Hyppää pääsisältöön

Hari Nagarajan: New machine learning methods for co-modelling design and manufacturing

Tampereen yliopisto
PaikkakuntaKorkeakoulunkatu 8, Tampere
Hervannan kampus, Festia, Pieni sali 1 ja etäyhteys
1.12.2022 6.30–9.30
Kielienglanti
PääsymaksuMaksuton tapahtuma
Ihmishahmo tohtorinhattu päässään, musta siluetti violetin kuultamalla taustalla.
In his doctoral dissertation, Hari Nagarajan studies methods for fast development of simulation models by encoding pre-existing knowledge about products and manufacturing processes. The method supports conjoint modelling of multidisciplinary aspects in design and manufacturing. This new framework introduces tools for knowledge integration using graph modelling for simulation and decision support in manufacturing.

Understanding how a product and manufacturing process function together is important to support efficient and cost-effective production. Current process modelling approaches are highly dependent on data for developing models with high level of detail for their use in decision making.  Different models may also be difficult to combine and evaluate together as a whole.

Hari Nagarajan’s research explores methods to develop graph models of manufacturing processes which support simulation through machine learning. The research aims to provide engineers the capability to visualize linkages between mechanism and phenomena using graph representation with varying levels of detail. The approach combines different types of models, knowledge from experts as well as experimental data.

“I have always found it easier to understand how something works when I can visualize it. I wanted to bring this aspect forward to improve understanding of advanced manufacturing from an engineering context using graphs,” Nagarajan says. 

In his study, Nagarajan proposes several tools for modelling and simulation of additive manufacturing processes. He sees developing integrated modelling approaches as a potential means to move additive manufacturing technology forward in industrial settings.

“The developed framework is a tool that serves that purpose and provides the necessary attributes to integrate several engineering domains and concepts under a single approach. My method is generic and can have a vast potential for application in various fields and domains like systems design, smart manufacturing, artificial intelligence, and cybersecurity,” he adds.   

The doctoral dissertation of Hari Nagarajan in the field of Mechanical Engineering titled Development of a graph-based metamodelling framework for additive manufacturing and its simulation using machine learning will be publicly examined in the Faculty of Engineering and Natural Sciences at Tampere University on Thursday 1 December 2022 at 8.30 at pieni Sali 1 of the Festia building in Tampere University, Hervanta Campus. The Opponents will be Professor Ola Isaksson from Chalmers University, Sweden and Associate Professor Tom Vaneker from University of Twente, Netherlands. The Custos will be Professor Eric Coatanéa from the Faculty of Engineering and Natural Sciences.

The doctoral dissertation is available online.