Statistical Sample Size for Construction of Soil Liners

Abstract

A method is described for selecting the number of samples (i.e., the sample size) to be collected and tested during construction quality control of compacted soil liners. The sample size is selected to ensure that enough data are collected so the probability of excessive equivalent hydraulic conductivity (i.e., overall hydraulic conductivity) is greater than or equal to a predefined maximum value is below a specified value. The method requires computations that can be performed using a spreadsheet program. Charts are provided to select the sample size based on these computations. The sample size depends on the properties of the soil, their spatial variability, and the number of lifts in the liner. Regression models are used to relate spatial variability of construction‐quality‐control measurements (such as compaction data, Atterberg limits, and particle‐size measurements) to variations in hydraulic conductivity at point scale. A three‐dimensional stochastic model is then used to estimate the equivalent hydraulic conductivity of the soil liner for statistical parameters describing spatial variability of point‐scale hydraulic conductivity. An asymptotic method is used to determine the precision of the estimate of equivalent hydraulic conductivity and the probability of excessive equivalent hydraulic conductivity.