thermal recovery testing

[mobile_bob]

i have been negligent of my duties here, and with the current discussion on heat flow measurement
figured it might be time to open the discussion on developing methodology to attempt to get accurate
numbers for those that have interest.

so to that end i am proposing the following as a basis to work out a test protocol for the measurement of heat recovery
beit from the exhaust or the engine coolant

i will be using what i have witnessed and measured on my test stand using the changfa idi s195 engine

from published data from the manufacture and with some further testing i have determined that the engine is approx
32% efficient at converting fuel into mechanical work, and the st7.5 is approx 80% efficient at converting this mechanical
power into electricity.

this leaves the engine with approx 68% losses to predominately heat, although there are losses due to windage, external frictions,
some excitation current for the genhead, drive losses, heat that radiates off the unit, its hoses, piping etc. and other factors like noise.

so we have to make some assumptions that later can be recapped and cross checked to determine if it is likely we have overestimated
or underestimated the assumed numbers.

starting with the assumption that 8% will cover all the extraneous losses, this leaves us with approx 60% as heat that is available to harvest
30% exhaust and another 30% from the coolant.

so starting with our assumptions for 30% lost to the exhaust we make a series of test runs to determine how much fuel is burned in an hour
at some predetermined load level, in my case we go for a 8 kwatt load and burn one gallon of fuel over the course of that hour.

this gives us a burn rate of 1/8 gallon per kwatt hour which is acceptable

we also accept that our fuel has approx 135k btu's per gallon, so the exhaust carries away 30% of those btu's that were
burned in that hour, so

135k x 30% = 40,500 btu's available to harvest from the exhaust

now we know we cannot capture 100% of this heat, so we set a goal  of perhaps 75% or

40,500 btu x 75% = 30,375 btu's

now we install our exhaust heat exchanger, and connected a pump, hoses and a tank to circulate to.
we make some effort to insulate the exchanger, its related plumbing, pump and tank so that we don't lose
heat recovered which would lower our results and efficiency.

we then add water to the tank, and we measure how much water of course in gallons which weighs approx 8.35lbs per gallon

so if we put in 30 gallons we have 250.5 lbs of water.

we circulate the pump to stabilize the temperature of the water to that of the ambeint room temps,

if we start at an ambient temp of 70 degree's F, and start our engine and run it for an hour under the 8kwatt loading
we should end with a tank temperature of approx?

our target was to recover a measured 30,375 btu's, so

30,375 / 250.5 =  121.28 degree's of temp rise

121.28 = 70 (ambient startup) = 191.2 degree's F, should be the final temperature of the reservoir in the tank.

if however the end result is less than 191 degree's F, we would therefore conclude the heat exchanger is less than
75% efficient, but why? there are a couple of reasons we can explore later, or

if the end result is higher than the 191 degree's F, we could conclude the heat exchanger is more than 75% efficient.
if so we might want to check the temperature differential across the exchanger to make sure that the outlet temp is still
above the boiling point of water so that there is no condensation issues?

it is my belief that if the exchanger can harvest the available heat and still keep enough heat in the exhaust to stay above
~212F on exit, at the engine design load range and above, its done its job.

now going back to try to explain why the exchanger efficiency might be lower than our target of 75%, we might be able to explain
this because of the rise in temp of the coolant flow reduces our temperature differential across the exchanger, certainly the exchanger
can remove more heat with 70 degree water than it can with 140 degree water, so i would stand to reason that as the coolant tank
temperature rises the efficiency of the exchanger lowers.

anyway, just wanted to start a thread on the topic and get some discussion going

it seems like we are trying to quantify heat flows of the coolant, and and out, when perhaps there is a simpler way to determine
relative efficiency of a system.

maybe with a microprocessor, with programmed timer, factoring start temp in the coolant tank, engine load, a look up table for fuel consumption
at that load, etc, perhaps we could get the chip to crunch the numbers and give us a readout of efficiency? heat gain in btu's? or other info
needed?  rather than looking at temp in and temp out and trying to determine what it taking place?

to me it all comes down to, how much fuel is consumed? how much heat is going out the stack? these can be calculated accurately enough
so the bottom line is , how many btu's are getting stored for use?

once i got to the 30k btu level, i can then sit back and tweek flow rates, add insulation and do other modifications to make incremental
increases, by changing one thing at a time, then testing the results.

any thoughts are welcome here guys

btw, don't be to critical with my math, its late an i may have made a mistake or three
this thread is basically  one to spur some dialog, share some idea's and maybe we can establish a test protocol
that will be useful for others wanting to add a heat exchanger.

bob g

[Crumpite]

Bob g,

Great summery of the problem !

I'm afraid that it'll be kind of difficult to get a good mass/energy balance on a small system, at least on a short term basis.

Finding the temperatures is not too hard, but measuring the flow rates is a miserable business !
I worked in a chem plant for 20+ years and flow measurement was one of the worst problems to solve.
I'd sure like to find an inexpensive accurate way to measure flow...(I'd be rich then !!!)

I'm pretty sure that a lightweight tank on some type of scale would be the easiest to measure the mass flow of the fuel.
I don't know about the high speed diesels, but my 6/1 draws in spurts and then returns some back to the tank, making quick measurements pretty hard.

The heat gain of each part could be measured by circulating into a smaller holding tank and measuring the temp rise.
One of those little circulating pumps would would work fine - with the pressures at both the inlet and outlet, you could get a good idea of the flow from the pump curves.
You could measure the heat loss by shutting off the engine and drawing a curve of heat loss vs. temp.
With the known heat loss in hand you could then calculate the BTU's added vs. time vs. engine load.
That would tell you if your flow rate estimation was in the right ballpark.
And if *that* were true, you could then do changes in the part under test and by measuring the in/out temp you could directly figure the BTU's added.
(at least that's what I'd do, but I'm a nut on getting good data - arm waving doesn't do much for me, I need everything cross checked)

There would be a whale of a lot of data recording to get a good idea of the energy flows, but it would be pretty accurate.

Some of it could even be applicable to other engines of the same size/make.
I know of only a few folks who've done data collection and analysis on their CHP systems, it's a whole project of it's own.
It really would be worthwhile though, not much exists to plan a system from...

(I'm still trying to get my 6/1 put back together... sheesh...)
Daryl

[mobile_bob]

Daryl:

i guess my point was, i don't really care about determining a specific flow, and even a relative measurement of flow
is interesting but not necessary in opinion

to me what is of most interest or most important is the following

1. how much electrical power in kwatts am i putting out per hour

2. how much fuel in gr/kwatt/hr, that i can determine with a high degree of accuracy in testing and then build
a graph, after which a lookup table can be generated for a microcontroller to use for calculations, and

3. what is the heat gain in degree's F per hour of operation, at a specific load, again taking from the lookup table
and factoring against actual measured temperature differential/gains.

bottom line is if i burn a gallon an hour to produce 8kwatts of electrical power, i should have approx 75% of the exhaust
heat recovered as evidenced by x amount of degree's gain in y amount of water in lbs.

in testing all the graphs can be built and overlaid to get an overview of what to expect, and have the lookup tables
to compare against, if everything is within range i am good to go, if it falls out of range over time, then it is time to look
at various factors to determine why i am out of range.

if i can make 55-60 gallons of 140 degree water while running at peak load per hour of run time, i have done a good job
of heat recovery, and who cares what the flow rates are other than for the purposes of tweaking the system to see if
it can be improved?

my point being that i am not sure i see the need for measurement of flow, at least in a classical sense of being highly accurate

relative?  yes,, accurate?  for me it is unneccesary.

as for fuel flow rates, here again it is surprising how consistent these engine's are, if you set them up well

i know my changfa will burn x amount of fuel in gr/kw/hr at Y load, from test to test, day after day, over a fairly broad ambient
temp range, and the variance is so slight when it is observed i am not sure that it is not within the margin of error.

so testing the unit at various load levels from low, medium and full load, and generating a graph, i can very accurately calculate
within a couple of grams per kwatt/hr what a specific test will result in, at any point over the load range of the unit.

Ronmar ran some tests on his 6/1 about a year ago, i crunched number and forecasted with a high degree of accuracy what
his fuel consumption was going to be from one load level to the next, given only 2 data points.

given that degree of accuracy, i see little or no need for actual real time flow measurement when it can be calculated so accurately
over the full load range and over a fairly wide temperature range.

i suspect with a bit more testing over a broad temperature range, a set of graphs and formula's could be generated that would
give you extremely accurate fuel flow rates, probably more accurate than any affordable flow meter is going to deliver.

my opinion only, and based on lots of testing, number crunching, and confirmation of the results

ymmv

bob g

[Crumpite]

Ok, I understand now... 

That's the thing about writing - it's sometimes hard to get the exact meaning across.
It probably taken five min. to talk it out. (lost somewhere in the hours of other talk !!)

I agree wholeheartedly with your analysis, it would be very easy to program a micro to act an an "energy gauge".
It would just need to know the amount of HP being pulled from the engine, and then give you BTUs, fuel consumption, etc.

I'm curious how you determined the exact amount of Hp and efficiency of the engine and alternator ?
Like you say, if we can get a standard method going, we then can compare apples to apples.

Very interesting thread !
There are *so* many interesting and valuable threads going on in this forum !!
Lord, what a gold mine we have here so far. 

Daryl

[mobile_bob]

Crumpite:

let me try to explain my analysis in determining the efficiency of the engine and then the genhead
bear in mind i am stoned on pain meds, having a spinal problem (ongoing) and a upper canine tooth
absess (current) that is much like someone driving a railroad spike into my skull with a sledgehammer.

so if i ramble or are unclear it is somebody else writing this.


i had to have a place to start "mathematically" basically an assumption had to be made to eliminate one variable
so that testing could be done and results crunched to fill in the other blanks, after which time i was able to cross check
mathematically my original assumption.  the end result was a catch 22 where in order for the engine to be higher in efficiency
the generator had to be lower, and visa versa.

to begin with i used changfa's fuel consumption in gr/kw/hr mechanical as advertized, from that i could mathematically determine
that the engine was approx 32% efficient at converting fuel to mechanical torque/power.

because the generator is direct driven, there is no chance of belt slippage to account for, an because of the design of the drive couple
and the lack of temperature gain during the test, the losses are assumed to be negligible in the drive, so i did not account for them.
belt manufactures commonly state 2% as a loss figure, and we know they slip, and they do get warm and in some cases hot in use
so it would seem plausible that a direct drive coupler that does not gain in temperature would have far less loss than the 2% attributed to belts.
certainly under 1% and probably under the margin of error in any case, so i dismissed this part from the equation.

now after running the engine at peak rated power (the point that the manufacture rates its best bsfc numbers) and driving the generator
i measured the kwatt/hrs produced from a gallon of fuel,

measured the weight of a gallon of pump diesel and then checked the bsfc per kwatt/hr electrical, factored against the rating of the oem
leaves the st head at very close to 80% efficient. which btw falls within the range of efficiency of single phase heads of this class and size.

now going back and rerunning the numbers i determined that for the genhead to be even 2% better in efficiency the engine would have to be approx the same in reduced efficiency, and in order for the engine to be even a point or two better in efficiency meant the genhead would have to drop in efficiency into the mid to upper 70's percentile,  either seems unlikely. from that i accepted that the manufactures advertized bsfc in
gr/kwatt/hr mechanical was probably very close to reality, and the resultant 32% (calculated) engine thermal efficiency is likely very very close
to reality, certainly within +/- 1%

this is a rough overview from memory of the many steps mathematically i took to derive these numbers, and several fuel consumption tests
taken over many separate days with remarkably similar results.

from memory the math pretty well proves the engine to be 32% and the st head to be 80%, and for either to be much different would require
the other to be similarly much different, which is unlikely based on similar engine's and generators of this class.

if anyone is interested, and after i get this tooth fixed, i will look and see if i can find all the notes and calculations i made in the determination
of these numbers.

one final piece of the puzzle that helped me to confirm the numbers as being accurate, was in testing of my high output/high efficiency
alternators on the same engine. the alternator consumes exactly the same amount of fuel per kwatt/hr produced as does the st head
per kw/hr. (by exactly i mean right down to the same amount of fuel in grams/kwatt/hr)

now it is very unlikely that the modified alternator is over 80% efficient, and for it or the matching st head to be much under 80%
the engine would have to be 34-35% efficient, which is equally as unlikely in my opinion.

in other words if the st head is 78% efficient the engine would have to be approx 34%, or for it to be 82% efficient the engine would have
to be under 30% efficient,

it is far more likely that my numbers are damn close, either that or i have the holy grail in a modified alternator for 24/48volt charging, which is
very unlikely. 80% efficiency for my alternator is very good, and as good as any of the oem's i have seen, not likely it would be even one percentage point more than that.

rambling, here i know,

just not sure how else to present the findings/results without some detailed explanation and reasoning.

bob g

[Crumpite]

Bob,

Not bad for being on pain meds. 

That was a very interesting method of cross checking your accuracy.

One of the key problems, as you stated, is getting one good data point to start from.
I'm not sure the Listers have engine curves to get to where you started from though !

I wonder just how hard it would be to build a Pony brake to measure the Hp and torque of my Metro brand 6/1 ?

Anyone ever heard of engine curves for Listers
I think I'll cross post this on the genuine lister forum too.

There is nothing as bad as tooth pain, hope you can get to the dentist soon !
Daryl

[mobile_bob]

9:30 am tomorrow, and the dentist can have his way with me!



bob g