Virtual fuel flow estimation in shipping: how to formulate and validate a model for time-dependent momentary fuel flow estimation by combining high-frequency GPS and meteorological data with low-frequency noon reports - a case study

Loading...
Thumbnail Image

URL

Journal Title

Journal ISSN

Volume Title

School of Business | Master's thesis

Date

2017

Major/Subject

Mcode

Degree programme

Information and Service Management (ISM)

Language

en

Pages

89

Series

Abstract

This thesis validates a novel approach for virtual fuel flow estimation on commercial vessels, such as container ships, tankers, and bulkers. A model is fitted onto high-frequency GPS and meteorological data, and onto low-frequency crew-generated noon reports. The research design is a quantitative case study and utilizes two months of real-life vessel data from a single containership. Based on publicly available Python packages, the validation study examines the benefits of the modelling approach presented by Antola, Solonen, and Staboulis (2017). Stakeholders in the marine shipping industry are under increasing economic, environmental, and regulatory pressure to improve the energy efficiency of commercial vessels. A growing number of 3rd-party providers has entered the market and offers data- driven solutions for vessel performance monitoring and optimization. Most of them require tight integration to onboard systems to collect the necessary data. In a study on barriers to energy efficiency in shipping, Rehmatulla and Smith (2015) found that principal-agent- problems prevent a widespread diffusion of energy-efficient measures, even if they are cost- effective. Especially for small and chartered vessels, costly integration-based systems remain unfeasible. Consequently, a lightweight system is needed that can monitor vessel and crew performance, without being subject to the barriers to energy efficiency. Antola, Solonen, and Staboulis (2017) proposed a novel approach that uses virtual sensing techniques to model some of the quantities that require costly onboard integration in traditional solutions: vessel speed-through-water and instantaneous fuel flow rates. My validation study confirms that the proposed modelling approach provides reasonably accurate speed and fuel flow estimates that deviate on average by ca. 10% from the reference data. However, the modelled output is noisier and can be affected by prolonged periods of erroneous crew reporting. Yet, model-based speed-fuel-curves allow stakeholders to demonstrate a vessel’s operational profile for various speed and draft levels. The theoretical main contribution of my study is a transparent validation that highlights the benefits of the novel modeling technique by Antola, Solonen, and Staboulis (2017). For most customers, the alternative to the (slightly inaccurate) lightweight system is not a more accurate integration-based solution, but a sole reliance on crew-generated noon reports. The managerial implication of this study is therefore that the lightweight system offers reasonable accuracy, significant customer value, and a way to reduce the information asymmetry between charterer and ship owner; at a fraction of the price of established integration-based solutions.

Description

Thesis advisor

Malo, Pekka
Liesiö, Juuso

Keywords

virtual sensor, noon report, commercial shipping, marine shipping, vessel performance monitoring, ship, performance monitoring, modeling

Other note

Citation