Does fuel consumption between engines "scale"?
IOW, does a 6/1 (or a 6/10) running at say.....2000 Watts use about the same ammount of fuel per hour
as a 20/2 running at 2000 watts?
I understand a 2 cyl might use a bit more fuel, and that there are additional frictional losses,
but curious if part loading of a larger engine can be economical.
I want to run economically at light loads, but still have plenty of reserve power available.
I guess another possible solution would be 2 single cylinder engines coupled with a clutch that
could be engaged to 'bring in' the second engine?
I would also say 2 generators. The first KW costs the most to create. Each successive KW up to the load limit cost less and less in terms of fuel. IE: doubling the load from 1 to 2 KW does not double the fuel consumption. My experience with my 6/1, and the fuel consumption graph and research done with Mobile Bob thru load testing bears this out. Partial loading is less efficient. Here is my 6/1's consumption graph at 350' above sea level. There are of course a few factors that also effect efficiency such as engine temp. I found while running these tests that the 6/1 is more efficient when ran hot, than when cooler.
(http://i270.photobucket.com/albums/jj85/rmarlett/th_graph.jpg) (http://s270.photobucket.com/albums/jj85/rmarlett/?action=view¤t=graph.jpg)
Quote from: Ronmar on November 16, 2011, 10:12:03 PM
I would also say 2 generators. The first KW costs the most to create. Each successive KW up to the load limit cost less and less in terms of fuel. IE: doubling the load from 1 to 2 KW does not double the fuel consumption. My experience with my 6/1, and the fuel consumption graph and research done with Mobile Bob thru load testing bears this out. Partial loading is less efficient. Here is my 6/1's consumption graph at 350' above sea level. There are of course a few factors that also effect efficiency such as engine temp. I found while running these tests that the 6/1 is more efficient when ran hot, than when cooler.
(http://i270.photobucket.com/albums/jj85/rmarlett/th_graph.jpg) (http://s270.photobucket.com/albums/jj85/rmarlett/?action=view¤t=graph.jpg)
Ron,
That looks like really good data. Very carefully measured. Nice work!
Hi Fellas. This is one of the great mechanical givens - a small engine run with a good load efficiently uses less fuel than a large engine with the same load running unloaded. Cheers.
So the $64 question is.....how much worse on fuel?
Enough that it is worth fussing with another set, trying to split loads, or
coupling 2 singles with a clutch?
It would be an interesting Rube Goldberg study to couple 2 singles with some sort of
air clutch so the second engine could be cut in under motion.
I just wonder if that would or could work feasibly ??
I wonder it 2 singles would play nice together, or if they'd need to be 'aligned' crank wise to
keep from tearing each other to bits?
dk
Ron,
Seeing a chart like warms the dark recesses of my cold and clammy heart. That's impressive data, and not all that easy to collect. Now, how much black smoke were you getting at Pmax? Seems you might expect to see an inflection in the curve as you get near to overloading the engine. Maybe you didn't take it that far.
[Edit: I see the 0.5 and 3.5 kW were estimates (extrapolations). So it's likely that at 3.5 kW with all the smoke (unburned fuel) the consumption wouldn't show a corresponding increase in power output, at least not to the same degree.]
Quinn
Nice to know information when coupled with honestly taken data. Thanks,
Casey
Hi Guys!
I have a procedural question here . . .
How was the fuel consumption measured to FOUR decimal places? 0.0001 gallons is something like half of a half of a drop - you would have to have an analytical balance to measure that!!!
I get kinda nervous when I see numbers that seem too carefully measured to be true!!!
Most of the guys who are running a Lister(oid) would do well to measure the fuel consumption to +/- 0.001 gallons on anything like a repeatable basis - Is the methodology for this test spelled out somewhere that I can look at? I really am curious how that precision of fuel consumption was measured.
Presumeably the time that elapsed during the run was measured with a stop watch, and how you press the start/stop button can make a differance of .1 second, just to start with, so unless there are forty runs for each data point, its all still sort of a WAG to about the second decimal place.
Not trying to throw rocks here, but the differances in BSFC (yea, I calculated them, so the data are in a form that makes more sense to me!) are VERY small and the temperature is specified as 60 - 70 F that would make enough differance in the fuel density to wipe out half of the differance in BSFC that is portrayed.
So what REALLY happened? Is this one datum for each point on the graph? And how were the values of gallons to FOUR decimal places derived? And how a long a time period is represented for each datum? One minute, ten minutes, thirty minutes, an hour??? Inquireing minds, and nosy ones, like mine, want to know! <grin>
Again, I not trying to throw rocks for the sake of throwing rocks, I really am curious about the methodology!!!
Edify me, please!
Regardz,
Wayne Stayton
Regardz,
Wayne Stayton
Yea, so....
A 10/1 at 5kw load burns xx ounces of fuel an hour.......(or a 6/1 and a 12/2 as their more common)
And a 20/2 at 5kw load burns xx + how many more 'ounces'? per hour?
0.6oz more, 6oz ounces, 16oz??
Any guesses?
Anybody ever compare this?
We have (questionable?) fuel data from the 6/1 engine, but not any 2 cylinder data.
Wayne,
Here's how it's done. You measure it with a micrometer, mark it with a grease pencil and then cut it with an axe. :D
I get wrongly reported significant numbers when I forget to round off the calculator or spreadsheet. I'm still completely comfortable with the study. No one claimed it was brain science or rocket surgery that I read.
Casey
having fought the testing battle this is what i finally got down to using to get accurate and most importanly
repeatable results.
i measure fuel consumption in grams weight and then calculate it to gallons for the sake of discussion here and elsewhere.
i gave up and the stop watch method, however it is acceptably close enough if the test is long enough in duration to dilute
any start/stop issues. i used 15 minute runs when testing "only" after a full load run and only then after the engine temperature
was at full operating temp and stabilized.
i ended up using a GE digital kw/hr meter, one of the residential units
what i found was a way of setting the meter to calibration mode wherein it can measure down to watt/hrs with a +/- 2 watt/hrs resolution.
using both the kw/meter as described and a good electronic gram scale where i could measure everything in real time, i then became very easy to get accurate and repeatable numbers.
i could then even measure and calculate exactly how much fuel (btu's) it took to cover base loads such as belt drives, auxilliary water pumps,
field currents, windage and all sorts of other stuff, right down to some pretty fine numbers.
before the use of the gram scale and the GE meterhead i was left to try and do all these things, like taking all sorts of measurements, do the calculations, over and over again and it just wasn't much fun, was not at all accurate or repeatable.
personally i won't ever setup to do this sort of testing again without both an electronic gram scale and the GE meterhead.
and yes i have 3 of those heads, set aside for when i get back to doing this sort of thing again.
now as for the consumption curve bending over as the engine max's out in power...
with the s195 changfa i never saw this sort of expected bend in the curve, at least not the typical sharp drop seen with other engine's.
the engine would pull up to 6kwatts into a resistive load, no problem with the graph trending up linearly
right through 7kwatts, then up to just over 8kwatts with significant smoke, then the head gasket would let lose.
the head gskt was the limiting factor keeping the fuel curve almost a straight line.
8kwatts is not at all bad for an engine rated at 12hp continuous or 13.2 for one hour.
upgrading to a better head gskt allowed me to operate at the 8kwatt load, however i could not push past this
as that is all my st7.5 head can muster.
best economy for the changfa S195 as tested was right at 10kw/hrs per US gallon of pump diesel fuel, that attained at
full bore output of 8kwatts into a resistive load.
iirc Bill Rogers reported near 11 kw/hrs per US gallon of diesel fuel with his 1115 changfa operated at full load.
as for significant digits, here is my take
the digits are only significant if they can be used to extrapolate or predict results in testing under differing conditions.
any good testing must be repeatable
Ronmar's testing was very good, accurate enough and the results were predicted with astonishing accuracy at the time.
fwiw
bob g
Wayne, it was indeed stopwatch and flow method. The numbers on the graph are the raw data, please feel free to round them off to whatever gives you a warm and fuzzy:)
I setup a burret type rig with a container up high, and a thin line feeding down to a T in the fuel line feeding the filter. The bypass line from the injector was also fed back to this point so all fuel entering the engine was consumed by the engine. Immediately ahead of this T was a valve to the tank that was secured when the test was started. I would run the engine get it up to temp, put some fuel into the test rig and close the fuel to the main tank. When the fuel in the test rig burnt down to my measuring mark, the watch was started, and a measured volume of fuel was added to the test rig. When this measured fuel was consumed down to the mark, the watch was stopped. I went thru this routine MANY times, as I discovered after doing individual runs, I did some successive runs and got different data. Initially I was not warming the engine throughly enough, so the end of a series of runs was giving me better numbers than the previous single runs at those same loads. Fuel was measured with a syringe, as I didn't have access to a good scale. The load was purely resistive and energy determined with amp and volt meters. It would have been nice to have a recording energy meter like Bob describes, but I was using good quality (fluke) test equipment all in current cal status. I took timed data rounds and averaged these to determine the energy produced. Sure would have been easier with a device that did that for me:) The test runs were a pretty busy live by the clock process...
It was the best I could do with the tools I had available, and I like to think I am a reasonably through man. At any rate, I am very comfortable with the numbers, and was able to get excellent repeatability. If I wasn't I would never have made them public. When I did successive runs at the same load, I could predict to within a fraction of a second when the fuel would reach the timing mark. At the time Bob was working on a formula to calculate this very thing. One of the things that really brought it all together was when I gave him my initial few readings, he was able to predict with the formula to within a very close tollerance my last I think it was 2 readings?
Bob, where did we discuss this? It was way before SOMRAD, was it on Georges old forum, or the LEF?
Oh, and RCA, I used chalk, not a grease pencil:)
I see no reason that chart can't be extrapolated into a 12/2 twin. Simply double the consumption numbers and the loads. The various fuel consumption specs don't vary much between Listeroid designs. Like they do when you compare to air cooled variants, or high speed diesels.
In my case, I have a 20/2 twin. I configured it with a high speed and low speed pulley. However, I've found that it's efficient enough to leave on "high speed" without excess consumption with my daily moderate loads. TV, computers, AC on low, etc.
QuoteI see no reason that chart can't be extrapolated into a 12/2 twin. Simply double the consumption numbers and the loads.
Not that I'm any expert, but you wouldn't have 2 belts, 2 gen heads, 2 field coil losses, etc., so I'm guessing just a little less than double fuel use. Since we were talking about 4 significant figures and cutting w/ axes.... :)
Ronmar
it was on George's old forum that we went through the exercise
i am comfortable with your numbers they are more than accurate enough for the women i run with.
no real reason to get any more accurate unless one is deep into the weeds in R&D and trying to get a thorough understanding
of just what the efficiencies of all the related crap we use to generate power.
things like belt drives, differences in genheads, engines, temperature effects, and all sort of other crap.
the only other reason to get all wound up in high degree's of accuracy comes when one is trying to increase overall efficiencies of a genset
when a single percentage point becomes the benchmark.
as far as i am concerned your methods and numbers are good.
bob g
Chris,
A couple of years ago, I extracted the trend line equation from the old forum 6/1 data. Very interesting, did it again yesterday. Incidentallu, since we both have the same GTC 20/2 enigine and down rated to 850 rpm for me and 750 rpm for you: Do you want to get together sometime and accurately measure the fuel burn in 1500 KW increments for 15 minute periods? Since reviewing the 6/1 data a few years ago on the other forum, I've often wondered about the actual fuel burn rate on our twins.
If you are willing to join me for this endevor, I was thinking about using a few of the extra electric hot water heating elements I have laying around. I have two 4500 watt elements (230V) and could pick up another two 1500 watt element to make: 1500, 3000, 4500, 6000, 7500, 9000, and 10,500 watt laods. We'll need to measure the actual power draw on each of the elements but it should be pretty close to the rated figure. These elements could be submerged in a bucket of water with a small cold water flow keeping it from boiling. Shouldnt take more than a couple hours one day or night.
I'm sure we could do it with a flow meter too. The flow rate, once stabilized on the load would be more than accurate enough for the listeroid world. I have a Floscan meter with digital display. I'd say an hour of running would do it to a high degree of accuracy. Really, you could probably do it in 5 or 10 minutes.
Could be fun! Since you have a single too, the data could confirm the existing data.
I stand by my initial "guess" though. The twins consume double, + or - a few percent. The belt drive has minimal losses so even a 50% increase in belt loading won't result in more than a percent change in fuel consumption. Same goes for the gen heads. The larger ST head certainly has higher parasitic losses.
Chris,
How will the flow meter be connected? I have marine hoses with snap fittings.
How accurate or what is the resolution of the meter? Under no to light loads, the folw will be very small.
Just thinking: I believe, we'll need a no load reading to measure the mechanical losses from the engine, ST generator head, electrict water pump, and to keep a standardized base line, I supose we should also engage one of the three electric cooling fans fulltime. This should eliminte the error from a fan running occasionally under light load and then running full time under heavy loads, which will occur later on during testing. I would like to get good, useable data, to allow a reasonable comparison to the 6/1 fuel data curve.
Back to creating a load bank. I mensioned the two 4,500 Watt hot water heating elements I currently have, plus I would need to buy a couple of 1500 Watt elements, then build fixtures to hold these elements in a bucket of water. Instead of doing all of this, I wonder if I should swing by the Faith Farm Thrift Store to see if they have any beat up, old, electric stoves / ranges. If I can get one for $20 or $30 dollars, this might make a nice load bank for testing. You thoughts?
Bruce
Quote from: cujet on November 20, 2011, 01:26:06 PM
I stand by my initial "guess" though. The twins consume double, + or - a few percent. The belt drive has minimal losses so even a 50% increase in belt loading won't result in more than a percent change in fuel consumption. Same goes for the gen heads. The larger ST head certainly has higher parasitic losses.
You mean double for the same load?
As in a 10/1 uses 1 quart an hour for a 2000W load and a 20/2 uses 2 quarts an hour for a 2000W load??
How about the twin spreads the load over twice as many cylinders, plus a bit more fuel for added friction losses, plus a couple more percent due to light loading inefficiency?
But surely the twin is NOT half as efficient at the same load.....
i will make a bold prediction
a 20/2 driving an st12 will come out to be just as efficient as a 6/1 driving an st5 under the following condition
make sure to test both at near full rated load
the little added frictional/pumping losses of the twin over the single will likely be offset by the slightly more efficient
st12 that it is spinning.
if both are running at 2000 watts the 6/1 should have the advantage because the load is really too light for a 20hp twin to efficiently
provide for, however it might not be too bad depending on ones needs.
there is always compromises to be made if you try to make one unit do everything, sort of a "jack of all trades, master of none".
if ultimate efficiency is the target, the use of two or more differing sizes of gensets, each tailored to the expected load will burn less fuel
than one larger or two smaller running in a tandem setup will end up using.
the compromise there is with money, size and everything related to multiple units, or
just buy a big one and compromise on the lower load end with consumption of more fuel?
depends on ones "real" needs i suppose
bob g
lol....
At the RPM's and fuel burn ratios these engines operate at, I have a hard time imagining the difference is
measurable. BUT...make like I didn't write that.....I'm enjoying the debate and varying views presented here.
dk
Quote from: DKMC on November 20, 2011, 07:15:26 PM
Quote from: cujet on November 20, 2011, 01:26:06 PM
I stand by my initial "guess" though. The twins consume double, + or - a few percent. The belt drive has minimal losses so even a 50% increase in belt loading won't result in more than a percent change in fuel consumption. Same goes for the gen heads. The larger ST head certainly has higher parasitic losses.
You mean double for the same load?
As in a 10/1 uses 1 quart an hour for a 2000W load and a 20/2 uses 2 quarts an hour for a 2000W load??
How about the twin spreads the load over twice as many cylinders, plus a bit more fuel for added friction losses, plus a couple more percent due to light loading inefficiency?
But surely the twin is NOT half as efficient at the same load.....
No, I meant to simply take the published chart and double the figures. Not double the consumption at all. In fact, under mid loads, the consumption would be similar.
So, at zero load I'd expect .126GPH X 2 on a 12/2 twin running at the same RPM, or .25GPH. OK, so that's double at "no load", sure (but that's the only time you'll see that). As you travel up the chart,,, the single makes 1.5KW at .219GPH, so the twin would likely produce 3KW at 0.438 GPH. Clearly a little less efficient at that load (but not double)
My reasoning is that the diesel engine has a known efficiency by load. So a twin simply has 2 of the very same cylinders as the single, and the combustion efficiency is related to the individual load in the cylinder. So, double the load on a twin will result in the exact same cylinder load as a single.
Sure, there are small differences such as a center crank bearing and so on. However, I humbly submit that they are minimal in nature and won't result in much of a change. The major frictional component is piston ring swept area. And in the case of a twin, we simply have 2!
Quote from: Horsepoor on November 20, 2011, 06:52:03 PM
Chris,
How will the flow meter be connected? I have marine hoses with snap fittings.
How accurate or what is the resolution of the meter? Under no to light loads, the flow will be very small.
Bruce
The sensor uses 1/4 pipe threads on each end. It's quite small and is simply a counter of flow. It's very accurate as it counts volume per pulse. I'm not sure what the exact volume per pulse is, but since the sensor is small the volume per pulse is incredibly tiny. It's upper limit is about 30 gallons per hour. There is no lower limit. 1 pulse per hour is possible. (good enough for a model aircraft engine!)
It will count 1 pulse, or 10,000. The flow rate is, of course, how many pulses per unit time.
It's accurate enough to bet your life on. Many pilots do...
Hi Guys;
A look at a different perspective . . ..
Admittedly this was with spark-ignition engines, not diesels, so the numbers are meaningless, but the trends are there . . . .
When I worked at Ford, yea lo many years ago, wse looked at a v-8 engine management system that would have used four cylinders for idle/low-load, six cylinder for medium-loads and all eight for heavy loads.
The first cut, since hardware to disable valves on the fly was expensive/complicated, was to just run several engines on the dyno with the push rods pulled out of the cylinder that you wanted to disable.
Anyhow, the net result wa that we THOUGHT that if you could build the mechanism to disable the valve train components on-the-fly, we could gain about 10% om the EPA fuel economy test cycle.
The project was scrapped when they determined that it was going to be something like $400 to $500 per vehicle to build the necessary control/actuating hardware - which was huge for a vehicle where the total variable cost of manufacture was about $4,000 across the product line to which the technology was applicable - i.e. it didn't count four cylinder Pintos and in-line-six cylinder vehicles.
GM has, since the advent of computer controls, used this sytem on some Cadillacs, and I presume it must be worth the cost, since EVERYTHING in the auto industry is driven by the bottom line! <grin>
A diesel is inherantly more fuel efficient than a spark ignition engine, so there is probably less to be gained by using fewer cylinders running at higher BMEP, but I would THINK that there must be 5% or so to be gained by running cylinders at higher BMEP - the real question is, is this more than offset by the fact that the twin has one less main bearing, one (two?) less cam bearings, one less bearing in the injector pump, one less drive belt/shaft, etc etc. My guess is that it would be about a wash, but that is a guess based on absolutely NO data! <smile>
Of course the numbers are going to depend on what BMEP, you measure at!
My opinion which is worth EXACTLY what you paid for it!
Regardz,
Wayne Stayton
The sensor reads 29,590 pulses per gallon!
Chris,
Good enough resolution for what we will be measuring for about 1 pulse for every 3 drops of fuel. Other considerations unique to my GTC 20/2 setup before we measure.
My 20/2 and 6/1 normally run on 4 to 6 year old waste jet fuel. For the test runs, I have some freshly purchased pump diesel. This should help reduce one source of error on the energy content. If others will be drawing comparisons to their twins, I wonder how some of my setup configurations will impact the test runs.
My 20/2 is down rated to run at 850 rpm with extra heavy flywheels and a very heavy gen head flywheel (about 80 lbs) on an ST 15 generator head. The gen head is connected to the engine with an 8 rib micro-v-belt. Now the ST 15 has much greater capacity than the maximum engine output at 850 rpm. During maximum load test runs, I have gotten sustained 11,200 watts output (15 minutes) but I consider 10,500 watts to be the upper limit. So I have a bigger gen head than most people will be using on a down rated 20/2 (850 rpm) with heavier flywheels than most people.
For the NO LOAD base line, do we want to include the electric water pump and the electric fan? I can isolate these loads from the generator head for just the bast line - no laod test run, which is probably best since it will provide more compatible data for others. Do you agree?
Lets talk via cell phone on how best (easily) to be able to switch the flow meter bewteen engine fuel supplies on my engines and yours.
.
Also, my Metro 6/1 is running at 800 rpm, 8 rib belt, ST 7.5 generator head (Overkill again), electric water pump and fans. I have a meticulousily tunned engine, 4,400 matts absolute maximum, 4,000 watts normal maximum load. Add sea level and 80 degree F