Decision support

for installation of offshore wind turbines

Prepared by:
Yngve Heggelund
with contributions from
Birgitte Furevik, Sigrid Ringdalen Vatne, Angus Graham,
Idar Barstad, John Dalsgaard Srensen, Joachim Reuder,
Rune Yttervik
Slide 1 / 10-Sep-13

Motivating problem

The cost of installing offshore wind turbines

must be distinctly reduced

Waiting for weather windows is a significant

cost contributor

Criteria to commence installation operations

are related to simple parameters

The physical limitation are however related

to response parameters

Significant wave height

Average wind velocity at reference
height

Motions
Accelerations
Forces

Uncertainties are currently not properly

taken into account in the decision making

Slide 2 / 10-Sep-13

General project idea

Couple weather forecast models to an advanced

dynamical model (SIMO) to obtain response
parameters

Local weather
measurements

EPS

Improve local weather forecasts by utilizing local

measurements

Calibrated
weather models
with uncertainty

Calibrate forecast models

Provide estimates of uncertainty

Use statistical models to capture uncertainty of

response characteristics

Integrate the above into an online risk based

decision support system

Clear and informed view of the risks and

potential costs of carrying out an operation in
a given timeframe

Slide 3 / 10-Sep-13

SIMO

Statistical
calibration

Models of
operational
phases

Statistical
models
Costs of
failed
operations

Decision
support system

Rational
limits for
responses

Research proposal
Title: Decision support for installation of offshore wind turbines
Research partners: CMR, met.no, Uni Research, UiB, AAU,
Marintek, UiS, UiA.
Industry partner: Statoil.
Associated partners: Reinertsen Engineering, Fred. Olsen
Windcarrier.
Proposal for competence building project was submitted to the
MAROFF program in the Research Council of Norway September
5th 2012.
Total budget: 8.4 MNOK over 3 years (80% by RCN, 20% by Statoil).
Project management by CMR.
Consortium agreement signed August 15th 2013.

Slide 4 / 10-Sep-13

Project overview

Slide 5 / 10-Sep-13

Installation test case 1

Integrated installation of offshore
wind turbines of gravity-base
type

Reduce installation cost by reducing

offshore heavy-lifting activities
Complete, or partly complete
structure transported to site
(integrated installation operation)
Operating phases:

Slide 6 / 10-Sep-13

Tow out
Mooring and positioning on site
Lowering of foundation to sea-floor
Setting foundation down into sea floor

Installation test case 2

Installation of wind turbine
rotor by floating crane vessel
Installation of one piece at a time
on site
Operating phases:
Transportation of rotor to site
Mooring and positioning on site
Lifting the rotor from the deck of
the transportation vessel
Placing the rotor onto the preinstalled nacelle

Slide 7 / 10-Sep-13

WP 1: Responses and requirements

given met-ocean conditions Marintek & Statoil
Couple a simulation tool (SIMO) to weather forecast models
Transform forecasted environmental parameters to input parameters
for SIMO

Description of critical responses for selected test cases

Describe the operating phases of the test case operations
Establish simulation models
Identify durations, dependencies
and point of no return
Identify critical response
parameters and their value

Slide 8 / 10-Sep-13

SIMO: Equipment response simulator

SIMO (Simulation of Marine Operations) developed and
owned by Marintek

Non-linear time domain simulation of motions and station

keeping of multi-body systems
Used in the oil and gas
industry:
Offshore crane operations
Subsea installation
Jacket installation and removal

Slide 9 / 10-Sep-13

WP 2: Improved forecasting from

models and measurements Universitetet i Bergen
Collect and organize observations
Existing observations and tailored measurement campaigns
Met-ocean instrumentation from NORCOWE and the NORCOWENOWITECH infrastructure projects EFOWI and NOWERI will be
available
2 buoy-mounted atmospheric
turbulence measurement
systems
1 oceanic turbulence
measurement system
1 scanning wind lidar
1-2 met-ocean buoy systems
2 WindCube lidar
wind profilers
2 ZephIR lidar wind profilers
1 scintillometer

Data will be available through the METAWIND portal

Slide 10 / 10-Sep-13

WP 2: Improved forecasting from

models and measurements (cont.)
Ensemble prediction system (EPS)
Uncertainties in forecasts occur because

Uni Research & met.no

Wave height

there are uncertainties in the initial conditions

numerical models approximates the exact laws
of physics

An EPS runs the same model many times

with slightly perturbed initial conditions

The EPS of the European Centre for MediumRange Weather Forecasts (ECMWF) is
currently tuned to develop a realistic spread
only after 48 hours
Part of doctoral study to modify the scheme to
yield one fit for purpose (under NORCOWE base
funding)

Downscaling of weather forecasts to

capture the effect of coastal
topography and bathymetry
2.5 km for wind
0.25 km for waves and currents
Slide 11 / 10-Sep-13

Temperature (yr.no)

Precipitation (yr.no)

WP 2: Improved forecasting from

models and measurements (cont.)
Calibrated probabilistic forecasting

Met.no & Uni Research

Wave height

Use measurements and raw ensemble forecasts to develop a statistical

method to calibrate the probabilistic forecasting of wind, waves and
current (example from deterministic model shown below)
Quantify forecast skill of using real time measurements for short term
forecasting (6-24 hours lead time)
Export weather forecasts with associated uncertainty in a format
suitable for SIMO

Wave
model
prognoses
Observations

January 2009

Slide 12 / 10-Sep-13

Original forecast

Calibrated
forecast with
uncertainty

WP 2: Improved forecasting from

models and measurements (cont.)
Aalborg universitet

Statistical models to capture the uncertainty of leading

response characteristics based on the uncertainty of wind,
waves and currents
Use the computer simulation tool (SIMO) to make models of the
response characteristics as a function of the geophysical variables
Develop methods to estimate the probability of exceeding critical
levels as a function of time

Slide 13 / 10-Sep-13

WP 3: Decision support system for

Christian Michelsen Research
operation planning
Map visualization
Of weather variables (with uncertainty)
Of response characteristics (with uncertainty?)

Plan and optimize the transportation route

Slide 14 / 10-Sep-13

WP 3: Decision support system for

operation planning (cont.)

Christian Michelsen Research

Compute and visualize

below critical time intervals
for operational phases
User defined probability of
being below a critical level

User evaluation of
presentation and interaction
Establish a representative user
group of potential end-users
Task the user group with
testing, and collect feedback
Compare existing methods to
the proposed method

Slide 15 / 10-Sep-13

Challenges
Cross discipline project between institutions
with little or no prior project cooperation
Do we speak the same language? Do we understand
each other?

Choice of project test case site

ECMWF ensembles are not stored in full in the
archives, making it difficult to use a historical test
case like Sheringham Shoal
Dudgeon will probably not be scheduled until after
project completion (virtual test case?)
FINO3?

Slide 16 / 10-Sep-13

Summary
Provide an objective foundation for decision support taking
into account
The real physical limitations of the equipment being used
The uncertainties in the weather-dependent data

Challenge existing practice of using simple parameters such

as significant wave height and average wind velocity
Enable evaluating different installation procedures

Ideas and principles can also be applied to the operational

phase
Main goal: Reduce the cost of installing offshore wind
turbines
Slide 17 / 10-Sep-13

Thank you for your attention!

Slide 18 / 10-Sep-13