Carbon fibre reinforced polymer composites have seen successful applications in aerospace, automotive and renewable energy industries over the past six decades. A number of processing and manufacturing technologies have been developed to make the composites but essentially they are all an additive process to some extent. From initial hand layup of carbon fibre prepregs (Composites 1.0) to later automated fibre/tape placement (Composites 1.5), there are increasingly released freedoms to manufacturing more complex, higher performance composites. The conventional manufacturing processes aim at scaling up the meso-scale material (such as lamina and tape) to the macro-scale composite structure, and as such the design and analysis theories for the manufactured composites are still within framework of composites laminated plate theory (CLPT), i.e., Design-for-Manufacturing. Although numerous design tools and failure criteria have been developed, there is still lack of confidence in adopting them in real-world composites design – consequently extensive and expensive experimental tests are still required for the certification of the manufactured composite structures. In this talk, I will introduce our recent advancement in real additive manufacturing of composites with be-spoke material system at multiple length scales to produce highly customised composites. Supported by the high-fidelity micromechanics based data-driven prediction and design, I will share my vision of Manufacturing-by-Design for next-generation digital, multifunctional lightweight composites in the context of Composites 2.0.

Manufacturing speed is very important in commercial applications of additive manufacturing. On the other hand, current additive manufacturing techniques are not readily applicable to on-orbit manufacturing (e.g., in space stations, due to microgravity) and unstable working environments (e.g., on vehicles in maneuvering, due to random vibration). Volumetric additive manufacturing is a newly developed high-speed approach. However, same as all other rapid approaches in current additive manufacturing, the materials to be cross-linked are liquid resins.

We have proposed a new way for volumetric additive manufacturing using solids instead of liquids in UV cross-linking. It appears to be similar to conventional laser engraving inside glass, but the un-cross-linked parts can be separated from the cross-linked parts to create the desired shapes. Since cross-linked polymers tend to have excellent shape memory effect, high dimensional accuracy is ensured after separation.

We propose a general framework to study the on-demandsharedride-sourcing transportationsystems and summarize the relevant researchproblems in four areas, namely, demand, supplyplatform operation, and system problems. We thenfocus on the online matching problem, in which theplatforms match passengers and drivers in real-timewithout observing future information, consideringmultiple objectives such as pick-up time, platformrevenue, and service quality. Given stationary andnon-stationary decision scenarios, we developefficient online matching policies that adaptivelybalances the trade-offs between multiple objectivesin a dynamic setting and provide theoreticalperformance guarantees for the policy. Throughnumerical experiments and industrial testing usingreal data from a ride-sourcing platform, wedemonstrate that our approach can arrive at adelicate balance among multiple objectives andbring value to all the stakeholders in the ride-

Many soft matter systems in manufacturing industry exist as solutions, e.g., polymer solutions/gels, and colloidal suspensions. A soft matter solution is made by dissolving one or more different soft materials in a liquid. Soft matter solutions usually show fascinating phase structure and dynamic properties. They have found wide applications in intelligent manufacturing technology such as functional coating and ink-jet printing. In this talk, I will present my recent works on the multiscale modeling and computations of some typical multiphase soft matter solutions such as simple binary solutions, polymer solutions, and diblock copolymer solutions. I will introduce some general major difficulties for the theoretical modeling and computations of these multiphase solutions, for example, free-interface dynamics, contact line dynamics, complex substrate topography, evaporation dynamics, and phase separation dynamics. I then present our methods of overcoming these difficulties. In addition, we have recently noticed a natural combination of the deep learning methods with promising “brute forces” with “intelligent” variational principles of physics. This combination provides a very powerful numerical method of solving various problems in soft matter physics. I will also give some examples to show how we use this method to study the dynamics of soft matter solutions.

Shipping alliance has become a prevalentcooperation form among major containershipping companies in recent decades. Byforming an alliance, shipping companies canbetter satisfy the shipping demand andimprove capacity utilization and profitability.At the same time, the shipping industry hasbecome an important source of globalgreenhouse gas emissions. Faced with theemergingworldwidemaritimecarbonemission tradingsystem, the shippingalliance needs to redesign its fleet to controlcarbon emissions and reduces its negativeeffects on the profit of the alliance members.This paper considers the fleet deploymentand slot exchange problem in shippingalliance with the consideration of carbonemission trading system. A mixed integernonlinearprogramming model is first

Li-ion batteries are commercially successful power sources for a diverse range of applications. However, the characteristics of Li-ion batteries make them susceptible to both swelling and thermal runaway, resulting in failed products, and occasionally in fires and explosions. In some cases, defects in lithium batteries are not detected prior to use. For example, recently both GM and Hyunadai had to recall their EVs due to defects in the LG Chem batteries. In other cases, some companies manufacture batteries with inadequate safety features or misrepresent the true nature of their battery. To mitigate safety hazards prior to the occurrence of thermal runaway, various strategies have been applied for battery cells, as well as battery packages. This lecture reviews the safety issues and safety strategies for Li-ion batteries.

Greater uncertainty in global trade flows and black swan events, such as COVID-19, have increasingly challenged the current supply chain business model. Digital technologies have been recognized as a key enabler for enabling resilient and responsive supply chains. In this seminar, I would like to share about the ongoing research program Supply Chain 4.0, which aims to research and co-develop digital and automation technologies with industry members to enable responsive, resilient and secured supply chain management processes to prepare company for the ever volatile, uncertain, complex, and ambiguous (VUCA) global supply chain environment. Details of the three research theme will be shared, which including Supply Chain Connectivity, Data-Driven Optimization and Smart Warehouse Automation for next generation logistics.

Therapeutic intervention for Parkinson’s disease (PD) via deep brain stimulation (DBS) represents the current paradigm for managing the advanced stages of the disease in patients when treatment with pharmaceuticals becomes inadequate. Although DBS is the prevailing therapy in these cases, the overall effectiveness and reliability of DBS can be diminished over time due to biocompatibility issues with the electronic implants. To achieve a lifetime solution, this talk envisions that the next generation of neural implants will be entirely “biological” and “autologous”, both physically and functionally. Thus, in this work, we set forth toward developing a biological brain pacemaker (BBP) for treating PD. Our focus is to investigate engineering strategies for creating a multicellular biological circuit that integrates innate biological design and function while incorporating principles of neuromodulation to create a biological mechanism for delivering high-frequency stimulation with cellular specificity. We engineer a 3D multicellular circuit design built entirely from biological and biocompatible components using established tissue engineering protocols to demonstrate the feasibility of creating a living neural implant. Furthermore, we show we can organize cellular materials to create potential functional connections in normal physiological conditions, thus laying down the foundation of designing a high-frequency pacing system for selective and controlled therapeutic neurostimulation. The findings from this work may lead to the future development of autologous living neural implants that both circumvent the issues inherent in electronic neural implants and form more biocompatible devices with lifelong robustness to repair and restore motor functions, with the ultimate benefit for patients with PD.

Inter-modal transfer efficiency is the key tosuccess for the development of Inter-modalpassenger transportation system, such ashigh-speed rail trains and last-mile transport.Long waiting time often occurs for passengerswhen they get off trains and transfer to taxi ore-hailing services. On the other hand, taxi ande-hailing service drives also suffer from longwaiting time for picking up customers at theover-crowded pickup points. In light of thetransfer' efficiency challenge, it becomesurgent for transport hub operators (e.gairports,railway stations) to designsystematic approaches to maximize thetransfer efficiency between rail trains andlast-mile transport. In this paper, we propose

Due to the rapid population aging globally, the current trend of robotic research has been shifting from traditional industrial robots that are separated from humans to human-centered robots that coexist, cooperate, and collaborate with humans. A major motivation for introducing compliance to human-centered robots is physical human-robot interaction.