Designing porous skins to passively reduce drag

James Ramsay¹, Mathieu Sellier¹, Wei Hua Ho²

¹Mechanical Engineering, University of Canterbury

²University of South Africa, South Africa

The energy used to overcome drag by vehicles in New Zealand produces about 8% of our national greenhouse gas emissions. Therefore, there is a strong incentive both monetarily and environmentally to improve the aerodynamics of vehicles, to reduce their drag and emissions. Traditionally, an object is made more aerodynamic by making it very smooth and changing its shape so that flow moves around it favourably - with a small wake and low turbulence/vibrations. However, for vehicles like trucks, their shape cannot be altered as they must carry as much cargo as possible. What’s more, there is a limit to how much improvement in drag can be achieved by merely changing an object’s shape.

A new option is to control the flow directly - to remove or add fluid (air) at strategic locations on the object’s surface to maximise the improvement in the flow. Even better, since all object’s that move through fluid disturb the flow and generate natural pressure gradients, we can use the pressure profiles inherent in the flow to drive these secondary flows. This can be achieved in practice by applying a skin of porous material over an object, and manipulating the porosity and internal connections of this skin, we can direct secondary flows which improve the drag characteristics of the object substantially - and without needing any power. This turns traditional thought on its head. Instead of making objects smooth and streamlined to make them aerodynamic, we keep them bulky and make them rough. Through computational fluid simulations with optimisation, we have determined the best flow control for a reference case of a cylinder, achieving over 50% reduction in drag. And now, the implementation and design of the porous skin to achieve this goal is underway.

Virtual reality enhanced haptic teleoperation with virtual fixture for intelligent robotic welding

Yunpeng Su¹, J. Geoffrey Chase¹, Christopher Pretty¹, XiaoQi Chen¹

¹Mechanical Engineering, University of Canterbury

This paper presents an integrated scheme based on a virtual reality (VR) and haptic feedback approach for intuitive and immersive teleoperation of robotic welding systems. By incorporating VR technology, the user is fully immersed in a virtual operating space augmented by real-time visual feedback from the robot working space. A predictive control approach is implemented to accurately track the welder’s hand movements. The proposed robotic tele-welding system features direct mapping from the user’s hand movements to the robot motions and enables the spatial velocity-based control of the robot tool center point (TCP). The system allows the user to intuitively and precisely manipulate the position and orientation of the end effector to adjust the corresponding welding parameters including travel speed and travel and work angles. The proposed VRVF integration approach implements haptic constraints to guide the operator’s hand movements following the path of a specific task and constrain the operation within a collision-free area. Robot movement experiments and proof-of-concept welding experiments demonstrate the effectiveness of the proposed methods. The VR and haptic feedback integrated scheme reduced undesirable variation in the two directions perpendicular to the weld by 90% and 72% respectively, while generating the smooth movement in the welding direction.

Multi-electrode chips and neuromorphic networks of Sn nanoparticles for unconventional computing

Edoardo Galli¹, Susant Acharya¹, Saurabh Bose¹, Joshua Mallinson¹, Shota Shirai¹, Matthew Pike¹, Ford Wagner¹, Samuel Cleland¹, Zachary Heywood¹, Stephen Weddell², Philip Bones², Matthew Arnold¹, Simon Brown¹

¹School of Physical and Chemical Sciences, University of Canterbury

²Electrical & Computer Engineering, University of Canterbury

The race towards smarter and more efficient computers is at the core of our technology industry. However, due to limitations such as the increasing costs and inability to indefinitely keep shrinking CMOS transistors, novel hardware is needed. Brain-inspired, or neuromorphic, hardware has attracted great interest thanks to its promising capabilities in computational tasks like classification and pattern recognition.

However, most of the neuromorphic systems proposed in the literature rely on regular architectures to integrate brain-like elements such as memristors or CMOS neurons and synapses. Due to the lack of complexity in the connections between elements, these systems are still unable to replicate the intrinsic complex structure and dynamics of the human brain.

Our research takes a different approach, focusing on devices made of Sn nanoparticles which are randomly deposited and subsequently self-organise into percolating complex networks where tunnel gaps between nanoparticles occur [1-2]. The networks have fractal geometries and scale-free topology [3]. Under electrical stimuli, the tunnel gaps act as atomic switches whose dynamics lead to brain-like properties such as avalanches, self-tuned criticality, power-law scaling behaviour, self-similarity, and long-range temporal correlations [3-5]. Thanks to these properties our devices are a promising candidate for neuromorphic computing.

Multi-electrode chips are required in order to interact with the network through multiple inputs and outputs and to physically implement many computing paradigms such as reservoir computing. However, traditional lithographic fabrication techniques produce electrodes which are weakly connected to the networks of nanoparticles, leading to device failure. Thus, a novel and simple fabrication technique for obtaining working electrodes was developed and optimised [6], paving the way for performing computational tasks in our devices.

This presentation will focus on the dynamics of our complex neuromorphic networks of nanoparticles and on our novel fabrication technique for improving the connection between networks of nanoparticles and electrodes.

[1]. S. K. Bose, J. B. Mallinson, R. M. Gazoni, S. A. Brown, (2017). Stable Self-Assembled Atomic-Switch Networks for Neuromorphic Applications. IEEE Transactions on Electron Devices, 64, 12, 5194-5201.

[2]. S. K. Bose, S. Shirai, J. B. Mallinson, S. A. Brown, (2019). Synaptic dynamics in complex self-assembled nanoparticle networks. Faraday Discussions, 213, 471-485.

[3]. S. Shirai, S. K. Acharya, S. K. Bose, J. B. Mallinson, E. Galli, M. D. Pike, M. D. Arnold, S. A. Brown, (2020). Long-range temporal correlations in scale-free neuromorphic networks. Network Neuroscience, 4, 2, 432-447.

[4]. J. B. Mallinson, S. Shirai, S. K. Acharya, S. K. Bose, E. Galli, S. A. Brown, (2019). Avalanches and criticality in self-organised nanoscale networks. Science advances, 5, 11, eaaw8438.

[5]. M. D. Pike, S. K. Bose, J. B. Mallinson, S. K. Acharya, S. Shirai, E. Galli, S. J. Weddell, P. J. Bones, M. D. Arnold, S. A. Brown, (2020). Atomic Scale Dynamics Drive Brain-like Avalanches in Percolating Nanostructured Networks. Nano Letters, 20, 5, 3935-3942.

[6]. E. Galli, S. A. Brown, S. K. Acharya, (2019). Electrical contacts to nanoparticle networks. New Zealand Patent Application 760520.

Optical resonant nanostructure lattice for lasing

Aran Warren¹

¹Electrical and Computer Engineering, University of Canterbury

Periodic lattices of nanostructures can be engineered to form high quality optical resonant surfaces. These resonant surfaces can be used as laser cavities where many properties of the laser, such as the wavelength, direction, and polarisation can be controlled by changing the geometry or environment of the lattice. I will be presenting recent work done in the UC Nanofabrication Lab on the fabrication and characterisation of these lattices and some of the challenges involved with pattering on the sub-100 nm scale.