maximize, as compared to a top
speed run made five or six times a year.
Are you starting to see why it really is all about the exit?