Corvus energy developed DNV– Accepted methods for data-driven state of health (SOH) testing.
SOH test marine battery system These are essential and provide a comprehensive assessment of battery condition, performance, and safety over time. These tests help determine the battery’s capacity, efficiency, and remaining life. Using a data-driven approach in these tests will improve the accuracy and efficiency of monitoring and maintaining marine battery systems. In addition, testing can be carried out almost without disruption to normal operations, significantly reducing vessel costs and off-hire time.
Lars Ole Valøen, EVP and CTO, Corvusexplains: “We realized very early on that if we could carry out annual SOH tests without taking the ship out of service, we could significantly reduce costs and unwanted out-of-service for shipowners. Due to the complexity involved, it took almost five years to collect sufficient degradation data from the field and develop a robust data-driven SOH algorithm.
“This work was carried out by our team of experienced battery experts in collaboration with world-class research institutes and class society DNV. We use a large amount of data from our facilities, lab test data and powerful digital twins. The approach can be used to simplify test procedure requirements without compromising test accuracy. It also makes more frequent SOH testing a realistic option in the future, especially at the end of life of battery equipment. This will lead to improved safety and more predictable behavior.”
“This innovative approach, based on digital twin technology, represents a step forward in operational efficiency. We are grateful for the opportunity to provide our input and are pleased with the results. Fjord 1 is looking forward to seeing the positive impact this new methodology will have on our fleet.” Sondre Oostreim, Discipline Lead Electro for the Norwegian ferry company Fjord 1.
Going forward, Corvus aims to further develop its data-driven approach and provide its customers with up-to-date SOH validation at any time upon request, based purely on real data from normal operations.


