FAA: Not dawdling on icing rule

Washington DC
Source:
This story is sourced from Pro
See more Pro news »

US FAA officials say criticism from the National Transportation Safety Board (NTSB) that the agency has not acted fast enough to revise aircraft certification rules for super-cooled large droplet (SLD) icing are not warranted.

FAA deputy associate administrator for aviation safety John Hickey, testifying at a US House aviation subcommittee hearing today, said a notice of proposed rulemaking (NPRM) on the topic will be issued in June.

SLD, which Hickey defined as droplets that are "larger in size than we have previously assumed was very typical in icing conditions", was a key factor in the crash of an ATR72 near Roselawn, Indiana, in 1994.

Certification rules today do not currently require manufacturers to show how aircraft perform with the SLD shapes. Due to their size and characteristics, the droplets can cause abnormal shapes as they form into ice behind the wing protection devices, a process that changes flight characteristics, including an increased stalling speed and decreased warning of an impending stall.

Hickey notes that SLD is a "very rare event", present in less than 1% of icing events. He says notions that the FAA is moving too slowly on new rules - complaints he says were aired during a roundtable meeting with House lawmakers in October 2009 - are incorrect.

"First we had to gather enough data," says Hickey. "At the roundtable, it was suggested we completed our data-taking in 2000 but failed to act."

Hickey says that though the FAA had gathered sufficient SLD data to move forward in 1999, NASA and others then looked at the data until 2001, after which an aviation rulemaking advisory committee (ARAC) was called on to develop SLD icing envelop parameters that manufacturers could design and test to.

That group "completed the majority of the envelope work" in December 2002, and then continued working into 2003. At issue was coming up with concrete methods that designers and manufacturers could use to comply with the envelopes, says Hickey.

The first ARAC report came out in 2005, and was then revised three times in the next four years "as we learned more", he adds. "After the third report, we had enough detail to move forward and did just that."

Hickey says the NPRM is "in final executive coordination" and will be issued in the June timeframe. Government rules then require the NPRM to be finalized within 16 months of issue.

Since Roselawn, he says the FAA has issued 200 icing-related airworthiness directives (AD) on 50 different models. An ongoing study by the Government Accountability Office (GAO) at the hearing revealed that during the past 12 years, there have been six icing-related accidents involving large aircraft, none fatal.

However there were more than 500 accidents involving small aircraft, causing more than 200 fatalities, says Gerald Dillingham, director of civil aviation issues at the GAO.

He notes that though there are "very few large commercial aircraft" involved in icing-related accidents, incident data "says it's still an issue". He notes that the FAA incident database contains 200 reports of icing problems with large aircraft between 1998 and 2007, and NASA's aviation safety reporting system has more than 600 icing-related incident reports for the same period.

NTSB chairman Debbie Hersman voiced frustration with the slow pace regardless of the promise of a forthcoming NPRM.

"The challenge is that there is always more to do," says Hersman, "but at some point [the FAA] has to pull the trigger. It's been 13 years since we issued a recommendation [for SLD certification rules]. It will be several more years before we see the regulation."

She says there are 15 open NTSB recommendations on in-flight icing, four of which are on its Most Wanted list.

"I empathize and agree with all the comments on length of time for the rulemaking," says Hickey. "We understand the frustration. In the case of SLD, we have pulled the trigger."

He says that though the FAA had the "raw data" on SLD early on, it did not yet have the knowledge to convert that data to actionable certification criteria.