Aarhus University Seal / Aarhus Universitets segl

Mechanical Engineering Scientific Talks

On the first Friday of every month, one of our mechanical engineering researchers presents his or her research to colleagues and students.

If you have any questions or comments, contact Associate Professor and organiser Gorm Bruun Andresen, gba@eng.au.dk.

NEXT TALK


Friday 27 September 2019 at 11:00-12:00 in room 04.081 at Navitas

The Chrono multiphysics package: a ground vehicle mobility perspective

We provide an overview of the design, features, and capabilities of the open-source multiphysics package Chrono, using various applications in ground vehicle mobility as a unifying theme.
After a short introduction to the Chrono structure, design, and philosophy, we provide brief descriptions of some of its salient modules with examples to vehicle simulation. We discuss various applications and problems, including mobility of legged robots, a co-simulation framework for ground vehicle mobility, UQ and ML for mobility assessment, and virtual simulation of connected and autonomous vehicles. We conclude with a list of ongoing related efforts and a look at future developments.

Presenter: Radu, Serban, PhD, Senior Scientist, University of Wisconsin-Madison, USA


Friday 4 October 2019 at 11:30-12:30 in room 03.070 at Navitas

Simulating the Real World for Validation of ADAS and AD Systems

The definition of meaningful simulation scenarios including road descriptions, vehicle and traffic parameterizations is a key factor for successful virtual validation of highly automated driving functions. The dSPACE simulation tool suite ASM (Automotive Simulation Models) offers a variety of possibilities for road and scenario definitions based on requirements, real world data, measurements and standards like OpenDRIVE®, OpenSCENARIO and Open-Simulation-Interface (OSI). The recognition of the environment is performed by sensors of different types and level of details, like ground truth and physical sensor models. For a reliable validation, an appropriate automation is essential, as well.

Presenter: Torsten Kluge, Engineering Group Manager, dSPACE Automotive, Germany


PAST TALKS

Friday 19 July 2019 at 11:30-12:30 in room 03.070 at Navitas

The Spectrum of Wind Power Fluctuations

Wind is a variable energy source whose fluctuations threaten electrical grid stability and complicate dynamical load balancing. The power generated by a wind turbine fluctuates due to the variable wind speed that blows past the turbine. Indeed, the spectrum of wind power fluctuations is widely believed to reflect the Kolmogorov spectrum of atmospheric turbulence; both vary with frequency $f$ as $f^{-5/3}$.

This variability decreases when aggregate power fluctuations from geographically distributed wind plants are averaged at the grid {\it via} a mechanism known as {\it geographic smoothing}. Neither the $f^{-5/3}$ wind power fluctuation spectrum nor the mechanism of geographic smoothing are understood.

In this talk, I will chart out the non-equilibrium character of wind power fluctuations, and explain the wind power fluctuation spectrum from the turbine through the grid scales. The $f^{-5/3}$ wind power fluctuation spectrum results from the largest length scales of atmospheric turbulence of order 200 km influencing the small scales where individual turbines operate. This long-range influence correlates outputs from geographically distributed wind plants over a range of frequencies that decreases with increasing inter-farm distance. Consequently, aggregate grid-scale power fluctuations remain correlated, and are smoothed until they reach a limiting $f^{-7/3}$ spectrum, which is confirmed with field data.

I will discuss engineering and policy implications of these results and chart out the future directions I intend to pursue in the broad area of energy research.

Prof. Mahesh Bandi from OIST, Japan.


Friday 14 June 2019 at 11:30-12:30 in room 03.070 at Navitas

Increasing UR robot performance by input shaping for suppressing mechanical vibrations

Current development of industrial robots includes higher requirements in terms of productivity, safety and energy efficiency. A light weight robot allows to perform rapid motions with low power consumption and low level of kinetic energy during motion, leading to a reduced risk in case of a collision.
The downside of introducing light weight designs, is that reduced mass comes with reduced stiffness and damping of the mechanical system. The reduced stiffness and damping introduce increased sensitivity to unwanted mechanical vibrations during motion, especially for high accelerations. These vibrations are unwanted, because they affect robot precision, accuracy, wear, power consumption and productivity in a negative way. This presentation includes the thoughts and efforts on increasing robot performance by reducing mechanical vibrations.

In the recent decades many different strategies have been investigated, in order to find a suitable method of reducing mechanical vibrations in light weight, i.e. low impedance, robotic systems. Generally, the different strategies can be divided into hardware design, trajectory optimization, feedback control, and feed-forward control methods.

One feed-forward vibration suppression method, which has gained a lot of attention for its simplicity and efficiency is called Input Shaping. The basic principle of Input Shaping is to convolve a reference signal with a vibration free impulse train in order to obtain a new (shaped) reference signal, which is sent to the system.

I will cover the challenges and proposed solutions in implementing Input Shaping in industrial robots, which has configuration dependent dynamic behavior.

PhD Student Dan Kielsholm Thomsen, Robotics, Mechanical Engineering.


Friday 3 May at 13:00 at the Deep Tech Experimental Hub in Skejby

Maker Space Opportunities for MSc in Engineering Students

Assistant Professor Marcelo Azevedo Dias, Mechanical Metamaterials and Soft Matter, Mechanical Engineering.


Friday 5 April 2019 at 11:30-12:30 in room 03.070 at Navitas

Numerical fracture mechanics using XFEM

The so-called extended finite element method or XFEM has received much attention recently. It is a method for modelling discontinuities such as cracks, voids or interfaces in an FE context without conforming the mesh to the discontinuity. It is therefore a very efficient technique for cases where the discontinuity is changing over time, such as in the case of crack propagation, because the mesh does not need updating. The method works by locally “enriching” the discretization and is thus conceptually rather simple. However, as is often the case, there is a price to pay in terms of extensive bookkeeping. We take a superficial look at the theory behind the method and implementation issues. The possibilities of the method is shown in various mixed mode crack propagation scenarios.

Assistant Professor Mikkel Melters Pedersen, Fatigue, Mechanical Engineering.


Friday 8 February 2019 at 11:30-12:30 in room 03.070 at Navitas

Selected topics on wind-farm modeling and optimization
Several topics on the optimization of wind-farms will be presented: (i) mod-el-based optimization of wind-farm power based on axial induction factors and yaw angles, (ii) model improvement with the statistical wake mean-dering model, (iii) wind farms with non-standard multi-rotor turbines, and (iv) wind farm layout optimization by treating wind turbines as repelling charged particles.

Professor Martin Greiner, Sustainable Energy Systems, Mechanical Engineering.


Friday 7 December 2018 at 11:30-12:30 in room 03.070 at Navitas 

The Shape of Green Mankind is in a new epoch—the anthropocene—where human activities have made a significant impact on the natural environment. At the same time, we have reached a new level of economic and social prosperity. De-coupling development with resource consumption and waste creation re-quires progress in two key areas, (1) quantifying the environment impacts of products and production systems, as well as (2) developing data-driven design strategies to mitigate the resulting impact. This talk will discuss ex-isting research challenges in these areas and ideas for addressing them through lifecycle information modeling . The potential of such modeling approaches will be illustrated through visual analytics tools that support sus-tainable design and manufacturing via interactive exploration of lifecycle information. Finally, we will discuss the open problem of incorporating the shape complexity of mechanical parts into environmental life cycle assess-ment and the potential benefits that it can offer.

Assistant Professor Dev Ramanujan, Lifecycle Design and Manufacturing, Mechanical Engineering.


Friday 9 November 2018 at 11:30-12:30 in room 03.070 at Navitas 

Computational surface mechanics: what, why and where
Mechanics of contacting surfaces is of critical importance in many appli-cations, including tribology and machining processes. Friction and wear occurs everywhere – all around us. From an industrial viewpoint, friction and wear mean waste of energy and material. During sliding contact, surfaces lose material at all scales, ranging from atoms to debris. However, in-situ observation and analysis of the process of material removal have been a century-old dream for scientists. In this talk, I would like to present our recent advances to use computer simulations to further understand the process of surface material removal during contact.

Assistant Professor Ramin Aghababaei, Surface Mechanics, Mechanical Engineering.