Discussion in 'Aircraft Design / Aerodynamics / New Technology' started by Apollo, Jul 7, 2014.
News | Joby Aviation
You are very insightful - we hadn't thought about this, until we saw the results from our initial wind tunnel test. If you look at the CTOL/STOL/VTOL slides I sent, you'll see on slide 9 that I discuss exactly this characteristics that we pleasantly discovered. Typically with advanced concepts as you move forward to more detail, the majority of 'discoveries' are the concept getting worse - but this concept has been the contrary where we keep running into very nice surprises! Humpback whales have leading edge bumps (tubercles), which helps them achieve a higher CLmax for tight turning that's required for feeding. I've attached a research paper on this effect. From my presentation you can see the effect 'we think' is coming from these distributed nacelles (essentially acting as vortilons) where our CLmax is almost completely flat out to 40 degrees angle of attack with almost no drop off (a very nice characteristic to have - especially if we wanted to use this for a STOL or tilt-wing VTOL aircraft!). We need to do more investigation to prove this is what's really happening. In that first (quick and dirty) wind tunnel test the nacelles were not nearly as distributed as we have now (that was 8 vs now we have 20) - so the turbercle effect should be even better now.
You are a bit presumptuous aren't you It would be better to phrase this as a question instead of a statement please. Of course we were well aware of this characteristic. But you aren't taking into account the new degrees of freedom present. We can vary the spanwise distribution of power quite a bit, without changing the total lift. Look at slide 9 of my CTOL/STOL/VTOL presentation and you'll see the spanwise distribution with power-on and all motors getting the same power during landing approach. It is an relatively awful lift distribution, which gives us some extra induced drag. As I indicated previously we can decide to put more power outboard and less inboard (while still generating the exact same total lift). This will increase the induced drag even further as we create a truly horrible lift distribution (but ensure we don't experience tip stall and instead can have root stall). So we can create more drag to control our glide slope. But we really need to do some time detailed step trajectory simulation on this - which we haven't had time to do (another thing on this list). But I assure you we have plenty of degrees of freedom to take care of this issue - this concept is full of new degrees of freedom!
Coaxial counter-rotating props might be a way to go if the lift engines were switched to full-time duty since you couldn't fold them then. Aside from the improvement in efficiency it's really easy to implement when the props are driven by electric motors since you just need to make the drive shaft of the front engine hollow and need not fiddle with heavy gearboxes.
If the front and rear props had different twists you could vary the type of thrust produced by running them at different speeds during different parts of the flight envelope.
The mixed phase multi-prop sound is, for lack of a better word, bizarre. Not loud, and not unpleasant, but more like the sound of a flying saucer from a B-rate science fiction movie than that of a multi-engine plane.
Were either or both of these files generated synthetically or did someone actually build a physical setup to record the sounds?
That's about how I would expect a bunch of slightly out of phase sounds would come across. Perfectly matching sounds or ones that are specific multiples of each other tend to reinforce while ones like these that are really close but don't quite match up are blunted and stretched as they clump together in one big wobbly lump.
The Aviation Week article is now available on line without a subscription for those interested in reading that article.
NASA Plans Tests Of Distributed Electric Propulsion | Commercial Aviation content from Aviation WeekNASA Plans Tests Of Distributed Electric Propulsion | Commercial Aviation content from Aviation Week
Mark, we often lament the high cost of aircraft components due to low production rates (in part). One of the things I like about this concept is the potential for lower costs due to higher production rates. DEP increases the propulsion unit volume by 20 fold! Smaller, lighter components in greater quantities are cheaper to produce. Is everything in the propulsion units custom built or can we eventually use COTS components to lower the costs?
I believe you also said the battery, prop, motor and battery control electronics were integrated into each tubular unit. That would make the propulsion units very modular-like. I imagine you're shooting for "plug-n-play" replacement of the prop units onto the wing? Also, twenty propulsion units will distribute the battery heat quite nicely! Do those 20 motors/batteries act as wing de-icers too?
On the topic of redundancy, I'm not worried about the 20 motors (lots of redundancy there). I'm worried about the sensors and power distribution electronics. How do you sense lift conditions and control all those motors? Will there be three voting computers to measure and control lift, performance, motor speed, etc? If only one or two CPUs are used, how fault tolerant can that system be?
I would have presumed that any production version would use a dual channel CAN bus system. There is plenty of fault tolerance built into this and each propulsion unit could be a node on the system.
Yes we are using a CAN bus on the initial test this year. We'll see how well it works in terms of letting us talk to all of the motors, a CAN bus is essentially a network that lets each Engine Speed Controller talk to the others (without a host server). NASA Armstrong is doing this work and this is one of the major research questions we hope to have answered with this years experiment. There is a lot of confidence built up in CAN bus since it's been used on cars for many years for on-board diagnostics. Essentially this protocol capability lets us distribute the control and have built in redundancy exactly as you said. This is an area I need to learn much more about, but our controller guys seem confident that this is a lower risk area (but one that needs lots of experimentation to build up experience and confidence in the system design).
I answered the redundancy/controller question already from the other post. Your cost question is something we care a great deal about as we're not merely trying to provide advanced technologies for new aviation products, but also provide an incentive for early adoption of this new technology - and that means the ability to achieve low cost. As long as the utilization of these vehicles is reasonable (not the current 150 hours per year of small aircraft, but more like 800 hours of a shared ownership aircraft) then the operating costs are going to be fantastic from the 10x reduction in energy cost (5x when you amortize the batteries). But as you point out, we think we can also deal with the production cost in a unique way. Low production volume of unique parts plagues current small aircraft with absolutely insane costs. How can a 30 year old design, 300 hp aircraft engine cost as much as a new BMW I3 (which has a full carbon composite, electric car including lithium batteries and a range extender engine)???? The answer is that as long as 500 aircraft per year are being made (and only a few thousand engines across the entire industry per year) then costs are outrageous. But people are so used to the fact that the Cirrus SR-22 costs $600K+ that they simply think new small aircraft will always be insanely expensive. I may be picking a fight on this board, but there is no reason that this 4 place, 200 mph, 400 mile range LEAPTech aircraft will cost more than $200K if we can manufacture it at production volumes of 3000 units per year. Certainly I need to prove such a statement and we are just beginning a detailed costing study - with industry participation because everyone knows how good NASA is at predicting costs . As you point out, these motors/propellers/controllers/nacelles/batteries are all identical and have the intent of being line replaceable, and they will be made in production volumes of 60K units per year (a healthy number to push way down into the economies of scale). Plus there is little touch labor in these components (which makes them push down the curve even faster). Getting to these higher volumes is critical to be able to use advanced tooling - if we were only going to buy 500 a year, then tooling costs are not justified and its cheaper to use lots of touch labor (at that low production volume). But not only do we think the production costs will be dramatically lower, the maintenance costs should also be substantially lower (there is so much extra equipment related to a reciprocating engine and fuel system). But most importantly this lower cost should be coming along with improved safety as well - which is essential if we're going to get new people flying. I'll say one more thing to pick a fight here - my goal with this research is to invalidate every GA aircraft that is being produced right now and all of the 208,000 GA aircraft that are out in the fleet now. If we can show this lower total operating and ownership cost, while providing high performance, while drastically lower community noise, while eliminating lead and other environmental emissions, while dramatically improving ride quality, and while improving aircraft safety - then suddenly there are 200K units that need to replaced within a 5 to 10 year period, and perhaps we can be seeing production volumes of 20K aircraft per year. Based on the past 50 years you would be correct in calling me insane - but we are talking about introducing a new technology here that is more different than when the turbine was introduced, and has the disruptive characteristics to make such dreamy statements a reality. It saddens me that as we've reached out to the US small aircraft manufacturers they've expressed little interest, while the European and smaller companies such as Joby are begging to do this research (and paying their own way just to be part of it). But this is the Innovators Dilemma, where US small aircraft manufacturers are beholden to their current customers and way of doing business. They think that they can always just catch up (as Cessna did when the Eclipse 500 made all its claims and then Cessna just came out with the Mustang to knock the wind out of their sales). But what the US aircraft companies are completely missing is that when a fundamentally new technology comes in and displaces their old products, they won't have the time or opportunity to catch up and their reputation and loyalty to the old customer base will be an anchor holding them down. I'm sure if the decision makers at these companies were to read this they would laugh at my arrogance for making such statements - but they will be displaced by others who have more vision and see there incredible new opportunities in this new technology.
Regarding the de-icing issue there is great potential to use that waste 5 to 7% of power heat to help. However, we only have those motors active in in the low and slow portion of flight. So I'm not sure how effectively this will meet de-icing needs. There are a lot of permutations to this design to look at - others have already brought up counter-rotating props to be able to keep running the props at cruise (and eliminating the swirl issue effecting the lift load distribution). Joby is ahead of us in this regard of thinking of many possibilities, and they have already been analyzing many permutations (because they are simply more agile and because they have their own skin in the game compared to us NASA guys who are encumbered in many ways be the govt way of doing things). I'll just say this - that because electric aircraft don't experience power lapse with altitude they have the potential to be screaming fast vehicles and if you start looking at taking them up to 30K or 40K feet, these aircraft are not only screaming fast, but also highly efficient (i.e. 300+ knots). I do think we're achieving the best of both worlds with this research, where we have 2 highly agile companies (ESAero and Joby Aviation) quickly getting experiments completed, while we also have the benefit of incredible disciplinary experts from NASA (because we have thousands of engineers who are some of the most experienced in each disciplinary area). Combining these two characteristics lets us be both fast and smart - something that NASA can't do by itself. So I'm extremely thankful for the excellent partnership we have with these small companies.
I just returned from three days at the Arlington Fly-in. Mark flew his Joby motor powered E-Gull daily (see photo).
The flights were short (5-8 minutes) twice around the small pattern. The motor makes a loud growl/grind sound at each startup. His assistant said the motor sound was some sort of resonance issue.
Mark and Hot Wings, thanks for the education on the CAN bus. This appears to be a data bus system and I'm wondering if the batteries are interconnected to provide power to other motors or does each battery power only the motor/prop in that particular module? The reason I ask is, let's say I'm at cruise speed and the battery for one of the tip props goes bad (for whatever reason). Is there no capability to "cross-feed" the tip motor with power from the other batteries? Or does the system start up some leading edge motors, which are not efficient at cruise speeds. So I have to slow down or have reduced range? Just curious.
LOL - well that's ambitious to say the least! For most pilot-owners, I don't think the adoption rate will be that high even if the DEP objectives were achieved. For one thing, battery technology is going to limit DEP range for a few more years. Then there's the regulatory hurdles to overcome. I know there's progress being made on both fronts, but every vision for advancing general aviation seems to get throttled by the FAA. Witness the current state (or lack) of UAV regulations. The FAA was caught off-guard by the hordes of people wanting to fly quadcopters for personal or commercial use. But I agree, once those hurdles are resolved, the business case for buying a new DEP aircraft rather than "old technology" will be quite strong, assuming cost and performance goals are achieved.
At that price point, there would be a real demand for 4 passenger aircraft with the performance you specify. Most of us will take a "wait and see" attitude. We've been disappointed too many times by past predictions of revolutionary developments. OTOH, every advance in aircraft performance has been preceded by advancements in propulsion technology. I think electric or hybrid-electric is such an advancement but others disagree. Your paper on Misconceptions About Electric Aircraft points out how ingrained our thinking is, and many doubters are going to have their preconceived notions blown away.
Did you say BLACKBERRY? Yes, some companies are going to be displaced because they didn't understand what was happening in their own industry. We've seen so many disruptive advancements in personal electronics and other technology fields. That's one reason I'd like to get past the ICE era and into the electric (or hybrid) propulsion era. Electronics & software are advancing at an incredible rate. OTOH, the great equalizer is still the FAA, which may slow down new advancements enough that existing companies CAN get up to speed. There are huge barriers to entry in the certified aircraft market. But we seem to be on the cusp of some real advancements and I love watching it unfold. Thanks for sharing your insight and related papers.
Excellent question. They will be interconnected, but we have yet to design to what degree and exactly how. There are many advantages to an interconnected system, but the question is to what degree and how complex that makes the wiring paths and other bus management issues. If the tip motors/prop fail that is the worst case (both at takeoff/landing/cruise). If it happens at cruise, yes we could start up one of the other leading edge props, but there would be a yaw differential caused by a difference in induced drag (but this would be manageable). There are many failure modes we need to investigate, and that is one of the reasons we're pushing so aggressively to this full scale moving rig test this year. We could still find showstoppers, and you are quite right in taking a wait and see perspective - because there are still some very high uncertainties and risks present. If it were going to be easy, others would be jumping on this already... But I'm pretty confident that the core integration ideas make sense, and quite a bit of analysis is backing up our statements.
I don't know how Mark's project will handle this. One option is to configure the battery for each module with separate battery packs grouped into a complete battery. Failure of a complete battery pack will then be very unlikely. If one of the battery packs fails or is showing signs of failure it can be isolated from the rest. This will naturally reduce the endurance and/or maximum performance, but it's not a total failure. The battery could be integrated with the motor and share a CAN address or each battery could have it's own address on the CAN buss. This decision depends on how the batteries are physically distributed and if there is a main power buss that can feed each of the motors or if each pod will have it's own battery module.
This is, IMHO, one of the biggest advantages of distributed electric/hybrid propulsion systems. Failure modes can be designed in that are progressive and benign. In many cased the failures can be detected before they are a problem and time to failure predicted.
Just a quick note to see if NASA is going to have a booth or information about this Distributed Electric Propulsion project at AirVenture next week. Thanks for sharing your research.
Raven ReDrives Inc.
The hardest part of the distributed electric propulsion must be able to design an aircraft that is beautiful :ermm:
As always, in the eye of the beholder.
Mark, once again thanks for talking to us on this forum. Your statement about the yaw differential made me think about the Rotordyne that was said to have a yaw differential by varying pitch on the twin engines on either side of the fuselage.
You guys probably already thought this out but cant you incorporate roll and yaw by differential thrust/lift from these numerous props and an actuator for the rudder(or maybe eliminate it??)
This get's rid of the ailerons and simplifies wing construction(=cost+time+reliability) a lot.
Joby had the pitch control in their sparrowhawk driven concept with top and bottom mounted props but I do not know how you can accomplish that in your current concept. Heck, you are already doing so much with this and with all the "good" side-effects maybe all this also becomes possible for a true autonomous flight experience ......
Keep up the good work.
Separate names with a comma.