Automation and Standardization of Measuring Moisture Content and Density of Soil Using the Technique of Time Domain Reflectometry

Abstract

The methodology developed by Siddiqui and Drnevich (1995) for measuring soil water content and density using Time Domain Reflectometry (TDR) was extended for routine use in the quality control testing of compacted soils. The objectives of the study were to develop computer software to automate data interpretation and data reduction, develop prototype equipment for field use, compare the results of the TDR method with the sand cone and nuclear density methods on actual construction sites and develop draft specifications for ASTM and AASHTO. New prototype equipment was developed for compaction quality control testing. Waveform interpretation and data reduction were automated by incorporating developed algorithms into a WindowsTM based computer program that is used on a palm top computer connected to the TDR apparatus. Over 150 laboratory and field tests were performed to evaluate the TDR equipment and procedure for use in the determination of moisture content and density of compacted soil.Under field conditions it was shown that the TDR method is more accurate than the nuclear density gage in estimating water content when compared to oven drying, with standard errors of 1.1% and 1.8%, respectively. The ability to assess the accuracy of the TDR method for determining density was limited because the true density of the compacted soil is not known. The sand cone test was used as a basis for comparing the TDR and nuclear density tests on several sites while accepting that the sand cone test is not an exact method. The time required to perform the TDR test under field conditions is approximately 15 to 20 minutes. The sand cone test can also be performed in approximately 15 minutes, while the nuclear density test can be performed in approximately two minutes. The time required to complete the TDR test relative to the nuclear density test, is viewed as its greatest limitation. In general, the nuclear gage provides estimates of density and water content much more rapidly than the sand cone or TDR tests, is less accurate than the TDR test in estimating water content, and must be calibrated for a specific soil using sand cone tests. Moreover, the nuclear gage uses a hazardous source requiring operators to take safety training and leading to expenses associated with equipment maintenance and disposal. The sand cone test is time consuming, and can not provide estimates of water content. The TDR test is also time consuming, but does provide accurate estimates