Aircraft Monitoring Student Project

[Del Herring]

Hi folks,
I'm currently an aerospace engineering student/professional, and for one of our courses we are doing a conceptual design of an aircraft health-monitoring system for the general aviation market. Who knows, if the project goes well, maybe we'll actually be able to make it some day, for now we're still conceptual.

For the project, I'd love to get some input from people experienced in flying or maintaining general or experimental aviation aircraft. The general idea is something non-intrusive, that would use existing, or potentially new systems to monitor and log information about the aircraft and let pilots know when a problem is starting before it becomes catastrophic. We'd hope to improve safety and reliability of aircraft at the lowest cost.

So, if you could have the system however you wanted, what would be most important to you? What are some of the most common aircraft mechanical/electrical problems that people experience? What would be some major concerns for the pilots/owners, or from the ever helpful people at the FAA? I'm looking for any input you guys and gals have on this. Let me know what you would want if the system was built just for you.

I appreciate any input you can give. It'll help us out with school, and maybe it'll be the start of something useful for the general aviation community down the road.

Thanks!

 

[AggieMike88]

Warning... Tough Crowd ahead


Soooooo an aviation version of automotive "check engine" idiot lights?

 

[Mason]

EDM-930

 

[MAKG1]

Warning... Tough Crowd ahead


Soooooo an aviation version of automotive "check engine" idiot lights?

Except it would be a "pull red handle" idiot light.

Yup, tough crowd. It's not "conceptual" unless you have a concept. And it's poor form to ask the board to do your homework.

I always thought a knock sensor might be nice even without feedback (which you can't do anyway without adjustable spark advance). As pilots we have several ways we try to prevent detonation, yet we have no way to tell if it's happening aside from watching for high oil temperature.

 

[Sundancer]

Hi folks,
I'm currently an aerospace engineering student/professional, and for one of our courses we are doing a conceptual design of an aircraft health-monitoring system for the general aviation market. Who knows, if the project goes well, maybe we'll actually be able to make it some day, for now we're still conceptual.

For the project, I'd love to get some input from people experienced in flying or maintaining general or experimental aviation aircraft. The general idea is something non-intrusive, that would use existing, or potentially new systems to monitor and log information about the aircraft and let pilots know when a problem is starting before it becomes catastrophic. We'd hope to improve safety and reliability of aircraft at the lowest cost.

So, if you could have the system however you wanted, what would be most important to you? What are some of the most common aircraft mechanical/electrical problems that people experience? What would be some major concerns for the pilots/owners, or from the ever helpful people at the FAA? I'm looking for any input you guys and gals have on this. Let me know what you would want if the system was built just for you.

I appreciate any input you can give. It'll help us out with school, and maybe it'll be the start of something useful for the general aviation community down the road.

Thanks!

Yes, tough crowd. . .most of what worries us has been addressed by engine instruments/analyzers, etc. You might be a bit late to the fair. One thing I miss from military days is a Master Caution - more or less a light that says "scan your indicators, something just broke/burned/blew. . .but I don't miss it that much.

If your focus is predictive, as in "number three cylinder average temp has been trending upwards over the previous 20 tach hours", some guys who like widgets and geeking might buy it. Some of that is already availble in some glass installations. A flight data recorder for the "retentive", so to speak, but again, I think outfits like Appareo are way ahead of you.

For non-glass, maybe find a way to detect primary flight instrument failures, though how you'd do that non-intrusive is a tough problem. Much bigger market than glass, though. . .

 

[AggieMike88]

A foundation question for the OP.... You say you're an "aerospace engineering student/professional". But the way the question is phrased, you don't have any real world flight experience. Is this correct?

 

[Del Herring]

Aviation equivalent of check engine light is a good, if somewhat simplified, version of what we had in mind. Unlike a check engine light, we'd like to be able to provide more specific detail to the pilot/mechanic. Instead of just saying, something's wrong, we might for instance like to say, "hey we've noticed that over the past few flights your engine temperatures have been slowly, but consistently increasing, time for some preventative maintenance" or whatever other information might be relevant.

That EDM 930 looks like a solid basis that would have a lot of the information we'd be looking at, but it doesn't seem like it contains much in the way of logs or predictive capability, which is really where we'd like to head with this. I think it would be an excellent example of the type of data collection we could expect to be available though, so I'll try to keep that in mind going forward. Thanks

We're certainly not asking the board to do our homework. On the contrary, our homework IS to ask the board. It's when engineers don't ask the relevant customers that they all come back complaining about those engineers who have no idea what they're doing. We're in the "figuring out what we should be doing" stage. No point designing a whole system around a need that isn't there.

Sundancer, The predictive is really where we're trying to head with this. I'm aware of the numerous instruments that exist to tell you when something has gone wrong. We're hoping to find a practical, affordable way to start figuring out well before so we can avoid the really big problems.

Aggie, you're dead on. I'm not a pilot, it's a dream for down the road, but I've got student loans to knock out first. I do work in the aviation field, but my experience lies in commercial airliners and large turbine engines 15,000 lbs of thrust and up, for which we do actually have very solid predictive systems, we're regularly able to tell customers that their combustor fuel nozzle #x has begun to clog and will need to be replaced within a given number of hours to prevent problems, and various other common problems. It prevents a lot of downtime and helps avoid serious failures for the commercial airline industry. To my knowledge (and correct me if I'm wrong), there's nothing quite like that for general aviation. I figure the concerns in general aviation will be a bit different from the commercial world. Taking a plane out of service for a week might not be a big deal for most of you, or maybe it is. So I'd like to get a feel for the priorities that are important to the general aviation crowd, so we can design around that.

I'll look into detecting instrument failures in steam gauges. It'll take some research into how they all function on my part, but if that's something you think would be useful, it's something I'll look into.

This Appareo system (I'm looking at ALERTS and Vision 1000) looks like it has the data logging side of things down pretty good. We might look to tweak that a little for our objectives, but in reality that wouldn't be much of an issue. It doesn't look like it does any predictive analysis on mechanical systems conditions/problems though. We'd essentially be looking to add that feature on. It may be that we could simply consider a piggyback system that analyzes the data from a system like that and uses the information to detect problems with instruments, engines, or control surfaces.

 

[SixPapaCharlie]

You need a pilot monitor. That's where most failures occur.

 

[GeorgeC]

I always thought a knock sensor might be nice even without feedback (which you can't do anyway without adjustable spark advance). As pilots we have several ways we try to prevent detonation, yet we have no way to tell if it's happening aside from watching for high oil temperature.

Or an O2 sensor/lambda gauge...

 

[Mtns2Skies]

Hi folks,
I'm currently an aerospace engineering student/professional,

What school is your degree through?

Also as stated above what is your experience with general aviation?

I'd love to help answer your questions, just trying to get some background.

Perhaps some vibration analysis can be used. One of my professors owns a company which does this and based on vibrations and loads experienced at any given component can reliably predict when it will fail. Typically two transducers are used pointed in the X and Y direction at the given bearing.

I bet this can be applied to multiple components in an aircraft.

 

[Everskyward]

Trend monitoring is common on turbine aircraft. In aircraft without a way to automatically monitor, we would write down various readings on each flight. This information was sent away to some company which specialized in trend monitoring. In newer airplanes, the information is on the FDR. I am not a mechanic or an engineer, but I know they analyze this information to spot trends that might cause problems. But I believe you are thinking of something in the cockpit which would alert a pilot in flight?

Whatever you develop, it needs to be reliable enough and not prone to hypochondria. When people get enough false warnings it's hard to believe the real one.

 

[MauleSkinner]

I'll look into detecting instrument failures in steam gauges. It'll take some research into how they all function on my part, but if that's something you think would be useful, it's something I'll look into.

Gyros have power failure indicators, but maybe something to predict an impending failure?

Pitot-static instruments, on the other hand, don't have any type of failure indicators...the most common failure mode is probably an iced-over pitot or static port. Even my gee-whizzy glass jet doesn't tell me when that happens...it warns me that I've got a disagreement, but doesn't tell me which one's bad.

From the instances I've heard of, I think a lot of instrument malfunctions become far more overwhelming than they really should, and proper potty training would solve most of the problems. But then Boeing thought potty training was the answer for Dutch roll in swept-wing airplanes, but the yaw damper had to be developed.

 

[Del Herring]

Thanks for the input so far guys, it really does help.

Since it was asked: I did my undergrad in Aerospace engineering at Embry-Riddle in Arizona, I'm now working on a master's degree at University of Cincinnati. This project is for a system's engineering course I'm taking there. I've been in several GA aircraft, but am not myself a pilot. I'd like to be eventually, and have started planning finances for taking lessons. I took a particular interest about a year ago when a roomate came back from a family trip to Oshkosh for the EAA event there and was sharing stories with me, but I've always had an interest in the aerospace field in general.

I'm assuming by lambda gauge you're referring to Air/fuel ratio? I actually would have thought those were standard. That would definitely be an important parameter to monitor, though it may be hard to use predictively given that pilots can manually alter it.

The vibration analysis seems like a very good route, and I agree it could be used in multiple areas. That may be a good option for items like control surfaces, and anywhere with a bearing.

I've done a little more research, and found the "Nall Report." It shows that 70% of mechanical failures are between the engine and prop, and 15% are at the brakes and gear. With the other 15% split between various other items. So it looks like focuses on engine and propellor reliability would be the best area to focus. I would also think that makes sense economically for pilots. I've seen what a Lycoming costs, I'm sure you'd all like to replace them or repair major damage as little as possible.

Vibration monitoring, air/fuel ratio monitoring, knock sensors all seem like they'd be directly applicable to engine reliability. It may also be possible to detect specific maintenance issues in brakes and gears this way too. I'll continue digging and see if I can find more specific information. But I imagine that vibrations probably develop in most landing gears long before they actually fail, and just aren't detectable by the pilot until the problem is much more... severe.

Indigo, yes. You understand what I'm aiming for. For people who do more of their own aircraft maintenance I think it would be especially useful. I could see it being helpful for the experimental crowd, but hopefully it could be useful to others as well.

From what I've researched, it looks like common failure points on the piston engines are bearings, connecting rods, and valves. Hopefully you guys haven't been involved in too many, but does that sound about right to the crowd?

I think detection of bearing problems ahead of time is probably quite achievable. And effects due to valve problems could probably be detected downstream of the engine in the exhaust, which would probably keep things a little simpler. These seem like good candidates to focus on to me.

Pitot icing also seems interesting. I'll think on ways you could pinpoint which failure is occurring there.

I'll also look at that power loss question. Electrical isn't my strong point, but I know some people I can talk to.

Thanks again, y'all are throwing out lots of good ideas.

 

[cowman]

You need a pilot monitor. That's where most failures occur.

Good lord where would you mount the probe?

 

[MAKG1]

Keep in mind, bearings and connecting rods do not fail by wearing out. The usual reason is oil starvation. The airplane already has oil pressure and temperature gauges, required by regulation.

Similarly, brake wear comes with its own warning system -- it makes noises, and you can see the pads (even in flight in some airplanes). Gear extension failures have indicators. Gear breakage can be caused by a hard landing or hitting an object.

I think your scope here is getting excessively large. You'll need to understand the whole system in detail to engineer something useful.

 

[TonyG]

I've got some familiarity with the kinds of system you're talking about -- where a large volume of system wide data (with years' of historical background) is used to identify sets of indicators that will ID a future problem. That has actually been made to work pretty well in the part 121 world.

To apply that to GA, you've got three real challenges:
1) There's no huge repository of historical data like there is in the airlines. Think of the reams of operational data that Boeing, Airbus, GE, Rolls, or the airlines have available. You just don't have that in GA.
2) Airlines fly only a few types of aircraft compared to GA -- that kind of "amplifies" the value of all that data they've already collected, as it's not spread across a huge number of types.
3) For any given type flown, the flight profile of an airline operation is pretty much the same from flight to flight, at least as compared to a GA flight. My airplane may spend 3 hours in the pattern, then go on a 4 hour cross country, then go up for spin practice. My tires/gear/brakes see operations on smooth asphalt, smooth grass, rough grass, and gravel, all in a week. That diversity of operation further clouds the usefullness of all the data collected.

But, maybe I'm full of s$%t. It's certainly worth exploring, and I'd love to be able to replace crap before it broke.

Good luck!

 

[Sundancer]

Not to be discouraging, but for fixed gear, six seat or less, single engine GA planes, you might be putting a $50 saddle on a $10 dollar horse. . .with complete electrical and instrument failure, VFR, the airplane isn't in real danger; and that's not a failure mode I've ever heard if. Even total or catostrophic engine failures aren't common, at least not without some warning. . .many folks have oil anslysis done, for example.

The avionics in GA generally "stand alone", including glass - the black boxes don't operate/control any flight control or other systems, with the exception of the auto-pilot. No FADEC, no automated warnings on flaps, slats, etc., no arming of spoilers, or adjusting V or approach speeds for weight, just not much required of systems to maintain controlled flight.

So not much need to monitor, or predict, beyond engine health. And that's a luxury more than a necessity.