New Artificial Intelligence Technology Improving Fuel Efficiency and Reducing CO2 Emissions of Ships through Use of Operational Big Data

摘要

Fuel cost and CO2 emissions in operating ships are major challenges for the maritime industry.A large marine transport company spends more than 2.6 billion U.S.dollars on fuel every year.In order to reduce fuel consumption as well as CO2 emissions from ships, it is crucial to be able to accurately calculate the impact of winds and waves on ship speed and fuel efficiency.Normally, existing ship performance estimation technologies rely on experiments with model ships in tanks of water, or on physics model simulations.However, they do not take into account the complicated interactions of winds, waves, and sea currents that influence the state of ships at real sea waters, resulting in large margins of error.Against this background, Fujitsu Laboratories has developed a technology to visualize ship performance.In addition to weather and sea conditions including winds, waves, and sea currents during actual ship operation, it collects engine log data and ship operation data such as location and ship speed, and posts it to the cloud, and then analyzes these data using high-dimensional statistical analysis that we developed.Applying this technology to a university-owned test ship and some merchant ships resulted in highly accurate estimation of ship speed and fuel consumption for each of the ships, with error of 5 or less.We also evaluated this technology through simulation and verified that it can improve fuel efficiency significantly.This paper describes this technology to predict ship performance in real sea waters, which is key to reducing ship fuel consumption, with some examples of system configurations.