card in standard column format and can be used later for retrieval and
easy import into conventional processing and plotting software like
Excel.
• The designed set-up can be used for both single ring falling head and
double ring constant head DRI by simply adopting the Arduino coding
for each method.
• Since we measure the flow of water that passes through the pump,
unlike other methods, a single reservoir can supply the water for
both inner and outer rings.
• Automated pump shut downs when no water needs to be added to
the inner ring extend battery life well enough to survive for the test
duration.
• Compared to the complication of adding additional water toMariotte
systems due to their requirement of being “sealed”,water can be easily
added to the current system.
• As the measurements of flow are done using the Hall effect sensor
attached to the pump, adding water to the reservoir does not affect
the calculations. To provide continuous supply of water to both rings,
water can be added to the reservoir, whenever its level drops, without
a need of correcting the measurements.
• Using the Hall effect sensor and counting the number of pump rotations
tomeasure the flowof water also has the benefit of constantmeasurements
in different temperatures. Unlike some other methods that
require precautions and calibrations for temperature or sunlight effect,
measurements of this set-up is not a factor of temperature or barometric
pressure.
• Using the Hall effect sensor also enables precise measurements of low
infiltration values. Since any rotation of pump corresponds to specific
volume of water, even small volumes can be detected and measured
at high precision by the sensors.
• The systemcan detect when the steady state infiltration has happened.
This way, the test automatically stops and concludes data recording.
• This system can easily be stored in a waterproof, ruggedized storage
container. This is very desirable for remote data logging in an outdoor
environment.
4. System testing and results
In order to check the accuracy of the automated system, three sets of
tests were conducted at different locationswith different soilmaterials.
For comparison purposes, at each test location, a manual and an automated
DRI testwere performed. In each location,we ran the automated
andmanual testswith enough of a separation distance to avoid interference
of thewet fronts. The DRI field experiments, as shown in Fig. 7, had
20 cm inner and 40 cm outer ring diameters with ring depths of 10 cm.
Themeasured steady state infiltration rates and initial infiltration values
for bothmethods were measured afterwards using the recorded results
of the flow. The values were compared for each set of tests. These data
are plotted in Table 2. The automatic test results showed no irregularity
in the infiltration from the pump response. After each experiment,
granulometric analyses were performed on the field soil samples
according to ASTM D 422-02 (ASTM, 2007).
The first testing location was the FSU Reservation area in Tallahassee,
Florida. The light brown sand of the test site could be described as
well-sorted, fine to coarse sand. It was categorized as SW based on
Unified, and A-3 in AASHTO soil classification systems. The test site
had a gently sloping surface (2–4°). A nearly constant infiltration rate
over the 40–60minute test runwas produced by relatively high permeability
of this soil. Graphs of cumulative infiltration rate versus time for
automated and manual tests performed in this location are provided in
Fig. 8. The second test locationwas in the vicinity of Tallahassee regional
airport area. The surface soil material included mottled dark and light
brown medium to fine sand that was classified as SP-SM based on
Unified and A-3 in AASHTO soil classification systems. The test site