基于CPT资料的土性参数随机场特性研究

通过CPT资料研究土性参数随机场特性是一种比较理想的方法。土性参数随机场的平稳性是后续研究的先决条件之一，可以通过Bartlett统计量进行检验，但采用不同的模型拒绝域相差比较大，这样给实际应用造成困难，新模型LGS由于适应范围广可以解决这一问题。从随机场的基本理论出发，介绍了新相关模型LGS的数值模拟方法。基于LGS相关模型，采用修正的Bartlett统计量对随机场平稳性进行检验，统一了拒绝标准。用两组数据进行了实例分析，通过编制的程序进行计算，结果表明，这种相关模型对不同形式的相关函数有较好的适应性，可以解决拒绝域不统一的问题。

Study of random field characteristics of soil parameters based on CPT measurements

Cone penetration tests (CPT) measurements are widely used in soil characterization. An important precondition for subsequent study is the stationarity of the soil random field, which can be tested by Bartlett statistics. But the rejection criteria is not accordant when using different autocorrelation models, which brings difficulty in practical using. The new model, LGS, can solve this problem because it can be adaptive with different autocorrelation functions. The simulation of the new autocorrelation model, LGS, is presented based on the random field theory. The stationarity of soil characteristics is tested using modified Bartlett statistics based on the autocorrelation model, LGS, which unifies the rejection criteria. Two examples are analyzed; the result calculated by the program shows that this new model is adaptive with different autocorrelation functions; and the problem of rejection criterion can be solved well.