Jv
J.B. van der Hoorn
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
1 records found
1
Magnetic Field Data in Signal Matching Models
Feasibility Study of Using Magnetic Field Measurements to Estimate Soil resistance Properties Using Multiple Predictive Models
Master thesis
(2026)
-
J.B. van der Hoorn, P.C. Meijers, E. Kementzetzidis, E. Sulollari, Robert Hasselaar
Reliable soil resistance estimates are essential for the safe and efficient design and installation of offshore monopile foundations. These are currently obtained by feeding Pile Driving Analyser (PDA) data into a signal matching model, but installing a PDA offshore requires sensors to be bolted onto the pile and later removed, a slow and costly process. This thesis investigates a non-contact alternative: using the change in magnetic flux around a steel monopile during driving as a substitute for the PDA's force signal. CAPWAP was selected as the basis for the model, since it produces a full soil resistance distribution rather than a single capacity value. Several CAPWAP driving configurations were tested with magnetic flux, including the standard wave separation formulation and a "Pulse" approach, in which the flux signal drives the model only during the initial hammer impact, after which the model matches purely on velocity. The Pulse approach proved substantially more tolerant of the flux signal's deviation from true force than wave separation, which produced unreliable estimates when driven by flux. Using Pulse, both the radial and axial flux signals produced shaft resistance and soil resistance distribution estimates in close agreement with the force-driven Pulse model and the wave separation approach. The main remaining discrepancy is in toe resistance, attributed to the Pulse model's reliance on velocity alone, which is more sensitive to the errors present in the available velocity signal. Repeating each model's optimisation from many independent starting points further showed that the shaft and toe resistance, and most other parameters under the Pulse formulation, converge consistently to well defined, unique solutions. These results show that magnetic field data can be used to estimate soil resistance parameters to a meaningful extent, closely comparable to a conventional PDA-based model. The flux signal must still be scaled to a measured force peak, so the approach is not yet fully non-contact; removing this dependency, along with closing the toe resistance gap, depends primarily on obtaining more accurate velocity measurements. A preliminary comparison against Laser Doppler Vibrometer data showed a promising alignment between the measured and impedance-based velocity peak, though the velocity signal could not yet be fully validated beyond this initial peak. A genuinely non-contact alternative to conventional PDA-based pile driving analysis therefore appears within reach.
...
Reliable soil resistance estimates are essential for the safe and efficient design and installation of offshore monopile foundations. These are currently obtained by feeding Pile Driving Analyser (PDA) data into a signal matching model, but installing a PDA offshore requires sensors to be bolted onto the pile and later removed, a slow and costly process. This thesis investigates a non-contact alternative: using the change in magnetic flux around a steel monopile during driving as a substitute for the PDA's force signal. CAPWAP was selected as the basis for the model, since it produces a full soil resistance distribution rather than a single capacity value. Several CAPWAP driving configurations were tested with magnetic flux, including the standard wave separation formulation and a "Pulse" approach, in which the flux signal drives the model only during the initial hammer impact, after which the model matches purely on velocity. The Pulse approach proved substantially more tolerant of the flux signal's deviation from true force than wave separation, which produced unreliable estimates when driven by flux. Using Pulse, both the radial and axial flux signals produced shaft resistance and soil resistance distribution estimates in close agreement with the force-driven Pulse model and the wave separation approach. The main remaining discrepancy is in toe resistance, attributed to the Pulse model's reliance on velocity alone, which is more sensitive to the errors present in the available velocity signal. Repeating each model's optimisation from many independent starting points further showed that the shaft and toe resistance, and most other parameters under the Pulse formulation, converge consistently to well defined, unique solutions. These results show that magnetic field data can be used to estimate soil resistance parameters to a meaningful extent, closely comparable to a conventional PDA-based model. The flux signal must still be scaled to a measured force peak, so the approach is not yet fully non-contact; removing this dependency, along with closing the toe resistance gap, depends primarily on obtaining more accurate velocity measurements. A preliminary comparison against Laser Doppler Vibrometer data showed a promising alignment between the measured and impedance-based velocity peak, though the velocity signal could not yet be fully validated beyond this initial peak. A genuinely non-contact alternative to conventional PDA-based pile driving analysis therefore appears within reach.