Fig. "Failure potential of O-rings": The mounting and dismounting of components with an O-ring seal (e.g., magnetic plugs) as well as the exchange of an O-ring is a definitely demanding procedure. Typical causes for maintenance required problems are

• Unfavourable assembly conditions: To these belong bad visibility (Fig. "Eelastomer seal overheating") and hindered accessibility, for example at mounted aeroengines (Fig. "Technicians unfamiliar with aircraft").
• Lack of experience and expertise (Fig. "Nightshift problems" and Fig. "Ill advised maintenance work").- Damage during assembly process (Fig. "Assembly of O-ring seals" and Fig. "Influencing the O-ring sealing effect")
• Material problems, confusion, SUP (chapter 20.2.1) with early aging. Then typical results are unusual remaining deformations, shrinking and crack formation (Fig. "Problems due to change of the oil type").

To these problems belongs also the reassembly of O-rings, which rather should be exchanged against new parts. With this, the scheduled life time could be markedly exceeded and so aging deteriorations can get effective.
Missed exchange, e.g., after an `mothballing' of the aeroengine/aircraft (Fig. "Failing mechanisms of O-rings"). This can also apply for the long time storage of a component with O-rings (e.g., gear, control unit) at unfavourable conditions (e.g., too high temperature).

O-ring problems can from experience act dangerously. During leakage of a medium like lubrication oil or hydraulic fluid, important components (e.g., bearings, control units) and the aeroengine can fail. This danger increases, if during the same maitenance procedure, at several aeroengines of the aircraft the problem arose (Fig. "Maintenance error leading to engine failure").
Escaping leakage fluid can ignite and endanger the aircraft. Aggressive media or decomposition products at hot parts (e.g., hydraulic fluid) can trigger dangerous corrosion failures (Fig. "Contaminated Hydraulic fluid").
Also unnoticed scheared parts of an O-ring can block/clogg narrow, close tolerated (e.g., in a control unit) crosssections or oil nozzles.

Fig. "Assembly of O-ring seals" (Lit. 23.4.1-5 up to Lit. 23.4.1-10 and Lit. 23.4.1.16): To the layman, the assembly of an O-ring seems simple and unproblematic. This is a mistake. As well the mounting of the O-ring before the joining into a groove, respectively on the seat demands attention and skill. Must the ring slipped over edges of the groove, steps or threads (Fig. "O-ring damages by assembly"), it can be damaged by a cutting process. This is promoted by the usual elastic expansion and correspondent high cutting forces.
An especial critical assembly phase is the pushing together of the sealing faces.

If the O-ring sits already in the key seat („A”), it can be observed in the delicate moment of the pushing together.

Sits the O-ring in a groove of the bore („B“), it can not be observed. Unsuitable handling and/or a too sharp edge can shear the ring unnoticed. To indicate this, experience and sense are needed.

Shifted O-rings („D”) can no more fulfill their sealing function. Thereby the O-ring can be shifted out of the groove as a whole or partly. (Fig. "Failing mechanisms of O-rings"). It can even during the assembly shift as far, that its missing will not be noticed (Fig. "Technicians unfamiliar with aircraft").

As an earnest problem, hardly to experience, the forgetting of O-rings has proved („E“, Bild 19.2-2.2). Such a situation seems more likely for a sealing with two O-rings (Fig. "Maintenance error leading to engine failure"). Not seldom a lacking consideration of the `human factors'(Fig. "Maintenance error leading to engine failure") is also concurrently causative.

Fig. "O-ring damages by assembly" (Lit. 23.4.1-6): Already the preparation of the sealing surfaces of the O-ring like groove or the opposite face (sleeve), is of importance for a continuing safe sealing effect. They must be clean and must not have rough or sharp edged damages like wear spots, scores, notches or burrs.
Before the insertion, all contact surfaces of the O-ring must be suitable lubricated. This is especially important, when not sufficient guaranteed by the medium to be sealed (e.g., aeroengine oil). For example this applies for the whole length of a driving rod. The lubricant must be suitable for the service conditions. It also must not inacceptable influence the medium to be sealed. Therefore, only lubricants corresponding the specifications must be used. In case of doubt the OEM must be consulted.
For movable gas seals (pneumatic) often barium saponified grease/lubricant is used.
During the insertion, the ring shoud not be twisted and not pulled over sharp edges (sketch above left). Dangerous are spline toothings, slots, connection openings and threads. To prevent a damage, dangerous edges can be covered with a plastic cap or a plastic tape (sketch above right). During sliding on, the expansion should not be too large. Till the pushing together, the O-ring must have enough time to contract again. This is due to the fact, that elastomeres do not resile like metal. In the groove, the O-ring should not more expanded than 5% of its inner diameter. During sliding into the `sleeve' rotary motions and back and forth motions are not recommended. With it, the risk of a deterioration rather increases.

Fig. "O-ring failures and its causes" (Lit. 23.4.1-2, Lit. 23.4.1-3 and Lit. 23.4.1-15): To prevent failures at O rings at the beginning or with remedies, it is necessary to identify failure causes and failure mechanisms. Requirement is the evaluation of the failure mode. Specific problems shall discussed in the following more in detail:

Chemical decomposition is primarily determined from the service temperature. It should lay below the longtime temperature for elastomers, mostly between 135 and 150°C (Fig. "Aged O-rings"). Thereby it must be suggested, that concerned is an endurance temperature and not a short time maximum temperature (see section „thermal deterioration”). During too high temperature, a change of the molecular strukture (aging) will occur. The thermal stability of an elastomer is also for the same type markedly depending from the individual quality. Thereby contaminations and the interlacing of the molecules play an important role. For example a decomposition can be triggered also with trace metals. This means, that the quality assurance must finally rely on the producer and the supplier/source. This can be important undre the view of SUPs (chapter 20.2.1).
Typical consequences of an aging are:

• The embrittlement is the main problem and can be seen as the result of a temperature caused aging process. It makes especially prone for failing and crack formation during shock like load. To embrittlement tend „nitriles“.
• Drop of strength.
• Material change/decomposition.- Crack formation and crack growth. A special form is inner crack formation, caused by diffused gases into the material (Fig. "Inner crack formation at O-rings").
• Blister formation: Also these failures are traced back to gas diffusion (Fig. "Inner crack formation at O-rings").
• Lasting deformation, creep. In this connection the drop of resilience/spring back must be seen.
• Carbon dioxide can be fully dissolved in certain elastomers and gets them to swell with a sponge like structure.
  

Thermal deterioration is understood as caused by overheating. It develops, when the temperature, during which the elastomer decomposites short-term was exceeded. Such a situation exists when after the shut down of the aeroengine, the cooling by oil and air lacks. This leads to heavy heating by the neighboured hot parts („heat soaking”, Fig. "Eelastomer seal overheating").

Compression deformation is a plastisc (enduring) deformation of the O ring under the compressive contact forces. In connection with temperature changes and the so arising volume change of the ring, a leak can develop (Fig. "Failing mechanisms of O-rings"). Especially fluorine elastomers (fluororubber) tend to creep deformation at the typical high service temperature.

Explosive pressure drop: Into the O ring elastomer diffused medium, especially gas, gathers at flaws below the surface, like micro pores. During a sudden pressure drop, the gas epands or it comes to vapour formation. The result are deteriorations like blisters, cracks and holes.

Extrusion traces back to an overload of the O ring by the pressure deviation in the sealing gap (Ill.23.4.1-7).

Assembly damages: These occur in different types. Most frequently are cuts and sheared areas, which develop during carelessness during pushing together of the seal faces (Fig. "Assembly of O-ring seals" and Fig. "O-ring damages by assembly").
To these count so called „spiral damages“. They stay in connection with a twisting of the O ring. It occurs during pushing together with a rotative moverment.

Outgassing and/or dissolving from components, shows as shrinking because of the volume loss. During high temperatures it can come to a chemical decomposition process. It makes components of the seal material liquid or gaseous. Especially thermoplastic urethans are susceptible for such a deterioration.
Gets water vapour or overheated water into the pores which form during outgassing, a sponge like structure develops. Is the O-ring not sufficiently supported, a fast seal failure must be expected.

Fig. "Inner crack formation at O-rings" ( Lit. 23.4.1-12): The development of blisters in innner cracks, is traced back at a diffusion of the medium to be sealed. This diffusion proceeds to flaws like tiny pores. These (about 3 Vol %) must be expected in elastomers. The there accumulated medium leads to pressure rise and to tensile stresses vertical to the surface (not net modal flow). Then the crack development takes place parallel to the surface. A widening of the cracks forms a blister at the surface.
Susceptible for blistering are especially soft duromers (low shear modulus respectively low hardness). The propensity to blistering can be markedly reduced in a tight temperature range with a filler.

Fig. "Eelastomer seal overheating" Fig. "Eelastomer seal overheating" (Lit. 23.4.1-5): During shut down of an aeroengine, also the oil flow and the cooling air come to an stand still. With this, the heat from the hot parts can advance to the seals (heat soaking). Are the temperaturees too high for the used seal elastomer, it comes to a permanent deformation and/or deteriorations like embrittlement and crack formation (Fig. "O-ring failures and its causes"). At pumps for shaft seals, usually a minimum leakage of about 1 drop per minute (correlates 3 ml/h) is tolerated. Also this will be prevented if possible, because painting/lacquerung and coatings can be damaged (especially annoying because visible from the outside!). However thereby it should be considered, that an aging of the O ring leads fast to an unacceptable leakage of several drops per minute.

Fig. "Failing mechanisms of O-rings" (Lit. 23.4.1-18): O-rings are from a further mechanism mechanical loaded, besides the load, due to the pressure difference in the medium.
Usually an O-ring is embetted in a circumferential groove. It is elastically compressed by the sealing face and contacts this as well as at the groove face („A”).
Does it come during service to a heating, the elastomer expands markedly more (about a magnitude) than the metallic components. This volume expansion leads to a rise of the compression of the ring („B1“). Under this compression the ring forms plastically (creep). Does now the region of the seal cool down, for example duriing shut down of the engine, the ring volume shrinks correspondingly. Is the plastic deformation so large, that the elastic spring back of the ring material is no more sufficient to shut the gap, a leak will occur („C1”).
This process can also lead to a further failure mode, shear. In this case the seal material is extruded through the sealing gap into the groove edge at the low pressure side („B2“). This is promoted from an aging caused loss of strength. Thereby the ring can at least locally sheared off and a leak form („C2”).

Fig. "X-raying for an O-ring failure" (Lit. 23.4.1-14): Failure relevant O-rings usually are positioned inside badly or not visible cavities. During the disassembly, the danger of a damage or a change of failure explaining position features exist (Fig. "Influencing the O-ring sealing effect"). This suggests wrong conclusions. Therefore it is of high importance, to identify and document the unchanged position of the O-ring in the as assembled condition. Against expectation, the practice showed, that this is possible with X-ray, at least in casings from light metals. The O ring material (possibly depending from the filler) weakens the suitable adjusted X-ray sufficient, to produce enough contrast in the pictures.

Fig. "Influencing the O-ring sealing effect" Fig. "Influencing the O-ring sealing effect" (Lit. 23.4.1-14): During former flights, the crew noted fuel/oil smoke in the cockpit during flight. This lead about 1 week before the flight accident to the exchange of the Hydromechanical Metering Unit (= HMU) at the side of the aeroengine. The work was carried out from the operator. The dismouting took place as specified by a mechanic. The work was controlled and signed by an inspector. After this, an other mechanic mounted the change component. To this work belonged the installing of the cable to the bypass switch at the cover. To the HMU belongs the fuel filter. Also this was signed by an inspector. After this followed a test run of the aeroengine with a fuel leak clontrol. Two inspectors monitored the run and signed it. A mechanic cleared the aeroengine for flight. Obviously, at this time no noticeable fuel leak has occurred.

The HMU of the suspect right accident aeroengine was dismounted for a test run. Already at below 10% of the operation rotation speed of the fuel pump an extreme leakage occurred at the casing of the fuel filter. After this the test was cancelled.
Already during the first inspection of the aeroengine on the wreckage, an almost 0,3 mm wide gap between the filter cover and the contact face of the casing was noticed (sketch left). After about one day, fuel leaked further out of the gap.

Before the removing of the filter cover, the filter was X-rayed. This was carried out although it was not certain if the O ring material weakens sufficiently the X-rays in the metallic casing for an analysis. However the O ring was perfectly visible in its position. It could be seen, that the O-ring was partly pressed out of the groove to the cover (sketch below right). This enabled the exit of the fuel at the aeration bores and the drainage bores as well as the gap at the cover. It can be supposed that the usual pressure fluctuations in the high pressure fuel system loosened the unsecured filter cover and caused the dislocation of the O-ring. At the opinion of the investigator the cover was already not mounted as specified at the producer supplied component before the assembly.

After X-ray the filter cover was unscrewed. Now the O ring was again in the designated groove (!). Obviously the O-ring had displaced back. Merely in the before displaced region, small traces of abrasion and a little cut could be seen. This shows the high valuue of the X-ray indication without which inportant findings would have lacked. The cover of the brushes from the starter generators of both aeroengines have not been fixed as specified in the manual. These are belts which are laid around the generator. The corresponding instructions could neither found in the work handouts of the operator, nor in the maintenance instructions of the airplane manufacturer. The gaps at the fasteners have been unfortunately so distorted, that they have been positioned over the casing openings to the brushes. So during a fuel leak, here an ignition could occurred.

Conclusion of the investigating authority (analogous): Cause for the flight accident was probably at the OEM the faulty installation of the cover from the fuel filter. Thereby it came to a gap between cover and supporting face. The operator personnel did not notice this obvious failure. So the large fuel leak was possible. The fuel could ignite at the incorrect bare generator brushes. The consequence was an explosion inside the aeroengine nacelle. This brought the fire extinguish device out of function. So the fire was possible.

Fig. "Aged O-rings" (Lit. 23.4.1-9): Three years after the already about thirty years old aircraft has been „mothballed“ three years and the rebuild to a fire extinguishing airplane, it came again into service. Before it underwent an extensive inspection. Thereby no alarming changes have been determined. Particularly during the examination of the fuel system about three months before the accident, no fuel leaks have been observed.

Anyway it was known, that it can come after a three years storage to „drying-out“ and shrinking of the O-rings. However the O-rings are `on condition' parts, these are not life time limited. In this case an exchange applies for flattened and shrinked O rings (Fig. "O-ring failures and its causes"). The producer of the aeroengine indicated, that a used O ring must no more incorporated again. Already at other airplanes of the same type, during military operation fuel leaks occurred as result of O-ring failures. However, a fire in connection with a leak was not observed. Anyway there was a warning of the military operator in case, that the fuel leak develops near an aeroengine.The hot part of the primary failed aeroengine is positioned exactly below the nacelle region (dry bay). In it, leakage fuel can be gathered and flow on the hot aeroengine. The used fuel (Jet-A1) has a relatively low ignition temperature (flash point, ignition of the fuel vapour with a flame, Ill. 22.2.1-2) of about 40 °C.

These failures have been traced back to bending or thermal expansions of fuel pipe lines. They occurred more frequently at screw connections (couplings) as at a fuel valve.
The leaks had different intensity. Over a longer period of time `bleedings' or drops developed. The end phase is a sudden failing of the seal with a spray cone.

Fig. "Causes for leaks at O-rings in lines" (Lit. 23.4.1-9): Through thermal expansion differences between pipe lines and its screw connections, markedly movements can occur. Thereby bending and/or axial displacement is possible. Are so the sealing O rings deformed in the screw connections/couplings and/or underlie wear leaks can develop. In such application specific areas especially at the O-ring condition must be payed attention.

Fig. "Influences of O-ring materials" (Lit. 23.4.1-2 and Lit. 23.4.1-4): O-rings can be produced from many different elastomers (cross column). These materials can get specifically deteriorated by certain media (vertical column). Thereby it must be considered, that besides the material to be sealed, like aeroengine oil, hydraulic fluid or fuel further come into consideration. Naturally the corrosion protection oil must not be critical for a long time storage. During maintenance and overhaul rather short time contace must be expected. Here media like rust remover, degreasing, solvents and lubricants come into question. Also these must be used only correlating the specifications. In case of doubt the OEM should be contacted.

Fig. "Identification of O-ring material" (Lit. 23.4.1-2): The identification of the O ring material, at least a assignment to larger `families' can be of interest.
For this, the shown sinple device is suitable. It consists of a tube in which a weight drops at the specimen from a specified height. The height of the rebound will be assigned the calibrated range.

The following gradiation with decreasing rebound height is material typic:

• EP (Ethylene propylene),
• Nitrile (Buna-N),
• Perfluor rubber/elastomer (similar PTFE),
• Fluor rubber, Fluor carbon rubber (Viton ®).