Context:
The statically-unstable fly-by-wire flying wing discussed previously.
Problem statement:
(1) Modeling the post-stall behavior of this aircraft is very challenging.
(2) The post-stall behavior is likely, but not certain, to be an unrecoverable pitch-up, given how far back the CG is.
(Insufficient) geometric mitigation:
Although the wing is slightly swept (15° at 25% chord) and washed out such that aerodynamic stall is expected to begin towards the root, I have no confidence that this is sufficient to prevent pitch-up.
Primary mitigation:
Envelope protection will be used to maintain a safe margin between maximum commandable AoA and aerodynamic stall.
Limitations of primary mitigation:
(1) The AoA reserve between maximum commandable AoA and aerodynamic stall is set by physical limits including maximum pitch rate, maximum expected wind shear, etc; and the AoA "left on the table" in this margin represents a reduction in low-speed performance, particularly visible in landing speed and landing roll-out distance.
(2) The envelope protection system, which is independent from the primary control system (which includes the stability augmentation system), is the only part of the control system that relies on air data for safe operation, and I'm somewhat uncomfortable with the reliability of air data inputs for flight safety purposes.
Secondary mitigation:
In the event that the envelope protection system fails or is overridden by the pilot, and the pilot commands an AoA that exceeds the aerodynamic stall AoA; or in the case that the AoA reserve is insufficient and aerodynamic stall AoA is exceeded with the envelope protection system operational -- if sufficient altitude is available, a BRS-style whole-aircraft parachute can be deployed.
The next step:
Is it possible, and if possible is it a good trade-off, to try to either eliminate pitch-up at stall (if it indeed happens), or alternately to reduce the rate and severity of pitch-up at stall? If aerodynamic stall caused a nose-down pitch of any magnitude, the envelope protection system would become either of greatly reduced safety impact (e.g. hazard reduction from Major above BRS deployment altitude and Catastrophic below down to Minor), or completely unnecessary. Obviously, a traditional stable aircraft design achieves this, but I will continue to assume that all the design decisions that have led me this far down the road still hold, and investigate ideas in the context of minimal changes to the FBW design so far discussed.
So, a hat (hatless on left, simple plane showing the proposed hat sizing and placement, but not 3D geometry, on right):
The hat is a highly-swept low-aspect ratio (Delta) T-tail. This is a fixed stabilizer surface, aligned with the airflow at cruise to target zero lift at cruise AoA. This is a small surface -- about 1.5% the area of the wing, significantly smaller than a traditional tail.
Why I think this might be promising:
The idea here is to use a surface that has a much shallower lift curve slope than the main wing, but whose lift curve extends significantly beyond that of the main wing -- more so than the couple of degrees difference of a traditional tail. Due to the shallow lift curve slope and small area, this has a negligible effect on stability when the main wing is working. However, there are two effects that start as the main wing AoA increases. First, the presence of any aft surface maintaining lift during a stall event will slow down pitch-up, by preventing the potentially-sudden stall of the main wing from being the only contribution to total aircraft moment. Secondly, the lift of the hat is, during normal flight, reduced further below what its lift curve and area would imply because it is sitting in the downwash of the main wing. However, as flow detaches from the main wing during an aerodynamic stall, this will cause a corresponding increase in the AoA observed by the hat -- as soon as the main wing stops generating as much downwash (and thus lift), the hat will generate a bit more (while still, by design, being well away from it's aerodynamic stall). The hope is that this second effect is sufficient to reduce or eliminate the pitch-up.
The challenge:
The problem, of course, is that the same difficulty of modeling accurate moments in the presence of detached flows at the stall also prevents assessing if, and how much, this would help. A subscale wind tunnel model is likely to be informative, but that's a major project of its own, and there's likely significant Re-dependent effects here, making investing in that sort of test discouraging.
Request for feedback:
Thoughts? Ideas of how to see if/how much this helps? Comments that "if you're putting a tail on it why are you calling it tailless"? Other ideas to mitigate the pitch-up problem?
The statically-unstable fly-by-wire flying wing discussed previously.
Problem statement:
(1) Modeling the post-stall behavior of this aircraft is very challenging.
(2) The post-stall behavior is likely, but not certain, to be an unrecoverable pitch-up, given how far back the CG is.
(Insufficient) geometric mitigation:
Although the wing is slightly swept (15° at 25% chord) and washed out such that aerodynamic stall is expected to begin towards the root, I have no confidence that this is sufficient to prevent pitch-up.
Primary mitigation:
Envelope protection will be used to maintain a safe margin between maximum commandable AoA and aerodynamic stall.
Limitations of primary mitigation:
(1) The AoA reserve between maximum commandable AoA and aerodynamic stall is set by physical limits including maximum pitch rate, maximum expected wind shear, etc; and the AoA "left on the table" in this margin represents a reduction in low-speed performance, particularly visible in landing speed and landing roll-out distance.
(2) The envelope protection system, which is independent from the primary control system (which includes the stability augmentation system), is the only part of the control system that relies on air data for safe operation, and I'm somewhat uncomfortable with the reliability of air data inputs for flight safety purposes.
Secondary mitigation:
In the event that the envelope protection system fails or is overridden by the pilot, and the pilot commands an AoA that exceeds the aerodynamic stall AoA; or in the case that the AoA reserve is insufficient and aerodynamic stall AoA is exceeded with the envelope protection system operational -- if sufficient altitude is available, a BRS-style whole-aircraft parachute can be deployed.
The next step:
Is it possible, and if possible is it a good trade-off, to try to either eliminate pitch-up at stall (if it indeed happens), or alternately to reduce the rate and severity of pitch-up at stall? If aerodynamic stall caused a nose-down pitch of any magnitude, the envelope protection system would become either of greatly reduced safety impact (e.g. hazard reduction from Major above BRS deployment altitude and Catastrophic below down to Minor), or completely unnecessary. Obviously, a traditional stable aircraft design achieves this, but I will continue to assume that all the design decisions that have led me this far down the road still hold, and investigate ideas in the context of minimal changes to the FBW design so far discussed.
So, a hat (hatless on left, simple plane showing the proposed hat sizing and placement, but not 3D geometry, on right):
The hat is a highly-swept low-aspect ratio (Delta) T-tail. This is a fixed stabilizer surface, aligned with the airflow at cruise to target zero lift at cruise AoA. This is a small surface -- about 1.5% the area of the wing, significantly smaller than a traditional tail.
Why I think this might be promising:
The idea here is to use a surface that has a much shallower lift curve slope than the main wing, but whose lift curve extends significantly beyond that of the main wing -- more so than the couple of degrees difference of a traditional tail. Due to the shallow lift curve slope and small area, this has a negligible effect on stability when the main wing is working. However, there are two effects that start as the main wing AoA increases. First, the presence of any aft surface maintaining lift during a stall event will slow down pitch-up, by preventing the potentially-sudden stall of the main wing from being the only contribution to total aircraft moment. Secondly, the lift of the hat is, during normal flight, reduced further below what its lift curve and area would imply because it is sitting in the downwash of the main wing. However, as flow detaches from the main wing during an aerodynamic stall, this will cause a corresponding increase in the AoA observed by the hat -- as soon as the main wing stops generating as much downwash (and thus lift), the hat will generate a bit more (while still, by design, being well away from it's aerodynamic stall). The hope is that this second effect is sufficient to reduce or eliminate the pitch-up.
The challenge:
The problem, of course, is that the same difficulty of modeling accurate moments in the presence of detached flows at the stall also prevents assessing if, and how much, this would help. A subscale wind tunnel model is likely to be informative, but that's a major project of its own, and there's likely significant Re-dependent effects here, making investing in that sort of test discouraging.
Request for feedback:
Thoughts? Ideas of how to see if/how much this helps? Comments that "if you're putting a tail on it why are you calling it tailless"? Other ideas to mitigate the pitch-up problem?
Last edited:
