Numerical model validation and benchmarking

Home Forums Session 109: ARCTIC VI: Ice Loads Numerical model validation and benchmarking

  • This topic has 2 replies, 1 voice, and was last updated 6 months ago by Abolfazl Shojaei Barjouei.
Viewing 2 reply threads
  • Author
    • #2470 Reply
      Ekaterina Kim

      There are many numerical techniques to model ice loads and ice-structure interaction problems. How should our ice research/engineering community proceed w.r.t to the model benchmarking and validation?

      Should the ice load models demonstrate convergence? should there be a minimum set of tests that a model has to pass? or how and what empirical data to use for model validation?
      There are standards for conducting ice uniaxial compressive tests but no standards or recommendations for conducting numerical simulations of ice loads. Is there an opinion on how (if at all) should our community proceed further with this?

    • #2491 Reply

      Dear Ms. Kim,

      Could you please specify which paper and author you are referencing with your question? Thank you.

      Claude Chung

    • #2492 Reply
      Abolfazl Shojaei Barjouei

      Dear Ms. Kim,

      As you mentioned there are many ice simulation techniques in the literature. In our study, we utilized the recently developed model, MINCOG, that is mainly based on vessel and wave interactions rather than the ice from the atmosphere, since the developers of the model, Eirik Mikal Samuelsen et al., suggest that this (i.e., vessel and wave interaction) is the main water source in vessel-icing events. Consequently, the model provides higher verification than the other models, particularly using the data obtained in Arctic-Norwegian waters. Accordingly, the convergence to the observation from a large coast-guard vessel type, the KV Nordkapp class, was used for verification of the model. Further details about the model are available at
      In our study, we focused on the input parameters of the MINCOG model (e.g., wind speed, air temperature, atmospheric pressure, humidity, etc.) from a statistical viewpoint to enhance the outcome of the model. Moreover, we tried to estimate the parameters by emphasizing the recently sampled data in the last 5 years from the old, perhaps, uneven data from the previous 27 years. Therefore, ship observations, NOrwegian ReAnalysis 10km data (NORA10), were considered to verify the results. Furthermore, a test of hypothesis was conducted using the Anderson-Darling test at the 5% significance level, where the null hypothesis (i.e., H0) is that the parameter is from a population with a normal distribution, against the alternative hypothesis (i.e., H1) that the parameter is not from a population with a normal distribution. Accordingly, and based on the data over 27 years from 1980 to 2006, the null hypothesis could not be rejected in the majority of the days.

      I hope the above is useful to you. However, should you need any further questions, please do not hesitate to contact me.

      Best regards,
      Abolfazl Shojaei Barjouei

Viewing 2 reply threads
Reply To: Numerical model validation and benchmarking
Your information: