Technical information

3-23
Early synthetic work by Taylor (1945) and Monfore (1950) have shown that
measurement errors can be reduced by minimizing the thickness to diameter ratio of the
cell and making it as incompressible as possible. Based on experimental work, Peattie
and Sparrow (1954) have shown that if these criteria are fulfilled then the measurement
error relative to the ‘true’ stress level (in percent) equals 0.6 times the thickness to
diameter ratio. Geokon earth pressure cells are classified as hydraulic type gauges; by
design they have a thickness to diameter ratio of 0.026 (=6/230 see Figure 2.4.2) and are
relatively incompressible. According to Peattie and Sparrow (1954), they are expected
to record pressures that are higher than the ‘true’ stress levels by about 1.6%. This is a
relatively low error level for a geotechnical application.
More recently calibration chambers have been proposed as means for pressure
cell calibration (Theroux et al., 2001; Labuz and Theroux, 2005). For pavement
applications a more feasible method of calibration would be to apply a known load to
the pavement at increasing distances from the gauge and calculate the resulting
‘volume’ of stresses. In theory this ‘volume’ should equal the applied load. This in situ
type of calibration is best done with a single tire since the shape of the stress trace is
symmetric and thus only a few measurement points need to be considered. In the
current study dual-tires were used which resulted in an asymmetric stress trace. Also,
the location of the tires relative to the gauge array was only measured in the
longitudinal direction and not laterally. Hence it was not possible to perform such
calibration herein. Similar reasoning precludes in situ calibration of the pressure cells
for the NCAT study. With respect to the strain gauges, analysis of near field strain
disturbance and resulting measurement errors could not be found in the literature.
Subsequently, the stress and strain gauge readings were used as-is without applying any
correction or calibration factors. In addition, it should be mentioned that both types of
gauges experience drift over time due to temperature sensitivity (see Tesarik et al.,
2006) and other reasons; this makes them suitable for capturing dynamic responses only
and not for monitoring permanent changes.
A plan showing the embedded instrumentation aimed at capturing mechanical
responses in test section n1 is provided in Figure 3.5.1. The loading centerline is
denoted in the figure by the Y-axis and the transverse direction by the X-axis. The