Closing Out DT With WDA Completions

Production milestones, roll-outs, test flights, service introduction and other milestones.
User avatar
Forum Veteran
Forum Veteran
 
Posts: 868
Joined: 02 Mar 2013, 04:22
Location: Texas

by smsgtmac » 18 Aug 2017, 02:07

From Edwards today: http://www.edwards.af.mil/News/Article/ ... oward-ioc/

Key bits:
“This was kind of a cleanup, or a closeout, of (System Development and Demonstration). It’s the closeout of JSF developmental test for Block 3F, which is a big deal because it’s for Air Force IOC, and Navy IOC,” he said….

…“I like to think of these as the last for Developmental Test,” he said. “This is like our graduation exercise before we hand the aircraft off to the operational test organizations so they can go prove it’s ready for combat. That’s very significant for us.”…

…There is a single WDA event remaining for 3F, which marks the completion for all three variants and will pave the way to the declaration of IOC for all F-35s.


And for those with poor memories, some history:
http://elementsofpower.blogspot.com/201 ... t-dot.html
--The ultimate weapon is the mind of man.


User avatar
Elite 3K
Elite 3K
 
Posts: 3654
Joined: 12 Jun 2016, 17:36

by steve2267 » 18 Aug 2017, 02:33

Sarge,

From your blog:
...excluding the multiple gun scoring events, which must also be completed.


Can it be inferred then, that those vibration and/or sighting issues that allegedly interfered with correct aiming or sighting of the cannon, have all been solved / put to bed?

I vaguely recall an issue (DOT&E screamed about it) where firing the gun caused a yaw effect or something involved with firing the gun caused a yaw that pulled the cannon fire off target. I think there was possibly also a vibrational issue with the gun. I never did see a press release or story (certainly not from the critics) stating what the resolution for those issues were.
Take an F-16, stir in A-7, dollop of F-117, gob of F-22, dash of F/A-18, sprinkle with AV-8B, stir well + bake. Whaddya get? F-35.


User avatar
Forum Veteran
Forum Veteran
 
Posts: 868
Joined: 02 Mar 2013, 04:22
Location: Texas

by smsgtmac » 18 Aug 2017, 03:30

steve2267 wrote:Sarge,

From your blog:
...excluding the multiple gun scoring events, which must also be completed.


Can it be inferred then, that those vibration and/or sighting issues that allegedly interfered with correct aiming or sighting of the cannon, have all been solved / put to bed?

I vaguely recall an issue (DOT&E screamed about it) where firing the gun caused a yaw effect or something involved with firing the gun caused a yaw that pulled the cannon fire off target. I think there was possibly also a vibrational issue with the gun. I never did see a press release or story (certainly not from the critics) stating what the resolution for those issues were.


That's a very good question I hope some enterprising reporter asks someone about, but it would seem so. The latest bit about:
…There is a single WDA event remaining for 3F, which marks the completion for all three variants and will pave the way to the declaration of IOC for all F-35s.
does seem to imply all issues have been resolved doesn't it?

The only 'gun' thing of concern to me was the yawing effect compensation, and that is just a matter of dialing in the control laws. I don't see that a a hard fix compared to what they could have had to deal with. Personally, based upon what I've seen in the public domain (nothing lately) I think in the end they still may have to adjust the gun A2G lethality objectives unless the ammo absolutely performs as well as advertised and desired.
--The ultimate weapon is the mind of man.


User avatar
Elite 3K
Elite 3K
 
Posts: 3654
Joined: 12 Jun 2016, 17:36

by steve2267 » 18 Aug 2017, 05:20

smsgtmac wrote:The only 'gun' thing of concern to me was the yawing effect compensation, and that is just a matter of dialing in the control laws. I don't see that a a hard fix compared to what they could have had to deal with. Personally, based upon what I've seen in the public domain (nothing lately) I think in the end they still may have to adjust the gun A2G lethality objectives unless the ammo absolutely performs as well as advertised and desired.


The haze gently lifts, and my memory slowly returns... wasn't the gunport "door" allegedly what was creating the adverse yaw? I do recall somewhere one story mentioning dialing in the control laws to compensate slightly with rudder. Of course, on a Cessna 150, it would be mostly pure rudder (maybe some slight opposite aileron to counteract yaw-induced roll). But what with all the flight controls on the -35, and the fact that the tails are not vertical, it might mean several control surfaces are simultaneously activated. And the degree of actuation is sure to change at all airspeeds, and with all different combinations of pitch / yaw not to mention rates.

I never worked control systems. Does the state-of-the-art today enable an engineer to compose a table of desired "yaw effects" vs airspeed, and have some piece of software (possibly Matlab based), compute the optimum control deltas for all flight conditions? If so, would this be something that they would only need to test at a handful of flight data points to verify, or would they need to test the "fix" across the entire flight regime? Or has the flight test program in general so thoroughly verified and/or tweaked simulation/analysis/design performance data, that it can be done entirely with existing flight test / wind tunnel test / CFD analysis these days?
Take an F-16, stir in A-7, dollop of F-117, gob of F-22, dash of F/A-18, sprinkle with AV-8B, stir well + bake. Whaddya get? F-35.


User avatar
Forum Veteran
Forum Veteran
 
Posts: 681
Joined: 30 Apr 2015, 03:44

by rheonomic » 18 Aug 2017, 05:31

steve2267 wrote:The haze gently lifts, and my memory slowly returns... wasn't the gunport "door" allegedly what was creating the adverse yaw? I do recall somewhere one story mentioning dialing in the control laws to compensate slightly with rudder. Of course, on a Cessna 150, it would be mostly pure rudder (maybe some slight opposite aileron to counteract yaw-induced roll). But what with all the flight controls on the -35, and the fact that the tails are not vertical, it might mean several control surfaces are simultaneously activated. And the degree of actuation is sure to change at all airspeeds, and with all different combinations of pitch / yaw not to mention rates.


The F-35 uses a control allocator to convert yaw, pitch, and roll pseudo-controls into effector commands. The F-35 CLAWs are based on dynamic inversion, in which the basic approach is to solve the equations of motion for the control and replace the state derivative with desired dynamics (the equation in my signature is a DI CLAW for a linear system). For a linear system, xdot = Ax + Bu with control variables y = Cx, the output dynamics are ydot = CAx + CBu, which leads to the DI CLAW u = pinv(CB)*(ydot_des - CAx), where pinv is the pseudo inverse. The matrix CA is the onboard model of the dynamics (f(x) for the nonlinear version) and the matrix CB is the control effectiveness. The desired dynamics are transfer functions that map the pilot inputs to commanded rates such that MIL-STD-1797A Level 1 flying qualities are met. The description of the DI CLAW in the "Appendix B" link in this earlier post of mine is pretty close to what F-35 has; several of the authors of that book worked X-35. The paper "Control Allocation for the X-35B" also explains the basics pretty well.

steve2267 wrote:I never worked control systems. Does the state-of-the-art today enable an engineer to compose a table of desired "yaw effects" vs airspeed, and have some piece of software (possibly Matlab based), compute the optimum control deltas for all flight conditions? If so, would this be something that they would only need to test at a handful of flight data points to verify, or would they need to test the "fix" across the entire flight regime? Or has the flight test program in general so thoroughly verified and/or tweaked simulation/analysis/design performance data, that it can be done entirely with existing flight test / wind tunnel test / CFD analysis these days?


MATLAB/Simulink for model-based control design is the current state of the art (there was a good PDF of a talk on Simulink + F-35, but MathWorks removed it from their site). You can automatically tune gains, but for DI this isn't generally done. You do a LOT of CFD and wind tunnel tests to get the onboard model and control effectiveness terms, and verify and adjust in flight test. It's an iterative process and very expensive. There's some tuning work on the desired dynamics and a robustness outer loop around the DI as well; these sometimes need to be gain scheduled.

For the gun thing, they'll probably do a majority in simulation and then fly a few test points to verify, especially at "weird" parts of the envelope.
"You could do that, but it would be wrong."


User avatar
Elite 3K
Elite 3K
 
Posts: 3654
Joined: 12 Jun 2016, 17:36

by steve2267 » 18 Aug 2017, 06:40

Thanks for the mini-lesson / insights. Makes my head hurt. I thought CFD could be hard.

rheonomic wrote:
steve2267 wrote:For the gun thing, they'll probably do a majority in simulation and then fly a few test points to verify, especially at "weird" parts of the envelope.


Sounds like most of the hard work (i.e. model development etc) has been done and the "yaw solution" to the gun problem is (was) more of a PITA / minor annoyance (in the grand scheme of things).
Take an F-16, stir in A-7, dollop of F-117, gob of F-22, dash of F/A-18, sprinkle with AV-8B, stir well + bake. Whaddya get? F-35.


User avatar
Forum Veteran
Forum Veteran
 
Posts: 681
Joined: 30 Apr 2015, 03:44

by rheonomic » 18 Aug 2017, 06:43

steve2267 wrote:Thanks for the mini-lesson / insights. Makes my head hurt. I thought CFD could be hard.


I'll note that the "trick" of DI is that when you apply the CLAW to the plant you get y_dot = y_dot_desired, so you replace the dynamics. (Assuming perfect model cancellation, actuators with infinite bandwidth, no sensor noise or exogenous disturbances...the hard part is the robustness, which comes with a very good OBM and outer loops.)

steve2267 wrote:Sounds like most of the hard work (i.e. model development etc) has been done and the "yaw solution" to the gun problem is (was) more of a PITA / minor annoyance (in the grand scheme of things).


Pretty much, although model development is never truly done. F-16 for example still has model updates (for high fidelity sims; the F-16 CLAWs are classical given the flight computers available at the time).

The F-35 flight control is really, really good.
"You could do that, but it would be wrong."


User avatar
Elite 3K
Elite 3K
 
Posts: 3654
Joined: 12 Jun 2016, 17:36

by steve2267 » 18 Aug 2017, 15:17

This is probably discussed somewhere, but I'm going to ask anyway...

As a pilot my experience has been limited to piston poppers mostly with one fan. When I command a roll, or bank or yaw, I move the controls, observe the result, and then adjust accordingly.

For roll, I am typically looking for a certain value of roll (phi?), not so much the rate of roll. For pitch, if I'm climbing, I look for a certain pitch attitude or angle. Although if I'm turning, I may want more pitch rate to increase turn rate. Yaw seems to be more of a "feel" thing.

In the classical stick-and-rudder sense, what "variables" is the pilot commanding? Roll angle? Pitch rate? Yaw rate? Or just pitch angle, yaw angle?

The reason for my statement, is (I suspect - or read somewhere - that) the F-35 pilot is commanding all rates, or the F-35 FLCS is solving for rates.

Which prompts my question: Does the F-35 deliver "rates" rather than flight axis "angles"? Or how does the F-35 FLCS differ from a classic flight control system? I'm *guessing* that the "feel" is initially different, but quickly adjusted to, and it results in more precise flight control.
Take an F-16, stir in A-7, dollop of F-117, gob of F-22, dash of F/A-18, sprinkle with AV-8B, stir well + bake. Whaddya get? F-35.


Forum Veteran
Forum Veteran
 
Posts: 919
Joined: 26 Oct 2010, 08:28
Location: Canada

by alloycowboy » 18 Aug 2017, 18:47

smsgtmac wrote:From Edwards today: http://www.edwards.af.mil/News/Article/ ... oward-ioc/

Key bits:
“This was kind of a cleanup, or a closeout, of (System Development and Demonstration). It’s the closeout of JSF developmental test for Block 3F, which is a big deal because it’s for Air Force IOC, and Navy IOC,” he said….

…“I like to think of these as the last for Developmental Test,” he said. “This is like our graduation exercise before we hand the aircraft off to the operational test organizations so they can go prove it’s ready for combat. That’s very significant for us.”…

…There is a single WDA event remaining for 3F, which marks the completion for all three variants and will pave the way to the declaration of IOC for all F-35s.


And for those with poor memories, some history:
http://elementsofpower.blogspot.com/201 ... t-dot.html



smsgtmac, someone should ask Mr. Gilmore if he would like some bacon with eggs because he seems to be wearing quite a bit of it on his face. I think Navy Rear. Adm. Mat Winter should send him some in the mail.

https://www.zingermans.com/Category.aspx?Category=bacon


User avatar
Forum Veteran
Forum Veteran
 
Posts: 681
Joined: 30 Apr 2015, 03:44

by rheonomic » 19 Aug 2017, 02:43

steve2267 wrote:As a pilot my experience has been limited to piston poppers mostly with one fan. When I command a roll, or bank or yaw, I move the controls, observe the result, and then adjust accordingly.


In this case, you're the CLAW...

steve2267 wrote:For roll, I am typically looking for a certain value of roll (phi?), not so much the rate of roll. For pitch, if I'm climbing, I look for a certain pitch attitude or angle. Although if I'm turning, I may want more pitch rate to increase turn rate. Yaw seems to be more of a "feel" thing.


You're doing bank and pitch captures there. (And yes, bank/roll is phi.) Yaw is going to be mostly to regulate sideslip or ny for coordinated flight, or to crab/etc. (I only pilot simulators when testing things; I have maybe 30 seconds of real time on an AT-6?)

steve2267 wrote:In the classical stick-and-rudder sense, what "variables" is the pilot commanding? Roll angle? Pitch rate? Yaw rate? Or just pitch angle, yaw angle?


In that case, the pilot is directly commanding the position of a control effector; we generally call this "stick-to-surface" or similar terms. Most small GA aircraft have what we call a reversible control system, in which the control inceptor (stick/yoke) is directly connected to the surfaces. So, in a reversible control system, if you walked to the tail and moved the elevator you'd see the stick move. The other type is an irreversible control system, in which case this isn't true. "Fly-by- wire" aircraft (really, aircraft with an active feedback control system) are irreversible, but even aircraft without active controls can be irreversible (think hydraulically boosted surfaces).

If you look at a linear model for a conventional aircraft, you'll see that the aerodynamic control effectors (ailerons, elevators, rudders) have the effect of driving the angular rates (note that this doesn't mean they effect *only* the angular rates). The aircraft attitude (yaw, pitch, roll) are indirectly effected as they're essentially the integrals of the rates.

steve2267 wrote:The reason for my statement, is (I suspect - or read somewhere - that) the F-35 pilot is commanding all rates, or the F-35 FLCS is solving for rates.

Which prompts my question: Does the F-35 deliver "rates" rather than flight axis "angles"? Or how does the F-35 FLCS differ from a classic flight control system? I'm *guessing* that the "feel" is initially different, but quickly adjusted to, and it results in more precise flight control.


Well, it's a little more complicated than that, as the control variables are mixed a bit. Modern fighters generally use what's called a Command Augmentation System (CAS), where the pilot commands a value for a state variable or function of the state variables, and the FLCS/CLAWs track the commanded values. Also, for the STOVL version there is mixing in the transition and hover flight conditions. For example, in hover, the left hand inceptor ("throttle") accelerates the airplane forward/backward around the detent, the rudder pedals rotate the aircraft about the vertical, and the right hand inceptor ("stick") controls altitude via longitudinal stick and lateral velocity via lateral stick.

During conventional flight, the traditional system generally commands stability axis roll rate (ps) (i.e. the airplane rolls about its velocity vector, not its longitudinal axis), body axis pitch rate (q), and sideslip angle. The pilot generally commands ps and q, and sideslip is regulated to zero for coordinated flight, which means the pilot doesn't need to use rudder and can fly "feet on the floor". The control variables also sometimes mix in inertial effects due to gravity to better decouple each axis. In the longitudinal axis sometimes is mixed between q and a mix of AOA/nz commands.

The F-35 FLCS is similar to other modern high-performance fighter aircraft. The main difference is the use of more modern methods for the CLAWs, although there are arguments both for/against this.

One nice thing about a rate command is the following: suppose you're trying to do a pitch capture (say, 15 deg for a climb). You can command pitch rate to get to 15 deg, release, and then the CLAW will hold the pitch angle (assuming wings level) since if pitch rate is zero, the pitch stays constant.

I'd ask one of our resident meat servos (err, pilots :D) to comment on the feel difference between an aircraft with a CAS and an unaugmented a/c.
"You could do that, but it would be wrong."



Who is online
Users browsing this forum: No registered users and 2 guests