NTSB Factual Report

HISTORY OF FLIGHT

On January 10, 2001, at 1025 central standard time, a Saab 340B, N394AE, operated by American Eagle Airlines, Inc., as Eagle Flight 301, experienced dual engine power fluctuations (PF) while in cruise flight at flight level (FL) 200 (20,000 feet). The airplane was registered to First Security Bank of Salt Lake City, Utah, and did not sustain damage. The captain, who held an airline transport pilot certificate, the first officer (FO) who held a commercial pilot certificate, the flight attendant, and the 24 passengers were not injured. The flight was in instrument meteorological conditions (IMC) at the time of the incident, and an instrument flight rules (IFR) flight plan was filed for the 14 Code of Federal Regulations Part 121 scheduled passenger flight. The flight originated from the Dallas-Fort Worth International Airport (DFW), Texas, at 0912 and was destined for Laredo, Texas (LRD).

According to NTSB-conducted interviews with the flight crew, documents provided by American Eagle that were completed by the captain, and the digital flight data recorder information, the flight departed from DFW without incident. The captain was the flying pilot for the DFW-LRD leg. The engine anti-ice switches for both engines were turned to the ON position at 3,550 feet during the initial climb-out and, subsequently, the flight entered IMC conditions while climbing through 6,000 feet. The flight continued climbing to its cruising altitude of FL 200 and the engine anti-ice switches remained in the ON position for the duration of the flight.

The flight had been airborne approximately 30 minutes, and was 30 miles north of the San Antonio Very High Frequency Omnidirectional Range (VOR) navigational facility (024 degree radial), when both pilots heard "one loud pop," a change in the sound of an engine, and observed that the white ignition light for the left engine had illuminated. Following the loud pop, the captain observed the left (#1) engine's torque gauge needle drop from 65% to 60%, and then return to 65%. One passenger, who was seated in 12A, observed smoke discharge from the left engine. At the time of the event, the outside air temperature (OAT) was -18 degrees Celsius (0 degrees Fahrenheit) and the airplane was configured as per the following:

Airspeed: 190-205 knots

Torque: 65%

Propellers: 1,230 rpm

Continuous Ignition Switches (#1 & #2 Engine): NORMAL

Engine Anti-Ice Switches (#1 & #2 Engine): ON

Propeller Anti-Ice Switches (#1 & #2 Engine): OFF

Following this first event, the FO referenced the American Eagle Aircraft Operating Manual (volume one) and began to read the SYSTEM section, page 30, titled "Powerplant Phenomena."

Approximately 2 minutes had elapsed since the first event, when both pilots heard "continuous popping noises from both engines" and observed that the white ignition lights for the left and right engines were illuminated. The captain initiated a descent to 8,000 feet. At 1026:14, the flight contacted the Houston Air Route Traffic Control Center (HOU ARTCC) and requested a descent to 8,000 feet. HOU ARTCC approved the request, and subsequently, at 1029:44, the flight crew reported to HOU ARTCC that ice had ingested into the engines, and that the engines were experiencing "rollbacks." HOU ARTCC asked the flight crew to report on the type of icing that they were experiencing, and the flight crew replied "light rime" ice conditions. At 1031:17, the flight crew added that "they haven't had any problem in the past but we uh we were having both engines at the same time trying to flame out and relighting." The FO reported that engine operation returned to normal "just after [they] had initiated the descent." The airplane descended (descent rate of 1,000 feet per minute) to 8,000 feet in 11.5 minutes. The flight remained at 8,000 feet for 16 minutes before beginning its final descent into Laredo, and subsequently, landed without further incident at 1114.

The flight attendant reported that she contacted the flight crew after she heard the first loud bang, and the FO stated that "it was a rollback." The flight attendant asked the FO why a rollback happens, and the FO replied, "water gets in the engine and causes that." The flight attendant reported that following the second set of "rapid fire popping noises," the captain made an announcement to the passengers during which he stated that, "we had engine rollbacks and they automatically restarted just fine. We have descended to a lower altitude, so we shouldn't have that problem again."

In the American Eagle Flight Department Debrief and an American Eagle Saab 340 Power Interruption Report, the captain reported that at the time of the events, "minor rime ice accumulation [was observed] on the wing boots" and "on the windshield wipers."

PERSONNEL INFORMATION

The captain held an airline transport pilot certificate and was type rated in the Saab 340. He had accumulated a total of 13,300 flight hours, of which 9,200 were in the Saab 340. On November 13, 2000, he was issued an Federal Aviation Administration (FAA) first class medical certificate with no waivers or limitations.

The first officer held a commercial pilot certificate and was instrument rated. He had accumulated a total of 2,630 flight hours, of which 1,099 were in the Saab 340. On October 24, 2000, he was issued an FAA first class medical certificate with no waivers or limitations.

AIRCRAFT INFORMATION

The Saab 340B was equipped with two General Electric (GE) CT7-9B turboprop engines. The airplane was operated under a transport category airworthiness certificate and maintained in accordance with American Eagles' Approved Inspection Program.

On January 10, 2001, following the power fluctuations, American Eagle maintenance personnel inspected the left (#1) and right (#2) engine inlets in accordance with the GE maintenance manual 72-00, and no defects were noted.

FLIGHT RECORDERS

The airplane was equipped with a digital flight data recorder (DFDR). Following the in-flight events, the data was downloaded by American Eagle and sent to GE for analysis.

The data confirmed that the initial event occurred on the left engine. At the time of the event, the outside air temperature (OAT) was -18°C and the following values/trends were recorded over a 4 second time period:

Inlet Turbine Temperature (ITT): 15 Degree Rise

Torque (Tq): 5% Decrease

Gas Generator Speed (Ng): Less than 5% Decrease

The auto-ignition system was automatically activated during the event.

The data revealed that 3 minutes after the initial event on the left engine, the right engine experienced its first event. At the time of the event, the OAT was -17°C, and the following values/trends were recorded over a 3 second time period:

ITT: 25 Degree Decrease

Tq: 16% Decrease

Ng: 6% Decrease

The auto-ignition system was automatically activated during the event.

The data revealed that just before power was restored after the right engine's first event, a second, and final, event occurred on the right engine. The following values/trends were recorded over a 2 second time period:

ITT: 5-10 Degree Rise

Tq: 8% Decrease

Ng: Decrease

The auto-ignition system was automatically activated during the event.

The data revealed that as the right engine experienced its final event, the left engine experienced a second event. The following values/trends were recorded over a 2 second time period:

ITT: 12 Degree Rise

Tq: 1.5 % Decrease

Ng: 1.5 % Decrease

The auto-ignition system was automatically activated during the event.

The data revealed that 5 seconds after the left engine's second event, it experienced a third event. The following values/trends were recorded over a 2 second time period:

ITT: 17 Degree Rise

Tq: 10% Decrease

Ng: 3% Decrease

The auto-ignition system was automatically activated during the event.

The data revealed that 3 seconds after the left engine's third event it experienced a fourth, and final, event. The following values were recorded over a 2-3 second time period:

ITT: 15 Degree Rise

Tq: Decrease

Ng: Decrease

The auto-ignition system was automatically activated during the event.

TESTS AND RESEARCH

According to GE, during the approximate 9 million flight hours of Saab 340 service, even when operating in accordance with published requirements for activation of the engine anti-icing system, very short duration (1-4 seconds) engine power fluctuations (PFs) may occur. These PFs occur when snow/ice accumulations in the engine inlet are shed and ingested into the engine. Pilot reports indicate that the PFs occur at altitudes of greater than 10,000 feet, while operating at temperatures between International Standard Atmospheric values (ISA+0°C) and ISA+20°C. GE reported that there have been no pilot reports of PF's during the takeoff, initial climb, or approach phases of flight.

Based on past incidents and test data, GE has identified and categorized three types of ice induced PFs.

Type I: Occurs when the snow/ice is large enough to both stall the compressor and cause a partial or full quenching of the combustor flame. The event may be characterized by a popping noise, engine tailpipe flash, illumination of the cockpit engine ignition light(s) and the following changes in the engine parameters:

ITT- Drop

Tq- Drop

Ng- Drop

The auto-ignition is activated, and it is required for restoration of power. Engine power is restored within 4 seconds.

Type II: Occurs when the snow/ice is sufficient to cause only a compressor stall. The event may be characterized by a popping noise, an engine tailpipe flash, illumination of the cockpit engine ignition light(s), and the following changes in the engine parameters:

ITT: Increase

Tq: Drop

Ng: Drop

The auto-ignition is activated; however, it is not required for restoration of power. Engine power is restored within 2 seconds.

Type III: Occurs when minimal snow/ice is ingested into the engine and causes perturbations of the ITT, Tq, and Ng. The auto-ignition is activated; however, it is not required for restoration of power. This event will not cause an audible popping noise or engine tailpipe flash; however, the cockpit engine ignition lights will illuminate. Engine power is restored within 1 second. Additionally, since these events occur within a 1 second time period they have a sampling rate which precludes their data from being extracted from a flight recorder.

A comparison of the data recorded on the DFDR and GE's PF categorizations revealed that the airplane's left engine experienced four Type II (compressor stall) PF events, and the right engine experienced one Type I (partial/full quenching of the engine's combustor flame) PF event and one Type II event.

ADDITIONAL INFORMATION

On June 27, 1984, the Saab 340A was issued a type certificate by the FAA. In December 1985, following five occurrences of engine flameouts in icing conditions, the FAA issued Telegraphic Airworthiness Directive (AD) T85-25-52 prohibiting revenue operation of the Saab 340A in icing conditions. Saab immediately began researching the problem and determined that the flameouts were caused by ice/slush collecting on the engine inlet's bird catcher area and then breaking free and ingesting into the engine compressor. On August 18, 1986, the AD was amended to permit operations in icing conditions provided continuous ignition was selected, and further amended on June 17, 1988, to allow the use of a newly developed auto-ignition system as an equivalent alternative to continuous ignition. A new engine inlet design (Cox) was introduced; however, tests of the Cox inlet proved that it was not successful in preventing the accumulation and ingestion of ice/slush. On August 15, 1989, the Saab 340B, a derivative of the Saab 340A, was added to the existing type certificate. The Saab 340B was equipped with an identical engine inlet as the Saab 340A. The major differences between the two models were the substitution of the General Electric (GE) CT7-9B engine for the CT7-5A used on the 340A, and a standard auto-ignition system on the Saab 340B.

Saab and GE continued research and, in late 1992, a joint Saab/GE team was formed and a program initiated which revaluated the power interruption phenomenon. The teams actions included flight testing in icing conditions, engine altitude cell testing, and inlet icing tunnel tests. GE presented the team's findings to the FAA on May 19, 1994. The tests indicated that, while the engine inlet frame and inlet guide vanes were adequately anti-iced (heated), the inlet protective device or bird catcher collected snow/slush "under certain mixed conditions and in snow." Ice/slush was observed to build up on the rear wall of the bird catcher and shed from the splitter lip into the engine flow path.

On December 4, 1995, Saab aircraft briefed the FAA on the status of the PF program. Saab reported that a modified inlet design, utilizing increased electrical heat in the splitter lip area, "gave promising results" during additional inlet ice tunnel testing. They also reported that a Controlled Service Introduction Program (CSIP) was in progress to evaluate the effectiveness of the modified inlet in reducing power interruptions. The documented plan called for 42 modified inlets to be installed and flown on revenue service aircraft over the next two years.

On June 23, 1998, the Luftfartsverket (LFV), Swedish Airworthiness Authority, transmitted a copy of Saab's final report on the PF program to the FAA. The report summarized Saab's actions and included the CSIP program results. Saab's report stated that after testing various inlet designs, none resulted in any improved characteristics involving the PFs, and furthermore, changes to the inlet design may have adverse effects on the bird strike protection and the inlets' capability against ice related damages. Saab's report confirmed that PFs occur "infrequently and only at altitude normally above 10,000 feet." As a result of the program, Saab added a caution statement in the airplane flight manual which stated the following:

CAUTION: Engine power fluctuation may occur in icing conditions or shortly after exiting these conditions. Ice build-up might not be visible on the aircraft. Normal engine function will be retained by the auto-ignition system without any significant power loss. To aid in avoidance of these occurrences, activate the engine anti-ice systems prior to entering icing conditions, and maintain ON for at least 5 minutes after exiting icing conditions.

Saab also introduced information to the aircraft operations manual, further defining power fluctuations. Saab concluded that, as long the auto-ignition system performs its intended function, which is ensured by a mandatory daily check of the system (AD 96-01-04), safe operation prevails during icing conditions with the present design of the air inlet. The LFV stated in their letter that they were in agreement with Saab's position.

In August of 1998, the FAA responded to the LFV. The FAA stated that although AD 96-01-04 requires a mandatory daily check of the auto-ignition system, and the auto-ignition system has proven sufficiently reliable, this check was intended to be an interim action and does not obviate the need for a final design solution to address the root cause of the power interruptions.

During June and July 1999, Saab and GE convened a Power Fluctuation Flight Standards Review Board (FSRB), to evaluate inappropriate crew response associated with PFs combined with an in-flight engine shut-down. Power Fluctuation Pilot Surveys were distributed to pilot's flying the Saab 340, which included air carriers in the United States. The purpose of the FSRB was to determine the line pilot's opinion of the PF phenomenon regarding distraction and workload, and to determine the pilot's opinion of training material available regarding PFs. To further evaluate the distraction issue, all three types of PFs, including applicable sound effects, were incorporated into a simulator scenario. The pilots were placed in the simulator and evaluated by the FSRB and FAA. The FSRB determined that "a distraction exists, but not at a level w

NTSB Probable Cause

the ingestion of ice/slush into both engines, which resulted in dual engine power fluctuations.