Cruising in darkness at 41,000 ft., on July 31 near Chengdu, China, the crew of an AirBridge Cargo Boeing 747-8F were beginning to prepare for the descent into Hong Kong when they deviated to avoid a thunderstorm clearly depicted on the weather radar.

Even if they had been able to visually check their surroundings, they would not have noticed anything unusual about the area they penetrated in the outflow region of the anvil cloud trailing the relatively distant storm. There was no sign of airframe icing, nor any echoes from the radar.

Yet the cloud was full of undetectable ice crystals that—within minutes of the encounter—caused significant damage to three of the aircraft's four engines, one of which lost thrust while another surged. The AirBridge Cargo (ABC) crew had unwittingly come face-to-face with core engine icing, a poorly understood phenomenon that has been striking a wide variety of aircraft and engines on a growing scale since the 1990s. As well as surges and mechanical problems, the previously unrecognized form of icing inside engines causes thrust loss, or power “roll-backs,” with virtually no warning.

According to Russian federal air transport authority Rosaviatsia, chief investigators of the 747-8F event, the crew saw at least one typical clue to the phenomenon. Entering the area of ice crystals, the total air temperature (TAT) rose by 20C to -34C for 86 sec. The crew reacted by switching the engine ice-protection system from automatic to manual for about 10 min. But approximately 22 min. after flying through the warmer sector, the aircraft's No. 2 (inboard left) engine surged and automatically restarted. The No. 1 engine then experienced a speed reduction of 70% of N1 (low-pressure rotor speed). After landing at Hong Kong, inspections revealed damage to the high-pressure compressor blades of the Nos. 1 and 2 engines, as well as to the No. 4.

Within weeks of the latest event, Boeing and General Electric flight tested an engine software upgrade specifically designed to counter the ice-crystal buildup. GE says the software changes to the GEnx-2B full-authority digital engine-control unit will help the engine itself detect the presence of ice crystals when the aircraft is flying through a convective weather system. If detected, the new algorithms will schedule variable bleed valves to open and eject ice crystals that may have built up in the area aft of the fan, or in the flowpath to the core. The modification to the GEnx control logic leverages similar changes made to improve the ability of the CF6 to operate in similar icing conditions.

The AirBridge Cargo event is the latest in a growing number of engine-icing incidents, which have triggered recent changes in international certification requirements. Unlike traditional engine icing, in which supercooled liquid droplets freeze on impact with exposed outer parts of the engine as the aircraft flies through clouds, engine core ice accretion involves a complex process where ice particles stick to a warm metal surface. These act as a heat sink until the metal surface temperature drops below freezing, thereby forming a location for ice and water (mixed-phase) accretion. The accumulated ice can either block flow into the core or shed into the downstream compressor stages and combustor, causing a surge, roll-back or other malfunction.

Until relatively recently, it was assumed that ice particles would bounce off structures and pass harmlessly through bypass ducts, or melt inside the engine. Now, there is evidence of an environment where a certain combination of water, ice and airflow is susceptible to accreting ice. Like many of the other known core icing events, the ABC 747-8F incident occurred near convective clouds. When incidents were first reported, investigators initially assumed supercooled liquid water, hail or rain were responsible because they had been lifted to high altitudes by updrafts. Yet most events have been recorded above 22,000 ft., which is considered the upper limit for clouds containing supercooled liquid water.

According to investigators studying flight data recorders and crew observations from previous engine-loss events, all took place at high altitudes and cold temperatures. Incidents struck regional jet aircraft at median altitudes and temperatures of 29,000 ft. and -32C, while for larger jet transports, medians for most events were at altitudes and temperatures of around 26,000 ft. and -21C. All events occurred near convective clouds and/or thunderstorms, in air significantly warmer than the standard atmosphere and in clouds or visible moisture. Common to all were anomalous TAT readings with no significant airframe icing and no weather radar returns.

To find out exactly what is happening inside the convective systems that most frequently cause core icing, particularly in mid-latitude and tropical regions, an international team plans to conduct the High Ice Water Content (HIWC) test campaign in Darwin, Australia. The team includes NASA, FAA, Environment Canada, Transport Canada, Airbus, Boeing, the U.S. National Center for Atmospheric Research and the Australian Bureau of Meteorology. Also joining the effort will be the European Union's High Altitude Ice Crystals (HAIC) project, which will be contributing a specially configured Falcon 20 research aircraft.

The European effort also builds on the European Aviation Safety Agency's (EASA) High Ice Water Content program, which itself used data collected on a series of flight-test campaigns conducted by Airbus in 2010 in the wake of the Air France 447 A330-200 accident in June 2009. The investigation determined the chain of events leading to the crash began with the “likely” obstruction of the pitot probes by ice crystals. As part of its safety recommendations, the French air accident investigation agency, BEA, proposed in July 2012 that EASA “undertake studies to determine with appropriate precision the composition of cloud masses at high altitude,” and based on these results, modify certification criteria for air data probes. The HIWC/HAIC campaigns are therefore intended to provide better understanding of glaciated icing conditions that could also affect air data probes.

Originally planned for early 2013, the timetable for HIWC was slipped to 2014 after delays to the modification of the NASA Gulfstream II—originally designated as the primary test platform. However, further delays to the Gulfstream modification have forced HIWC planners to consider contingency plans under which the Falcon 20 will become the primary aircraft, possibly flying with a scaled version of a research instrument originally intended for the larger NASA aircraft.