Saturday, November 22, 2014

Concrete Runways and Rain

Two recent Phenom overruns have an interesting variable in common: both occurred landing on a wet, concrete, ungrooved runway.  Ungrooved concrete runways represent less than one in five of all paved runways in the US, but can be particularly treacherous when wet.

The graphic below comes from a nearly 50 year old study conducted by NASA.  The aircraft in question had a higher touchdown speed than would be common for a business jet, but its absence of speedbrakes, spoilers, or reverse thrust, and the fact it has a single main wheel on each assembly compare to many light jets.  Notice the massive difference in stopping distance required on an ungrooved concrete runway, versus a grooved asphalt runway.

Taking into account that brakes were not engaged until approximately 1700' from the runway threshold, the ground roll portion of a landing on wet ungrooved concrete was 250%-290% that of a dry surface (depending on surface texture of the concrete, as represented by surface "A" and "D" above).  A wet, grooved, asphalt surface, in contrast, required only 3% more stopping surface than a dry runway.

Thursday, July 24, 2014

Use Vertical Guidance on a Non-WAAS GPS Approach

When IFR GPS broke onto the scene nearly two decades ago, it caused a sea change in the way approaches could be flown into airports too small to have an ILS. Rather than conducting an approach referencing a relatively inaccurate VOR or NDB signal, every airport in the country could be approached with localizer accuracy. What the first GPS did for lateral navigation, WAAS has done for vertical navigation. Yet some pilots still aren’t taking full advantage of WAAS capability, and some training providers are teaching procedures that hobble a major safety benefit that WAAS presents.

The issue is the ability of WAAS-capable Garmin navigators to create vertical guidance as part of most GPS approaches without published WAAS minimums. This advisory vertical guidance turns an LNAV-only approach into an "LNAV+V" approach. Where "+V" vertical guidance is available, the autopilot is fully capable of coupling to the vertical path, as if the approach was a full-blown LPV or ILS.

Yet many pilots conduct this type of approach in the same manner they would a VOR or NDB, setting the MDA in the altitude selector and using vertical speed mode to reach MDA. It is true that the presence of vertical guidance on an LNAV approach doesn’t change the MDA, nor make it an official APV (approach with vertical guidance) in the eyes of the FAA. However, there’s no reason not to use the same procedure to fly the approach as if it were an LPV, coupling vertically via the GP (glide path) mode, and never leveling off at MDA.

Flying an LNAV+V in the manner of a precision, rather than non-precision, approach offers several safety benefits. The most important is that the aircraft can be configured to land and be on-speed outside the FAF, with no further configuration or power changes required. Wherever the pilot may be when visual contact with the runway is acquired, the aircraft will be on a stable path to the touchdown zone.

In contrast, a pilot who dives down to MDA at a high rate of descent may visually acquire the runway before the aircraft is in a position to begin further descent. Deciding when it’s safe to leave MDA can require quick thinking at a time when the aircraft is very close to terrain –– not a desirable situation.

Another benefit is that, as level-off at MDA is not required, the altitude pre-selector can be set to the level-off altitude of the missed approach once the FAF is passed. Forgetting to set missed approach altitude after level-off at MDA is one of the most common errors made by transitioning jet pilots, and beginning the missed without the correct altitude set can lead to undesirable autopilot behavior or an ATC violation.

You can employ a trick to turn an MDA into a DA. Built into a published DA is consideration that if a pilot looks up at DA, the aircraft will still be traveling downwards while the pilot makes the decision to go missed, adds power, and transitions to a climb. An MDA, however, is exactly what it says –– a minimum altitude below which the pilot must not go unless landing. The solution is simple: add "padding" to the MDA so that even if it takes a few seconds to get the plane climbing, MDA will not be violated.

A light jet flying on a three-degree descent path will be descending about 550 FPM, or a little over 9 feet per second. Assuming a worst-case scenario of five seconds to transition from descent to climb and rounding the result up, you get 50 feet. So, adding 50 feet to the published MDA results in the new functional DA –– the altitude at which the missed approach will commence, and which should be bugged as "minimums."

Wednesday, April 16, 2014

Understanding V2



Many of the new V-speeds first time jet pilots are exposed to are only used in the event of an engine failure.  V2, VFS, VENR, VAC- depending on the aircraft manufacturer and phase of flight (takeoff, go-around, etc.), a pilot will calculate several speeds for each takeoff and landing that he will almost certainly not need to use.  While most pilots quickly learn to compute the speeds in initial training, and use them nearly every simulator session during one simulated engine failure after another, many pilots harbor misunderstandings about what the speeds represent.

V2, for example, is often mistakenly thought to be the jet equivalent of VXSE, or single engine best angle of climb speed.  Pilots of piston and turbo-prop multi-engine aircraft learn not only the speeds for best rate and angle of climb when both engines are working, VY and VX respectively, but those for best rate and angle with one engine failed.  For most non-jet twins, these speeds are given for one weight and altitude combination, and the pilot never ventures beyond these default values.

In contrast, before every takeoff in a jet, the pilot will typically calculate four V-speeds based on criteria such as weight, altitude, temperature, and flap settings.  The first two of these speeds, V1 and VR, determine how late into the takeoff a pilot may abort the takeoff, and when the pilot begins the transition to flight.  The other two speeds, V2 and a "final" climb speed (which goes by different names depending on the aircraft manufacturer- we'll call it VFINAL), are only used if an engine fails during the takeoff, after the point at which an abort can be safely completed.

During initial sim training, for nearly every simulated engine failure on takeoff, the pilot is presented with a textbook, that is to say, worst-case, event.  The engine is programmed to fail just before V1, the pilot reacts, rotates the aircraft at VR, and climbs to a safe altitude above obstructions at V2.  Once above immediately threatening obstacles, the pilot accelerates to the final climb speed and retracts flaps.  From this sequence many pilots develop a logical, but fallacious, correlation of V2 to VXSE, and VFINAL to VYSE.

In fact, it is VFINAL that correlates much more closely to VXSE.  So what does V2 represent then? In short, a compromise.  For most circumstances a jet will encounter, certification requirements define takeoff distance as the longer of the distance to either bring an airplane up to V1 and initiate an aborted takeoff, or continue on one engine so that the aircraft reaches V2 speed at 35' AGL.  Clearly the higher the V2 speed defined by the manufacturer, the longer the distance that will necessary to accelerate from V1 to V2 on only one engine.  Due to the desire to minimize published takeoff distances, the manufacturer often sets V2 to be the minimum allowed by certification requirements, based on minimum allowable ratios of V2 to stall speed and VMC. 

So what's the implication of V2 being a bit lower than VXSE? A brief review of basic aerodynamics will illustrate the drawback to this minimum V2 approach.  While rate of climb is determined by the excess power an aircraft has available, angle of climb is determined by the amount of excess thrust available.  For given atmospheric conditions, the thrust output of a jet engine is nearly constant if plotted against airspeed, while the thrust required by the aircraft follows the familiar "J" curve.  Given these two curves, it is apparent that the greatest distance between the two lines, or the point of maximum excess thrust, occurs at minimum thrust required speed, which is to say, the minimum drag point of the curve.

At any speed slower than this point, the increase in induced drag means that climb angle will suffer.  So by selecting a V2 which lies below minimum drag speed, the aircraft manufacture is giving up some possible climb gradient in favor of a reduced takeoff distance.  As most jet aircraft have an abundance of extra thrust, this is an acceptable tradeoff, and even during a non-optimal climb at V2, most jets will turn out adequate, or even impressive, single-engine climb angles.

While not common in light jets, some larger jet aircraft have software or performance charts that allow for an optimized, rather than minimum, V2.  If takeoff distance is calculated as 4000', but 9000' of runway is available, it is easy to see that by increasing the V speeds until optimal V2 is reached, extra runway can be converted into better engine-out climb performance.

What is common in light jets is the ability to depart with a reduced, or even zero, flap setting when runway available is not a limit, but engine-out climb performance may be.  By departing with reduced flaps, the aircraft must accelerate to a higher speed before rotation, which brings the aircraft closer to true VXSE.  Also, the reduced drag of a lower flap extension means that the thrust-required curve shifts down- for any given speed less thrust is required simply to maintain level flight, so more thrust is excess, and can contribute to a relatively greater climb angle.

Beyond the performance implications of knowing what V2 really represents, there is also a practical flight application.  While nearly all engine failures in the simulator occur just prior to V1, a real engine failure can occur at any point in the takeoff roll.  When I conduct in-aircraft training, for example, I typically retard one thrust lever to idle just as VR is reached.  With the two to three second spool down time of a jet engine, by the time the pilot recognizes the engine has failed, the airspeed is often five to ten knots above V2.   Most pilots just out of sim training exhibit a strong desire to pitch the aircraft up until V2 is reached, then hold V2 until clean up altitude.

Keeping in mind the drag curve, it becomes apparent that a pilot who has attained a speed higher than V2 before recognizing an engine failure would be better served by maintaining that higher speed until at clean up altitude, provided the aircraft hasn't accelerated so much as to be faster than VFINAL.  Doing so will result in lower drag, and thus a better climb angle.