My involvement with ‘ice crystal icing’ started in the late 80’s with engine ‘rollback’ events involving the BAe146 aircraft; but only with the ALF 502 engine. At that time, the cause of the problem was not known.
As events accumulated, numerically and severity, investigations and tests associated the issue with ‘weak’ engine anti-icing, large Cbs, high altitude (>FL300), and large ISA+ deviations.
Ground tests in engine/icing tunnels together with modeling provided a consistent theory linking ice particles (seen in other work), from which modifications evolved.
Flight tests in European icing conditions repeated the normal icing certification parameters (Appx C) and checked that very severe SLD was not the cause.
Flight tests in / near large Cbs were conducted in Panama to measure atmospheric conditions and induce a rollback; no rollbacks occurred. Subsequent tests in the US in similar conditions did induce a rollback on an unmodified engine and proved the effectiveness of the modifications on another.
These tests are summarized in the ‘Strapp’ report (The Ice Particle Threat to Engines in Flight -
http://www.pprune.org/tech-log/37634...ml#post5027449) under ‘Commuter Transport Aircraft’.
The essential elements of the theories were substantiated with the work by Strapp, and that of Stromberg (University of Manchester/UMIST), Maltby (BAe/now RR?), et al – Honeywell engines etc, using the BAe146 test aircraft. Fully instrumented engines including one with cameras inside the engine identified ice formation and severity over a range of ‘ice crystal’ conditions.
The relatively small 502 engine suffered ice blocking of the first stage compressor stators which reduced the airflow, rolled back the N2 and thence N1 with increasing EGT. The rollback conditions were only susceptible in a narrow ranger of atmospheric conditions, depended on the duration of the encounter, and engine air off-take – airframe anti-ice, bleed-band setting etc. The conditions were associated with TAT probe anomalies due to freezing, but no pitot problems were seen.
No airframe icing was encountered during these tests – little or no liquid water – classic icing theory. Some light icing was seen on the windshield and wipers where the crystals melted on the heated screen and refroze at the edges; the screen appeared wet / very wet like flying in heavy drizzle.
There were some extremely damaging hail encounters in the tops of the Cbs, but not much fine ice crystal content, this was more in the outflow and anvil regions. Hence aircraft avoiding, or ‘cutting the corner’ around Cbs were more likely to encounter the problem. Interim restrictions on the 146 required Cb avoidance by a wide margin (20 nm IIRC) and subsequently no flight above FL280 in ‘icing conditions’ (< -10C in visible moisture). Problems still occurred due to HF and ‘interpretive’ issues, misjudging the distance to the storm (core centre / edge of the ‘red’) – use of WR, and lack of an airworthiness mandate. Also, the conditions were very difficult to detect. The ‘ice’ cloud structure was very wispy, almost non existent, and gave little or no return on radar.
Following sever multiengine events, ADs were issued and a mandated modification program instigated; all engines were modified by improving the aerodynamic design of the first stage of the compressor and heat flow in the inlet guide vanes, amongst other minor improvements.
The very similar ALF 507 engine was not affected; it had two stages of at the front of the compressor vs one, and was further reinforced by the introduction of FADEC (bleed control).
BAe/Honeywell did liaise with other engine manufacturers who also reported high altitude ‘unexplained events’ – (manufacturers do talk to each other about safety matters). However the nature of the problems on the ‘big’ engines were not the same or severe (at that time) as the 502 engine; the differences are described in the report under Large Transport Aircraft.
Although the ‘threat’ does not appear to be universal to all engines or aircraft, the issues of severity and exposure to ice crystal conditions, the inconsistent nature of the atmosphere, and changes with the ‘high tech’ evolution of engines should not (IMHO) lead to any complacency; it may be that we have yet to see a large enough storm or for someone to misidentify the hazard applicable to thier aircraft.