| Real-time knowledge of soil moisture content and its variability during earthwork construction operations could have tremendous impact on process control (i.e. fill placement, disking, compaction, etc.) and the resulting fill quality. A means of rapidly determining soil moisture content using an off-the-shelf microwave sensor (Hydronix Hydro-Mix VI) is described in this report. The sensor provides an analogue output of 4 to 20 mA and is scaled to report zero in air and 100 in water. The sensor is placed in contact with the soil and has a measure up to about 100 mm. The sampling rate is 25 Hz, but usually takes 2 to 3 seconds to stabilize. The operating temperature is 0 to 60°C.;The purpose of this phase of the study was to develop relations between the microwave value (MV) and gravimetric moisture content of the soil in the laboratory, although some field tests were also performed. Tests were performed using several different soil types at different compaction efforts and at a wide range of moisture contents on the wet and dry sides of "optimum" moisture contents. The MV values from the sensor are correlated with oven dry moisture contents. In short, low values of standard deviation, standard error and coefficient of variation in the microwave data indicate that the precision in the measurements is high. Microwave sensor proved to be a very useful instrument for fast and accurate soil moisture content determination. The findings are promising and warrant further evaluation and development.;Some of the key findings and observations from the study are as follows: (1) The standard laboratory mold dielectric is found to have a significant effect on the MVs and should not be used for laboratory calibration. (2) The MV value is sensitive to small changes in contact area of the sensor. The maximum allowable change in surface area of a specimen compacted on the wet of optimum is found to be 3cm2. (3) The height up to which the steel plate dielectric affects a microwave value of an extracted soil specimen resting on the plate is about 50 mm. (4) The suitability of the microwave sensor for five different soils, namely Edward Till, Kickapoo Clay, Kickapoo Topsoil and FA6 and CA6G were studied both at ISU laboratory and in the test beds at Caterpillar's soil mechanic lab. Regression analysis showed that R2 values from linear relationships ranged from 0.87 to 0.98. (5) Statistical models were developed based on soil type using the laboratory data. MV and MV2 terms proved to be the most significant parameters affecting the models---dry density and various soil index parameters were also considered and in some cases were significant. Using just the MV terms in the statistical analysis results in predication models can be improved. (6) Accuracy and precision tests on Edwards till samples compacted at -3%, 0%, and +2% of standard Proctor optimum moisture content produced standard deviations of 0.4 to 0.6%. The standard error of the mean was 0.06 to 0.08%. For Loess samples compacted at -3%, 0%, and +2% of standard Proctor optimum moisture content, the standard deviations varied from 0.2 and 0.3% and the standard errors are from 0.03 to 0.05. At a 95% confidence interval the predictions are within +/-1%, which meeting the target established for this research.;The low values of standard deviation, standard error and coefficient of variation in the microwave data indicate that the precision in the measurements is high. Microwave sensor proved to be a very useful instrument for fast and accurate soil moisture content determination. |
