cheap heat exchanger for genset exhaust (air/air)

[shustrng]

I am new to the forum and not an engineer, just a science guy, but did see several threads on HXs so thought this might be considered.

Scenario: Hot air (<150F & 1600-2000 cfm) various engines/brands from the radiator of ~20 KW genset flows through the HX, which then uses exhaust heat 150 cfm (750-1000F)to increase the temp and use for dryer air for sand in drum tumbler. Max bk press on exhaust is 3 in Hg but only 0.5 in water on fan so may have to use bypass and thus get less flow from fan across HX. Quotes were efficient and got air to 197F but can't use because of cost. Must get cost to less that genset or <$3-5K. Granted, throughput through dryer may be slower to get to same drynesss but cost is a real limit. Higher price would not be justified on a payback basis.

What best might be done to lower cost of a quoted $12-20K SS double pass air/air HX? Lose fins? What is maj cost components? labor? Should I try just several bends of exhaust pipe in a insulated sheet metal box? Any estimate of what gain (temp) might be obtained?

Thanks, Mike

 

[Compositepro]

If you are only drying sand you may be able directly mix the hot exhaust gases with the air. This is 100% efficient and the lowest cost option.

 

[shustrng]

Since the exhaust gas will contain lots of water from the combustion, i was reluctant to do that. If the temp is too low will the water then re-wet the sand? Will soot or unburnt hydrocarbons make the sand sticky? Fine screening is the next step and the whole point of drying is to assist screening by reducing caking etc when wet.

 

[Compositepro]

The amount of water in combustion gases is very minor compared to the heat content (i.e., the relative humidity is very low). Most gas fired dryers, including laundry dryers are direct fired.

The possible contamination of your sand is a determination you must make. If your exhaust is that dirty then a heat exchanger may be hard to maintain from fouling. Either way the exhaust gases eventually end-up in the environment where people will breathe it after much dilution.

 

[zdas04]

I think the above understates the amount of water vapor in the exhaust stream. I did this calc a while back and I seem to recall that at the end of the stack that I was evaluating, the stream was over 90% RH, which is a bunch of water (somewhere around 0.032 lbm/SCF).

The evaluation I was doing was for a jacket that went around the outside of the exhaust stack (very low added dP) and the gas to be heated was inside the jacket (outside the exhaust stack). When I insulated everything I was able to reach my target 350F, but I had to cut flow way back to get enough retention time in the jacket. It looked like this arrangement was able to recover about 15% of the heat in the stack, which was about all I wanted to recover because I had a minimum exhaust temp to minimize condensation in the stack.

This stuff wasn't cheep. I think the jacket was on the order of $10k, and the labor to get everything installed was about the same. This was on a 500 hp gas-fired piston engine. For your $3-5k I would expect that you won't get much.

David

 

[MikeHalloran]

Rule of thumb: Combustion produces about 1-1/4 to 1-1/3 lb of water for every lb of fuel consumed.

Because a Diesel is not typically throttled, the water is carried in a considerable volume of hot air, but any interior exhaust surface that gets under ~212F _will_ condense water. There will also be some soot, so it doesn't make good tea.



Mike Halloran
Pembroke Pines, FL, USA

 

[shustrng]

At 12#/hr that would ADD 1/4 # H2O per min to the 150 cfm.
So yes, I believe that an HX must be used because the evap of 6% H2O gfrom the sand down to a 1% level is a cooling effect and I do not want to increase the humidity of the drying air or condense water back into the sand as the temp drops. Direct fired kilns operate at much higher temps and dont worry about such condensation. Waste heat recovery units operate differently. Nor do I want to burn more fuel to effect the drying. Rather, recover the most possible that makes economic sense.

So I am back to the orig ?s, what improvement(s) can be done with the exhaust other than a finned HX that costs 2-4 times what the gen set costs.? It has to have a better payback? Anyone know the relative costs contributions? I don't think SS vs CS will be much difference. So it must be the fins or the labor? Should I learn how to weld SS?

Are there any cost modling air/air HX softwares that I could do a couple of what ifs ? Again that don't cost an arm & leg?

 

[MikeHalloran]

I wouldn't use plain tube in a box. Spiral finned tube is surprisingly inexpensive, barely more than the cost of the bare tubing.

Fabrication is expensive. You can learn to do it yourself, but you need serious tools; stainless fights back, hard.



Mike Halloran
Pembroke Pines, FL, USA

 

[shustrng]

Thanks Mike, at least now i know where the costs are weighted. maybe I can find other quotes. Welding the finned tubes would likely be real tricky, but I might could do the housing and insulation myself and save some. CS might corrode quickly too, so I guess my choices are now pretty limited.

 

[btrueblood]

Long coils (>100 ft) of stainless tubing are surprisingly cheap (compared to 20-ft or 10-ft sticks). While Mike is correct that finned tube can also be fairly competitive with plain tube, the number of welds with a continuous coil is reduced to two, which can save quite a bit of cost in a HX. Winding it is a b***h, but is do-able with purpose-built tooling.

 

[MikeHalloran]

One finned tube is no harder to weld than ordinary tube. You just cut off part of the fins, or order it with plain ends.

Fitting and welding a hundred or so between a pair of tube headers could be a challenge, unless you hire or become an experienced TIG welder. I have drawn 'em up, and I have watched them welded; I don't think I see well enough to do it entirely by myself anymore. Not that I ever could.

Given the gas flows that are required and the allowable pressure drops, I don't think one long tube is an ideal solution for a gas to gas exchanger.



Mike Halloran
Pembroke Pines, FL, USA

 

[shustrng]

You guys are teaching me a lot. Way to go! But here is the unanswered question: What length of a tube is equal to the lenght of the more efficient finned tube. heat transfer question?

 

[MikeHalloran]

I think a first, very rough, approximation, is to compare the developed surface area.



Mike Halloran
Pembroke Pines, FL, USA

 

[ssn596]

WHAT IF
1. He is using screw conveyer to move sand
2. Screw shaft is hollow
3. Screw shaft is inclined
4. Put engine exhaust in lower end of screw shaft and it exhausts at upper end. Allow for water to drain at lower end.
5. Heat is transferred from exhaust to shaft wall to the screw to the sand.
6. Is efficiency really that important when the diesel exhaust is normally dumped to atmosphere and wasted anyway? At least he is getting some benefit for his fuel cost dollar.

Dan Bentler

 

[MikeHalloran]

That's a very creative idea. It has the virtue of constantly mixing the freshly heated sand (at the core) with the remainder of the sand, so it will do better than a simple conduction model would indicate.

The next interesting challenge is getting adequate bearing life, given the hot gas going through the screw core.

The challenge after that is mitigating corrosion of the screw core from condensed water flowing down inside it.





Mike Halloran
Pembroke Pines, FL, USA

 

[shustrng]

The hollow shaft auger does not allow the other half of the recovered btu to be used, namely the air through the radiator. Again that 1600+ cfm has doubled in temp from say 80 F ambient to about 160 F and will blow through the dryer (it's a big blow dryer, but not quite hot enough for shorter lenght/residence time).

OK, here is another design question: How much pressiure is created by each "U" of a serpentine (Long) tube. You can tell I have never seen the inside of one of these HX's.

My idea now is to buy some std finned tube lenghts and bend them to be parallel for 18-24 " with u turns (not 2 x90) and thus only have a header with 16 welded fittings done at a local shop. I could then experiment with backpressures vs lenghts and diams, etc. What is the finned (od) diam for a 3/8 tube? Place a thermister here and there during initial tests and a P-gage too. OK, to be sure the header gives no backpressure, how many 3/8 holes is equal to the area of a 2 in exhaust? let's see now, where did I put my slide rule.....

 

[MikeHalloran]

You can't bend finned tube. The fins are delicate, and the edges are sharp. You have to use elbows or unfinned u-bends. It's normally a custom product, so you can specify the diameter over the fins and the pitch of the fins, within limits.

The fins are made by bending a strip the hard way, flanging or wrinkling the inner radius. A diameter over the fins of three tube diameters is probably pushing what can be done, but talk to the fin-maker about it.

Fin pitches of 8..12 fins/inch are common, but what's achievable is also dependent on the material, so again, talk to the fin-maker.

A smooth u-tube will have a pressure drop that you can calculate, or at least estimate, based on geometry and assumed mass flow, using equations in books you probably already have. Sometimes heat exchanger tubes will be dimpled, or have turbulence inducers inserted, in which case the equations won't give you accurate results. After that, you use the estimated pressure drop to compute a new mass flow, and iterate until the numbers stop changing.

It may exist, but I haven't seen stainless finned tube as small as 3/8", and I wouldn't want to deal with the number of branches you need to get an equivalent to a 2" exhaust pipe. I'd guess 4 or 5 tubes of 1" od would be about right.

The tubes have to be paralleled, so you need two headers, of say 2-1/2" pipe.





Mike Halloran
Pembroke Pines, FL, USA