Cylinder Shock Cooling

[kestrel]

Water cooling would be sooooo nice! Tighter temperature control, no concern about shock cooling, better and safer cabin heat!

[Bcone1381]

we can all agree that thermal cycles are hard on engines...... the larger the cycle, the greater the risk of accelerated wear. So anything we can do to minimise thermal cycles (number, magnitude, and rate of change) has to be a good thing. Even if nobody cracks a cylinder - I think improved engine management is a worthy pursuit.

I flew Continental IO-520's in a C-310R in the 1980's in the upper midwest. (all airports about 1000' msl) With 6 of them flying, we has about 280 individual engine thermal cycles a week on the fleet. To prevent cylinder cracking due to shock cooling we did the following procedure because in order to (academically) flattened the thermal cycle. We had very course poor CHT data with slow moving gauge and nothing ever recorded. Our cockpit data was very poor but we knew.....

-Takeoff using full power mixture rich.
-At first power reduction we left the throttle and mixture full forward and reduced RPM. (to keep mixture richest keeping cylinders coolest)
-At top of climb (6-8000' typically) we reduced to cruise power and leaned the engine (75F rich of peak...Goverment paid for fuel). (so now we had a smaller change in CHT)
-At top of descent we left the mixture alone and reduced power an inch HG/minute. (to flatten the thermal cycle)

The data showed a large decrease in cracked cylinders. Like 75% decrease. Real Data...old data.

[kestrel]

Waiting to read it!

[Battson]

My original reason for installing the cowl flaps had as much to do with keeping the engine warm as it did with keeping it cool.

Very relevant comments Tim - thanks for adding that.

I think there's an important point in your post - we can all agree that thermal cycles are hard on engines. They are a proven cause of accelerated wear and tear. To Kestrel's point - the larger the cycle, the greater the risk of accelerated wear. So anything we can do to minimise thermal cycles (number, magnitude, and rate of change) has to be a good thing. Even if nobody cracks a cylinder - I think improved engine management is a worthy pursuit.

[Nev]

There's definitely interest on that article Tim.

[dramsey]

I think thats a good point. I believe its why the turbo engines are more prone to the issue. They run the hottest at altitude then you push them over and cool that irregular shape quickly and introduce all the subsequent stress in the metal. The starting point would seem to matter a lot. My Buddy who set a lot of crazy records flew over the North pole in January, The hottest he could get his cylinders was 150 F and the oil was 75 F. That engine was seeing a heck of a gradient, But it did not start from a high temp and cool. The engine now has over 2000 hrs on it and is still doing well.

[svyolo]

Yeah. Plus add in the steel cylinder liner inside an aluminum cylinder can't help when there is temp gradient. I also think I remember from the distant past that # of thermal cycles affects air cooled cylinders more than water cooled ones.

[Nev]

Good logic Tim.

[TimTall]

Makes sense. The gradient between the super hot internal parts of the engine and the cooling air will be steeper the hotter the engine is.

[TimTall]

My original reason for installing the cowl flaps had as much to do with keeping the engine warm as it did with keeping it cool. Bear with me for the logic.

I live in the mountains, so I often must climb out steeply to clear terrain. If I'd built the cowl outlet large enough to handle the cooling demand of the steepest climb outs, I'd be way over cooled in cruise. The downside to that is unnecessary drag which costs speed and/or fuel. Not ideal but not damaging anything. Descent would be where the problem is. It would mean painfully slow descents to avoid shock cooling and dangerously low CHT’s at low power settings. Sometimes steep descent profiles are unavoidable in the mountains so I wanted to give myself as much as possible to work with in terms of control.

I didn’t reinvent the wheel for this. I crawled under all of the planes at my airport and copied what I thought would work. I’m basically running the same size, shape, and opening angle as a Cessna 206. I figured if it works in that thing with 300 hp it should work for me.

It works very well. I may write up an article or post explaining exactly how I did it and why if there was interest.

[kestrel]

One piece that seems to be missing from this conversation is the absolute cylinder temps. I've read numerous articles about how Lycoming's 500F limit is crazy because the aluminum loses some large percentage of its strength when that hot. The more common recommendation these days is more like 400F or maybe 420F in climbs. It has been my understanding that a hot cylinder is much more likely to crack from rapid cooling than one that is not so hot.

[Battson]

Metal is a very good conductor. If the whole engine starts at room temperature and heats up together steadily - even if it heats up quickly - then everything expands at a similar rate. There is limited cooling airflow. The difference in temperature from the outside to the inside of the cylinder is not that large. It changes shape relatively uniformly, roughly speaking, based on its thermal coefficient of expansion. This means there is little strain within the metal, so there's not enough stress to crack the metal. Typically, we heat our engine up over 5 or 10 minutes from cold, allowing the heat to soak through the whole thing before applying full power.

When shock cooling happens we have a very hot cylinder inside, where the combustion and friction is happening, with all the cooling originating around the fins. The cooling at 130kts is very efficient, but the engine is still producing a lot of power / heat. So the same part is very hot inside and relatively cold on the outside, across the space of an inch or less. Remembering we don't measure the temperature at the outside edge of the cooling fins. This creates a huge thermal gradient across the cylinder, and some parts of the cylinder try to change shape in accordance with the material's thermal coefficient. That creates strain within the metal. That strain creates intra-material stresses. The stress is proportional to the thermal gradient, and can be powerful enough to cause crack growth if the thermal gradient is large enough, and if the stress is applied often enough (regular shock cooling). This is kind of similar to low cycle fatigue, except in this case it's caused by thermal strain. This force is felt most keenly at any sharp corners in the material - especially corners with sharp edges, like a thread - or where different materials (with different thermal coefficients) are forced together tightly. That is why the spark plug holes, with a torqued metal part inside the aluminium threaded hole, are one of the main areas prone to cracking. We borescope ours every 100 hours.

Because the metal is relatively hot, certain kinds of crack formation can occur more easily. This is quite a different cracking process at a microscopic level, compared to the cracking we are more familiar with on the airframe - such as high cycle fatigue cracking or stress corrosion cracking.

[Nev]

Yeh the turbines are different for sure, ATC aside a normal descent profile is to reduce the power to idle at top of descent, effectively a glide approach until spool up when gear and flaps are out. Shock cooling is not an issue!

[Nev]

Right - that's a lot of force then. I wonder what that equates to in inches of pressure differential. I don't see that on mine with the fixed exit lip even in a fast descent. The cowl flaps must be adding even more differential. I'm wondering if the location they are installed is having more effect than I'd anticipated.

[Battson]

I definitely rely on the cowl flaps to get that initial heat out of the cylinders pre-approach. As TimTall says, the cowl flaps create a huge suction force, compare to the smooth cowl with flaps shut. With the cowl flaps shut, there is very little cooling suction, we rely on the ram air from the prop.

With cowl flaps shut, and not having a cooling lip, I have to reduce power a loooooong way out to cool the engine slowly, which makes for a slow last 10 miles - it just feels laborious chopping power that far out. I am sure the ATPLs will laugh at that!

It is possible to supplement cowl flaps by leaning the mixture further (if LOP), or for the hottest summer days when I cruise with cowl flaps open - then mixture is all I have left. I am sure you are already familiar with that trick. In NJB, I cannot open cowl flaps and lean the mix at once without great care, otherwise I get the shock cooling alarm in a big way (i.e. across 5 or 6 cylinders). More often, it's just #2 or #5(?) which shock cool, being the hotter cylinders.

It would be interesting to rig a manometer so we can compare, however I won't have time to learn how to do that, and then do it, around holiday season.

I think the crux of the issue is: We are comparing a smooth cowl, to open cowl flaps. That is like comparing night and day. I believe you are comparing a cowl with a huge lip, to a cowl with a huge lip + open cowl flaps. The effect of the flaps is lost compared to the lip. More like comparing a day and a cloudy day, it is still bright outside.