Is there a way, as far as anyone knows, of measuring coolant flow without actually opening the circuit ?
Here is where I am coming from ...... the coolant loop for the engine goes from the engine hot end to a 205 degree thermostat, into the house, through a heat exchanger that cools the coolant, through the exhaust heat exchanger and then into the cold end of the engine.
I have been looking at temperatures in the system and I have come up with an oddity. The coolant, as measured at the input of the cooling heat exchanger in the house, is at about 80 C when the engine temperature exceeds the setpoint for overtemp shut down. I am expecting the thermostat to be wide open and I expect to see a temperature close to 100C at the cooling heat exchanger (instead of the 80C I am seeing). The engine shuts down at 220F so in theory, even if one assumes a small loss of temperature from the uninsulated pex run, 210F/100C or thereabouts seems reasonable.
I measured the output as the coolant exits the engine enclosure (about 2 feet past the thermostat) and I see about 85C there at the moment.
I am having a bit of an issue trying to think what is causing this temperature. I can see the temperature gradually heating up in the coolant run but anything coming from the engine should be (according to my theory anyway) be close to the rated temperature of the thermostat in the engine.
There is one possible issue in that the coolant hose past the thermostat takes a big dip before exiting through the top of the engine enclosure. Could it be that all the heat is trapped there ? I never thought this would be a problem since this is a pumped circuit and not a thermo-cycling type setup. In addition to the droop in the engine enclosure, the coolant hose makes another large dip a bit further and I can't eliminate that at the moment.
Anyhow, I am trying to figure out why I would see 220F just past the thermostat and only 85C only 2 feet further along the coolant path.
The only possible explanation I can think off is that the flow through the system is very slow. I am using a Taco 007 pump which is located between the exhaust heat exchanger and the cold side of the engine but I have no idea how fast the coolant is actually circulating. I will be installing some more sensors to check temperatures at various points which might help explain things a bit but flow is an unknown. I suppose I could install a Taco 012 which has a much higher flow rate and head but I would rather not just guess at this issue.
I would welcome any thoughts about the possible causes of the observed temperatures and/or thoughts about how to track down if there may not be enough circulation going on.
I do have a spare Taco 007 so that I could put two in series but I would hate to just try things because I use straight antifreeze and I always end up with spillage.
Jens
Taco XXX - presumably this is an electric central heating pump? If so there will be published data on head/flow/power requirement/etc.
Regards, RAB
Oh, and your comment about antifreeze. I do hope that is propylene glycol and not ethylene glycol. The former is a food grade chemical, the latter is very poisonous. Any hint of a leak into water, that might possibly be used for drinking, is a risk not worth taking.
Install a drain-down point, then there will be no need for anything but the smallest spillage.
Regard, RAB
I assume your using an IR thermometer to read temps. If so are the materials your shooting the same? Same color? Same texture?
For a simple flow indicator try to find one of those sight glasses from a decommissioned gasoline dispensing pump. It has the petalled butterfly thing inside that spins around.
I'm pretty sure you all know this... ;D
In order to use most flow spec charts, you need a discharge pressure at the outlet of the pump, and a suction pressure at the inlet. Subtract the inlet pressure from the outlet pressure, this will give you a poormans' net positive suction head pressure (NPSH). Multiply that pressure by 2 (roughly accurate for water) to give you head in feet, and use that number on the flow spec chart to get a more real-time flow measurement.
This will also give you a fair(er) representation of the restriction to flow in your system.
I know it's not really realistic to do in a DIY system - but all exchangers should have a pressure gauge on the inlet and outlet, or the pressure should be measured every now and then to see if anything's changed that would indicate fouling or plugging or a pump going bad...
I would find it difficult to believe you are losing 15°C in 2 feet of hose... I would supect an air trap at the measurement point - you're measuring the temperature of the air instead of the liquid, or a faulty measuring element...
(I know you must have a good reason, but I'm still trying to figure out why you bring your return flow from the domestic ex back through the exhaust heat exchanger just before reinjecting it back into the engine... I would have thought that the exhaust heat ex should be just after the hot water comes out of the head, just before the exchanger that you use to pull out the heat for domestic use. In your system, you are preheating engine coolant... Is the coolant so cooled by the domestic ex that you are worried about thermal shock to the engine? Or - would passing it through the engine, then the exhaust ex raise the coolant temp above the operational limit of the piping/equipment/boiling point of the coolant?)
Ah! Okay. I understand why you did it that way.
But - I think ther might be a flaw in your logic. It doesn't matter what temp the inlet is - hot or cold makes no difference as to how much heat you will pick up - as long as there is no state change...
What I mean is, if you heat water from 40°C to 80°C, it gains as much heat as if you heated it from 50°C to 90°C - the net change is 40°C in both cases. As long as the flow stayed the same.
So - if you ran the outlet of the coolant from the engine into the inlet to your exhaust ex, there would be more heat in your coolant available to be picked up by your domestic ex.
All things being equal - flows - there will be a net benefit to your heat exchange process, I think... if you have the ability to vary the flow on the cool side of your domestic exchanger, you can pick up more heat there, instead of sending hot pre-heated water back to your engine - which may in fact be confusing your thermostat in the configuration you have right now... which might be causing your shutdowns on high temperature...
Here's my reasoning:
Current Setup -
On engine startup, water is slowly warmed in the head, thermostat is closed until the head comes up to temp. There is a small flow through a hole drilled in the thermostat so flow is minimal. Yes?
Head comes up to temp slowly.
At the same time, the exhaust temperature rises very quickly. But, there is very little flow of coolant through the exhaust exchanger, so the water is being heated here to quite a high temp...
When the thermostat opens, flow increases through the cooling system. The first thing the head sees is a slug of high(er) temperature water that has been heated in the exhaust exchanger. This causes the thermostat to behave as if the engine is hotter than it really is... then, as flow increases, and the cooled water that has gone through the domestic exchanger cycles back, the water entering the head is cooler, causing the thermo to close... and the cycle repeats.
I would expect in this system that the cycle repeats in ever-shorter cycles... One of these cycles could introduce water that has been preheated in the exhaust exchanger to the point that when the hot coolant slug hits the high-temp shutdown sensor for the head, it triggers it.
Just guessing here...
With the exhaust exchanger in line right after the head, this situation couldn't happen. You would do most of your heat exchanging where you want it - for domestic uses. You would tailor the fluid flows through the domestic exchanger to control your coolant return temperature. Pull out as much heat as you can here.
In this case, there would be less chance of dropping the outlet temp of your exhaust below 100°C, since the coolant temp would be higher... Of course, this sytem will be limited by the temperature that you can heat your coolant fluid to without boiling it or causing a high-temp failure of something down the flow line.
I could see a lower temperature thermostat being indicated here, say 190°F... keep the engine slightly cooler, pick up more heat in the exhaust ex...
Just some musings from very far away... ;D
my simple comments:
I would go with the coolest water into the exhaust heat recovery as the gas retention time is shortest and the highest delta T is required for the best heat transfer (without condensation problems, of course).
I would, however, want the engine coolant circulation reduced to a minimum - ie first heat exchanger pretty well adjacent to the engine to reduce the hysteresis in the system, which may well present a problem.
Questions: there is presumably a bypass fitted in this system, so does this circulate through the exhaust heat exchanger and engine? Is it simply a resticted flow (orifice plate) or what?
A schematic diagram might be useful, but I doubt that would help me to help you unless there are glaring errors!!
Regards, RAB
Jens,
Two things come to mind regarding your pump flow rates....
1] System head:
The taco pump tops out at around 9.8ft. of head, at which point the flow is minimal.
Depending on the size of your coolant lines running the H20 storage tank, I would be surprised if you had less that 8' of head resistance. So the system may be limiting your flow to as low as 1 gpm.
2]Thermostat:
One thing I found with my 6/1 is that the 3 tiny air bleed holes I drilled in the thermostat are enough to cool the engine when I use a pump to force the liquid around. The engine is very slow to warm up (loaded) and the therm doesn't open. I will be replacing the thermostat with a single bleed hole version.
Where am I going with this ?.....
Perhaps your thermostat would not have to open very much to keep the engine cool.
I could be modulating in a partially open position because the incoming coolant temp is low enough that minimal flow is needed. This too would add resistance to you pumps efforts.
Just a few ideas from what little I know of your system. ;)
Veggie
I am beginning to understand.
I now know - there is a variable load on the coolant loop due to demand on the engine, and there is a variable load on the exchange loop, due to the fact that it is possible that heat demand can be met, and the exchanger side heat load can be saturated...
And plenty other variables, like line length, pressure drop, etc...
As far as optimising and controlling this loop, without some pretty extensive control valving and pressure/temperature sensing, I have to change my answer to D: I Don't Know.
This would be easy in my world - but expensive.
;D ;D ;D
Jens
A diagram of your system would be real helpfull...
As for the plumbing of the exhaust heat exchanger coolant output back into the engine, I would have to say that is probably going to be a problem. If the water entering the engine is already hot, it is going to have a hard time cooling the engine. I would liken that to plumbing the hot coolant out of one engine into the coolant input of a second engine. The second engine is going to overheat being fed the heat from the first engine... If your exhaust heat exchanger is anywhere nearly large enough, it should be putting as much heat into the coolant as the cylinders are(rule of thirds)... If the engine thermostat is controlling the flow, then when it is closed, it could be allowing the slow flowing coolant to superheat in the exhaust heat exchanger... What are the temps like at heatex coolant output/engine coolant input?
To maintain the best deltas, for greatest heat transfer, I would reccomend splitting the flows thru the two heat sources(engine and Exhaust heatex), then re-combining them at the outlets to feed into the home. A thermostat on the output of both engine and heatex should help balance the flows. Kind of like this maybe? the recirculation pipe and valve at the pump will allow you to taylor the pump output to the system. The open impellar circulation pumps don't have any real issue with being dead headed(when both thermostats are closed), but I never personally like the idea of deadheading a pump... The recirc valve allows you to control the head pressure that the pump creates, so you don't overpower the system. basically it is setup when the system is at full load/output, to deliver enough flow to carry away the required ammount of heat.
(http://i270.photobucket.com/albums/jj85/rmarlett/th_listeroidjens.jpg) (http://s270.photobucket.com/albums/jj85/rmarlett/?action=view¤t=listeroidjens.jpg)
Note the flow thru the heatex. The coolant flows in the opposite direction as the exhaust gas. This helps maintain the greatest overall delta between the two fluids and the greatest heat transfer.
As for measuring the flow thru the system, if you know the ammunt of fuel you are burning(energy input), and factor in the rule of thirds, measuring the ammount of temperature rise of coolant passing thru the system should allow you to roughly solve for flow...
Quote from: Jens on December 19, 2009, 03:50:52 PM
Quote from: Ronmar on December 19, 2009, 03:16:22 PM
A diagram of your system would be real helpfull...
I probably asked this before but forgot - how do you create these beautiful drawings ? All I can do is scribble something on paper, take a picture and then post that :(
Jens
I use the paint program that comes with windows. They are bitmap images that I save-as a .JPG's for a smaller upload file. The program is easy to use, but a little slow. I have quite a library of drawings now, so it is usually pretty easy to cut and paste different bits from other drawings to illustrate a concept. That last drawing the only thing I drew were the arrows...
Quote from: Jens on December 19, 2009, 03:50:52 PM
I probably asked this before but forgot - how do you create these beautiful drawings ? All I can do is scribble something on paper, take a picture and then post that :(
Over on another board I frequent, that technique is known as Crap-o-Cad, and is widely regarded as the fastest way to outline an idea...
Ron beat me to his reply, so my theory that it was MS Visio (or similar) is blown out of the water. However, that's what I'd use. OpenOffice Draw claims to have the same sort of functions as Viso, with the advantage of being free.
if it were me i would move the heat ex pump to the inlet side, that way there would be no cavitation? maybe?
and the water flow would be more turbulent, which might be of benefit with your exchanger?
just kickin sand with you
:)
btw, have you ever measured the btu recovery rate of your heat exchanger? if so i would be interested in hearing about that.
bob g
Better check your drawing again, Jens, by my weary eyeballs your proposed change doesn't do what Ron recommended, and seems to eliminate any input to the cold engine return. You should be able to reduce this to a single pump, as Ron shows so nicely.
Seems you might also need a cooling loop with radiator, for when your return water temperature gets too high, unless you are satisfied with an auto shut down at that point. It would be nicer to do it based on return water temp, rather than waiting for the inevitable engine overtemp.
As someone already pointed out, you have to do a differential pressure measurement at the pump inlet and outlet to be able to then calculate flow rate from the pump specs. It's the sort of basic info you need to design or tweak something like this. Veggie may be right- your head may be gone. Or not- we need more data on pipe length and number of (especially 90 degree elbows and tees) fittings to compute the head loss, even if we have to WAG the heatex head losses.
WAGs (wild ass guess) and TLAR (that looks about right) designs sometimes thump up against reality in an alarming way.
I don't quite understand your change either Jens. Are you adding a second pump? If so, the basic issue I got from your initial description remains. The coolant supply to the engine is still from a heat source(ex heatex). If this is a pump plumbing change, and not an addition, then as Bruce noted, where is the engine getting it's coolant supply from?
Jens
You probably should install a gauge on each side of the pump you are now using to give the flow determination a shot. That should not cost to much.
My vote is to install some sort of bypass so the coolant is moving all the time and then when the thermostat opens it dumps heat to the house until the house can take no more then the over temp murphy shuts down the listeroid. But the pump keeps moving heat at least until thermo closes.
Billswan
Here's how I think I would do it.
I would have the exhaust heat exchanger be a fixed flow loop, always full flow. It could be adjusted with a bit of back pressure with a ball valve on the coolant inlet, to provide a minimum coolant flow through the engine, through the hole drilled in the thermostat.
The engine coolant loop would be on thermostat; it would open/close on engine cooling demand. The flow through this loop could be fine-tuned so that with the thermostat fully opened - maximum cooling demand from the engine - there would be adequate flow 'robbed' from the constant flow loop[ to satisfy demand.
This loop would join the constant flow loop just before the domestic water exchanger(s).
Since both loops dump their heat in the domestic heat exchanger, then return back to the coolant pump, there will always be cooled coolant available to the engine - and maximum heat would always be picked up in the constant-flow exhaust exchanger loop.
Just my take on it, I think others are recommending something similar that would work, too.
Cognos's proposed configuration is the same as Ron's proposed fix, both returning the coolest water (storage heatex) to the engine. It's the only configuration that makes any sense to me.
Jen's existing setup, which returns exhaust heatex hot water to the cold return to engine is...not going to provide the best heat transfer efficiency, and will have serious problems as storage temps rise, as the engine sees "cold" return water temps of storage temp PLUS the rise in temperature from the exhaust heatex .
Some flow rate data and/or data on pipe runs and fittings would show whether adequate flow is being provided.
There is no way this system is performing at anywhere near optimum, as is.
Well since the engine was providing 220F or less, and the exhaust is providing 400F or more under load, I would hazard a guess that you are making steam in the exhaust heat exchanger... Your temp measurements between exhaust heatex and engine should show this.
Splitting the loops and putting a 90C thermostat on the output of the ex heatex, would allow the ex heatex to regulate it's own temp and stay below boiling. That of course is dependent on the pump being able to provide enough flow to transfer enough BTU away from the heatex...
Oops, missed the secondary loop part. How is that loop circulated? Got a drawing of that side? If it was circulating, and the shutdown limit out of the cogen is 220F, the secondary shouldn't have boiled unless it was completely heat saturated... I think you mentioned that the secondary loop provides ALL your engine cooling? If so, I would think you would have hit engine shutdown from lack of cooling long before then..
A little dab of heat sync compound where your sensor contacts the pex might help with your temp measurements...
It might be worthwhile to immerse the thermocouple in boiling water and an ice bath to verify their accuracy.
Quinn
Quote from: Jens on December 22, 2009, 06:13:22 PM
Maybe I am seeing thermal lag and that is all (the thermocouple has a very tiny thermal mass compared to the DS18S20).
Jens
Jens, I think that's exactly what's happening! I'm using LM35 for sensors rather than the 1-wires, but the packaging is essentially the same. I'm also seeing a lag in temperature readings.
could you put the sensor into a piece of copper tubing?
crimp one end shut, solder it off to seal it
drop the sensor down in and epoxy fill it, or pot it with whatever you like
then fit the copper tube into a ferrule T fitting?
that would get the prob down into the flow, and the ferrule would give you positive sealing
what i am thinking of using anyway.
bob g
Jens, maybe youve seen this:
http://www.instructables.com/id/Waterproof-a-LM35-Temperature-Sensor/
I didn't do it exactly like that, but it gave me the general idea.
I also used a RG59 crimping tool to seal the wire end off, filled it with heat compound and epoxy the open end closed.
Two of them that will screw in, I just soldered the copper tubing into NPT threaded fittings before fitting the sensor and wire.
Mine now looks like the automotive ones more or less.
I like your method, WJ, more than the link's method of using PVC and copper, and silicone.
The "instructables" article has a guy crimping the end of the tube in a vice- with the LM35 in place???!!!
Wow. And silicone that can never cure in a tube with wires. Hmm.
Steel filled epoxy is decent for thermal conduction, and a decent permanent plug in many fluids.
Once I land back on my feet after all the stuff I'm busy with at the moment, I'll write-up a small paper with pictures of how I did it, it's quite simple though, the crimping tool does a great job.
Afterward I used a short piece of heat-shrink tubing over the wire-end, just to neaten it up.
Jens
A comment and then a question.
In your quote below you say "the exhaust heat exchanger seems to be recovering significantly more BTU's than the engine coolant" gee that seems to fly in the face of the old rule of thirds, 1 third power 1 third heat into radiator and last 1 third out the exhaust as lost heat. Not arguing with you just wondering out loud? Wounder if any smart guys here would comment on that? Oh and please excuse me if I am not following what you wrote.
You also say "This unit is outside the engine enclosure and not yet insulated" I thought it was inside the enclosure? We are talking about the exhaust heat exchanger aren't we or am I missing something ?
Billswan
Quote from: Jens on December 23, 2009, 12:52:45 AM
The one thing that has me puzzled is the fact that given the same coolant flow rate, the exhaust heat exchanger seems to be recovering significantly more BTU's than the engine coolant. This unit is outside the engine enclosure and not yet insulated and I expected considerably less output from it. Some plumbing changes in the coolant system will be in order.
Jens
Re: the unexpectedly low heat value of the cooling water, maybe thermal resistance is the culprit. Heat transmission through relatively thick cast iron (cylinder head) and steel (cylinder liner) is going to be slower than through the thinner sections in the exhaust gas heat exchanger. It might be that the heat is being scavenged away by the coolant faster than it can pass through the metal in the cylinder head and cylinder liner, whereas in an exhaust gas heat exchanger you have thinner metal sections for the heat to pass through.
If the t-stat is cycling, it sounds like maybe you need some flow restriction. Consider closing a valve just enough to keep the thermostat fully open at max load.
Quinn
my .02 and its worth about that, :)
i would not want the Tstat cycling, i would want the engine up to temp and have the tstat running open
thermal cycling in the engine rarely is a good thing.
heavy trucks run with 195 tstats, and the engine runs over 195 to about 205, with the fan used to moderate
the temps and keep things running at a fairly stable temp under load.
now i suppose one could moderate the temp in much the same manner by regulating coolant flow, so that the tstat
could find its equalibrium point and run at some opening and stay pretty much there at that load.
i will defer to fellars on the virtues of coolant over exhaust heat flows, but it has been my observation that exhaust heat
being of higher quality seems more efficient at transfer to the coolant medium, if it is because of thinner wall sections or higher
heat differential or both, i don't know for sure,
but i suspect a bit of both myself
bob g
Jens,
I think you'll find that the total amount of heat you can recover from the coolant and exhaust will be more than you expect. But I wouldn't obscess too much looking for a nice clean 1/3 distribution.
To continue the earlier discussion, (I don't mean to argue; I think it's just an interesting point), you said: "In both the engine and the exhaust heatex the coolant removes the heat at the same rate it is generated.
Not quite. :) In both cases, the coolant removes heat at the same rate is is conducted, not generated. And the total heat of combustion, less work done, is divided at least three ways, not two, as the 1/3 rule would have it.
Last point first: The 1/3 rule of thumb assumes heat is evenly divided between coolant and exhaust gas and ignores heat lost to conduction and radiation (and vibration/noise, material flexure, etc). Recall that these engines burn very little fuel for their mass. There is a lot of metal to conduct heat away from the combustion chamber.
The 1/3 rule is a really rough rule of thumb. It should probably be stated:
Some heat is translated into work, (18-20% best case)
Some heat is conducted into the coolant.
Some heat passes into the exhaust stream, and,
Some is conducted away into the metal of the engine and is lost to radiation and convection. (I've read ~10% several sources)
My earlier statement about casting thickness didn't convince you. :-[
Imagine two identical engines: One with thick metal castings, and another one with thin castings.
Fourier's Law* holds that the quantity of heat transmitted through conduction is inversely proportional to the thickness of the material through which the heat is conducted. So of the two engines, the one with the thicker castings will conduct less heat per unit time into the coolant. That means more heat will remain in the combustion chamber, resulting in higher temperature and pressure than would be the case in the engine with the thinner castings. There would be no continuous heat buildup leading to a "meltdown" because after each combustion event, the exhaust valve opens, and the combustion products pass into the exhaust stream, where more heat will be available to transfer into the coolant.
On the other hand, the engine with the thinner castings would transmit a greater proportion of the available heat of combustion to the coolant, leaving less heat remaining to pass into the exhaust stream.
So, given the relatively heavy construction of the 6/1 type engine, I'm not surprised that more heat than was expected can be found in the exhaust.
*Q=kAdT/s
Q Heat transfer per unit time
A Heat transfer area
dT Temperature differential
k Thermal conductivity
s Material thickness
Merry Christmas, Everyone!
Quinn
Jens
You mind publishing your graphs? Of especial interest to me would be the input and output temps of he EX heatex and the engine... Also what is the size difference between your exhaust heat exchanger surface area and your engine coolant surface area? That could explain some of the differences you are seeing.
You are saying that the Exhaust heat exchanger is collecting twice the BTU as the engine coolant, but you have stated several times that you don't have the ability to measure the BTU output? Are you basing this assumption on strictly the thermal rise across the devices? If you go ahead and measure the actual BTU output, you wll probably find that the EX heatex is not collecting twice as many BTU as the engine. But in reality, the engine coolant is only collecting 1/2 the BTU that the exhaust heatex is...
Heat transfer is about thermal coefficient of the barrier material, surface area, flow, contact time, turbulence, and probably most important to this, the temperature difference from one side of the barrier to the other... In your case, The coolest fluid arrives at the exhaust heat exchanger and is exposed to the highest heat source. The exhaust gas also passes into the heat exchanger and expands and slows down, so it is in contact with the heatex surface area longer than in the combustion chamber. The wall thickness as Quinn mentioned is also thinner. All this and the high temp difference leads to a very high heat transfer rate for your given fluid flow, so no wonder you see a much larger temperature gain. You then take this heated coolant out of the heatex and pump it thru the overall lower temp of the engine cooling passages. Lower Delta = lower/slower heat transfer.
If you swapped the flow so it entered the engine first, then went to the EX heatex, You would probably see a reversal in temperature gains for a given flow. If you split the system and fed both "heaters" with the same temperature fluid, you would probably see comparable temperature rises for a given flow.
I think you are comparing apples to oranges...
Quote from: Ronmar on December 23, 2009, 06:23:30 PM
If you swapped the flow so it entered the engine first, then went to the EX heatex, You would probably see a reversal in temperature gains for a given flow. If you split the system and fed both "heaters" with the same temperature fluid, you would probably see comparable temperature rises for a given flow.
I think you are comparing apples to oranges...
That's what I've been doing for a 1000+ hours and it does exactly as Ron says. I see nearly equal temp gains in both of my heat exchangers. The coolant temp return to the engine is steady at ~140°F after running all day. I do have a thermostat regulated fan on the coolant return to the engine but it rarely turns on.
Ron and others helped me work the bugs out of my design and it made a big difference in all aspects. What has being advised here works, is the most efficient, time tested and easiest to implement design.
Thanks, Geno
If you maximize your heat transfer to fluid, you should actually have to run the engine less(more BTU transfered in a shorter time)... Since your fuel consumption is fixed with load and time, the faster the transfer, the better. What storage temperture are you looking for? Your engine is capable of providing up to 220F before it hits it's shutdown point, and probably pretty close to 195F over the long term average. You mentioned 3 tanks, 80 gallons each, for storage and some of that is used for domestic hot water, so surely you do not want over 120F-130F for safety reasons, or do you have a TRV on the domestic outlet?
Well 195F to 120F is 75F of delta at the worst case. 80F in the tanks would be 115F of delta. The transfer heat exchanger should be sized to deal with these working deltas to transfer the heat in the time allowed. Based on your storage sizes, we can figure roughly what you should see for run times based on a known energy input.
Say 5KW electrical load @ .125 gallon per KW/HR. Diesel is around 140,000 BTU per gallon, so a 1 hour run at 5KW will consume roughly 87,000 BTU in fuel. 2/3 of that is around 57,750 BTU that you should be reclaiming. You have 1920 pounds of water in storage, so that reclaimed heat should raise the temp of that water roughly 30F(1BTU raises 1LB, 1 degree F). How quickly it does it depends on the size of the heatexchanger and the delta across it. Got low delta? Need a bigger heatex to compensate for that design parameter...
I think you mentioned that the 3 tanks were plummed in series? This also complicates things somewhat as the tanks tend to separate vertically by heat. So you are pushing hot water into one which is pushing it's, I assume coldest water into the next tank and so on, so you will have cooler pockets in each tank, till ALL the water gets circulated. Ideally you would feed all 3 tanks in parallel and feed the hot water into the top and draw the coolest water out of the bottom to provide the coolest water to the heat source/exchanger to maintain the highest delta. Parallel tanks however provide their own set of flow issues...
Don't forget there is considerable energy imparted as kinetic energy to the gas flow. Turbos recover a fair part of this for boosting engine outputs. Yes, knock on effect, extra boost equals more exhaust velocity eqals more recoverable energy, etc. But it is there and is normally lost on a 6/1 engine as pumping losses.
Only added this as an extra way that the energy of the fuel is transferred out of the engine, lesser losses would be as sound energy. So the total energy balance of the system is complex to say the least. Rule of 'one thirds' is a pretty approximate one. Just like the rule of 2HP per kW.
Regards, RAB
Jens, Nice Graphs! Are those output from your data collection software? What are you using to collect the data?
I would say you do indeed have some sensor error/calibration issues. If you have a 96C/205F thermostat in the engine and the highest temp seen on the engine output graph is 90C... That is a big difference. I have no difficulty reading the coolant temp at the engine outlet fitting with an IR temp gun on my engine, and it is pretty much within a degree or two of the thermostat temp, so the heat is available there for measurement. Have you compared the sample points with an IR shot of the same area? The sawtoothing on the engine graph is indicative to me of the thermostat cycling, so the temp at that point is probably around 96C. That is maybe a 15C error from what is seen on the graphs? Since the thermostat controls overall flow, you can also see the fluxuations on the heatex graphs, so you are capturing those changes nicely. Right up to the point, that the engine inlet temp rose high enough that the thermostat went full open, and the temps started to climb away smoothly from there.
The fact that you are able to see the thermostat cycles tells me that you are getting pretty good thermal transfer to the sensor. Is there a calibration factor adjustment in your data collection device/software to calibrate the sensors output to the collection device input?
What was the data sample rate timing on these graphs?
Measurment consistency can, as you see now, really make a big difference in your results. By that last graph, I would say that the engine is delivering approximately a 50% greater heat rise than the EX heatex is.
Well if the temp IS accurate, you are missing out on another 15C of possible delta on the generator output to the secondary heatex. But the temp being accurate dosnt make sense after seeing the graphs. I have found thermostats to be pretty consistent devices. Take the temp gauge in your vehicle as an example. The engine output graph shows a steady temp increase at first(water pushed thru thermostat vent hole?). Then it starts a series of rapid cycles and the temp graph mostly levels off. That indicates to me that the thermostat has reached it's temperature and is cycling to maintain it. Now if it was a 80C thermostat, I would say the temp readings would back that up. But a 90C thermostat isn't even open at 80C...
Yes, you saw 96C on that last graph, but not untill the very end of the run. The thermal cycling that appears to me to be thermostat modulating the temperature stopped long before that, back in the low 80C range...
A couple of other observations I was thinking about. The short cycling on the engine output graph might indicate a little too much flow on the primary loop. A little less flow might allow the thermostat to open to a point and remain there. A low pressure gauge on the pump output might also show these thermostat cycles as slight pressure spikes every time the thermostat closes.
Boiling in the secondary loop with 100C input indicates perhaps not quite enough flow in the secondary loop. With 100C input, you should have something less than that on the output. The secondary loop enters at the storage tank temp. But with enough flow, the secondary loop heatex outlet should still be less than the primary loop inlet. No temp difference/little delta, no heat transfer... With the storage temp still well below 100C, you are not moving enough fluid to keep the heatex from boiling. This low flow is costing you potential delta/transfer rate near the end of your run..
Yep, the losses from a high storage temp are a problem. A larger tank at a lower temp will store the same ammount of energy, but have less loss for a given insulation thickness. It is also hard to get that last heat into the tank when the tank storage temp is approaching that of the heat source:)
Quote from: Jens on December 25, 2009, 10:02:56 PM
Fair enough ... but remember that I don't know the exact coolant temperature at which the secondary on the heat exchanger starts boiling.
Jens
Well if there is any secondary flow at all, the coolant temp reaching the heat exchanger must be somewhere over 100C.
Do you have graphs of the secondary loop inlet and outlet temps?
Folks,
Just a quick heads-up on measuring flowing liquid temperatures:
Even with the most modern (as in $$$) equipment available, chemical engineers need to resort to a lot of hand waving when trying to get an accurate energy balance on small scale systems (like ours)
Technically, the best way to get a temperature of a flowing liquid it to use a very small temperature sensor inserted into the center of the flow.
Usually the leads up to the sensor are very thin to minimize heat loss through the leads.
Then you apply a correction factor that includes the thermodynamic data on the liquid flowing, it's flow speed, distance of the sensor from the edge of the pipe, what type of pipe it is, phase of the moon and whether you got a haircut that day.
So don't feel bad if you can't make hide nor hair of what's exactly going on inside your CHP unit !
I've seen some engineers fresh from collage getting the edges of their diploma's dogeared from chewing on them due to a 'simple' energy balance on a small system.
Don't get discouraged, it *will* make sense after you do enough playing around, (at least it usually does...)
Daryl