Summary

Experimental data show, that the investigated FD-sensors cannot compete with the accuracies of the TRIME-TDR technique. Even with a specific material calibration the FD-sensors show a higher deviation than TRIME using the universal calibration! Using a material specific calibration the TRIME-sensor measures twice as accurate as the specifically calibrated FD-sensors.

The FD-sensors are also stronger influenced by the electrical conductivity of the pore water causing significant deviations at conductivities of more than 1 dS/m.

Moreover the FD-sensors have difficulties in determining higher water contents, which results in a significant underevaluation of the existing moisture. This is due to influences of electrical conductivity and the fact, that accuracy and resolution of FD sensors are best at high impedances (i.e. at low water contents) where the transmitted voltage is high and thus easy to measure. Generally applies: the higher the water content becomes, the lower is the accuracy of FD sensors.

The measurements of FD devices are sensitive to changes in the geometry of the probe rods and to mechanical stress. A slight bending already distorts the measurements, whereas the TRIME rods can even be bound together at the tip without influencing the measurement value. Another limitation in this context is the difficulty in constructing FD sensors with rod lengths of more than 15 cm which limits both the measurement volume and the application range of the sensor.

Finally there is a higher influence of electromagnetic perturbing radiation on the FD-sensors.

 

Introduction

TRIME time domain reflectometry (TDR) uses a voltage pulse signal (pulse rising times < 200 ps) for determining the dielectric constant, which contains a broad frequency range from approximately 1 MHz to 1 GHz.

Beside TDR capacitive devices exist which are also dielectric moisture measuring systems. They are named „capacitive" because this method is based on measuring the capacity between the two electrodes of the sensor. Capacitive systems only use one single measuring frequency. The method is also called frequency domain (FD) technique for this reason.

The broad frequency range (1 MHz to 1 GHz) is the big advantage of TDR, because - despite its name - the dielectric constant is not a constant at all. In an alternating electric field, it depends mainly on the applied frequency. This frequency dependence can be described by the complex dielectric constant e *:

[1]

The real part e ‘ denotes the dielectric constant, from which the water content can be deduced. The assumed part denotes the dielectric loss, consisting of the dielectric absorption e ‘’_d and the bulk soil electrical conductivity s _dc (w is the angular frequency and e ₀ the permitivity of vacuum). e ‘’_d is a measure for the energy absorption and increases with increasing frequency, whereas at lower frequencies the assumed part is dominated by the bulk electrical conductivity s _dc .

 

TRIME-TDR

Because of the relatively high frequencies of a TDR-pulse, the real part of e * is a function of the water content in porous materials. The assumed part can be neglected, as the frequencies are either so high that the ionic conductivity does not have a significant influence or they are so low, so that the energy absorption does not play an important role. Due to a coating of the metallic probe rods, the TRIME-system has the advantage, that it can be used for water content measurements in high conductive media. With a thicker coating, as for example for the TRIME-IC, measurements in materials with very high electrical conductivities (> 15 dS/m) are possible.

 

FD-sensors

Due to electrical reasons the capacitive sensors can only achieve a single frequency with a maximum of about 100 MHz. Mostly they are far below that (for example: Hydra probe Vitel = 50 MHz; TFDL-DLO-sensor = 20 MHz; Delta T = 100MHz). But the lower the frequency, the higher the influence of the ionic conductivity so that the assumed part of the dielectric constant can no longer be neglected and distorts the water content measurement. That is the reason, why there is a greater dependence of FD sensors on the bulk electrical conductivity or salt content and - consequently - on the soil type. Also Dirksen & Hilhorst (1994) found that the sensitivity of FD devices against soil types present a greater need for calibration and that at lower FD frequencies effects occur which are rather difficult to interpret.

 

Measurement results

A TRIME P2 two-rod-probe and a FD-sensor with four rods were tested in different materials.

Fig.1 shows a comparison of water content measurements in glass beads. The beads were gradually mixed until saturation with different pore water solutions ranging from distilled water up to solutions with an electrical conductivity of 3 dS/m.



Fig.1: Comparison of moisture measurements with TRIME (universal calibration) and FD sensor* (specific calibration) in glass beads and pore water of various electrical conductivity.
Line indicates 1:1 correlation. The reference is thermogravimetrically determined.

The measurements were both carried out with the commercially proposed universal calibration functions and with a specially determined calibration for glass beads.

The root mean square deviation (rmsd) was determined according to the following equation:

[2]

The universal calibrations gave a rmsd of 3.5 % by vol. for the FD sensor and of 2.3 % by vol. for the TRIME-P2 probe. With the special calibration for the glass beads, the accuracy could be increased, yielding a rmsd of 2.7 % by vol. for the FD sensor and of 1.2 % by vol. for the TRIME probe.

In a second experiment the influence of electrical conductivity on the measurement was tested. The glass beads were saturated with solutions of different electrical conductivity ranging from tap water (0.5 dS/m) up to 5 ds/m. Fig. 2 shows the results.

Whereas the TRIME probe shows no influence on the pore water electrical conductivity except a slightly higher value at 5 dS/m (3 % by vol.), the measured water content of the FD-sensor decreases significantly with increasing conductivity, resulting in a drop of 6 % by vol. at 5 dS/m.

 

 



Fig.2: Comparison of TRIME and FD measurement values* for glass beads which were saturated with pore water solutions of different electrical conductivity.

 

 

Finally the two sensors were tested in clays with high water content. The following figure 1 shows a comparison between TRIME-TDR, FD and thermogravimetrically determined moisture contents of three different illitic clays. It can be seen, that the FD-sensor has difficulties in measuring high water contents showing an underestimation which can reach more than 15 % by volume. This may be due to the high electrical conductivity that corresponds with the high water contents in clayey materials.



Fig.3: Comparison of moisture measurements in illitic clays with high water contents using TRIME-TDR, FD-technique*, and oven drying method.

 

 

 

References:

DIRKSEN, C. & HILHORST, M.A. (1994): Calibration of a new frequency domain sensor for soil water content and bulk electrical conductivity.- Symp. on TDR in environmental, infrastructure, and mining applications, 07.-09.09.1994, Evanston, Ill., U.S. Bureau of Mines Spec. Publ. SP 19-94: 143-153.

HYDRA soil moisture probe user’s manual version 1.2

Delta T soil moisture probe user’s manual.

Further information:

E-mail info@imko.de

Web page www.imko.de