Performance Monitoring Insight using Predictive Analytics: A Step towards IMO’s GHG Emission Goals 2030

The International Maritime Organization (IMO) and Marine Environment Protection Committee (MEPC) regulates the GHG Emissions from ships through operational and technical measures. According to the survey, thousands of owners will be affected which may initially bring inflation in marine transportation economics. Many people believe that installing Energy Saving Devices could help the ships perform. On the contrary, many others are limiting the engine power, fuel consumption and emissions from highly complex network of sensors as well. The present study could benefit the ship owners who would wish to adapt a less complex methodology using Machine Learning which utilizes ship performance database of several voyages in order to do prediction of Fuel oil consumption in derated power in collaborating with the forecasted weather data. The study proposes a method to estimate the Power consumption of ships by predicting the ship operational performance at various environmental conditions, and to provide Engine Derating measures to improve the EEXI. An oil tanker is selected as the candidate vessel and a predictive model is developed with the vessel’s Noon Report data. The predictive variables considered are engine brake power, engine RPM, draft, trim, speed over ground, speed through water and environmental factors. The machine learning model is trained over the analysed data and relations between parameters are established. Data is categorized into test and train datasets and a comparative analysis between combination of various multiple regression models, classification models and feature engineered datasets was conducted. The effectiveness of the models is measured based on a number of key performance indicators. The developed model can predict the power consumption of the vessel with excess of 95% accuracy. Moreover, a preliminary version of mobile application (speed log app) is also developed to digitally enter the noon data on board the ship.