TU Yi-liang, LIU Xin-rong, ZHONG Zu-liang, et al. The unity of three types of slope failure criteria[J]. 2018, 39(1): 173-180.DOI: 10.16285/j.rsm.2016.0056.
The energy mechanism of material damage in thermodynamic theory was introduced to slope engineering. An energy conservation equation that was applicable in strength reduction period was deduced. Then
a calculation procedure for the slope energy was developed with FLAC3D
which was applied to a traditional slope example. Compared with the result from Spencer method
the slope’s energy change was closely related to its stability in the strength reduction method
hence four new slope failure criteria-energy catastrophe criteria which were theoretically in unity were proposed. By comparison
the criterion of the kinetic energy catastrophe is in correspondence with the criteria of numerical calculation non-convergence
the loss of gravitational potential energy catastrophe is in correspondence with the criteria of dramatic increase in the marked nodal displacements
and the dissipated energy catastrophe is in correspondence with the criteria of a plastic zone going through the slope. In summary
the aforementioned correspondences demonstrate the unity of the three common failure criteria. Several examples demonstrate that the different results from the three common failure criteria are because of artificial factors such as mesh generation precision and numerical convergence criteria. In essence
the different results stem from the fact that numerical calculation is a kind of approximate solution. In application
the accuracy of safety factor can be evaluated by the consistency of the results from various slope failure criteria. The better the consistency is
the higher the accuracy is. The basic method of improving the accuracy of safety factor is strict convergence standard and fine mesh precision
but achieving strict convergence standard and fine mesh precision may require extremely long numerical calculation time. Therefore
a moderate numerical calculation time should be chosen to maximize the computational efficiency.