Gill fuel flow sensor failure modes

[zcar]

I didn't see any in-depth discussion of the fuel flow transmitter in this subforum on the top couple pages, though I read several pages of the topical discussion due to the ruling concerning RB's Australian GP infraction. Please link to a existing tech discussion if it already exists elsewhere, thanks.

The Gill ultrasonic fuel flow sensor did experience an acknowledged calibration shift over the course of RB's practice runs according to the FIA explanation of Ricciardo's disqualification. Given this, what would be the failure modes causing this kind of shift? I work on industrial rotating machinery instrumentation and my company employs a panel of senior engineers to shoot holes in proposed changes to the machine's design, so that's the sort of conversation I'm imagining would have happened at the FIA for the introduction of this sensor.

Looking for ideas, here are a couple to start:
- Ultrasonic fluid flow directly measures velocity of the stream, with mass flow being derived based on the fluid density and the geometry of the measurement flowpath. What could "change" about the sensor with racing abuse that would alter any of the measurements required for accuracy?
- Could intense vibration knock loose or shift the ultrasonic emitter/receiver in its mountings causing a shift in its response?
- Could vibration damage the delicate RTDs employed to measure the in/out temperatures of the fuel, altering the density calculation?
- Could fuel aeration or contaminant over the course of a session affect assumptions made in the calculation of density?

Some background reading I did:
http://www.f1technical.net/news/19212
http://www.gillsensors.com/content/data ... r-2014.pdf
http://en.wikipedia.org/wiki/Ultrasonic_flow_meter

Let me know if the topic isn't very interesting...first post

edit: http://www.fia.com/sites/default/files/ ... eter_0.pdf
Lifted from the Aust GP post...FIA sensor spec

[Jef Patat]

I'm also a bit intrigued with this issue. I made a post about it in another topic, here would've maybe been a better place:
http://www.f1technical.net/forum/viewto ... 75#p495875

[Tommy Cookers]

creative and convenient use of aliasing, surely that's not within the spirit of the rules ??

and this is not a recirculation fuel system with an FIA homologated fuel pump
there must be a lot of pulsation and rather a high pulsation frequency at high power ? (as high as one might choose)
so what does a calibration lab type calibration mean anyway ?

many offer similar sensors but make much more limited claims than Gill does

[Jef Patat]

Spirit of the rules and F1 just don't go in one sentence. I'm just trying to think outside the box. I don't really know how the fuel flow happens in an F1 car. But isn't the flow only determined by the engine? It's the only component that can consume fuel. And it is consuming/injecting fuel at a much higer rate that the 100 Hz.

I just think this might be a hole they are trying to exploit.

[marcush.]

My First question is :
How is the sensor fixed
to the Car and is the calibration adjusted to the Installation stiffness ? This sensor surely is affected by right everything around - be it Vibration Temperature magnetic fields You Name it...
if I remember correctly Gill struggled Big Time to get the measuring accuracy demanded .the sensor Itself is available for some Time...

- for good measure: the datasheet above says : EMC :not rated More questions ?
- the sensor Drawing does show no means of Mounting Hardware let alone Mounting points....
- the sensor is specified for 85c Ambiente temperature.....What were Red Bulls Main Problems the other Day?

[Jef Patat]

The datasheet only raises more questions than it solves. 1kHz measurements with internal filtering and 100Hz output on CAN and then 5Hz lowpass by FIA again. How the hell can you detect peaks with such a system?

[idfx]

If the fuel tank is heated to extreme heat, the stored fuel can evaporate. The sensor can be recorded in high vapor pressure and volume given a wrong fuel? is a hypothesis

[marcush.]

If the fuel tank is heated to extreme heat, the stored fuel can evaporate. The sensor can be recorded in high vapor pressure and volume given a wrong fuel? is a hypothesis

8 ) cavitation and entrained Gas.......GILL are aware of installation issues

[WhiteBlue]

The datasheet only raises more questions than it solves. 1kHz measurements with internal filtering and 100Hz output on CAN and then 5Hz lowpass by FIA again. How the hell can you detect peaks with such a system?

AFAIK the analogue core sensor package is sampled by a very fast digital package and the values integrated. The values transmitted to the FiA server are communicated on a much lower frequency than the internal measurement. I have no idea where the bugs are. They could be anything. IMO it makes more sense to wait for the tribunal to investigate and report.

[Tommy Cookers]

Spirit of the rules and F1 just don't go in one sentence. I'm just trying to think outside the box. I don't really know how the fuel flow happens in an F1 car. But isn't the flow only determined by the engine? It's the only component that can consume fuel. And it is consuming/injecting fuel at a much higer rate that the 100 Hz.
I just think this might be a hole they are trying to exploit.

fuel events at the engine end are affected by fuel compression eg 4% density change typically at 500 bar
momentary accumulation effect to avoid loss of mixture strength or boost over the 'through the gears' rpm range ?
only 1-2 gm for 1-2 sec is needed for this, maybe the rulemakers intended this flexibility ?

there will be a lot of face-saving and apparent consensus over fuel rates
but the races will anyway be rather dominated by the fuel total permitted or loaded
if the qualifying becomes a bit controversial, that will suit the general purpose

[erikejw]

The accuracy is 0.1%?
The hiccup is that the flow is measured for a full minute.
And the flow is probably constant.

In F1 the period is 0.2 seconds.
Anyone see a problem with accuracy?

Anyone at FIA heard of Nyquist theorem?
What is the sample frequency of the sensor?
What is the algorithm to measure?
What underlying assumption do they have?

Temperature, pressure. What do they really measure?
How do they measure it?
Speed of sound? It varies with pressure, temperature, solvent.

8:)Cavitation and entrained gas can cause meter damage and spurious measurement
results, this must be avoided by appropriate system design and flow meter operation

[zcar]

My First question is :
How is the sensor fixed
to the Car and is the calibration adjusted to the Installation stiffness ? This sensor surely is affected by right everything around - be it Vibration Temperature magnetic fields You Name it...
if I remember correctly Gill struggled Big Time to get the measuring accuracy demanded .the sensor Itself is available for some Time...

- for good measure: the datasheet above says : EMC :not rated More questions ?
- the sensor Drawing does show no means of Mounting Hardware let alone Mounting points....
- the sensor is specified for 85c Ambiente temperature.....What were Red Bulls Main Problems the other Day?

FIA spec (added to first post) states that for F1, the sensor is to be submerged in the fuel tank. The spec states also that the device should meet EMC immunity. Interesting that they quote an SAE standard rather than EU, ISO, DIN... Then the fact that the Gill spec sheet lists no EMC compliance.

I guess I'd prefer to stick to thinking about what could cause legitimate error in the FIA flow sensor, rather than propose how teams may be attempting to fool it. I'm of a similar mind to many of those who commented elsewhere about the difficulties of measuring flow with the degree of accuracy and certainty demanded by the FIA. I'm also sure that a lot of smart people have accounted for most (?) of the basic pitfalls in implementing this sensor...or so I would hope!

[natehall]

Right, if im not mistaken, and considering im not an engineer just a formula 1 fan with a computer science backgroud I probably am

So the car can use a maximum of 100kg of fuel, with a maximum peak fuel rate of 100kg an hour..

Surely there must be a very simply way to restrict the flow, the engines are forced to run a maximum fuel pressure of 500bar if I remember correctly, so could they not calculate the minimum diameter of pipe that the fuel must go through to equate to 100kg/hour at 500bar and have a restrictor pipe.. therefore eliminating going over the peak flow rate as it is impossible?

Or is this to simplistic?

[bill shoe]

The dueling press releases don't say how much the discrepancy or alleged discrepancy was. This lack of information is the elephant in the living room. From this I conclude everyone is guilty. I assume the FIA sensor was inconsistent, and I assume Red Bull should have just turned their damn fuel down a bit.

In the lead-up to Australia, the FIA thumped their chest about no tolerance on fuel limits. Then after the race we hear that RB would have been allowed to get away with a rate violation if they had simply lowered the rate after the FIA detected the initial violation. So some tolerance and flexibility is allowed on the rate.

I suspect the discrepancy was only a percent or two. If that guess is right, then over the course of the season that needs to be improved, but Red Bull was weirdly stubborn to keep the rate up despite written procedure in the technical directive.

Red Bull's confidence in the Renault sensors over the FIA sensor is also weird. This reminds me of the 80's turbo-era when many drivers claimed their onboard computer said they had X liters remaining when then ran bone dry short of the finish. At least then we knew the outcome at the end of the race.

I'm not anti-RB, but on this fuel issue, either the FIA stands up hard-ass for their written procedure, or every race result will be open to protest and arguing after the fact.

[ecapox]

The more I read about this, the more I side with Redbull. Not about their exclusion, because I believe they DID break the rules, but in regards to their fighting the penalty.

It looks like this fuel flow regulator (FFR) was suspect really early in the engine development process. Luca Marmorini brought it up when he saw significant differences in readings based on fuel mixtures. Accordingly, some things were revised with regards to the software. This is what led Montezemolo to write that open letter. Looks like Ferrari, and other teams were resigned to run lower than 100kg/hr. I think Omnicourse article said 98 or 96. If Ferrari were going to detune their engine, which for all intents and purposes this is, they were going to hope the FIA came down with a very heavy hand if these rules were not followed.

I assume all this will come out during the appeal....if true. And if they are true, we might get a lot more powerful engines in quality trim, where they aren't limited to a set amount of fuel.