Shedding the hoary myth of ice bridging

NTSB advises immediate activation of deice boots on entering icing conditions.

By Todd Gunther
Air Safety Investigator NTSB

Senior airmen work to deice a US Air Force Lockheed C130 Hercules on the ramp at RMS (Ramstein AFB, Germany) in Apr 2008.

One of the toughest aviation challenges for pilots is to unlearn guidance and techniques that have been in existence for years even though better methods have been developed and proven.

For 60 years, pilots have been taught to wait for a prescribed accumulation of leading-edge ice before activating deice boots. The intent was to prevent ice bridging and provide for a cleaner shed of ice.

In theory, ice bridging could occur if the expanding boot pushed the ice into a frozen shape around the expanded boot, thus rendering the boot ineffective at removing ice. In practice, this is not the case.

Ice bridging-just a myth

A review of NTSB investigations has disclosed no instances in which ice bridging has led to an accident or incident.

In fact, recent research has shown that ice bridging is extremely rare, if it exists at all. Ice bridging may have occurred with very early boot technology on piston-engine aircraft equipped with an engine-driven pump that had wide deicer tubes and slow inflation and deflation rates.

However, modern turboprop and turbojet airplanes use engine bleed air to operate deice boots, which results in much quicker operation. Consistent with accident investigation data, FAA, NASA and deice boot manufacturers have determined that there is no evidence that modern pneumatic boots have ever experienced ice bridging.

So where is the risk? Significantly, a review of 30 years of NTSB investigations has disclosed numerous cases where thin accumulations of leading-edge ice (sometimes far less than the manufacturer-recommended boot activation threshold) have led directly to higher stall speeds and stall-related accidents.

Thin, rough ice accumulation can occur prior to deice boot activation. This example was obtained in an icing tunnel after only 33 seconds of icing exposure. See FAA Advisory Circular AC 20-73A, Appendix R.

These accidents were caused, in part, by delayed activation of the deicing boots. Further, FAA and NASA research has shown that the first 1/4 inch of ice accumulation on the leading edge of a wing can result in significant performance degradation, and that the next several inches of ice only increase that degradation incrementally.

Early activation of the deice boots limits the effects of leading-edge ice and improves the operating safety margin. While it has been shown that deice boot activation with a small amount of ice on the leading edge will not shed the ice as cleanly as with a thicker amount, the results of current research, as represented in FAA advisory circular AC 25.1419-1A, show that residual ice remaining from initial deice boot activation does not act as a foundation for ice bridging and is shed with subsequent deice boot activations.

Best practices

NTSB investigations have shown that the correct advice is as follows. Leading-edge deice boots should be activated as soon as the airplane enters icing conditions. There is no evidence that early activation leads to ice bridging.

Correcting erroneous beliefs is even more important as the icing season approaches in many parts of the country. Today more than ever, corporate, fractional, charter and private aircraft are flying with pneumatic deice boots.

And, as regional carriers try to reduce costs, they may be using more deice-boot-equipped turboprops in the future.

A recent accident

An NTSB accident report, released in late August, drives the point home. A Citation 500 was substantially damaged during landing at BVY (Beverly MA) on Mar 17, 2007.

Cross section of preactivation ice shape, with graphs demonstrating the changes in lift, drag and pitching moment for this type of preactivation ice shape. See AC 20-73A, Appendix R.

None of the 2 flightcrew, 2 medical staff, or 2 passengers was injured. IMC prevailed and an IFR flightplan had been filed. The investigation found that a thin coating of leading-edge ice led to a higher stall speed and contributed to the accident.

NTSB determined that the probable cause of the accident was inadequate guidance and procedures provided by the airplane manufacturer regarding operation of the pneumatic deice boots.

Also causal were FAA's inadequate directives, which failed to require manufacturers to direct flightcrews to operate pneumatic deice boots immediately on entering icing conditions. This is an account of what happened.

During the landing descent, the copilot of the Citation noticed that the windscreen was picking up a trace amount of rime ice. Neither crewmember saw any ice on the wings, so they did not activate the deice boots.

The approach seemed normal until around 100 ft agl, when the crew experienced what the copilot described as a "burble" and the airplane rolled steeply to the right. The pilot stated that there was "no buffet and no warning."

During the NTSB investigation into the Nov 2004 crash of a Bombardier Challenger 600 on takeoff from YUL (Trudeau, Montreal QC, Canada), the board issued an alert to pilots regarding the hazard of ground icing, specifically targeted at bizjet ops.

The crew's attempt to recover was unsuccessful, and the aircraft's right wingtip struck the runway overrun area. After landing and taxiing to the ramp, the flightcrew conducted a postflight inspection of the airplane.

They noted that the right wing was bent upward about 10° and that "light rime ice" was present on the leading edges of the wings, horizontal stabilizer and radome. A customer service agent observed an accumulation of ice on the leading edges and nose of the airplane.

He described it as a rime ice strip, about 2 inches top to bottom, covering the entire leading edge from wing tip to wing root. The white of the ice was "highly visible," he said, although he could see the wing's rubber boot through the ice in some areas.

He estimated that 90-95% of the wing deice boot was covered by the 2-inch strip, which suggests that the ice was less than 1/2 inch thick and most likely less than 1/4 inch thick. On the nose of the aircraft the agent also noticed a solid coverage of ice 10-15 inches in diameter and 1/16-1/8 inch thick.


1 | 2| Next