EDIS - Effective Decision Making via Intelligent Data Integration for Conceptual Ship Design

Det er store mengder data involvert i den totale verdikjeden av skipsdesign. Som en følge av de store datamengdene er det en stor utfordring å drive effektiv datainnsamling, og å håndtere disse dataene på en slik måte at det kan gi økt forståelse av prosessen.I tillegg til info fra skip og utstyr, består datene av et annet og bredere omfang som inkluderer interessentenes forventninger, subjektive stilistiske preferanser, eiers krav, tidligere konstruksjonsdata, forskrifter, leverandører, skipsverft, prøvetur- og driftsdata. Antallet variabler som er involvert gjør at databehandling og deling mellom avdelinger, faser og interessenter er en vanskelig oppgave. Vanligvis mistes viktig informasjon når slike data er konvertert til lysbilder og regneark. EDIS Prosjektet utviklet innsikt for en intelligent og forbedret dataintegrasjonsrammeverk som danner grunnlaget for et effektivt beslutningsstøtteverktøy for skipets designprosess. Prosjektet utviklet avanserte analytiske metoder for å kombinere data fra PLM-programvare til tekniske verktøy, samt undersøke potensialet for åpen og samarbeidende teknologi, for eksempel webbaserte applikasjoner og virtuell realitet. Publikasjoner om open modelling and simulation og datadrevet skipsdesign er også konkret levering.

Results on data integration prototypes, producing learning from a ship lifecycle, which included data repository, as well as models for finding useful patterns and relationships among design parameters and existing fleet real performance data. Impact is the industrial use of modern data mining techniques, such as big data, clustering concepts, web-based simulation, open and collaborative libraries, which were achieved and applied in this project.

Effective gathering, understanding and handling the large amount of data involved during the ship design value chain is a challenge. It consists of not only data from the ship and its equipment, but a wider scope that should include stakeholders' expectations, subjective stylistic preferences, owner's requirements, previous design data, regulations, suppliers, shipyards, sea trial and operational data. The number of variables involved makes the data handling and sharing between departments, phases and stakeholders a difficult task, usually losing important information when such data is converted to slides and spreadsheets rather than properly manipulated. Nowadays, it is usual to have each process of the whole project treated independently by different groups, in a way that the information generated in each process is not effectively related and spread between different value-chain actors. This affects both ends of the value chain, increasing the number of hours during design (and consequentially cost), leaving room for communication issues, as well as delivering a lower value robust product, where the final constructed ship can be optimized to an different operational profile than it was originally designed for, operating super- or sub-optimised during its lifespan. Being able to deal with these issues in an efficient exploration of existing and related information across the whole lifecycle is essential to ensure the most value robust ship design value chain. This project will develop a intelligent and improved data integration framework during the ship design process, aiming an effective mapping of data produced during ship design (model-based system), during all value chain phases (from conceptual design thorough operation and scrapping) that can be used as an effective decision support tool during early stages of design. Such framework will be based on modern data mining techniques, such as data-driven methods, web-based tools and artificial intelligence.