14

Although diesel runaway is not so common...

... the consequences are spectacularly catastrophic.

On old-school engines, I can understand that there isn't much you can do about it except try to cut off the air supply.

But what about modern engines?

Since modern engines usually have a computer orchestrating between various sensors and actuators, it sounds quite plausible to have the computer detect the presence of a runaway condition and turn off the fuel injectors and (hopefully) avert disaster.

Of course, cutting off fuel supply doesn't guarantee that runaway won't occur (if the engine is running on oil vapor) but it handles a possible failure mode.

Questions

  • Is it possible to define a logical condition based on certain engine operation characteristics (e.g. coolant temperature, speed, throttle) that would serve to represent an engine in runaway-mode?

    One important factor to consider is that the logical condition used should not give a false positive (e.g. it cuts off fuel supply because the engine is under severe load while a fully-loaded vehicle is going uphill)

  • How are vehicle manufacturers doing it today, if at all?

  • Assuming runaway can be reliably detected by an engine computer, what actuation mechanisms could it deploy to do its best to stop runaway?

    I'm thinking along the lines of cutting off air supply, fuel supply, oil supply and preventing the engine from combusting.

  • The only sure fire way to stop a runaway is to starve it of oxygen or wait for the oil to burn away. Locomotive are my favourite to watch on runaways. youtube.com/watch?v=u8RvoppZT0Y – RomaH Nov 10 '16 at 19:25
13

The common way that effective overspeed protection is provided on diesels in the oil and gas industry is way more simple than using electronics to effect a shut down.

The only effective and reliable method of shutting down an overspeed diesel is to block intake air. A simple valve is installed on the intake air passage which closes when air flow through the valve exceeds what would be normal for normal engine operating speeds. It is actuated by the air flow itself, it does not need any sensory inputs or external power. It can be configured to shut upon loss of oil pressure as an added feature (generally unrelated to overspeed), and to be shut manually by cable from the operating panel.

These devices work on adjustable spring tension:

the closing force on the valve is provided by the intake air flow passing through. As the air flow increases, the closing force builds up. this is resisted by the valve springs, the pre-load on which is adjustable such that at a given air flow the resulting force overcomes the spring resistance and causes the valve to close. once closed the valve will not reset to the open condition until the engine stops.

API or NFPA (can't remember which) requires that mobile emergency response diesels used in the oil & gas industry (refineries), such as refinery fire engines, are equipped with hydrocarbon vapor detection that triggers the same valve mentioned above. I believe the primary reason is NOT to prevent runaway (even though it would), it's to shut down an active ignition source in a vapor cloud, which could be more catastrophic than a runaway diesel.

  • 2
    Nice Answer!!!!! +1 – DucatiKiller Nov 10 '16 at 21:03
  • There is something like overrun detection in the ECU but afaik the deviation of driver request from engine reaction will only be detected upon actuating the brakes during this condition. – AnyOneElse Nov 14 '16 at 8:51
6

It would presumably be very easy to detect a runaway condition based on throttle position, engine speed and possibly rate of acceleration or overspeed RPM. Shutting the fuel off won't help because when this condition occurs, the engine is either running on it's own oil or from ambient gas in the atmosphere.

There is a device known as a Shut Down Valve which is fitted to certain engines but I don't think these exist in road cars as such.

  • I think you're right. It isn't just a case of whether the condition is detectable, but whether there is something that can be done about it. I'm still curious about whether manufacturers bother doing anything about mitigating runaway though. – Zaid Nov 10 '16 at 15:22
4

A runaway happens when the engine starts getting fuel from a source other than the injectors – turbo oil leaks, oil pooled in the intercooler, oil being pulled from the crankcase. The only way to stop it is to cut the engine's air supply.

The runaways that I've seen / heard of happen fast – really fast – the increase in speed is unlike anything that would/could happen in normal operation (well maybe similar to free revving the engine. So, it sure seems like the ECU could detect this fairly easily. Besides RPM, and throttle position, the ECU probably has access to road speed and the gear the transmission is in.

Stopping it could be done by actuating an anti-shudder valve or something like that, such as a dedicated shutdown flap/valve in the intake.

That said, it puzzles me that this doesn't seem to be a feature of the ECU on any of the VW diesels that I'm familiar with – even though they seem to have the necessary sensors and an anti-shudder valve that is capable of shutting down the engine (at least from idle, I've never tried to get it shut down an engine at wide open throttle). Since it seems like it could be done completely in the ECU firmware, I would think that this would have been done – the development cost seems like it would be relatively low and a one-time thing (beyond maybe tuning some parameters if the design wasn't self-tuning), there would be no per-unit cost that I can see.

So this leaves me wondering if there is some other reason for not implementing the feature. If there was additional material costs, I could see trading the production cost against the likelihood of a runaway, but since the per-vehicle cost would be zero (or very close to zero), it doesn't seem like the rationale for not doing it would be economic.

  • 1
    What if I drive downhill? – JimmyB Nov 10 '16 at 15:09
  • That's where the code gets interesting. You could look at the rate of change. – dlu Nov 10 '16 at 23:50
  • You could trigger it only when the rpm's surpass redline. Also you can check if the current rpm's are explainable by vehicle velocity or injection amount/rate – Martin Nov 11 '16 at 8:52
2

All engines will reduce fuel injection when the red line is surpassed. That's what rev limiters are for, and they've been with us for many decades.

Some engines (most modern?) have a throttle valve, although not strictly necessary for basic operation. That valve should start to close when the engine needs, well, throttling, as is the case when you release the accelerator or when over-revving.

By the way, I once could not turn off my gasoline engine because it was running on self-igniting oil sucked in through a hole burned into the piston by a broken spark plug. It did not over-rev, but it wouldn't stop for a minute or so after I stopped and turned ignition off. So the issue may even happen with non-diesels.

  • Diesel engines typically do now have a throttle flap. – Steve Matthews Nov 10 '16 at 16:23
2

you can define a runaway engine based on if its too far over redline using cam or crank sensors.

To stop it you would do one of two things or both. fuel supply would be cut and or the intake butterfly valve will close. You can also use a brake and clutch or a trans brake and a high resistance torque converter.

You could push the cams sideways keeping the engine closed.

you can do something similar to a jake brake and open the exhaust ports on the power stroke.

you could block the exhaust pipe.

you can have a tank of inert gas that releases

  • Shutting off the fuel supply will not necessarily end the Runaway condition, it will run on the fumes of evaporated oil , One needs to plug the intake physically. – Shobin P Nov 10 '16 at 12:56
  • that is true but if the oil recovery rings were i good enough shape I wonder if it would still run away. but yes this leaves then with shutting off the intake and exhaust. – Cc Dd Nov 10 '16 at 13:06
  • There are other ways for the engine to have oil to run on, such as a leaking turbo or from the atmosphere (external of the engine itself). I like your idea of using a brake, clutch, or trans brake to control the runaway, though. – Pᴀᴜʟsᴛᴇʀ2 Nov 10 '16 at 13:27
  • @Pᴀᴜʟsᴛᴇʀ2 no doubt there are many ways for it to get oil to run on. Thanks I had remembered some of the larger truck transmissions have a downhill brake setup where hydraulic pressure locks the transmission up by trying to spin it backwards IIRC. – Cc Dd Nov 10 '16 at 19:35
1

Diesel engines use engine breaking to stop, but it is not merely a matter of shutting down air and fuel to the engine. The engine has to inhale air, compress it, then release it at the end of the compression stroke. If the air is not actively inhaled and exhaled then the engine will recover the energy it spends on the compression stroke on the expansion stroke. An engine that has not been designed to brake can't act as a brake. Stopping fuel and air will not stop a runaway on an incline.

It is possible to detect vehicle operation outside a safety regime, but the actions of an automatic system might prove hazardous. For example, aggressively braking a vehicle on a slippery surface might result in total loss of control.

Computers are probably better at reducing the likelihood of a runaway than they are at braking one. They can monitor critical temperatures, pressures and levels, and raise an alarm before the brakes fail.

1

My understanding of diesel runaway was that it's caused entirely by lubricating oil finding its way into the combustion chamber air charge, usually but not always through a failed oil seal in a turbocharger.

Most modern engines (by which I mean 1990s onwards), especially fuel injected ones, operate an "overrun cut-off" mode to save fuel and enhance engine braking, where if you're off the throttle and the RPM is above a certain level (typically mid-1000s), they don't deliberately inject any fuel into the cylinder or air charge, only resuming injection (or carburetion) when the speed falls below that threshold in order to prevent a stall (some particularly twitchy models, such as hard-boosted small diesels, are actually set up to start injecting a little bit, at slightly higher revs, if the engine speed is falling rapidly, as an "anti-stall" tactic, but that still requires rpms in the generally 2000-or-less range, and for them to be falling, not rising).

Therefore no change would need to be made to how they operate in order to implement your idea; all being well, it's what they already do by default. If you're not touching the throttle, and the revs begin to rise, the system will first progressively reduce the fuel being injected in order to try and regulate the idle back down to normal speed and, failing that, will cut it off entirely once it's above the overrun threshold in order to implement engine braking. If that's not enough, because combustible fuel is being added from another source, well ... unless the engine has some other specific anti-runaway system installed (e.g. a solenoid activated flap that will catastrophically strangle the air supply), or the driver is able to take sufficiently swift and brutal action to forcibly stall it whilst the anomalous fuel flow is still quite small (as I once had to do) ... you're stuffed. The ECU can't do anything.

On top of which, as engine computers generally don't do a great deal to integrate the information from the ABS wheel sensors and/or the gearbox (and particularly not the clutch), beyond the level of controlling the speedo needle and activating the ABS/ESP where needed or turning on the reversing light, or cutting fly-by-wire throttle power if wheelspin is detected, the ECU pretty much has no way of determining whether an uncommanded rise in engine speed is due to anomalous runaway, a bit too much lingering easy-start spray introduced on a freezing morning, or the engine being mechanically overdriven by the roadwheels due to a steep downgrade or use of downshifting to act as an engine brake. It'd certainly have difficulty knowing what's what in the case I experienced where the runaway kicked in whilst I was doing 80mph along an undulating motorway, and the rise could have been due to just coasting downhill, or riding the clutch whilst downshifting...

((incidentally, maybe it's since become a thing in everyday passenger-car diesel engines, but having any kind of flap, throttle valve or restriction in the intake was very definitely NOT the case in any that I've owned; part of the efficiency of the system in yer typical small TDi is based around the absence of such things, with engine power and speed being reliant entirely on the amount of fuel injected just prior to TDC on the compression stroke. If there's a safety shutdown flap, then it should be stowed safely outside of the usual airflow until it's actually needed, only then slamming closed. I suppose it'd be likely most useful, reliable, and simple to implement if it just opened when you turned on the ignition, and closed when you turned it off...?))

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.