Calibration Techniques for Improved Moisture Content Measurements Using a Hybrid Nuclear-Electric Density Gauge

ABSTRACT Rapid and accurate measurement of soil moisture content in situ during earthwork construction is critical, ensuring soils are compacted to their optimum condition. Consequently, field control of compaction moisture content is common in most existing quality assurance / quality control (QA/QC) specifications. Within the United States, the nuclear density gauge (NDG) has become the device of choice for inferring the in situ moisture content of compacted soils in a rapid and reliable manner. Recently, a relatively new test device termed the “EGauge” has been developed as a potential alternative for the NDG, as it emits significantly lower levels of radiation, consequently making it safer to use relative to the NDG. However, recent research has shown that the EGauge’s moisture measurements made using the manufacturer’s standard calibration may be inaccurate for some soil types. When these moisture content readings are inaccurate for a soil of interest, ASTM D8153-22, Standard Test Method for Determination of Soil Water Contents Using a Dielectric Permittivity Probe, recommends performing a soil-specific calibration. This manuscript examines the use of three calibration methodologies that were used to develop soil-specific calibration models to enhance the EGauge’s ability to estimate moisture content in situ. More specifically, the calibration approach proposed by ASTM D8153-22, a simple multivariate regression calibration approach, and a more sophisticated calibration approach involving the machine learning technique multigene genetic programming were implemented in this study. Results from this study showed that all three calibration approaches enhanced the EGauge’s ability to estimate moisture content in situ by up to 80 % relative to the manufacturer’s standard calibration. More importantly, the results from this study indicate that the EGauge may be a viable replacement for the NDG as a QA/QC tool to monitor the soil compaction process, with the caveat that efforts need to be made to independently develop a soil-specific calibration model for soil moisture content estimation purposes.