Diesel Fuel Flow Meter for mechanical engines data presentation and smoothing Q

[vas]

hello all,

a few years ago I started searching for cheaper/diy alternatives to Maretron regarding fuel flow monitoring.
After many setbacks, I ended up helping a company doing diesel fuel monitors for trucks open up to the marine market.
I hadn't realised that fuel monitoring in HGV is such a big business. (I guess mainly to help logistics companies sort out diesel costs and not be stung by drivers?)

Anyway, got two 250lph devices foc for testing and feedback. Each one is two separate measuring chambers with in and outs with a common electronic unit on top with NMEA2K cabling. Similar to Maretron and apparently v.accurate as well.

So to cut a longish story sort, and after 3-4 incarnations of the s/w in order to properly sent the right NMEA2000 sentences, identify port/sbrd engine, etc, I now have decent fuel flow data for both engines. Garmin (and all others that support fuel flow data) does some lt per nm calcs as well which are really helpful.
Following are three videos I took on Sunday evening coming back to port on flat seas, with a bit of growth at the places I've not put proper coppercoat this year (will do in spring), with trim tabs not functioning as in extending fully, and clean props (which really impresses me after almost 4months in the water, but I'll do another thread on that later). Also note that GPS vs Autopilot heading is almost 30degrees off (when going Northish, but spot on when going Southish...), need to do the a/p config thing again (turning around in flat seas etc).
Fins were parked in the middle and slightly offset in most videos (was mainly playing with them, but will do more testing during the next couple of weeks)

Now what I'd like your opinion is if you feel that the values reported should need any further smoothing. The guys doing this h/w never bothered with presenting the data realtime as it's basically a logger. So they are happy to do the calcs (and they are indeed accurate!) but not so sure how much to smooth the data.

Note smoothing or not, wont affect the overall values logged/reported and hence the overall trip fuel values.
Note2 during these videos, I had already been travelling on that speed for at least a couple of mins before taking the shot so flow and return was theoretically settled on the conditions...

https://youtu.be/CCP8yinDmyE (only two embedded videos allowed apparently, so third just a link to youtube)

there's another one at 6.5kn burning 1.1lpnm or so and I somehow missed my favourite speed of 7.5kn but not much point boring you to death...

so, should they smooth a bit more, or not?
If you do have fuel monitors how stable are the values? I'd guess that ECU based monitoring would be more stable, not sure on mech. engine ones though!

cheers

V.

PS. I personally think they should smooth them more

 

[MapisM]

so, should they smooth a bit more, or not?
If you do have fuel monitors how stable are the values?
I'd guess that ECU based monitoring would be more stable, not sure on mech. engine ones though!

I never had fuel monitors on any of my boats, 'cept the ECU-based one on the petrol Fountain.
But I've seen several of them (Cat, MAN, VP, MTU), and IIRC none had any smoothing algorithm, and just seem to react in real time, when for instance you go a bit up/downhill on a wave.
And regardless of how other systems work, I also don't see the need to stabilize those numbers, 'fiuaskme.
I'd rather include an average of the last X minutes/hours, if feasible.

As an aside, are those 250 l/h devices meant as 2x250, or 250 total?
If the former, they could cope also with the max fuel burn of my boat, and I might be interested!

 

[vas]

I never had fuel monitors on any of my boats, 'cept the ECU-based one on the petrol Fountain.
But I've seen several of them (Cat, MAN, VP, MTU), and IIRC none had any smoothing algorithm, and just seem to react in real time, when for instance you go a bit up/downhill on a wave.
And regardless of how other systems work, I also don't see the need to stabilize those numbers, 'fiuaskme.
I'd rather include an average of the last X minutes/hours, if feasible.

As an aside, are those 250 l/h devices meant as 2x250, or 250 total?
If the former, they could cope also with the max fuel burn of my boat, and I might be interested!

P.,

you have to consider that these things (as all electronics) work at rather fast sampling (probably not the right word for it) they can pump data at the NMEA2K bus twice a sec or a bit more (specs of the particular NMEA sentence, not their needs or wishes)
Within that time interval there are more than a few turns of these accurate small chambers they have, and if you add the temp compensation and all that, they are doing all sorts of calcs on board anyway.
I'd expect that ECU based systems do some data manipulation, call it smoothing if you wish. My only fear is that since the guys are going to market it soon I'd be interested in hearing (and passing over) any objections ppl in the asylum may have on the value fluctuation reported...

Average of the last X mins/h is not really feasible, not supported by NMEA2000, no way to present it anywhere. However, that's more or less what I mean saying smoothing. So maybe average of last 10secs would be appropriate...

Only two meaningful values is instant consumption (the two values in l/h on the top of the 4inch display) and then the calculated l/nm
and actually of the two, only the l/h per engine is NMEA2K the other is calculated within the display.

Regarding sizes, they have a smallish one (iirc 120l/h) a 250l/h and a 500l/h (and I think they can go larger...)
These are values PER CHAMBER, so your pump must not be able to flow more than the value for it to work.
I'm not sure yours will be within the 250l/h/chamber, is it? else you go larger, but you don't want to go too large on cost and accuracy on small flows grounds...
Couldn't find accurate value on how much my diesel pump flows, and ended up getting the 250lph on the grounds that my SEPAR filters wouldn't flow more than that
very scientific I know!

cheers

V.

 

[Whitelighter]

Think they look ok. It might be nice to be able to select instant or maybe 1 min average but I suspect it will be either or.
Given the choice of 1 I’d take instant.

Any idea when these might be commercially available?
I really want some fuel flow kit for the Detroit’s.

 

[MapisM]

All understood ref. the need to elaborate/smoothen the signal at least to some extent.
Obviously I have no clue about how that's handled in the OEM fuel monitors I mentioned before.
I was only telling it as I experienced it, with numbers whose decimals (in non-perfectly smooth conditions) kept changing constantly - sometimes, even more than once per second, as far as I can remember.

Ref. the sensors size, I couldn't find any "scientific" numbers for my engines, either.
All I know is that they are rated for a max fuel burn at full load of 162 l/h each, which would imply an 88 l/h backflow even at WOT, if the pump should keep flowing fuel at 250 l/h.
Not to mention that at normal cruising speed, the fuel burn is in the 70 to 80 l/h ballpark, which implies a 170 to 180 l/h return flow!
If I should consider the filter size, my Separs are rated for 1080 l/h, go figure.
But I think that the choice was just driven by the "better safe than sorry" attitude of the builder, because I can't for the life of me see a reason why the engine pumps used in engines burning 162 l/h max should be sized for 6+ times such flow!

 

[springer]

I may be talking from posterior, but does not fuel system also cool both injectors and injection pumps and therefore total fuel flow can approach double total fuel burn, and therefore capacity for this function would also be required. Thus the requirement for pre and post engine metering with associated electronic calculations.

Springer

 

[jrudge]

I may be talking from posterior, but does not fuel system also cool both injectors and injection pumps and therefore total fuel flow can approach double total fuel burn, and therefore capacity for this function would also be required. Thus the requirement for pre and post engine metering with associated electronic calculations.

Springer

you are correct and that is why it is tricky - return fuel is measured also - which is what this system does

 

[springer]

Certainly looks a good system and would have great aftermarket appeal assuming relative ease of installation, breaking into diesel lines is not always fun.

S

 

[jrudge]

I have fuel monitoring.

The values change - up a wave more, down less etc. But your seem to have very significant fluctuations. The values are not damped as far as the user can see but I imagine that there is some level of damping so maybe report once per second for example.

Have the tested the cumulative flow to confirm it is correct - I only watched one video but that was showing variations of 10lph ish and mine at steady state in flat seas is static apart from the odd decimal point moving.

 

[kashurst]

I think it needs more damping - I would average out over say 3 minutes and display that. Otherwise in a rough sea etc the values may vary a lot and may start to alarm you that something is very wrong - or you will just end up ignoring it.
another great job by the way - looking forward to more stuff on the stabs

 

[Bouba]

My electronic engine with its Smart Craft digital gauges give a very steady almost never moving readout. I have to adjust the throttle to get it to move

 

[jrudge]

I think it needs more damping - I would average out over say 3 minutes and display that. Otherwise in a rough sea etc the values may vary a lot and may start to alarm you that something is very wrong - or you will just end up ignoring it.
another great job by the way - looking forward to more stuff on the stabs

Too long - your full totals would be distorted and it would show 200 lph when you were already anchored up!

Max would be 1-2 secs I would think. You just want to stabilise the readout but still have it responsive so you can see the flow change.

 

[Portofino]

FWIW the MAN MMDS screen which page 4 shows fuel burn @ L/h only changes with throttle movements / rpm changes .
Waves make zero difference .Waves changes the boat speed by fractions say 27.3 , 27.2 , down then 26.9 ,26.7 or what ever up or in bigger waves - basically a bit more drag,
Fuel burn just sits at what ever the throttle position is .
At anchor it’s been ticking over then stopped so fuel burn follows .

I understand it’s measuring the opening time pulse sent to the injectors and knows the vol so the ECU(s) can work out how many L in an H the burn is and display this info .
Moving a throttle changes the pulse time .

I don,t think there’s any actual fuel flow through pipes to the pump less the return calcs .

Plotter tells me time to target so I can use the L/ h to calc the total burn , if I,am bothered
But really a glance at the sight glasses on the tanks in the morning doing the engine checks , ie mk 1 eyeball ,


Tend to find there’s a plateau in terms of range .
So there’s no point slowing the rpm down to say 1500 - 25 knots and watching the L/h drop from 1760 rpm
27/28 knots because it takes longer at the lower burn rate ,the net result is the same .
Faster less engine hrs but the same total burn .

So really guys nice as it is accessing L/ h , for me it makes little if any difference on how we use the boat .

I Just look at how many cm are left ,and how many used today ,on the sight glasses to work out when in days the next visit to the fuel pontoon is gonna be .

 

[jrudge]

Fuel burn just sits at what ever the throttle position is . .

Yours is a light over powered boat so you will probably see little change.

The throttle sets the revs only on an electronic boat. The throttle position is then managed by the ECU to keep the RPM static

A boat uses more energy to go up hill than down and the engine has to deliver more power by opening the throttle to keep the revs static ( which it does within 2-3 RPM). You can also see the boost change as it does this and the fuel flow with it.

It is also interesting that a tail wind vs a head wind makes a fair difference given these large lumps are not really that aerodynamic. Finally the difference a fouled prop makes is simply staggering.

On a manual boat then the throttles are of course directly controlling ... the throttle.

 

[Portofino]

Yours is a light over powered boat so you will probably see little change.

The throttle sets the revs only on an electronic boat. The throttle position is then managed by the ECU to keep the RPM static

A boat uses more energy to go up hill than down and the engine has to deliver more power by opening the throttle to keep the revs static ( which it does within 2-3 RPM). You can also see the boost change as it does this and the fuel flow with it.

It is also interesting that a tail wind vs a head wind makes a fair difference given these large lumps are not really that aerodynamic. Finally the difference a fouled prop makes is simply staggering.

On a manual boat then the throttles are of course directly controlling ... the throttle.

I,am not sure that’s correct .
I think the rpm or pulse signal and thus L/h stays the same , up waves , down waves , through folks wakes etc , The variable is the speed .
Or it just damped within reason so that in normal useage the screens are not flicking about say 90 L h , 89.7 , 89.8 etc
Just sit on 90 until I move the throttle .
Dry is 17 T fully loaded probably over 20 T ,weights in the hull :encouragement:

 

[Assassin]

FWIW the MAN MMDS screen which page 4 shows fuel burn @ L/h only changes with throttle movements / rpm changes .
Waves make zero difference .Waves changes the boat speed by fractions say 27.3 , 27.2 , down then 26.9 ,26.7 or what ever up or in bigger waves - basically a bit more drag,
Fuel burn just sits at what ever the throttle position is .
At anchor it’s been ticking over then stopped so fuel burn follows .

Totally agree, the engine ECU has to know when and how long to open the injectors for in a single opening injector, for the multiple opening injectors it has to know how long to open them for and how many times they pulse to inject the correct amount of fuel and additional information such as engine loadings.

I understand it’s measuring the opening time pulse sent to the injectors and knows the vol so the ECU(s) can work out how many L in an H the burn is and display this info .
Moving a throttle changes the pulse time .

I don,t think there’s any actual fuel flow through pipes to the pump less the return calcs .

Plotter tells me time to target so I can use the L/ h to calc the total burn , if I,am bothered
But really a glance at the sight glasses on the tanks in the morning doing the engine checks , ie mk 1 eyeball ,


Tend to find there’s a plateau in terms of range .
So there’s no point slowing the rpm down to say 1500 - 25 knots and watching the L/h drop from 1760 rpm
27/28 knots because it takes longer at the lower burn rate ,the net result is the same .
Faster less engine hrs but the same total burn .

So really guys nice as it is accessing L/ h , for me it makes little if any difference on how we use the boat .

I Just look at how many cm are left ,and how many used today ,on the sight glasses to work out when in days the next visit to the fuel pontoon is gonna be .

 

[kashurst]

Yours is a light over powered boat so you will probably see little change.

The throttle sets the revs only on an electronic boat. The throttle position is then managed by the ECU to keep the RPM static

A boat uses more energy to go up hill than down and the engine has to deliver more power by opening the throttle to keep the revs static ( which it does within 2-3 RPM). You can also see the boost change as it does this and the fuel flow with it.

It is also interesting that a tail wind vs a head wind makes a fair difference given these large lumps are not really that aerodynamic. Finally the difference a fouled prop makes is simply staggering.

On a manual boat then the throttles are of course directly controlling ... the throttle.

That is correct- the "throttles" set the desired rpm, the ECU decides the fueling. My old T50 with D9s (arguably a tad small for that boat) had l/h read out and when it was rough you could see it varying as you went up and down waves, which initially I found worrying, then later largely ignored it, hence my post about averaging for longer than 1 minute. Ultimately Vas, it's what works for you and its just a few lines of code so give it a try and see which you prefer.

 

[rgarside]

If you have logged a file of the readings from the present set-up, you can look at that off-line and try different smoothing times to see what works best.

 

[robrennie]

For me the issue is "what are you going to do with the information"..

- If it is just for engine monitoring, then a longish sampling size is fine because you are looking for long trends over time.
- if it is drive your boat more economically, then maybe a screen that shows litres per distance over water - in that case , I would think sampling over 10 seconds would be just fine - as I have a planing boat, I would be very interested in such a system.

Best of luck with your project.

 

[jrudge]

For me 3 things

1. does the consumption look right if not why not ( prop fouling makes an astonishing difference)

2. economical cruise

3. and save the best until last .... range if going a long way ( which I do)

I do agree that for all of the above a damped figure is fine - but not too damped!