During pulsating hydraulic fracturing (PHF), reservoir
rock can
be subjected to constant amplitude, constant mean stress (CACMS) cyclic
loading or constant amplitude, increasing mean stress (CAIMS) cyclic
loading. The influence of increasing mean stress on rock fatigue strength,
fatigue lifetime, fatigue damage, and energy evolution of shale is
rarely investigated, and which type of cyclic loading is more efficient
for PHF has not been determined and demonstrated. In this Article,
a series of uniaxial compression tests under these two types of cyclic
loading are first conducted. A fatigue lifetime model for CAIMS is
established. The shale strength and fatigue lifetime for CACMS and
CAIMS are then compared. Their differences are explained by their
dissipated energy density evolution and damage evolution. Finally,
a nonlinear damage accumulation model to predict damage evolution
for CAIMS cyclic loading is proposed. It is suggested that CAIMS is
a better cyclic loading type when its amplitude is higher than 30%
UCS (uniaxial compressive strength) of rock. This provides a meaningful
amplitude threshold for CAIMS parameter optimization in the PHF construction.
In this case, compared with CACMS cyclic loading, CAIMS cyclic loading
significantly decreases the shale strength by up to 20% UCS and the
fatigue lifetime from over 500 to 9. Different from an inverted-S-shaped
damage evolution for CACMS cyclic loading, damage evolution for CAIMS
cyclic loading exhibits a monotonic increasing trend. The damage variable
growth rates of CACMS and CAIMS both show a three-stage trend: (1)
their growth rates both decrease; (2) the growth rate for CACMS remains
stable, while for CAIMS its growth rate slowly increases; and (3)
their growth rates both increase sharply. Correspondingly, the dissipated
energy density evolutions for CACMS and CAIMS show a similar three-stage
trend. It can be concluded that during the second stage, for CAIMS
microfractures and plastic deformation inside specimens are developed
and accumulated more and faster than those for CACMS cyclic loading.
This explains the phenomenon that the fatigue lifetime and strength
of CAIMS cyclic loading are less than those of CACMS. The nonlinear
damage accumulation model proposed in this Article can well fit experimental
results. This model can be used for accurately describing reservoir
rock mechanical property degradation during hydraulic fracture simulation
induced by PHF.