Improved soil-structure interaction for offshore monopiles based on in-situ monitoring data

Onderzoeksoutput: PhD Thesis


In-situ measurements of the natural frequency of wind turbine structures with monopile founda-
tions have highlighted a mismatch between the as-designed and as-built natural frequencies of
these structures. This design inaccuracy can be mainly attributed to the lateral stiffness provided
by the subsoil. Accurate characterization of the lateral support provided by the soil is further
complicated by the spatial variation of soil properties and the non-linear and cyclic behavior of
soil when subject to loads of increasing amplitude. Recent research on monopile soil-structure
interaction has focused on onshore field testing and numerical modeling but there is a relative
paucity of back-analysis of the in-situ behavior of full-scale offshore wind turbine structures. In
this PhD research, in-situ monitoring data from two operational Belgian offshore wind farms is
used to investigate the applicability of the recently developed PISA guidance for monopile-soil
interaction to full-scale offshore wind turbine structures and to suggest improvements where nec-
essary. The sensitivity of the PISA method to the selection of geotechnical parameters is also
investigated. The research focuses on back-analysis of three monitoring data sources available
at Belgian offshore wind farms:
• Back-analysis of the bending moments in the monopiles under quasi-static thrust loading
based on strain measurements along the monopile;
• Back-analysis of the natural frequency of the offshore wind turbine structures during op-
eration based on farm-wide acceleration measurements and operational modal analysis;
• Assessment of the potential build-up of excess pore water pressures in the soil surrounding
the monopile based on measurements from earth pressure and pore pressure sensors.
These three tasks required an accurate geotechnical characterization of the ground conditions
at the monopile locations and efficient use of geotechnical and structural data in farm-wide
back-analysis. A database for managing the structural and geotechnical data for offshore wind
turbine foundations was developed to allow farm-wide back-analysis and sensitivity studies on
the geotechnical pile-soil interaction modeling.
As the small-strain stiffness of the subsoil is a governing parameter in the pile-soil interaction
models, the available methods for the selection of the soil’s small-strain stiffness were reviewed
and a novel data-driven approach for the derivation of stress-dependent stiffness was developed
based on in-situ shear wave velocity measurements in North Sea soils. This small-strain stiffness
was an essential input for the geotechnical back-analyses.
In the back-analysis workflow, location-specific structural and geotechnical data was extracted
from the database at each monitored location and bending moments or natural frequencies were
calculated and compared to the measurements. The sensitivity of the bending moment profiles
and the natural frequencies to various modeling choices was investigated. The method used
for calculation the pile-soil response curves had the greatest influence, with the API RP2 GEO
method leading to an overestimation of the bending moment and an underestimation of the
natural frequencies. The results highlights that the recent PISA design methods are able to
better match the in-situ monitoring data than the API RP2 GEO method. However, certain
areas for improvement still exist, notably in the effect of scour protection on the lateral stiffness
of the monopile soil-structure interaction. Modeling the effects of scour protection on the stress
and stiffness in the subsoil can further improve the match between the back-analysis and the
The choice of small-strain stiffness Gmax used for the pile-soil interaction calculations also has
an influence but the impact of using different CPT-based correlations for small-strain stiffness is
more limited. The location and magnitude of the maximum bending moment under quasi-static
thrust loading are almost identical for all CPT-based correlations. The difference in the first
natural frequency is 2.5% on average between the correlation providing the softest and stiffest
Gmax profile.
The pore pressure data shows that build-up of excess pore pressure is not observed for the
monitored location during the various loading regimes (parked conditions, power production at
rated power, storms). This is due to the permeable soil in which the pore pressure sensors are
embedded. Any excess pore pressures developed during the loading part of a load cycle appear to
fully dissipated during the unloading part of the cycle. Assuming fully undrained conditions for
cyclic degradation analysis during storms is therefore unrealistic for the instrumented location.
Originele taal-2English
Toekennende instantie
  • Vrije Universiteit Brussel
  • Devriendt, Christof, Promotor
  • Weijtjens, Wout, Co-Promotor
Datum van toekenning28 jun 2023
StatusPublished - 2023


Duik in de onderzoeksthema's van 'Improved soil-structure interaction for offshore monopiles based on in-situ monitoring data'. Samen vormen ze een unieke vingerafdruk.

Citeer dit