Machine learning based operational services saves VIKING time for sea trials test

Deployment and testing of the Viking LifeCraft™ during the mandatory sea trial tests, in which significant wave height is larger than 3 m. © VIKING Life-Saving Equipment A/S

Sea trial challenge

Life rafts sea trial is an important test before passenger and cruise ships can use them. Testing these life rafts are large and expensive operations. They must be tested at sea and with significant wave height greater than 3 m. One of the challenges is to determine the location of the best test site ahead of time.

Sea conditions are best suited for the test around the Shetlands islands in the Northern Atlantic but it is still a large area. Moreover, the actual conditions are often hard to predict and changes rapidly. Finding the most probable test site is normally a mixture of guesswork, experienced seafaring and comparing multiple forecasts. Doing this whilst on board the vessel at sea makes this job harder. One wrong move in the planning of sea route means loss of precious sea time resulting in hefty, additional operational costs.

Embracing innovation

To improve operations, VIKING Life-Saving Equipment A/S teamed up with DHI to pilot test a new generation of operational services. Code named Wave Hunter; the web service finds the most optimal location to conduct these sea trials. Wave Hunter utilizes new machine learning algorithms and 40 years of readily available high accuracy wave data in DHI’s Metocean Data Portal through an efficient API. It produces dynamic maps and wave forecasts of the most optimal test sites. The maps are forecast 3 days ahead and waves are forecast 5 days ahead. They are updated as frequently as  there’s updated weather data available and provides the basis for planning test site routes.

The test

On the 4th March 2020, this services was tested. VIKING scheduled a sea trial and the crew set sail. The test site selected was based on recommendations by Wave Hunter. In this position a significant wave height around 4.0 m was forecast in the morning hours. It was also predicted to gradually decrease to 3.0 m during the afternoon. The crew arrived on site in the morning and actual wave conditions were measured by an on-board wave meter. They recorded a significant wave height of 4.0 m. By the afternoon, the measurement showed 3.1-3.2 m, exactly what the Wave Hunter has forecast, and it was perfect sea trial conditions.

The forecast location was spot-on and saved us time and effort while at sea. We are happy to continue to work with DHI to further improve the service and recommend this service to anyone who wants to improve their sea operations.

By combining accurate long term hindcast data from existing data bases (or new MIKE Powered by DHI models) and machine learning models, we can achieve lightning fast computation.  This enables computing of multiple forecasts in a short time, uncertainty quantification and increasing the robustness of the service.

Contact us to learn more about applied research at DHI and find out how to get involved as early adopters in our innovation work.

Share this post

DHI are the first people you should call when you have a tough challenge to solve in a water environment – be it a river, a reservoir, an ocean, a coastline, within a city or a factory. Our knowledge of water environments is second-to-none. It represents more than 50 years of dedicated research and real-life experience from more than 140 countries. We strive to make this knowledge globally accessible to clients and partners by channelling it through our local teams and unique software. Our world is water. So whether you need to save water, share it fairly, improve its quality, quantify its impact or manage its flow, we can help. Our knowledge, combined with our team’s expertise and the power of our technology, holds the key to unlocking the right solution.