Lister Engine Curves ?

[Crumpite]

Folks,

I was wondering if there was such a thing as an engine curve or chart or some info on the torque and Hp vs speed vs fuel consumption of the Listers.
Any information would help a lot.

We're looking into developing some sort of standard measurement standard to evaluate system efficiency and having a good data point to start with would simplify matters greatly.

Thanks,
Daryl

[mobile_bob]

Crumpite:

it seems like i have seen such a curve posted somewhere for a lister 6/1, might have been on the guy over in
englands site, peter?

that might be of some value,

failing that there is this

we have published fuel consumption numbers from india in gr/kw/hr mechanical, if we can believe them?

we can accept that the st5 is 80% efficient, or the sake of an arguement

we can accept the published 2% belt drive losses,

run a 3 kwatt load and measure the fuel consumed over an hour

weigh the fuel to see how many grams per gallon

crunch the numbers and factor by the genhead 80% efficiency and get an engine mechanical efficiency
then take that number against the amount of fuel consumed, and if the result comes close to matching
the advertised gr/kw/hr the indians use, well?

bobs your uncle!

(guess i got lots of nephews)

if the numbers are way off, well then we need to rethink the math or determine which is more likely
the indians telling a lie?,or using old lister ad copy? or something else is wrong?

my bet is the numbers won't be far off, as long as we are careful to account for other factors like
elevation which can skew the numbers significantly.

aside from that, i am with your prior comment, maybe we need a prony brake? can't be that hard to setup

might even be very useful for verifying all or our other assumptions, and test results,   

more data is always welcome in my books, the more data points we have the better, the more factors we
can actually measure the better, leaving less to assumption.

bob g

[mobile_bob]

Crumpite:

i have developed several testing sheets, that are laid out in openoffice
where once the test results are taken from the sheets and inputted into the puter copy
crunch all the related numbers to provide most if not all the answers one needs, along
with graphing them.

i can dig them out, and would be happy to work with you and others to improve upon them
to develop a set of test procedures and protocols, so that we can get apples to apples comparisons.

instead of "bricks vs watermelons" 

what is interesting to me at least is the ability to develop equations to predict results for these engines under
a variety of conditions.

from that we can determine relative efficiencies of a proposed design, and determine what is likely to truly work
and what might well be a waste of time and resources.

along with providing answers to all sorts of other questions we might not even think to ask going in.

bob g

[dubbleUJay]

Guys, this was my "thinking" as well when I started the Data Acquisition project. I just wanted to get a uniformed method to take the readings with first. It seems that Jens has beaten me to it too a certain degree as he already had a module to get some of the figures from.
It would be a great tool once its finish IMHO.

[mobile_bob]

the way i figure it is as follows,

(a) what i can measure accurately i will,

(b) what i can't measure accurately i will calculate using data that i get from (a) above


then i will either find or build formula's that will allow me to get the answers i need, and they
must be verifiable either by actual measurement or by mathematical proofs

what i can measure very accurately

1. fuel weight, before and after a test

2. fuel weight per unit volume, (grams/gallon)

3. amps output from a generator

4.volts output from a generator

5. power factor, hz, and watt/hrs

6. water tank temperature, before and after a test

7. coolant volume, and weight

8. ambient room temperature

9. rpm

and a few other things that escape my mind just now


things i cannot measure accurately

1. fuel flow rates

2. coolant flow rates

3. exhaust gas temperatures, i can get close but not accurate

4. coolant temperatures, again i can get close but not accurate

5. mechanical load in hp, or torque, in real time

and a few more things that also escape me right this moment


the point is to recognize what one can measure accurately and use equations that provide the answers to those things
i cannot measure accurately.

for instance the following equation

hp =( rpm x torque)/5252

so if one has no hp meter, but has a tach and a prony brake he can calculate hp with a high degree of accuracy, certainly better than 5%

now if one were to want to know how many watt/hrs the engine produces in actual work and has a tach and a prony brake, and a stop watch he might use the following, and test over an hour.

watts/hr = ((rpm x torque)/5252) / .745

so basically we work with what we can measure accurately

weight is easy, there are very accurate scales the measure in grams,
there are digital kw/hr meters that can be set in test mode to read out watt/hrs to very good accuracy
there are digital tachs
amp meters, and voltmeters, hz, etc (all of which are incorporated in the digital kw/hr meter.
digital thermometers
graduated recepticals to measure fluid volume before and after a test
etc.

i see no reason why testing and results cannot be expressed in single percentage points with a high degree of integrity

if i had to accept a 5% margin of error, i would forego even trying to measure or work with the results, they would just be
about useless to derive anything meaningful from.

you know when your methods and numbers are good when the method/math stands up to proofs, and when the numbers replicate
accurately from test to test on different days, with different folks making the measurements and observations, 
i would expect that if the numbers were only good to 5% then the measured results would shift over that range as measured from test to test
day to day, and from observer to observer.

you begin to accept the higher accuracy when you can run back to back tests and predict mathematically down to a percentage point what the
results will be, before you even go out into the shop to run the test, and especially when you have your 15 yo daughter rerun the whole test battery herself and watch her begin to predict outcomes before the tests are run. and be within a percentage point of actual measured results.

about a year or so back Ronmar ran some BSFC tests on his 6/1 and provided a couple data points, from which i was able to develop a non linear equation to predict the next step in the load range, iirc the test confirmed the number i got well within the 1% margin.  i am no mathematician either! so the equations are not terribly difficult to develop or work with, or i couldn't do them.

the issue as i see it is one of perspective, i choose to view things ( in this case) from the end, and others choose to view things from the start

sometimes that works, but for me and seeing how i don't have lots of expensive equipment i have to work from whatever angle i can with what
i have to work with.

make sense?

bob g

[mobile_bob]

i would like to expand a bit on this subject if i may, sometimes what is clear in my head is mud to others once presented by me.

i have a gram scale that measures up to 1000grams, accurate to +/- 5 grams

now if i were to run a kwatt of load for one hour, i might be off in my measurement +/- 5 grams per kwatt/hr

if however i were to run the engine and produce 10kw/hr, the same scale would still be off +/- 5 grams but
once divided by 10 to get me 1kwatt hour the error is now 1/10th of the original test

same goes for my kwatt/hr meter, it is plus or minus iirc 25 watt/hrs on that scale, so i am best served to test over
a longer period and divide the error, also
i can reset the meter to test mode and read watt/hrs +/- 2.5 watt/hrs, that alone is an improvement of a factor of 10
and i can further improve that by increaseing the testing length from a single kwatt/hr to 10kwatt/hrs where the
end reading will still be off by +/- 2.5, but divided by 10 leaves me with an error that is rediculously low.

that is what i call testing with the benefit of resolution

now if you run your lister with a fixed load, and with a batch of fuel that will provide for several tests, (homogenous fuel stock)
and you use the scale to measure accurately, test over a 10kwatt/hr run, before and after, along with using the kwatt/hr meter set
to watt/hrs, and when it clicks of 10k watt hours you note the fuel weight, you will have a very accurate number of gram's/kwatt/hr once
you do a couple simple calculations,

now step away and rerun the test tomorrow, and next week

so long as the temperature in the room is somewhat close in all three tests, you won't have but perhaps a couple of grams difference in any of the three tests, and if all is stable you will be hard pressed to not find all three tests spot on in gr/kw/hr bsfc

once you have that number, the rest is all math
from that you can calculate the efficiency of the engine and generator
from those numbers you can calculate the amount of heat available to harvest in the heat exchangers
to a degree accurate enough to then test against

the heat exchangers and the storage tanks hold X lbs of water, a before and after temperature measurement and
some math will tell you how many btu's you have recovered, comparing that with the calculated btu's available will tell you
how efficient your exchanger is, within a couple points of course

you can certainly refine the testing to increase the resolution here as well, and in doing so increase the accuracy in the process.

the key is it has to replicate, and that is easier said than done "if" there is a flaw in measurement, procedures or the math/equations
if on the other hand the test results repeat over and over again and confirm your calculations you know your accuracy is getting better.

there are other factors that i have not mentioned for reasons of not wanting to cloud the point further, some of which are

excitation current losses
drive losses,
windage losses, specific drive losses
and several others that can be accounted for with some degree of accuracy

bob g

[oliver90owner]

As I see it a test looks like being hours long?  For one result?

Try this for fuel usage: Burette on a T piece into fuel line. Valve in fuel line between tank and burette.  One stopwatch, or opto-electronic timer (if you must).  Observe, measure and record time for a known volume of fuel to be consumed by closing the valve and then timing between two set points of the burette. 

A result as such can be generated within a run over a short period, be checked several times in a short or long time period, to find the standard deviation for the measurements, variability for other parameter changes, etc, etc.

Most other measurements would remain stable for the short time of each fuel rate check.

Any one engine system should not be difficult to set up in this way.  Transferring data from an ideal situation to an actual installation would always result in a drop in efficiency - and that might be a quite considerable amount.

Regards, RAB

[mobile_bob]

Jens:

starting from reverse working forward

heat exchanger temp in and temp out with temp sensors:

please note, that i have no concern for those measurements, the bottom line to me is what is the weight of the water in the system
and what is the start temperature and what is the ending temperature, this tells me how many btu's i have recovered from what was available

that i can determine with more than adequate accuracy "if" i don't use electronic sensors, but rather a standard mechanical thermometer
the glass filled mercury type, they certainly are no more than a degree off at any point on their scale. even if they were off 2 degree's we are talking about recovered waste heat and what we are planning to use it for. unless we are hoping for bragging rights who cares if we are off 1000btu's one way or another in this test if we have clearly recovered 20k, 30 or maybe 40kbtu's

also it is fair to note that the efficiency of the heat exchanger is a moving target starting at a higher efficiency with cold water and ending with a lower efficiency at the end of the test when the water is hot,  so does a stated or calculated efficiency of 75% for the exchanger really overstated?  i don't think so, it certainly started at over 75% to be able to end at a point where it averaged out even calculated at 75%.

moving on to the scale and grams

one excellent method of determining the accuracy over the range of the scale is first to narrow the range, and then with a constant load
one might see that the engine is burning for instance 100grams per 5 min from 1000grams down to 900, then in another 5 min we would expect the scale to read 800, 5 min later 700, and so on...
if by the time it has ran for 25 minutes the scale now reads 500 grams, it is as accurate a scale as we are likely going to be able to afford to work with.

if we then measure by volume a gallon of fuel and weigh it, and determine how many grams it weighs by our scale, when we end up with our
tests and recorrelate to bsfc in gallon/kw/hr, which is btw what everyone in our group likes to report it anyway then we have returned to an acceptably high accuracy.

at the end of the day we find that a few grams one way or the other really doesn't change the resulting efficiency number significantly, and if we
strive to get as accurate numbers as we can this only helps to firm up our calculations.

also i do know this much

at 7kwatt load on my set, i know specifically how much fuel my unit will burn per hour, down to a gram or two, and have it repeatable over and over again,  then
if i were to add one 60 watt light bulb i can see the difference that one light bulb makes in consumption over the next 15 minutes, this tells me that my accuracy is not only repeatable but pretty damned accurate or the 60 watt bulb would easily hide in the margin or error which one would expect to shift over that margin from test to test.
there is no shift from test to test, so long as the load remains the same, the rpm is held the same and the unit is up to a stabilized temperature.

also it is fair to note that even if there were several grams in error it takes a lot of grams to shift the calculated efficiency one percentage point on the engine or by extension the generator.

basically this is all relative anyway, and specifics are meaningless except for repeatability, we must have a standard so that all testing is repeatable, so that all the equations work and return meaningful results.

i know that my methods are not NASA quality and certified, but they are a quantum leap over what is generally reported as no more than
anecdotal evidence where the" first liar doesn't stand a chance". ( no i am not calling anyone a liar, it is a figure of speech)

all i can do is try to develop sound test procedures, that provide repeatable results, then use accepted equations to determine efficiencies and then compare those numbers with published results of oem's that hopefully have millions of dollars of extremely accurate test equipment and lots
of folks with pocket protectors.

so when i arrive at the changfa 195 being 32% efficient i can compare it with published data from the oems for engine's in the same class
they average 32-35% efficient, so it would seem reasonable that the changfa made the grade at the bottom of the class

from those calc's i can determine that the ST7.5 at rated capacity is 80% efficient, and i can find reference to efficiency charts for genhead
that are 500kw, 250kw, 100kw, down to 50kw with the 50 kwatt being approx 84% efficient, is it much of a stretch to think that my calc's for a chinese harmonic winding excited genhead comes in at 80%?  after all if it were better the changfa would have to be worse?

at 32% efficiency for the engine, that corresponds to the published bsfc from changfa, maybe they lied and fudge a bit and the engine is only
30% efficient, therefore the genhead will end up being 82-83%  is that likely?  i think not for two reasons.

in conclusion i think where we can get into trouble with number is useing too small a sample group, running very short tests, with scales and meters that are not very accurate really expands the inaccuracy when you stretch the time out

a 5 minute test where we might be off a couple grams over the test, becomes 24 grams when we calculate for an hour.

can we do better? of course!

we should discuss methods and procedures and strive to improve both, we should seek out more accurate and affordable scales
and verify their accuracy with a set of standards (i think we can buy a set of standards relatively cheap, that are accurate enough for our
testing).

the point it we got to start somewhere, otherwise how do you know when you have made an improvement or made a mistake?

comments?

bob g

[mobile_bob]

Jens:

you got my curiosity up

so i did ran some calcs

on an 8kw/hr test run, if the scale was off 5 grams, it would change the engine efficiency 0.15%
so it could be down from the calculated 32 to 31.85, or up to 32.15%

the same inaccuracy on a 1kw/hr test would change the engine efficiency a full percentage point
so it could be as low as 31 or as high as 33%

length of test has a significant impact on overall accuracy, or so it would seem.

bob g

[Crumpite]

As I see it a test looks like being hours long?  For one result?

Try this for fuel usage: Burette on a T piece into fuel line. Valve in fuel line between tank and burette.  One stopwatch, or opto-electronic timer (if you must).  Observe, measure and record time for a known volume of fuel to be consumed by closing the valve and then timing between two set points of the burette. 

A result as such can be generated within a run over a short period, be checked several times in a short or long time period, to find the standard deviation for the measurements, variability for other parameter changes, etc, etc.

Most other measurements would remain stable for the short time of each fuel rate check.

Any one engine system should not be difficult to set up in this way.  Transferring data from an ideal situation to an actual installation would always result in a drop in efficiency - and that might be a quite considerable amount.

Regards, RAB

RAB,

It's not quite that easy !
Most of us are using diesel engines, some of them running as low as 650 RPM.
As a result the fuel is being drawn in spurts and worse a variable portion is returned back to the tank.

This results in a very jerky movement that makes measurement very, very difficult - it's been tried.

We're try to develop a methodology that will work for many types of engines and that make finding a measurement method much more difficult.

Daryl

[Crumpite]

Folks,

There are two ways to measure something with great accuracy.

The first uses very precision sensors, sampled rapidly.
This gives you an 'instantaneous' measurement of your system.
Example, if you measure flow rate, and input and output temps on a heat exchanger, you will know, exactly at that moment, what the BTU gain is.
But every little niggling detail needs to be looked at to eliminate errors.
This has been Jen's method.

The other way is to use not high precision sensors, but you *integrate* the data over time.
Example, you forget about flow rate in the heat exchanger and just chart the temp. rise over time.
To erase the heat loss errors, you just shut the engine down, and chart the temp. drop.
Now you know the heat loss of the system.
With those two data points you will know what the BTU gain is, but, usually with a much higher accuracy !
This is what Mobile bob has been proposing.

Both systems work, but generally the second method is cheaper but more time consuming and usually more accurate.

If you want to measure the accuracy of a digital clock, you can use a very high precision frequency counter to set it's operating frequency.
Or you can wait 24 hours and see how much it's drifted and then calculate the appropriate correction to apply.
Which do you use to set your wristwatch ?
I've used both, but not everyone has access to a lab frequency counter...

Daryl

Daryl

[mobile_bob]

to be perfectly honest with everyone when i first started to do testing, my accuracy was dreadful, so much so
that it became first frustrating and then later laughable

i used a volt meter, and clampon ampmeter, and i used the graduated cylinder as well
i couldnt get consistent results to save myself, the clampon would shiift based on position, vibration would shift it
a few tenths of an amp, and the voltage would shift a bit too but not so much, and then i also had to have a stopwatch


then both would shift significantly when there was a draft of air come through and cool the load bank! that was always fun
to account for!  you can't account for all that, and put up with burps, and vibration traveling back up the fuel hose making the
meniscus virtually unreadable in the graduated cylinder.

so here i was one eye on the stop watch, one eye on the ampclamp meter, one eye on the voltmeter, and yet another eye on the
fuel level in the graduated cylinder dancing around.  its a wonder i can see anything today!

something(s) had to change

first was the purchase of a GE digital residential kw/meter, they can measure kw/hr to within a half percent, i thought wow
now i don't have to measure amps and volts and have them shifting about, account for drafts and all that, the meter measures
and displays on screen,,, well think again

its was a huge jump but, the screen update time was too slow to capture events but every 5 seconds or so iirc, so i started to dig
around and found the unit has in internal momentary switch you can press to change the head to test mode, in that mode it displays
in real time in watt/hrs instead of kw/hrs, this is a dramatic improvement in resolution, and the update is done ever 5 watt hours, with a bit
of experience one can cut that in half easily.  this turns out to be a very accurate method of measuring kwatt/hrs, along with over 200 other parameters relating to power, everything from amps per leg, volts per leg, hz, 5 different distortions, peak kwatt, average kwatt, on and on

that alone eliminated so many entry points for error and frustration it wasn't even funny.

the second addition was the use of a precision scale, made by a japanese firm, its calibration sticker was out of date, but it allowed me
to consistently measure fuel consumption

with these new upgrades, i went out and bought one of those 19" open frame telco racks, with some shelves, the lower section of which holds my load banks, 4x 1500watt heaters, the mid level shelf carries my fuel resevoir and scale, the upper level holds my meter head/breaker box combo

gone are all the other handheld meters, the only other i retained was a killawatt meter that i have plugged into the breaker panel within eyeshot of the kwatt/hr meter head so i can monitor hz

now with a glance i can basically do a countdown to my test period in watt/hrs, and then drop my eyes to one guage being on the scale
and note the fuel weight, the placement is close enough that i could use a camera and take a snapshot should i find the need but so far
i haven't found a need to do so.

another interesting note, the meterhead can display on the screen 6 different parameters that are programmable, allow with a power flow
indicator which for grid tie will show which way the power is going, to the load or back to the grid, also there is an option card
that is basically a relay that can be closed upon any programmed parameter, such as low hz, hi/lo voltage  or any number of other out of whack
things one might want to take a genset offline for.

there is a company on the east coast that will program these heads to read out anything we want on the front screen, up to 6 different parameter out of a list of over 200,  you pick what you want displayed and they will program it and ship it back to you. or you can buy one programmed to your spec from them directly and they will pull one of the shelf do their deed and ship it to you.

as far as i am concerned this one piece of equipment is likely the most accurate tool a diy'er can have, and it is affordable. both useful in testing
and later useful in service, especially with the option card.

with that option card one could monitor inverter power in watt/hrs and when a calculated setpoint was crossed a command could be sent to start the generator and recharge based on that setpoint,
for instance we could calculate how many watt/hrs left the battery through the inverter to the load, send the command and through another calculation within the programming of the microcontroller start the genset and put back the exact amount of watt/hr used plus the extra watt/hrs needed to offset battery charging inefficiency (plus 10-15% typical)

anyway, these are some of the major changes i made along the way in an effort to not only be more accurate but to get consistent test results
that are repeatable.

my next move will be a digital scale that is easier to read, along with some standards to test it with, so that i can find a range within it full scale range where it is most accurate, and that is where i will do my testing within.

bob g

[quinnf]

Daryl,

Here's what Lister published years ago for the 6/1 and 12/2.





FWIW, a few years ago Jeff Maier http://www.utterpower.com/jeffm.htm calculated the fuel efficiency of his 6/1 generator system at 0.125 gal/hr @ 2500 W electrical load.  If you assume 140,000 BTU/gallon, then that translates to 3412 BTU/17500 BTU or 19.5% efficiency.  If an ST-series generator (I think he was using an ST-3) is assumed to be 80% efficient, then that means the engine by itself under that load is 24.4% efficient.  

Seems I recall that Bob, tenacious Terrier that he is, after much growling and shaking of the rag, produced a number that was the same using his CF.

Quinn

[Crumpite]

Huh, I thought I had the Utterpower pages memorized, but I missed these pages. Thanks a lot !

I notice that the graph only goes up to 6 hp not the 6.5 that's specified in the clones.
Do you have the rest of the manual or know where I can find it?  I'd like to take a look the the other specs of the engine.

I've heard that the older listers were a bit different than the more modern versions.

Thanks again !
Daryl

[oliver90owner]

Not Utterpower.  These are Lister pages.

In addition note the straight line interpolation between 5 and 6HP.

The line should/would be a curve but a straight line was all they added when the power rating was increased by increasing the rpm.  That should be obvious to anyone with more than a little grasp of the underlying trends such as these.

So the graph shape should probably continue as a curve and if the 6HP figure was arrived at accurately the gradient should have been reducing by that point.

Different fuels, lubricants, etc, etc, etc.

Lister ratings were conservative, so accurate, in that the power was more than required etc,etc, etc.

That is all.  That is all you will find for any one installation.  Every installation will be different - by the small amounts you are trying to resolve to.

Look at the Nebraska tests for tractors.  Accurate comparison tests but in a lot of cases a 'selected' sample sent for test.

The top and bottom line?  Yes, develop tests which are verifiable and reproduceable by all, but don't expect all installations to be more than 'in the ball park'.  Individual results will be just that.  KISS principle must apply.  Accurate, yes - but simple.

Spill return is almost constant (for one injector at one speed).  It can be returned to the fuel supply for a short term test, or even collected volumetrically in an accurate burette-style collector or weighed separately.  The obvious advantage is that umpteen short term tests can be carried out to check the variability (standard deviation, etc) and thus design the test so that it is sufficiently accurate and reliable for all involved to obtain reliable and comparable results.

One long test will give one average result over that period.  Several short tests should give that same result over that longer period, but extra data will show all the expected variability in that time frame.

Precision is often more important than absolute accuracy, for individual installations, to answer the question - has the value changed from one set of parameters to another?  - which is the better mode of operation? - has the fuel consumption actually changed significantly? etc etc.

I think some may be expecting rather more than might be attainable.

RAB