A power model (ε = A. N^B) to estimate permanent strain (ε) from single-stage repeated load triaxial (RLT) tests can simulate both shakedown and incremental collapse. When the exponent B < 1, the ε versus load cycle number (N) plot concaves downwards, resembling that of a material that shakes down. When B > 1, the same plot concaves upwards, mimicking that of a material that incrementally collapses. A power model that relates A and B to the shear stress ratio (SR = 1/factor of safety), a measure of how far the RLT specimen is away from static failure, is proposed. Use of SR as an independent variable in a PD model is advantageous because it embodies the geomaterial’s gradation, physical state, suction, stress level, shear strength and margin of safety against static failure. No other parameters bar N and SR are needed, enabling the number of independent variables and hence model complexity to be kept low. This model was applied to a virgin and a recycled concrete aggregate utilizing 2 independent variables (N and SR) with A and B both functions of SR. A rational procedure to estimate the mobilized shear stress is presented and a methodology to quantify SR that considers non-linearity of the Mohr-Coulomb failure envelope and cohesion due to suction in the unsaturated RLT specimens is proposed; such considerations have been absent in previous models. For the materials tested, A was found to decrease non-linearly with increasing SR while the relationship between B and SR is quite linear. The resulting model was able to simulate the RLT test data quite well with four fitting parameters. The fitting parameters can be easily derived using linear regression on a spreadsheet and were all statistically significant with a confidence level exceeding 95%. Once B is known, the shakedown limit can be estimated by setting B < 1. An appropriate margin of safety can be incorporated on the shakedown load to limit the permanent strain.

从单级重复载荷三轴 (RLT) 测试中估算永久应变 ( ε )的功率模型 ( ε  =  A . N ^B )可以模拟安定和增量倒塌。当指数B  < 1 时，ε与负载循环数 ( N ) 的关系曲线向下凹，类似于摇晃下来的材料。当B  > 1 时，相同的图向上凹，模拟逐渐坍塌的材料。将A和B与剪切应力比 ( SR = 1/安全系数），建议测量 RLT 试样远离静态失效的程度。在 PD 模型中使用SR作为自变量是有利的，因为它体现了岩土材料的级配、物理状态、吸力、应力水平、剪切强度和抗静态破坏的安全裕度。不需要其他参数 bar N和SR，从而使自变量的数量和模型复杂度保持在较低水平。该模型被应用到原始和再生混凝土骨料上，使用 2 个自变量（N和SR），A和B都是SR 的函数。. 提出了一种估计移动剪切应力的合理程序，并提出了一种量化SR的方法，该方法考虑了由于不饱和 RLT 试样中的吸力引起的 Mohr-Coulomb 破坏包络线和内聚力的非线性；以前的模型中没有这样的考虑。对于测试的材料，发现A随SR的增加呈非线性下降，而B和SR之间的关系非常线性。由此产生的模型能够很好地模拟 RLT 测试数据，具有四个拟合参数。拟合参数可以使用电子表格上的线性回归轻松导出，并且都具有统计显着性，置信水平超过 95%。一次B是已知的，可以通过设置B  < 1来估计安定极限。可以在安定载荷上加入适当的安全裕度以限制永久应变。