DESIGN FOR PERFORMANCE

Materials development and design for high temperature applications has, for about seventy years, been based on isothermal constant stress (or more precisely constant load) creep tests. From such tests, the times for specific creep strains, and the minimum creep rates, may be obtained as functions of temperature and stress. Using this information and the times for these isothermal tests to fail (so-called rupture lives), and extrapolating these times to service requirements, has allowed the setting of design stresses with appropriate safety factors. Much has been made of the importance of improving the precision of extrapolation of such data, based on the assumption that this will lead to improved reliability in operation.

This approach, when applied to ductile materials, has proven to be very conservative, since design stresses are normally reduced to levels where a significant part of the creep strain is anelastic, i.e. removing the stress allows recovery of some nonelastic strain. Nevertheless, failures in such ductile materials do occur, and these are generally the result of overtemperature, overstress, cyclic deformation (fatigue), or environmental interactions. Anticipating such failures, i.e. assessing the remaining life of operating components, has become a high priority. Unfortunately, attempts to assess remaining rupture life on the basis of failure times of miniature samples taken from the components has revealed what has been termed A Remaining Life Paradox. This may be stated “When a component fails, the material of that component has a finite life in a creep rupture test sometimes approaching that of new material.” Thus, the rupture life of a sample taken from an unbroken part may have little bearing on the remaining life of the part.

In recent years as advanced materials of higher strength, but often reduced fracture resistance and greater sensitivity to gas phase embrittlement (GPE), have been introduced in energy conversion systems, the incidence of high temperature fractures has increased. A good example is in the operation of combustion turbine blades where the application of protective coatings is increasingly relied upon to prevent environmental attack and early cracking. As a consequence, the inherent conceptual flaws in the design philosophy are now becoming apparent.

It is often argued that long time creep tests are needed to simulate the evolution of microstructure and damage during service, and that the longer the creep test the better. Thus, unlike design procedures used in low temperature applications, the test is required to incorporate such changes, rather than be used to monitor their consequences. However, if we wish to simulate service complexity in our material testing, there is in fact a hierarchy of increasing complexity and expense. Thus, long term creep tests are clearly inadequate; we must additionally account for non-steady test conditions, complex stresses, cyclic stresses, environmental effects, and synergism among them. We see that, even with the most complex test plan, we cannot approach most service operating conditions. In fact long time creep rupture tests only allow more accurate extrapolation of the experimental test data

An alternative approach is to simplify the test methodology and develop tests to measure separately the high temperature creep strength and fracture resistance, ideally to evaluate the current state in terms of these properties. The consequences of microstructural evolution, induced in service or in laboratory simulations, can then be assessed using the same short time tests. Design is then based on minimum acceptable performance levels. This we term Design for Performance.

Whereas the creep to rupture test represents an arbitrary combination of deformation and fracture processes, the Design for Performance approach decouples the creep strength from the fracture resistance. Thus, a question such as: “What is the creep rupture behavior of a material in tests lasting 50,000 hours?” becomes: “What is the creep strength and fracture resistance for new material and what are their values after 50,000 hours service?” Short time tests are used to evaluate the current state of the material and set minimum acceptable values for these critical properties so that the end of part life may be defined without ambiguity. By subjecting material to suitable laboratory exposures, it is often possible to establish worst-case conditions for degradation of both creep strength and fracture resistance. The methodology then provides a basis for robust engineering design.

The creep strength is defined in terms of the stress vs. creep rate covering about five orders of magnitude and generated from stress relaxation tests (SRT) lasting about 24 hours. This eliminates the implicit use of time as a state variable in the current approach, and properly recognizes creep rate as a state variable. The challenge then becomes the generation of low creep rate data rather than long time data. Fracture resistance is measured in a constant displacement rate (CDR) tensile test at a temperature where the material is most vulnerable to fracture. The displacement at failure in this test is a sensitive measure of embrittlement.

This approach has been applied successfully to metals, polymers, intermetallics and ceramics. It enables rapid optimization of chemistry and processing variables, creep analysis for complex thermal-mechanical histories, and rational decision support for component life assessment. In particular, it has been shown that gas turbine blade superalloys may show little change in creep strength but profound degradation in fracture resistance due to environmental attack. This has allowed the end of safe component life to be based on a fracture criterion. By contrast, alloy steels used in steam turbine applications may show no reduction in fracture resistance but appreciable losses in creep strength. Remanent life decisions may then be based on creep strength evaluation.

Since actual failures of gas turbine blades may result from some combination of thermal fatigue, high cycle fatigue or impact, it might be argued that additional performance criteria will be necessary. However, when tested in vacuum, all classes of materials can be collapsed onto a single unique curve for both strain controlled fatigue (normalized with tensile ductility) and fatigue crack propagation in terms of stress intensity factor (normalized with elastic modulus). Thus, the CDR test provides a basis for measuring the sensitivity to environmental attack and for assessing fatigue behavior. For a cracked body analysis, quantitative fatigue data in the appropriate environment are needed, but for the Design for Performance objectives such data add little.

In summary, Design for Performance allows accelerated materials optimization during development and selection, provides a comprehensive framework for analyzing creep behavior, and offers a conceptually new and self-consistent approach to life assessment. This is not simply a way of accelerating testing to give imprecise but useful screening information. Rather, the approach addresses the fundamental flaws in the current method. By decoupling strength and fracture properties, and recognizing the importance of environmental attack (GPE), there is no longer a need to run long time creep tests, and the new approach becomes faster, cheaper, and better.