[jyl]

Just an idea, maybe a dead-end, but I'll ask.

I have what I suppose you engineers would call a "fluid dynamics" problem.

The task is to design a range hood exhaust system with make-up air (MUA) such that
(1) the smoke, grease, vapors, effluent etc from the range when cooking, grilling, stirfrying at full blast is all captured in the air exhausted
(2) the air exhausted is the exact same air that entered through the MUA. Repeat - the same air that comes in, goes out.
(3) the location and size of ducts, registers, etc works in the existing structure
(4) [bonus points] the MUA is passive, i.e. not fan-driven.

As you can probably guess, the goal is to have an exhaust system that is
- very effective at removing yuckies
- does not backdraft fireplace, water heater, etc
- relatively low-cost (existing structure, no interlocked MUA fan, no heating for MUA air)
- does not freeze [overheat] the house in winter [summer] with the exhaust is pulling 700 cubic feet per minute of air out of the kitchen and thus pulling 700 cfm of 25F [100F] air into the house. This is enough air to fill a box 7 feet x 10 feet x 10 feet, every minute, so you can see how the house could quickly turn into an icebox [oven].

So, I've found surprisingly little information about this. In the residential context, range and hood manufacturers mostly ignore MUA, some local codes require it, but there is nothing about how to design the systems other than a simple air balance. In the commercial context, I've found some helpful information (for restaurant kitchens) but it is at commercial scale/cost - 5,000 cfm systems, heated fan-driven MUA, etc.

I have come up with some schemes, but have no reason to think I have any useful intuition about airflow. So, I was thinking, how would you engineers handle this? And assume you didn't have a multi-thousand dollar software package.

Would you buy a student edition of some modeling software and set up the room and system?

Would you build a little scale model of the room - from clear plastic, say - and blow/suck smoke around?

Would you just pick a scheme and build it, sort of a $5K experiment?