Elastic materials store elastic energy upon loading. This energy is restored once the structure is unloaded. Inelastic materials dissipate some energy through a nonreversible process. This dissipated energy has been used to define fatigue criteria in several materials, since it can capture the stress-strain hysteresis effect. Depending on the application, the dissipated energy is defined in different ways (Ref. 4), such as dissipated creep energy, dissipated viscoplastic energy, dissipated energy of a specific creep process, or combined dissipations from different sources. In some fatigue models, part of the elastic energy is also used in addition to the dissipated energy. It is therefore of crucial importance to know which types of energy quantities that affect fatigue and which do not. This becomes even more challenging when, for example, a material has several creep contributions (primary, secondary, and tertiary), since each creep mode has an individual contribution to the energy dissipation.

Many of the Energy-Based Fatigue Models are expressed as an exponential relation between the number of cycles to failure and the dissipated energy. These models originate in the work by Morrow (Ref. 5 and Ref. 6), who introduced a fatigue model for ductile metals (see Morrow Model). A proposed relation has been applied with success in thermal fatigue modeling in the presence of creep or plasticity.

The definition of fatigue failure is not unique in the engineering and scientific communities. Commonly it is defined it as the crack initiation, but for some application all the history up to the final fracture is included. This ambiguity exists because certain structures are built to maintain operational integrity with cracks present, while some are considered as defective once a crack is initiated. The Morrow Model is frequently used to model failure defined as the crack initiation. Based on the numerical and experimental work on solder joint fatigue, Darveaux (Ref. 7) concluded that crack initiation only accounts for 10% of the number of cycles that are required to drive a crack through a solder joint. He proposed the Darveaux Model, which separates fatigue life into crack initiation and crack propagation.