Table of Contents
14.2 Risks and Problems of Technological Development
The first issue is the risk of defaulting. For example, a report is circulated in professional circles that states that a competitor was able to develop to maturity thermal barrier coatings with considerably better insulating properties than the ones used in serial operation at that point (Ref. 14-29). This development would mean higher gas temperatures with advantages in fuel consumption and performance concentrations. It could also be used to lengthen life spans and reduce operating costs. These decisive selling points could mean that this company would be able to control the market in future without serious competition. Therefore, if other companies without this technology minimize the development effort, it may be very costly for them later, and could even threaten their existence. For this reason, the risks of development must be weight against the risks of defaulting.
Technological development is typically marked by high risks. In addition, various risks influence one another. The development goal cannot be reached and/or the expense is too great (time, costs, resources). Glass can serve as an example. In general, it is considered absolutely resistant to corrosive influences in the surrounding environment. This assumption can be misleading under certain conditions that cannot always be completely excluded in technical applications. If glass is subjected to sufficiently high tensile stress (e.g. residual stress) and has a notch (small scratch, impact mark), normal atmospheric humidity can induce subcritical crack growth (stress corrosion cracking). In some cases, such as a windscreen with chips from rock strikes or a cracked drinking glass, crack growth can be observed over days and weeks until a critical crack length is reached and the glass shatters.
If an elaborate development is done in a part made of glass and this effect is not known or understood, there is a potential development risk. If the glass part were under high tension loads and in danger of being damaged, it is possible that the problem of crack initiation would only be discovered in serial operation. This means that even if a remedy were successful, the additional costs and loss of profit would threaten the existence of a company.
The greatest danger is in the case of a show stopper (Fig. "Show stopper during development"). In this situation, it is only after a considerable effort and expense has been invested that it becomes clear that a law of nature makes realization impossible.
The further a development process has proceded and/or the longer a system has already proven itself in operation, the more difficult and expensive it is to introduce a new serially implementable technology. Evolutionary improvements can only be expected in small steps (Fig. "Course of technological development"). Revolutionary improvements are usually countered by the high, already achieved requirements of operating safety and life span.
The development risk increases with the demands of modern engines. These include the use of ever harder materials and the increasing utilization of potential strength (Figs. "Resistance to crack growth by strength of materials" and "Technology for additional safety"). Even the very long guaranteed run times are a large risk factor for a development. High gas temperatures, which are considerably higher than the melting point of the materials currently used in serial applications, high efficiencies and inter-stage pressure ratios in compressors and turbines are a growing technological challenge.
New technologies usually have typical damage mechanisms and weak points. These sometimes reveal themselves only in serial operation with the large number of parts and accumulation of long operating times (Fig. "Residual stresses and thermal strength"). This makes it possible to realize only belatedly that improvements over the previous technology are cancelled out by drawbacks in the new technology.
If a technology has gained a bad reputation due to a spectacular accident or damage, a new attempt at implementation will probably only have a chance of success (Figs. "End of a revolutionary technology" and "Development time of fiber fanblades") with the next generation of professionals (Fig. "Technolocy development and maturity"). For example, an overspeed bursting test, which is designed to cause parts to fracture, can cause decision makers without sufficient technical knowledge to cancel development despite the part fulfilling all expectations.
These problems can lead to important safeguarding tests in the run-up to a technological development being left out, causing the development risk to grow uncontrollably. In addtion, pure show tests may be conducted which do not reveal a great deal about the actual required and utilizable potential, but are merely intended to positively influence the decision-making process in accordance with subjective interests. aA special problem occurs if a technology has a large improvement potential, but the required special properties can only be made useful with the aid of arranged design principles. A typical example is the optimal use of fiber materials such as fiber-reinforced synthetics (Fig. "Why a design did not go into series", 14-25, and 14-26).
Human problems are also inseperably connected with technological development. The required major engagement for the project and success, related prestige and cost responsibility, and a certain dependency of ones own existence, can compromise the necessary abililty to view projects critically. This means that the opportunity for sufficiently early withdrawal from a failed technological development may be missed. If there is a merely theoretic chance of success, by hoping for timely developments in the field (Fig. "Proof of concept tests"), it is likely that a path of no return is followed to the bitter end.
The reintroduction of technologies from earlier engine types only appears to make possible the use of immediately available experience with minimal risk. The demands for aerodynamically, thermo-dynamically, and mechanically highly loaded components of modern engines can usually not be met by older technologies, at least not without modifications. For example, the use of high-strength Al alloys (Fig. "Reintroduction of early technologies ") may fail because the sharp-edged, thin profiles of modern compressors cannot be realized, or the clearly stronger vibration excitements, relative to earlier compressors, cannot be controlled. In any case, new design principles, which must be determined, may be necessary. In Al blades, this could be roll bolt fastening of the blade root and the gluing-in of stator vanes with damping silicone gum in order to minimize vibrations, etc.
Technologies that also incorporate development of semi-finished parts can lead engine manufacturers to lose their focus on their actual task of producing serial parts, and instead concentrating on the development of foundations and semi-finished parts. However, this is the task of the suppliers, which is especially understandable in the case of cast and forged parts. Typical examples include parts made from high-strength ceramics or fiber-technical materials. It can even occur that the development focus shifts to technologies such as fiber production. This occurs because it evidently corresponds to a desire of the responsible scientist. In contrast, the actual task of the OEM, i.e. the testing of technologies and parts, is less pleasant. Result checking must be verifiably safety-relevant and takes place under special, usually well-controllable, time and cost conditions.
In conclusion, one more special problem must be mentioned. The selection of, i.e. investment in, personnel can determine success or failure. Their training and tendencies determine the course of the development. For example, engineers and technicians can be expected to focus on application-oriented development with testing of actual parts. Scientists, on the other hand, tend to foundation-oriented work such as material development and procedural methods. It is also important to realize these tendencies in the context of design project leadership.
Figure "Show stopper during development": The term “show stopper” refers to a situation in which, in the midst of the development process, a property that is absolutely necessary for a technology is discovered to be in violation of a physical law. If merely technical difficulties cause development to fail, then it may be possible to resume development at a later process when the anciliary conditions have improved (Ill. 14-3). Even an altered safety philosophy can justify a renewed attempt as long as only technically solvable problems remain. A typical historical example is perpetual motion devices. It is often difficult to recognize that these violate the first and/or second laws of thermodynamics. This became clear only after expensive and elaborate constructs were assembled and tested. Experience has shown that individuals external to a project have a greater probability of recognizing errors. In some cases, however, their objections were never accpted by the engaged developers. In such a case, costs can increase indefinitely.
Even in engine technology, developments have been undertaken even though they were doomed to fail under the conditions of the time. These include:
- Fan rotor blades made from fiber-reinforced synthetics
- Carbon fiber-reinforced ceramics for hot parts in civilian engines
- Turbine blades made from tungsten-reinforced nickel alloys
- Cr-based alloys (nitrogen embrittlement?)
- Niobium-based alloys (oxidation?)
- Effusion cooled turbine blades (Fig. "Success of cooled turbine blade types")
- Turbine blades made from directionally solidified eutectic materials with ceramic fibers that were formed in situ. In this case, which is depicted in the diagram, operation-realistic tests showed that the thermal strain differences between the matrix and ceramic fibers caused the fibers to break after very few thermal cycles. This eliminated the reinforcing effect. It is interesting that, in the multi-year, “scientific” development phase of these materials, the extent of the problem was either not recognized or not reported. The problem only became known to the users after they had already begun production of prototype parts. One should not ignore the personal difficulties that this caused for the individuals whose existence depended on these projects.
Figure "Resistance to crack growth by strength of materials" (Ref. 14-4): Although this diagram is based on older information from the literature, it clearly shows two tendencies that reveal a rising trend of failure risks in rotor parts of fighter aircraft engines. In one sense, the problem is increasing stress levels. These can be attributed to higher RPM (higher compressor performance) with increased weight reduction. The harder disk materials necessary for realizing the demands show a decreasing resistance to crack growth (structure, Figs. "Factors influencing crack growth" and "Trends of materials applications"). However, an increased resistance to crack growth is desirable in order to safely control any crack growth that occurs under the high LCF loads. A further complication is the fact that crack growth accelerates at higher stress levels (Fig. "Characteristic crack growth"). In addition, high stress levels allow smaller flaws to become capable of growth. Although this is an evolutionary development, the risk is increased. Quality assurance to compensate for the rising risk is correspondingly expensive.
Figure "Technology for additional safety": As discussed in Fig. "Material behavior depending on design and technology", it is of great importance for the development risk, if a technology is merely intended to increase the safety of a part, or to increase its stress resistance, as well. In the first case, the technology will have a safety-increasing effect, in the second, it will increase risk. Typical examples in engine technology include thermal barrier coatings, shot peening, and the inclusion of the crack growth phase. In all of these cases, the technology was used for many years to increase safety, and only used as part of the part design after evolutionary improvements were made.
Figure "Trends of materials applications": The trend towards increasing, highly utilized material strengths tends to lead to a potentially risky bevahior of the components (Fig. "Resistance to crack growth by strength of materials"). This tendency can only be balanced out through increased quality assurance eforts, better analytical control of all relevant loads, damage tolerant concepts (Example "AS HIP"), and extensive experience with tests in near-operating conditions.
Risks for part safety are influence by three main factors:
- design, i.e. loads
- material properties
- failure behavior
Design and loads: The design-conforming operating loads increase with strength. This, in turn, increases the crack growth rate, causes crack growth to occur in smaller initial flaws, and makes residual fractures occur at smaller critical crack lengths (Fig. "Characteristic crack growth"). The consequences can be seen in several examples:
Small material weak points become life-determining flaws. Requirements for serially implementable non-destructive testing methods increase. Finishing processes must be specified and controlled more exactly. Damage such as scratches created during assembly and maintenance work becomes less tolerable.
High mean stresses reduce the tolerable stress amplitude, thus increasing the danger of fatigue damage.
Material properties and failure behavior: The yield strength ratio is the relationship of the breaking strength to the yield strength. As loads increased, material variants with higher yield strengths and similar breaking strengths were introduced. The yield strength ratio fell.The greater the difference between the yield strength and the breaking strength, the higher the probability that plastic deformation will occur before the ultimate failure (fracture). Warnings, such as serious imbalances, enable earlier safe detection.
If the fracture strain decreases as strength increases, the material becomes more brittle. This is to be expected when maximizing strength. This makes it more difficult to break down notch stresses without damage through plastic flowing. In hot part materials, as thermal reistance increases, the probability of high residual stresses increases, as well. High creep resistance places tight limits on possible stress relief. The higher residual stress acts as a mean stress and lowers the cyclical load capacity (Fig. "Residual stresses and thermal strength"). Overstress can lead to brittle fractures even in thin cross-sections and low critical fracture toughness (Fig. "Forced fracture behavior by section thickness").
With higher strengths, the sensitivity of materials to environmental influences generally increases. These are corrosion types that initiate and accelerate cracking (Volume 1, Chapter 5.4).
If the increase in material strength and operating loads is paired with smaller grain sizes (e.g. powder-metallugrical materials), the unfavorable fracture-mechanical characteristic values can additionally increase risks. For example, the crack growth rate can increase (Ill. 12.2-8; Example "AS HIP", Excerpt 5) and the fracture toughness can decrease. This acts like a stress increase, causing smaller flaws to be capable of crack growth, and making critical crack lengths shorter.
Small grain sizes can have the advantage of better ultrasonic inspections (smaller disturbances, “grass”). Parts made from the same material, yet conventionally produced using casting and forging, can have larger grains than powder-metallurgical versions. This can result in considerably worse ultrasonic inspections, which increases the probability of larger flaws, thus worsening the strength utilization.
Figure "Residual stresses and thermal strength": High thermal resistance can considerably reduce the usable fatigue strength of parts in unexpected ways. The high thermal resistance and the structural influence sometimes prevent residual stresses from the forging process (Fig. "LCF fracture of a fan disk") and/or heat treatment to be sufficiently reduced through subsequent annealing treatments to make the mean stress increase is negligible (top diagram). The thicker the cross-sections, the greater the expected residual stresses. The residual stresses can have a life-determining effect on part zones in which this would not be expected (Fig. "LCF fracture of a fan disk"). If larger cross-sections, i.e. volumes, in a disk are under high tensile stress, it increases the probability of crack growth-capable flaws and thus also reduces the usable strength (Fig. "Causes of cracks below th surface").
Figure "End of a revolutionary technology" (Example "AS HIP"): This case is an excellent example of the risks of introducing new technologies in highly stressed serial components. The depicted fighter was undergoing testing. It crashed after the low-pressure turbine disk burst. The probable damage cause was an LCF fracture, which originated in a critical, highly stressed zone (arrows in right detail). The output power of this engine type had been increased considerably. The low-pressure turbine, which was especially highly stressed due to the single-stage design, had a disk made from ultra-high tensile strength material (Ref. 14-7). The material was an AS HIP version of a powder-metallurgical material, i.e. no subsequent forging. This material is distinguished by fine, even grains. Disadvantages include notch sensitivity, low crack toughness, and rapid crack growth. The dangers of this kind of combination of damage-promoting influences should serve as an important experience for anyone who works on the development of these technologies.
It can be assumed that the high loads made the crack growth phase extremely difficult to control. The major advantage of the powder-metallurgical material, aside from its high strength, was the exceptionally cost-effective AS HIP process, which resulted in semi-finished parts with dimensions that were very close to those of the finished part. This production process uses a hollow metal form (can), which is filled with metal powder in a vacuum and welded shut (Ref. 14-10). The filled capsule is then subjected to hot isostatic pressing (HIP) under high argon pressure (>1000 bar) and high temperatures, creating a dense material. The form is then removed.
This damage provides the following lessons:
- The informational value of an accumulated total run time of several engines and/or superstructural parts, especially if this is greater than several thousand hours, cannot be used as a sufficient argument for the serial application of a technology.
- New design concepts (in this case, a single-stage turbine) that considerably increase part loads demand more than mere increases in material hardness.
- The use of higher-strength materials reduces damage tolerance and demands additional fracture mechanical considerations (Fig. "Technology for additional safety").
- Fracture-mechanical charateristic values, such as fracture toughness, critical crack length, initial flaw size, crack growth, and notch-sensitivity, must be known, understood, and, if possible, optimized for the specific application.
- Safe life design alone is not sufficient. It must be combined with a damage tolerant concept. Therefore, the crack initiation and growth phases must be taken into account (Fig. "Configuration oriented safe life").
- A new technology must be understood in all its influences on part safety. This includes knowledge of flaws in the semi-finished parts: size, distribution, and probability.
- Proof-of-concept tests must directly verify the required properties. Indirect verification is not sufficient. The results of a burst overspeed test should not be used to draw conclusions regarding the LCF life of a highly-utilized part.
- An isolated case is usually only the first case (Volume 1, Ill. 2-3.2).
- A spectacular damage incident can easily mean the end of a technology, at least temporarily.
Example "AS HIP" (Fig. "End of a revolutionary technology"): “AS HIP”
Excerpt 1 (Ref. 14-6): “…fighter aircraft crashed shortly after departure from Farnborough Sept. 8, the day after the Farnborough air show ended. The aircraft, which had performed daily at the show, was en route to Spain for demonstration before the Spanish military…both (pilots) ejected from the aircraft shortly before it crashed…
The…(aircraft) departed Farnborough and reached an altitude of about 20,000 ft. when…(the pilot) experienced a loud explosion followed by a rapid rise in turbine temperature in the right engine. The engine was shut down…(the pilot) was unable to hold altitude due to control problems and also found that the throttle for the left engine would not move…
`The preliminary information we have received seems to indicate that there was a failure of the low-pressure turbine section…' Adding to this conviction of a possible explosion in the flight is the fact that two pieces of the engine's low-pressure turbine and an aircraft's speedbrake were found 20 mi. from the aircraft's crash site…
(The OEM declared) `We have accumulated 2,246 hr. in flight test program to date, and nothing in the history of the test program indicates a problem in the low-pressure turbine area'…“
Excerpt 2 (Ref. 14-5): ”…the …(fighter) engine will have new materials in the turbine section. Inconel DA718 is being used in place of Rene 95 in the turbine disks of this and all versions of the …(same engine type) being delivered to the Navy. The switch from Rene 95 was made after a low-speed turbine failed in a (fighter)…that crashed near Farnborough, England (Sept 1980)…
Use of the DA718 material for the turbine disks will save approximately $25,000 per engine,”…Rene 95 will continue to be used in high- pressure retaining plates and high-pressure rotor stub shafts.“
Excerpt 3 (Ref. 14-8): “Another …crash occurred on Nov. 14, 1980, but the navy opted not to restrict flight operations as had been done after the crash at the Farnborough Air Show. …The November 14 accident involved the first …(aircraft type) equipped with all of the proposed production fixes.
In December 1980, …(the OEM) indicated that while its technicians would probably never be able to identify the exact causes of the failure of the LP turbine disc in one of the …September 8…crash, the firm felt certain that such a occurrence would not again happen. According to the engine manufacturer, the fixes and changes being instituted were:
- Use of 150-mesh Rene 95 alloy in place of 60-mesh.(note: higher mesh value means finer grain)
- Low pressure turbine disks would be inspected at the 140-hour mark (half the number of hours on the engine involved in September 8 crash).
- The 6,000 hour-plus lifetime of parts made with the 150-mesh alloy had been downgraded to 2,330 hours.
- The number of bolt holes was being increased by 50 percent in the low pressure turbine disc (from 41 to about 60).
- The bore area of the LP turbine disc would be recontoured and bolt holes would be reshaped….(OEM) officials also confirmed that the company and its vendors had increased cleanliness in powder manufacturing by using completely sealed systems during the powder engine component fabrication.”
Excerpt 4 (Ref. 14-9): “A…design review team was formed soon after the…loss. …a copy of the report was first obtained by Defense Week and details were recently made public.
The report was based on a `worst case scenario' because only a portion of the low pressure turbine disc was recovered…The report's recommendations will be re-considered in view of the tests and evaluation of the test data.
According to the Navy officials, the low pressure turbine disc which fractured and failed in the September 8 accident had operated only 280 hours. The officials said that they never expected this particular component to fail since a similar disc underwent 4,000 to 5,000 hours of spin pit testing before failing. A second low pressure turbine disc, they added, has undergone 4,000 to 5,000 hours running tests and has yet to fail.
The report said that the low-time disc failure was caused by either
- a fracture emanating from a large undetected flaw or
- a fracture caused by higher than expected stresses, temperature or cyclic loading or
- a combination of both.Because the portion of the low pressure turbine disc which failed was not recovered, the project office is not able to pinpoint the cause of the low pressure turbine disc failure.
As a result, the Navy has implemented both manufacturing and design changes to guard against a re-occurance of the problem…“
Excerpt 5 (Ref. 14-11): The conclusions reached by the design review team …included:
- The Rene 95 material and the powder metallurgy process are not fundamentally suspect.
- Life analyses were overly optimistic and did not account for the developmental stage of materials and processes.
- A fracture emanating from a large undetected flaw or a low-cycle fatigue fracture caused by higher-than expected stresses, temperature or cyclic loading could account for the low time disc failure in the crashed …(fighter),
- Rene 95 notch sensitivity, manufacturing process evaluation and lack of confidence in …(OEM's) stress analysis suggest a combination of the above theories.
- Life prediction of …(the engines) rotors are in question as a result of this failure review…The team found that the …(engine's ) single stage turbine design required material of high ultimate strength and that considerations of material properties were state-of-the-art, but material selection placed little emphasis on fracture toughness and crack propagation and that … (the OEM) had compromised fracture toughness to achieve high ultimate strength…
A …(OEM official) countered these findings with:
`In the …(engine) our first criterion was to achieve overspeed capability in the neighborhood of 122. To achieve this goal you have to sacrifice some low-cycle fatigue values because they are not totally compatible. At …(the OEM) we judge time to crack initiation and not crack growth'…“
Comments: The problems related to the serial implementation of a new material can be seen as exemplary for this type of HIP process.
Evidently, an ultra high strength disk material was introduced and its strength potential was utilized considerably, even though the related risks were not completely understood. Because crack initiation during operation is to be expected at these load levels, the life span design of the disk must also incorporate fracture-mechanical approaches and use damage tolerant principles. Evidently, only the safe life concept was used (Fig. "Configuration oriented safe life"). The increased risk due to the notch-sensitivity of the part and poor crack behavior were not recognized in their full extent in the development process of the engine, which was still not completed.
Figure "New damage types of new technologies": With new materials and/or production technologies, specific damage forms must be expected. The introduction of directional solidified materials (left diagram, see Fig. "Grain boundaries influencing thermal fatigue") promoted radial crack initiation along the grain boundaries at the blade tip, followed by material break-outs and performance losses.
In single-crystal turbine blades (right diagram), during the finishing process (casting, annealing) or at high operating temperatures, recrystallization occurs near plastic deformations.. In this process, grains are created which tend to crack initiation and complete material break-outs under operating loads (creep, thermal fatigue).
Figure "Unexpected effects of new technologies": New technologies always come with surprises. Glasses and high-strength ceramics with glass phases (sintering aid) in atmospheric conditions exhibit extraordinary behavior. If these materials are under sufficiently high tensile stress, normal atmospheric moisture can incite stress corrosion cracking. In this case, cracking occurs, typically at small nicks (e.g. damage such as scratches or impact marks) with sub-critical crack growth to fracture (Ref. 14-26). A typical example is a car windshield crack (high tension residual stresses in safety glass) originating in a small impact mark from a rock. The crack growth to sudden shattering of the windshield can be observed over several days. A similar effect can be seen in household objects made of glass (bottom diagrams) and can also be found in medical technology (ceramic implants).
The failure of pyrometers (top diagram), which are used to monitor part temperatures in the turbine area, can be attributed to the described effect in several cases. If the fibers of the optical fibers to the sensor are subjected to high tensile stress in the fastening area, and moisture is present, individual glass fibers will break. This leads to gradual deterioration of the sensitivity of the pyrometer, and increases the danger of excessively high temperature levels.
Figure "New technologies unpredictable damage behavior": Unforeseen behavior under unusual loads, such as in the case of damage, heavy rubbing, or FOD, can prevent or restrict the introduction of a new technology. The following are several examples:
Typical situations with unusual behavior occur during damage incidents.
High deformation rates, such as impact loads, lead to brittle behavior in metallic and synthetic materials (Volume 2, Ill. 8.1-13). The degree of embrittlement is material-specific. Impact loads occur in situations such as containment incidents, bird strikes, or under the influence of other ingested foreign objects or fragments from within the engine itself. The embrittlement can be influenced considerably by the operating temperatures. Contrary to expectations, pronounced embrittlement can be observed at very high temperatures, as well as at very low temperatures (Fig. "Indications of turbine blade overheating"). The special tendency of some synthetic resins to embrittlement under impact stress can, for example, prevent their use in spinners made from fiber-reinforced synthetics. The reduced impact resistance of single-crystal materials at operating temperatures leads to considerably greater consequential damages than in less sensitive, but also less heat-resistant, multi-crystal materials.
The increasing risk with the use of high-strength materials can also be related to their unfavorable fracture-mechanical specific values. If this behavior is not correctly assessed, realization in serial applications is threatened (Fig. "Trends of materials applications", Example "AS HIP"). Increased notch sensitivity (e.g. in titanium alloys) can cause unexpectedly large consequential damages after damage due to foreign objects or during a damage process (e.g. haircut of a blade stage).
Notches and cracks in hard, brittle, and firmly-bonding coatings, which can be caused by foreign objects, can cause dynamic strength to drop considerably more sharply than it would in uncoated material.
Extreme rubbing at blade tips and in labyrinths causes a great deal of material removal from the contact surfaces. If this material is ignited, it can cause a dangerous dust explosion (Volume 2, Ill. 9.4-6).
Figure "Development time of fiber fanblades" (Examples 14-2 and 14-3): This diagram shows fiber-reinforced synthetic fan blades from two engine generations. The left diagram is based on a large, first-generation fan engine from the early 1970s. Despite similar design characteristics (long chord lengths, no clappers, dovetail blade roots), early development was a disastrous failure for the developer due to poor bird strike behavior.
The right diagram shows a fan blade from a third generation engine from the 1990s. The size difference is striking. The pronounced metallic front edge (titanium alloy) also attracts notice.
Serial realization was made possible primarily by the large mass, which means that equal bird strike energy will place considerably less stress on the large blade than it would on the small blade from the earlier development. In addition, 3-D reinforcements (against delamination) and improved fibers and matrix were used (impact resistance, strength). The integrated metallic leading edge guaranteed sufficient FOD behavior and erosion resistance. These properties have now proven themselves through very good operating behavior over long operating periods in a large number of engines. Their advantages include outstanding bird strike behavior and extremely high fatigue strength.
In order to use the now successful blade (Ref. 14-23), it was necessary to keep the circumferential speed of the blade below a threshold value (Fig. "Examples of successful fiber reinforced components"). Despite this, major problems occurred during verification of the part`s suitability for serial implementation. The blades, which were deflected considerably by bird strikes and containment, were overstressed by the sharp edges of the spacer segments that form the hub contour, causing the blades to fracture in the leaf section. The remedy for this was the use of less rigid spacer segments. The containment also had to be matched to the material behavior of the blades. This was aided by the demonstrated unique failure mechanism of the blades. Unlike fractures in metallic blades, the fractures in this case occurred well above the root.
Example "Failed serial implementation of fiber-reinforced synthetic fan blades" (Fig. "Development time of fiber fanblades"): “Failed serial implementation of fiber-reinforced synthetic fan blades”
Excerpt 1 (Ref. 14-12): ” …In another test work on Hyfil, the company has produced a theoretical analysis of bird impacts on metal blades and has applied this to the work on composites
…Hyfil has now been designed to withstand the impact of a 4-lb. bird, a size chosen from service experience which shows that chances of meeting a larger bird are small.
In the RB.211 engine, the Hyfil fan blade is nearly 3 ft. long and has a 1-ft. chord at the tip.
(Note: Hyfil is an early carbon fiber-reinforced synthetic material)
Excerpt 2 (Ref. 14-16): ”…it is now likely that 25-blade Hyfil fans will not be used until 1973 or 1974 for airline service even though one in four development engines is running with them. A decision whether to proceed with this development will be taken at the end of the year. The Hyfil blades are very stiff, require no midspan dampers and are stressed for 140 to 150 per cent overspeed, giving a 100 per cent excess stress level. Grit erosion and rain-induced de-lamination were overcome earlier this year, but the high stress levels in the root-generated by the steel inserts used to overcome these earlier problems,- posed another themselves. This has been overcome by a revised root splice, but the impact of large birds and the subsequent vibratory stresses still poses a problem.
Excerpt 3 (Ref. 14-17): ”…From the outset of the…programme it was decided that the fan blade should be specified in a carbon fibre composite known as Hyfil...Relative to the best performance that could be achieved with a titanium fan blade of acceptable weight penalty, the Hyfil blade represented approximately 1 % improvement in s.f.c and 300 lbs. reduction engine weight…
It was recognised from the onset that the introduction of a new composite material for the fan blade of a new engine could represent an extremely difficult development problem. The major difficulties that could be foreseen with the composite were those of protection against rain and grit erosion and doubts concerning the shear strength of the composite under conditions of large bird impact…
These potential problems prompted…(the OEM) to take out two insurance policies. The first was the initiation of a flight test program in which hyfil blades were installed in certain…(smaller engines). These blades are still flying…in normal civil operation,…the total flying time was 16,747 hours.
The second insurance policy was a decision to develop a titanium fan blade…in parallel…
experience on representative test pieces and also on glass reinforced composite blading…had indicated that a.g.r. composite material without any protection was quite unsuitable under conditions of rain and grit erosion…a perfectly satisfactory solution was developed in which the leading edge of the blade was nickel plated and the remaining surfaces protected by a layer polyurethane paint….under conditions of bird impact this treatment was inadequate in providing the necessary spread of the local impact at the leading edge…a hard leading edge was formed from a number of layers of thin steel which were than scarfed into the hyfil sheets…From a bird impact point of view this solution has proved extremely encouraging (and represented the basis for the standard blade).
…further problems…The first was a blade flutter …at approximately 80% L.P. speed…The traditional solution has been to twist a metal blade…A Hyfil blade is not at this time capable of accepting twist…
The second problem has been associated with the introduction of the steel laminates…Without the steel reinforcement the fatigue properties of the Hyfil blades have been excellent. Introduction of the steel laminates has led to some deterioration in fatigue properties due to the stress concentrations out of the steel sheets into the blade root…“
Excerpt 4 (Ref. 14-15): ”…the most famous name in British engineering, last week went into receiver-ship and was immediately nationalized by the British Government in a move patently designed to circumvent large cancellation claims for the…engine program, by the …(customer)…The official said that contributing to …(the OEM's) desparate position, in which there actually last week was no cash to pay salaries…“
Comments: Like Example "AS HIP", this example is very well-known among professionals, and is extremely instructive. The general impression is that important demands for the selected serial implementation of this technology were not understood in their full scope. This means that the developers ignored the fundamental rule, according to which a serial version should only be developed when all necessary properties have been verified and are available to the design engineer as directives (Figs. "Proof of concept tests" and "Time periods for series application of a technology"). Even after such grave problems became known, short-term developments such as metal sheets were applied, which led to new, unexpected damage mechanisms. This goes against the understanding that the potential of a new technology, which is even less understood than the problematic technology one is dealing with, will usually be positively overestimated. On the whole, one gets a feeling for the inevitibility of the entire process up to its catastrophic conclusion.
Example "Successful serial implementation of fiber-reinforced synthetic fan blades" (Fig. "Development time of fiber fanblades"): “Successful serial implementation of fiber-reinforced synthetic fan blades”
Excerpt 1 (Ref. 14-14): ”..Because the alternative test would be applicable only to composite fan blades, competing engines designed ..(by other OEM's) would be unable to benefit from the new test plan…The procedure … (the OEM) wants to change is the blade-off test. The present test calls for a blade to be explosively separated from a running fan at the blade root. The separated blade and resulting fragments must then be confined within the engine, mostly by the fan containment system - a ring of material, often reinforced with Kevlar or steel, that surrounds the fan - (the OEM wants) to separate the fan blade some distance away from the root because…tests and analyses indicate the composite blades - unlike metal blades - are unlikely to fail at their roots.
In addition to decreasing the weight of the resulting projectile, the modified blade off test would change the kinematics which would allow…to shorten the length of the fan containment system….the separated blade would be far less likely to strike other blades…the subsequent unbalanced loads placed on the engine, allowing …engineers to reduce the robustness of the engine's mounts, turbine exhaust case and bearing supports.
Excerpt 2 (Ref. 14-13): ” The …(OEM) has developed a solution to the fan-blade failure which has grounded…flight-test aircraft…The loss of three composite wide-chord fan blades during the birdstrike test was initiated by the affected blades coming into contact with aluminium platforms or flow-path spacers, located between the blades where they meet the hub. The first blade to be struck apparently flexed sideways, causing the aluminium platform to cut through the composite fibres at its base, weakening the blade and causing into break off when it rebounded. Two neighbouring blades were forced sideways and failed similarly, resulting in a fan imbalance exceeding design requirements. The fan is designed to withstand the loss of 2,3 blades. The …fan blade had already passed the same test at the rotational speed equating to the initial 377 kN-thrust rating and the test was being restaged at the higher fan speed required for 409 kN growth engine. In the first test the bird was fired at a point halfway along the blade. At the US FAA's request, the growth rig-test was staged with the bird aimed at the 75%-diameter mark.“
Comments: Even if the serial application of fiber-reinforced synthetic blades has evidently proven itself over a long time, the high residual risk during development can be sensed. Once again, it was high impact loads during bird strikes and containment tests that caused the unexpected problems. However, these problems were successfully solved because the properties of the blades were evidently sufficient for the chosen application. The following measures seem to have contributed significantly to this success:
- Development of the technology over many years, during which the decisive design and construction characteristics were evidently determined in a sufficiently realistic manner.
- A suitable containment test that took into account the advantages of the fiber-reinforced synthetic material. The OEM`s special accomplishment was establishing sufficient verifications of the advantageous fatigue and fracture behavior of the synthetic blades during the development stage. This information was deemed sufficient by the authorities to accept the blades.
- Determining design data, especially defining the circumferential speed limit of the fan blades that ensured satisfactory bird strike behavior (Fig. "Examples of successful fiber reinforced components").
Analysis of the cause of the problems in the failed bird strike test, including a specific practical solution, the effectiveness of which could be verified through the subsequent test.
Figure "Reintroduction of early technologies ": Contrary to expectations, technologies that have proven successful in older engine types cannot simply be used in modern engines. The high static and dynamic loads, as well as increased aero- and thermodynamic demands, require important adaptations that require development efforts approaching those necessary for the development of completely new technologies. Even the use of of traditional materials that are frequently used in older engines can lead to major problems in modern engine types with considerably higher part loads. A typical example are Al alloys in the fans or high-pressure compressors of fighter aircraft engines. The lack of a pronounced fatigue strength (bottom right diagram), combined with low strength levels and a sensitivity to erosion and corrosion, increase the risk of dynamic fatigue considerably. The thin, sharp-edged profiles of modern compressors further increase the problem.
Experience has shown that consequential damages following a blade failure and/or FOD lead to haircuts in large sections of the compressor (top diagram). This is primarily due to the large deflection caused by the low E modulus and the low blade weight (Volume 1, Ill. 126.96.36.199-5).
Another example is sheet metal constructions. The housings of early engine types differ from the cast parts of modern engines through especially complex welded constructions made from sheet metal and forged parts. Every so often, the suggestion is made to return to the older designs due to cost advantages they seem to offer. However, it is usually forgotten that these welded parts require a great deal of know-how. Conceptualizing sufficiently deformation-free weld patterns and training the necessary skilled specialist personnel is a long-term task. Even though the use of modern welding robots can minimize expensive labor costs, extensive finishing developments are indispensible. However, it may be worthwhile to consider these possibilities once again, since the possibilities offered by CAD, such as the simulation of joining processes, welding deformation, and strength calculation of complex structures, have opened up new perspectives.
Figure "Success of cooled turbine blade types": The cooling of turbine blades is a key technology that positively influences the fuel consumption, performance, and weight of an engine. Some concepts, such as effusion cooling and the use of monolithic ceramics, have not yet found their way into serial applications. In both cases, the reason was unsafe operating behavior due to an insufficiently safely controllable damage mode under influence of foreign objects (carbon impact). In addition, effusion cooling has problems with erosion, plugging, oxidation, and thermal fatigue. The blade with the “ceramic shirt” has problems with thermal fatigue and force transmission from the supporting metal core into the ceramic.
Therefore, intensively cooled cast blades have become the standard technology. These also have problems, but they are safely controllable within the required life span.
Modern engine types use turbine blades with ceramic thermal barrier coatings. This technology has thus far shown itself to be life span-guaranteeing. If the thermal barrier function becomes a component of the design, its reliability attains a higher degree of importance (Figs. "Material behavior depending on design and technology" and "Technology for additional safety").
Both serial technologies attained their current levels of maturity over years of evolutionary development, which makes the introduction of other, newer technologies very difficult.