Bounding surface plasticity has been widely used for capturing the stress–strain behaviour of geomaterials. However, it may require multiple sets of model parameters for constitutive modelling of sands with a wide range of initial states, because of the distinct critical state characteristics under low and high densities or pressures in the \(e-\ln {p}'\) plane. In this study, an enhanced bounding surface plasticity approach for sand with a wide range of initial material states is developed. A fractional plastic flow rule and a modified critical state line are suggested, which ensures that without using any predefined state indices, the developed model can consider the state-dependent dilatancy and hardening behaviours of sand subjected to low and high pressures/densities. The approach is validated by simulating the well-documented test results of Toyoura sand and Sacramento River sand. For comparison, the original state-dependent dilatancy approach in Li and Dafalias (Géotechnique 50(4):449–460, 2000. https://doi.org/10.1680/geot.2000.50.4.449) is also adopted and implemented. It is found that the two approaches can reasonably capture the typical stress–strain behaviour, e.g. hardening/contraction, softening/dilation, liquefaction, quasi-steady state flow, and non-flow, of sands with different initial material states, by using a single set of model parameters. However, compared to the current work, Li and Dafalias (2000) model relied on a predefined state parameter, for capturing the state-dependent behaviour of sand under a wide range of initial states

有界表面可塑性已广泛用于捕获土工材料的应力-应变行为。但是，由于在\（e- \ ln {p}'\的低密度和高密度或低压力下）具有不同的临界状态特征，因此可能需要多组模型参数来对具有广泛初始状态的砂进行本构模型。）飞机。在这项研究中，开发了具有广泛的初始材料状态的增强型边界表面塑性方法。建议使用分形塑性流动规则和修改的临界状态线，以确保在不使用任何预定义状态指标的情况下，开发的模型可以考虑受低压和高压影响的砂的状态相关膨胀率和硬化行为。通过模拟有据可查的Toyoura沙和Sacramento River沙的测试结果来验证该方法。为了进行比较，还采用和实施了Li和Dafalias中原始的依赖于状态的扩张方法（Géotechnique50（4）：449-460，2000. https://doi.org/10.1680/geot.2000.50.4.449）。发现这两种方法可以合理地捕获典型的应力应变行为，例如 通过使用一组模型参数，对具有不同初始材料状态的砂进行硬化/收缩，软化/膨胀，液化，准稳态流动和非流动。然而，与当前的工作相比，Li and Dafalias（2000）模型依赖于预定义的状态参数，以捕获在各种初始状态下砂的状态相关行为。