Engine Non-Containment
As part of my thesis at Cranfield, and at the request of the UK AAIB, I completed an analysis on the regulations pertaining to engine non-containment. The objective was to determine whether the failure modes demanded a change in the model as depicted by the then current regulations (1995). A central data base exists for the collection of data relating to each un-contained event so that regulatory design engineers and engine OEMs can assess what changes should be introduced to the regulations to maintain the probability of experiencing an uncontained event to the required level (which of course is ultimately what aircraft are designed around – probability that is !).
Un-contained events are more frequent that the general public would probably like to understand, and have occurred in almost every major engine type at some point throughout their production life cycle. The reason for the liberation of the material is though different for each event.
I am running from memory but I don't believe that the regulations, as they were back in 1995, demanded that there be an assumption that the fan blade, by itself, is more likely to fail than a turbine disc. However there is a higher probability that the fan blade will be damaged by an external influence (i.e. bird strike) and therefore fail - hence the containment ring.
The regulations are extremely precise about the quantity of mass, exhibiting specific energy levels, that need to be contained. Large pieces of turbine or compressor discs will not be contained as if they were the containment shield would be of such weight to make the aircraft un-economical. The fan blade shield for example already consists of a wrap of Kevlar a couple of inches thick. The regulations, for material that is liberated in an assumed dispersion zone, subsequently require that the airframe demonstrate a sufficient level of redundancy should such an event occur. This is achieved by the positioning of certain systems in physical locations that will demonstrate, when the material dispersion model is run, that a regulated level of redundancy is available, permitting the crew to retain control of the aircraft.
In addition the airframe and engine OEMs have to demonstrate that, with the current design and future maintenance and inspection philosophy (i.e. the maintenance and engine shop visit programme and inspection requirements) that the probability of the disc being released, is (from memory) and causing a fatal accident, in the order of 10 to the power of -6. This figure assumes that there are no flaws or errors built into the design and that all scheduled, mandated inspections and maintenance has been completed.
Most disc non-containments occur as a result of a deviation from the inspection or maintenance requirements or from the disc exceeding its operational limitations. The problem may be further compounded by an up-stream failure of a blade causing further damage downstream. Ocassionally design (CF6) and material flaws (Sioux City DC10) are also causal factors.
Some older types of aircraft, such as the B737-200, simply do not comply with the current EASA requirements in this respect, but modifying them would be too costly.
In short we should not be surprised when material, that exceeds the regulatory model, is liberated from an engine. Disks will continue to be ejected so long as we continue to operate aircraft.
Airbus should though be congratulated as they have demonstrated, very publicly, that the 380 has the required level of redundancy. Most events like this go un-recorded - much to the relief of manufacturer's around the world.
Now, can anyone calculate what is the probability is of a passenger being involved in an un-contained engine failure and then, once rescued by a passing QF744, be diverted for a second time, due to yet another engine failure of a different type, back to the same departure point - twice in succession ? I give up at that point.