Prediction of on-board energy usage combining physics-based modeling and machine learning

Mikael Manngård, Joachim Hammarström, Wilhelm Gustafsson, Jari Böling, Jerker Björkqvist

Research output: Chapter in Book/Conference proceedingConference contributionScientificpeer-review

Abstract

The global goal of environmental sustainability drives the maritime industry to reduce its emissions. Emissions are mainly caused by energy production, hence making both production and consumption of energy more efficient is essential. To use energy more efficiently, we need to fully understand how energy is being utilized. In this paper, machine learning is used, together with physical modeling, to both analyze and predict energy streams onboard a cruise ship. A physics-based model for the cooling and waste-heat management system on a ship is presented. To learn typical energy usage patterns, machine learning is performed on one-month operation data from a mid-size cruise ship. The combination of machine learning with physical modeling is used to predict energy usage and flows for the next 24 hours. The objective is to give officers onboard a ship the possibility to predict how their actions alter the total energy usage and efficiency.
Original languageEnglish
Title of host publicationProceedings of the 3rd International Conference on Modelling and Optimisation of Ship Energy Systems (MOSES2021)
Place of PublicationEspoo
PublisherAalto-yliopisto
Pages46-52
ISBN (Electronic)978-952-64-0555-1
ISBN (Print)978-952-64-0554-4
Publication statusPublished - 2021
MoE publication typeA4 Article in a conference publication
EventInternational Conference on Modelling and Optimisation of Ship Energy Systems -
Duration: 19 May 202120 May 2021

Conference

ConferenceInternational Conference on Modelling and Optimisation of Ship Energy Systems
Abbreviated titleMOSES
Period19/05/2120/05/21

Keywords

  • Forecasting
  • Ship energy systems
  • Physics-based modeling
  • Machine learning

Fingerprint

Dive into the research topics of 'Prediction of on-board energy usage combining physics-based modeling and machine learning'. Together they form a unique fingerprint.

Cite this