If you are looking for information on how the weather selector works when building cases, please see this article.

The weather is the primary source of uncertainty in an offshore construction project and a significant restriction on maintenance activities. The goal of running a weather downtime assessment is often to provide input to the estimation of probability-weighted master schedules.

  1. The traditional approach to planning master schedules has been to estimate weather downtime for different operations per month and multiplying this workability percentile with the corresponding operation’s duration to get the weather adjusted duration.
  2. With simulation modelling, it is possible to estimate the probability-weighted master schedule including the variability of weather in a month directly in addition to estimating the weather downtime per month, with higher accuracy and precision.

Both of these approaches are described in the first two sections and their difference.

Estimating a master schedule with weather window statistics

  • Identify all weather windows per month with required length satisfying the weather limitations
  • Calculate the workability by dividing the number of weather windows by the number of possible weather windows
  • For X years there will be X workability estimates for a particular month, this is the sample population of workability for that month
  • The analyst can calculate any percentile for workability for a month using the sampled data
  • This can be repeated for every type of operation to get the same distribution of estimated weather downtime per month for each operation
  • The project planner can derive the P50 schedule by adding the P50 weather downtime to each operation in the schedule.


  • Easy to calculate
  • Creates a smooth schedule because the same weather downtime is assumed the whole month


  • Assumes that the whole month will have the same Pxx weather (what happens if you have P50 weather the whole month)

Estimating a master schedule with simulations in Shoreline Design

  • Create a simulation model of how the operations in the construction project will be carried out including component logistics, pre-assembly, installation, commissioning, etc.
  • Simulations are run in time domain where every individual operation is simulated using the weather time series (hindcast or measurement data)
  • For each of the operations that has a weather restriction and minimum weather window the operation is not started before there is a weather window to complete the activity
  • When a weather window is not found record weather downtime
  • Record all start and end times of each operation
  • A large number of simulating runs (50+) starting the same date but in different weather years are run to include variation in weather during the same months in different years
  • Weather in a month in a particular year varies greatly and you never have e.g. P50 weather the whole month, hence, using the time series in a simulation will ensure that variability and uncertainty is taken into account
  • Estimate the P50 schedule by calculating the P50 percentile value of the start and end time of the same operation across all simulation runs, and repeat for all operations in the project
  • Separately estimate weather downtime by using all the recorded weather downtime per month from all simulation runs to estimate the percentile value for workability based on actual measured downtime and not only weather windows present in the weather data


  • A more accurate estimate of the master schedule and weather downtime taking the variability of weather into account
  • Separating the estimation of the master schedule from the estimation of weather downtime


  • More tedious and complex to make precise and accurate simulation models
  • The master schedule is not “smooth” because you see the effect of the weather variability in a month in the probability-weighted master schedule, which can be unfamiliar to some

Difference between the approaches

The main difference between these two approaches is that the simulations estimate a master schedule and weather downtime including the variability of weather in a month and the interdependence of all the operations (example follows in the next section) and is more rugged, while the traditional approach has a smooth curve because no variability in weather the whole month is assumed. The chart below shows the difference between weather downtime estimated traditionally (orange) and with simulation (blue).

Weather in Shoreline

To properly understand the risk and its implication on performance, there are four different approaches to analyse the weather; disable, historical, increment start year and Markov. More information on the different approaches can be found below in the section:  The four ways to use weather time series in the simulation. If you click on the Simulate tab you will see the below: 

The way it looks in O&M Design:

The way it is shown in ConDesign: 

How Shoreline Design use weather in a simulation run

Why use weather time series in simulations? The Shoreline applications use intelligent simulation technology that simulates the actual sequence and timing of operations in the time domain. This is why our apps require weather input in the time domain instead of statistical input such as e.g. weather windows or weather downtime. Weather data time series consists of one or more weather parameters such as significant wave height and wind speed (mandatory in Shoreline), current speed, wave periods, tide and visibility to mention a few, as well as a timestamp. In Shoreline, any weather parameter is usable as long as a time series is available. Below is an example:

Based on these time series, it is possible to generate weather window statistics and subsequently traditional weather downtime statistics. For this the following definitions are important: 

  • A weather state consists of one or more weather parameters, e.g. the pair (wind speed and wave height) is a weather state. 
  • The weather window statistics is a histogram of weather window lengths, for a specified weather state. 
  • Shoreline calculates the weather distribution from one data point to another data point (so from one hour to the next hour) using the "linear distribution".

When generating weather window statistics, the weather windows for a weather state with 1 hour, 2 hours, 3 hours, … 30 hours, etc. lengths in a weather data time series is count. 

PLEASE NOTE:  If you want to learn how to upload a file, see these articles might be able to help Weather data file, Input Library.

The weather data time series is used by the weather window calculation algorithm to check if the weather criteria for an activity is met for the required weather window length to carry out that particular activity. See an example below

We use the wind profile power law to extrapolate the wind speed at any reference height from the case's weather data. Shoreline takes the wind speed at the closest available height in your dataset and applies the power law exponent to derive a value for any wind speed reference height you set in your case.

Example: If your weather data has wind speed data at 10-m and 100-m heights, and you set a wind speed reference height of 40m, we will extrapolate the speed at 40m from the 10-m height using the power law exponent.


Cconstruction project case with 100 wind turbine installations in the Construction Design app. For all these 100 wind turbine installations, the heavy-lift vessel will go through a work process to assemble all parts. This work process has three steps in this example: lift the tower in place, lift the nacelle in place and lastly the three blades. Each of these steps has a duration of 7 hours but require a 9-hour weather window. These three activities have different weather criteria, as you can see below.

The weather checking algorithm will run for each of these activities to check if they are allowed to start. What happens during the simulation is that the installation vessel will transit to the wind turbine location and jack up (with their corresponding weather criteria), for thereafter to check if the weather permits the Tower installation activity to start. It verifies that wind speed is less than 14 meters per second at a 100-meter height for the next 9 hours. If these criteria are met, the activity can start.

When the tower installation is done, a new weather window check for the nacelle installation starts. For this, a weather window of 9 hours with wind speed below 14 m/s at 100-meter height is necessary. At last, the same weather window for the blade installation with a higher wind speed limitation will be searched for.

It's possible that a delay of the tower installation due to the wind speed affects a delay of the nacelle installation because the next weather window for nacelle installation was missed. This is called a negative feedback loop in the simulation theory. This interdependence between operations and vessels would not be easily calculated with the statistical approach.

Estimating weather downtime with a simulation approach is more accurate as it includes the variability of weather in a month and does not assume it is constant like the traditional weather window statistics approach.

The four ways to use weather time series in the simulation

One of the final steps before simulating a case is to choose how to create/use the weather series. The weather data can be generated with four different approaches, disable, historical, increment start year and Markov. These are described in the table below. One method must be chosen.

Input labelMandatoryDescription
Disable-No weather restrictions will be considered in the simulation runs.
Historical-The weather time series added to the case will be used and will start from year 1 in the weather data set for all simulation runs. ( PLEASE NOTE:  In ConDesign that yields the same result in every run.)
Increment start year-The weather time series added to the case will be used and will start every run one year incremented from the previous. ( Example: 5 simulation runs --> run 1 will start from year 1 in the weather data set, run 2 will start from the same date in year 2 in the weather data set, etc...)
Markov-A statistical model is used to generate a synthetic weather time series based on the weather time series added to the case using a Markov model (see below for a thorough explanation)  


The simulator simply starts looking at the weather data corresponding to the start date of your simulation case. E.g. you imported weather data from 1.1.1980 to 31.12.2016, and your simulation case starts from Jan 1 of year 1, which corresponds to 1.1.1980. This way you can isolate more of the uncertainty that comes from your strategy, vessel spread, manning, etc. 

Increment start year

If the weather data is over 10 years, and the simulation is over 15 years, the first ten years will be simulated and the loop will lead to the last year and simulates the next 5 years (last year - 4th year). The loop will be like: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 (first 10 years) 10, 1, 2, 3, 4, 5, 6, 7, 8, 9 (next 10 years).


The Markov model calculates the probability of going from a weather state (e.g. 1.5 m wave height and 10 m/s wind) to all other weather states that have appeared after that one in your weather data set. Based on these probabilities, it can generate a new string of weather states by drawing randomly generated numbers between 0 and 1 for each new state and picking the next state with the closest probability. 

Creating synthetic weather data time series is very useful when your weather data time series are incomplete, for doing longer simulation runs than years in the imported weather data time series, or for a big number of runs. The accuracy of the newly generated time series improves with the amount of raw import data. There is a lot of literature out there on this method, more information is also available in this article

NOTE: in order to get the Markov weather model to work, it is required to add the "Month" column when uploading the weather file.

Using multiple weather series in the same simulation case

  • The logic of multiple weather series

    Using different weather data time series is useful when the weather differs to such extent from place to place within the operating area. You can include several weather data depending on the specific requirements that need to be taken into account.
    Multiple weather series can be used in the same simulation case, by applying weather specifically to any location of interest. You can add a distinct weather series to any group of wind turbines and also add weather series for any base or port in Design.
    PLEASE NOTE: In case of Markov method is used, only one weather data file can be applied to all locations in the case.

    Used weather time series must overlap in time

    The weather time period which will be available for the simulation is the overlapping part of all the weather series used in the simulation case. The resulting weather start date will be the latest of the start dates for the used weather series. Likewise, the end of the combined weather data will be the earliest of the end dates of the individual weather series. 
    This is because the simulation needs to be able to check the weather for all the locations at any time in the simulation, and if a weather series period for a location has ended, this will then not be possible.

    How to force the use of non-overlapping time series

    If you cannot obtain weather data that overlap in time, re-import some of the weather series and adjust the start *years* for the weather series such that they all overlap for at least one year. The simulation will not accept if the available weather time is less than a year or the file is bigger than 50MB. 
    PLEASE NOTE: the simulation results will not be accurate/realistic unless the weather data is specified with the correct time information when importing it so that the simulation uses weather data which is synchronized as in the real world. If you adjust the start date of the weather import, at least not change the date and month, only change the starting year if necessary, so that the seasonal reference in the weather data set is correct.