CELLCOMP: Data-driven Mechanistic Modelling of Scalable Cellular Composites for Crash Energy Absorption

英国和世界范围内严格而雄心勃勃的排放目标正在促使运输行业向零排放车辆（例如电动汽车、氢汽车）发展。阻碍这一转变的最大障碍是安全问题、低里程范围和高购买价格。未来车辆的电池或燃料电池中储存的高能量会对乘客构成重大威胁，因为碰撞时可能会发生火灾或爆炸。为了加速向净零过渡，未来的汽车必须达到与传统汽油/柴油汽车相似的耐撞性水平，但要显著减轻重量，成本相对较低。这一极具挑战性的目标促使人们开发轻量化、可扩展和耐碰撞的结构（例如碰撞盒、保险杠），以保护储能设备。受天然材料（例如由纤维素纤维和木质素基质组成的木材）的启发，合成蜂窝复合材料（CC）是用于碰撞能量吸收的新兴轻质材料。CC是具有互连复合细胞壁的多孔细胞材料。由于多相材料成分和结构的独特组合，CC被证明可以吸收比许多单相蜂窝材料（例如聚合物泡沫、蜂窝）高出数倍的碰撞能量。迄今为止，大多数用于能量吸收的桁架式CC都存在吞吐量低和可恢复性差的问题，增加了制造和维护成本。考虑建筑特征、组成材料和变形机制的巨大设计空间导致预测吸能性能的复杂性大大增加。事实上，缺乏可扩展的制造技术和可靠的碰撞评估模型是增加批量生产运输车辆采用CC的主要障碍。该项目旨在通过整合计算力学模型和数据驱动的方法来预测和优化CC的碰撞性能，从而充分解决这一挑战。拟议的项目将解决两个基本问题：（i）微观结构拓扑结构，组成材料和应变率对CC的能量吸收性能的作用是什么？(ii)如何高效地对CC进行编程，以实现所需的崩溃响应，同时考虑可扩展性和可恢复性？为了解决这些问题，该项目将开发一种新的方法来理解，预测和优化CC的破碎响应。该项目将带来一个独特的团队，拥有可扩展制造，高应变率实验测试，计算建模和数据驱动方法的多学科研究专业知识。该项目将采用可扩展的制造方法来创建新的CC，以利用弹性屈曲不稳定性并最大限度地减少局部结故障，从而提高其可恢复性。材料成分的机械性能和它们的建筑CC将使用高速测试设备进行测量。将开发新的高保真计算模型来预测CC在破碎载荷下的屈曲、塑性和断裂响应。CC的结构-性能关系将通过先进的机器学习算法揭示，从而能够快速智能地识别所需应用的优化设计。将对优化的CC原型进行高速碰撞试验，以评估其能量吸收和可恢复性。该项目开发的数据驱动计算框架和可扩展的CC原型将改变未来的计算模式，使未来的零排放车辆更安全，更环保。通用数据驱动的设计工具还将为其他多孔蜂窝材料的有效设计开辟新的途径，从隔热泡沫，声学超材料到人工组织支架。

The stringent and ambitious emissions targets within UK and the world are prompting the transportation industries to move towards zero-emission vehicles (e.g. electric cars, hydrogen cars). The biggest obstacles holding back the shift are safety concerns, low mileage range and high purchase prices. The high energy stored in the batteries or fuel cells of future vehicles poses a significant threat to passengers due to the potential fire or explosion in a crash. To accelerate the transition to net-zero, the future vehicles must reach similar level of crashworthiness achieved with conventional petrol/diesel cars, but significantly reducing the weight and at a relatively low cost. This exceptionally challenging target is motivating the development of lightweight, scalable and crashworthy structures (e.g. crash box, bumper) to protect the energy-storage devices. Synthetic cellular composites (CCs), inspired by natural materials (e.g. wood, composed of cellulose fibres and lignin matrix), are emerging lightweight materials for crash energy absorption. CCs are porous cellular materials with interconnected composite cell walls. With unique combinations of multi-phase material constituents and architecture, CCs are proven to absorb multiple times higher crash energy than many singe-phase cellular materials (e.g. polymer foams, honeycombs).To date, most existing truss-like CCs for energy absorption suffer from low throughput volume and poor recoverability, increasing the cost for manufacturing and maintenance. The vast design space considering architecture features, constituent materials and deformation mechanisms leads to a huge escalation in the complexity to predict energy-absorbing performance. Indeed, the lack of scalable manufacturing techniques and reliable models for crash assessments are the main barriers to increase the adoption of CCs in volume production transportation vehicles. This project aims to substantially address this challenge by integrating computational mechanistic models and data-driven approaches for predicting and optimising the crash performance of CCs.The proposed project will address two fundamental questions: (i) what is the role of microstructural topologies, constituent materials and strain rates on the energy-absorbing properties of CCs? (ii) how to efficiently program CCs to achieve desired crash responses while considering scalability and recoverability? To address these questions, this project will develop a new methodology to understand, predict and optimise the crush responses of CCs. This project will bring a unique team with multidisciplinary research expertise of scalable manufacturing, high strain-rate experimental testing, computational modelling and data-driven approaches. The project will employ a scalable manufacturing method to create new CCs for exploiting elastic buckling instabilities and minimising the localised junction failure, thereby enhancing their recoverability. The mechanical behaviours of material constituents and their architected CCs will be measured using high velocity testing facilities. Novel high-fidelity computational models will be developed to predict the buckling, plasticity and fracture responses of CCs under crushing loads. The structure-property relationships of CCs will be revealed by advanced machine learning algorithms, enabling the rapid and intelligent identification of optimised designs for desired applications. High velocity crash tests of optimised CCs prototypes will be conducted to evaluate their energy absorption and recoverability. The data-driven computational framework and scalable CCs prototypes developed in this project will shift the future computing paradigm and make future zero-emission vehicles safer and greener. The generic data-driven design tool will also open new avenues for efficient designs of other porous cellular materials, ranging from thermal insulation foams, acoustic metamaterials to artificial tissue scaffolds.