Participants of the fourth research workshop in Copenhagen in February 2018 (from left to right: Jakob Trauer, Sebastiano Piccolo, Julian Wilberg, Professor Udo Lindemann, and Professor Anja Maier)

Engineering design and data science: Successful EuroTech collaboration with the Technical University of Munich (TUM)

fredag 20 jul 18
|

Kontakt

Sebastiano Piccolo
Ph.d.-studerende
DTU Management Engineering

Kontakt

Julian Wilberg
Gæste ph.d.-studerende
DTU Management Engineering

Kontakt

Jakob Trauer
Gæstestuderende
DTU Management Engineering

Kontakt

Anja Maier
Professor, Afdelingsleder
DTU Management Engineering
45 25 60 45
How can data support and improve engineering design? Finding the answer(s) to this question was the goal of a three-year collaboration with the Engineering Systems Division at the Technical University of Denmark (DTU). This fruitful cooperation included research visits, student exchanges, joint publications, and four workshops. The first workshop took place in Munich in September 2016 and for the most recent one, the research team met in Copenhagen in February 2018

Companies produce large volumes of data in a variety of ways. For example, an increasing number of technical products are constantly connected to the internet and sensor capture and transmit use phase data, such as errors. In addition to product data, digital data traces are also produced during the development process and project execution as an output of peoples’ work. This may include emails, project files, or activity logs. As such, the environment of engineering companies is changing. The availability of large and diverse data sources offers unprecedented possibilities to improve products, business models, and processes. However, even though data is often praised as a guarantor for additional value or turnover, experience indicates that data does not come without costs. A key challenge, for instance, is to extract value from data in a form that is suitable for obtaining insights on which to ground decisions. Undoubtedly, data offers great potential to support engineering design, yet, engineering companies require additional support in order to overcome the related challenges and to obtain the maximum value from data.

Bridging engineering design and data science

With the goal of using data to provide value to engineering design, Professor Anja Maier, Head of the Engineering Systems Division at the Department of Management Engineering at DTU and Professor Udo Lindemann, former Chair of the Institute of Product Development at TUM, jointly initiated the collaboration under the EuroTech Universities Alliance – a strategic partnership of four leading European universities of science & technology. The collaboration involved two PhD students, Sebastiano Piccolo (DTU) and Julian Wilberg (TUM). Over the course of three years, four workshops took place in Copenhagen and Munich to discuss research findings and future directions. The collaboration bridges Engineering Design and Data Science and makes two main contributions: Firstly, it addresses engineering companies’ needs by providing them with a process model to develop and implement a strategy for use phase data. Secondly, it leverages Data Science to exploit the various data archived by companies in order to improve products and processes. The collaboration has been one in a more than a decade-long lasting series of research collaborations between the professors and has provided important inputs for current and future research directions between the Chairs.

Two PhD dissertations

The collaboration under the EuroTech Alliances partnership has been an integral part during the research forming the PhD dissertations of Sebastiano Piccolo and Julian Wilberg. The joint workshops served as sounding boards to share, discuss and advance findings, enriching both researchers. Sebastiano Piccolo’s research focuses on the analysis and modelling of engineering systems to improve the design process and to provide insight for managing engineering projects. The research uses different data sources including non-technical data (e.g. emails) produced during a large development project. Julian Wilberg’s research supports companies in developing a use phase data strategy that allows a systematic exploitation of products’ use phase data. The main contribution of his research is a process model that guides companies through the steps required for the development of a use phase data strategy.

Scientific dissemination

Scientific achievements of the collaboration include two conference papers. The first paper, presented at DESIGN 2018, shows the relationship between engineering changes and peoples’ sentiments through analysing 50,000+ emails exchanged during a large-scale project. The paper shows that when changes occur, sentiments tend to become more negative and when changes are implemented successfully, sentiments are more positive. Additionally, the paper shows an application of how to leverage a non-technical data source (emails) to monitor and diagnose the health status of a project. The second paper, currently in preparation, combines emails, an activity log, and a document log to understand the completion time of tasks in relation to the multi-layered, structural complexity of the project. In order to help experts provide better estimates, the goal of the paper is to understand which factors experts should consider when they make time estimates for the project. The second paper builds upon the research work of Jakob Trauer who is currently writing his master’s thesis at the DTU as an exchange student from the TUM.

Further information

EuroTech Universities Alliance: http://eurotech-universities.eu/