CHARACTERIZATION OF WAVE IMPACT SPRAY GENERATION FOR BETTER PREDICTION OF MARINE ICING ON STRUCTURES

The SPRAYICE project is now completed. Sea spray icing is one of the ongoing problems in offshore operations because it poses safety hazards and operational issues. In order to reduce risk improvements to our scientific knowledge are needed to develop models that can better predict ice loads with reasonable accuracy. Sea spray is formed in subfreezing air temperatures by the impact of waves on vessels and fixed structures which break-up into water droplets that are deposited on surfaces and freeze. The size and scale of these droplets span a wide range from centimetres to micrometres. Larger droplet sizes can be measured but methods to measure smaller droplets (micrometer range) are complex and expensive and therefore current models lack empirical data. The project was a partnership between the thermo-fluids research group at NTNU and the flow technology group at SINTEF industry. At NTNU the research focus was on the development of a new time-resolved volumetric experimental method to measure sprays in the micrometre regime using only a single coherent light source and to apply this method to measure droplet sizes and other statistics to a spray formed by water impacting on a surface. This is the first step to providing valuable empirical data on the size of sprays formed under a variety of conditions which are needed as an input to engineering models to predict ice accretion. The research at SINTEF has focused on computational tools to predict wave impacts and droplet formation. The main objectives of the project were achieved and several new experimental and numerical methods have been successfully developed. There were some minor deviations in the project due to the need to overcome unforeseen challenges as is the nature of research. However, these also gave rise to additional outputs and new methods such as a new, robust particle tracking method which provides Lagrangian statistics about sprays. At NTNU, the PhD student, Vipin Koothur, has developed a new particle tracking and a particle sizing method with the support of a post-doc in my group, Dr Melissa Kozul. The methods were developed using a novel approach of synthetic experiments. This approach makes use of publically available Direct Numerical Simulation (DNS) databases of turbulent flows. Particles/droplets are then added numerically which can then be used to quantify the accuracy and robustness of the methods. This new way to develop experimental methods enables identification of the key parameters of the real experiment which can be optimised in advance. The main difference between existing particle sizing methods which are point or planar measurements, is twofold: 1) we require only a single coherent light source, in this case a high-speed laser and 2) we obtain volumetric data which provides significantly more information about droplet statistics and dynamics. The method has been validated in experiments conducted in a turbulent flow with mono-dispersions and poly-dispersions of spherical particles. A data set of an impact spray have been acquired but processing and analysis will not be completed within the timetable of the PhD but will be ongoing and published through new Masters students and internal resources when available. At SINTEF, the PhD student Son Tung Dang has developed and validated a generic two-phase flow solver using a novel method which combines Cartesian cut-cells in combination with a staggered grid to ensure mass and energy conservation properties which are problematic with alternative methods in multi-phase flows. A droplet break-up model was then successfully implemented into the software SIMCOFLOW which can simulate droplets produced as a result of wave impact against a solid structure. In the model, individual droplets are not resolved but are rather treated as Lagrangian droplets which are entrained from the resolved large scale interface. The entrained droplets physically drive mass transfer that erodes the liquid at the gas-liquid interface. The entrained droplet size is computed based on the instability modelling of liquid sheet break-up, and secondary drop break-up is accounted for as well. The droplets are treated as Lagrangian particles that affect the flow field through a two-way momentum coupling. Future work is to further develop the model to capture the freezing of these liquid droplets into ice, thus permitting the simulation of marine icing due to wave impact. In terms of output, two PhD students will submit their PhD theses by spring/summer 2020. New scientific knowledge gained by the project has led to several journal publications in high impact journals and numerous conference papers. At least 2 more journal papers are expected to be published in 2020. The particle tracking and sizing codes are in the process of being made publicly available on the wiki pages of NTNU. The simulation software is available at GitHub (https://github.com/lovfall/simcoflow).

A volumetric particle sizing method using a single coherent light source was developed using high-speed laser scanning, multi-camera set-up giving access to particle dynamics and statistics. The method will primarily impact research communities in fluid mechanics including, multi-phase flows, combustion and ocean science. Synthetic experiments were used to develop both a particle tracking method and particle sizing method and are a powerful way to optimise complex experimental techniques before implementation. The particle tracking method gives access to 3D Lagrangian statistics which are important in turbulent flows, particle/pollutant dispersion etc. Synthetic experiments are a powerful way to numerically test and optimise complex experimental techniques before implementation. The new numerical solver that was developed for three phases and will be implemented CFD tools that are widely used in the oil and gas, marine, automotive, aeronautical, renewable energy sectors.

Previously, we have been developing novel experimental techniques which enable detailed local information about spray droplets and the underlying physics. In parallel some of us have developed multiphase models and simulation tools for flow assurance in oil and gas transport lines. We further have applied multiphase flow modeling to predict icing potential on vessels and found that detail predictions are possible if the spray source due to wave impacts is known. The wave impact sprays are known as the major source of icing on structures in the Arctic, but cannot be predicted at the current stage. In this project we plan to develop new and quantitative understanding about the physics of droplet formation. By running well defined experiments in idealized geometries and measuring the details of the wave induce liquid jet, its break-up and resulting droplet flow, the physics will be revealed as function of the governing dimensionless numbers. The results from these experiments will be applied in the development of a multiphase computational method, which can predict the relevant physics on a sufficiently coarse numerical grid to allow reasonably fast 3D computations. Simulation methods for wave propagation and wave impact towards a structure in a wind field will be merged with the new wave induced spray breakup prediction, and the resulting model can be applied for predicting the complete droplet field due to a wave impact with a structure. The developed 3D model will be applied to develop two databases containing simulation data. The first database will extend the experimental data with simulated data and the result will be tried extracted in simplified models, using dimensionless number. The second database will be based on a realistic geometry and methods will be developed for fast extraction of the date for use in models which include the complete icing physics for a given construction or vessel.