Re-Imagining Engineering Design: Growing Radical Cyber-Physical-Socio Phenotypes

本提案的基本目标是重新想象设计工程，以便快速产生新的想法和概念，并且产品及其相关的制造系统（包括其供应链和人员）同时设计并完全相互定制。通过这样做，在概念设计阶段“锁定”的大约70%的生命周期和供应链成本可以被理解、最小化和验证。该计划的目标是通过可互操作的网络-物理-社会（CPS）服务实现设计工程的转型，其中：(i)工程能力和多尺度物理通过创新的数字能力集成；（ii）先进的分析支持知识捕获，增强弹性，并通过可互操作的“智能测试”和完全模拟的生命周期分析来预测合规性，以验证以模型为中心的设计；（iii）新颖的商业/供应链模式提供了从数字设计到制造的透明价值流和途径，以确保英国培养下一代数字工程人才。在我们的愿景中，未来的制造系统是自组织、自我意识和分布式的，这将带来一个与今天完全不同的制造业。这自然导致确定该计划的四个主要研究挑战：互操作性- CPS设计理论：我们如何快速产生想法和概念，使工件与制造系统同时设计，以创建具有弹性的扩展企业，并在整个系统中进行开放通信？网络世界- CPS建模设计与制造：我们如何虚拟地表示概念，使驱动预期行为的关键设计特征通过健壮的、智能制造的产品变体被理解、编码和实现？物理世界- CPS概念到现实：在网络-物理-社会制造生态系统的帮助下，需要什么样的验证和验证概念来找到最短和最有益的物理实现途径？扩展制造企业：我们如何在网络-物理-社会生态系统中转换和利用新组织结构中的概念，以加速扩展企业设计解决方案的演变？这四项技术挑战是综合的，构成相互依存的挑战。它们构成了本节目将要编织在一起的四条线。整个计划需要一系列建模、评估和预测的方法，处理从模拟模型到个人和商业组织的各种系统实体需要多种技术方法。“网络基因”和“网络种子”的概念可以在进化方法中使用，形成了提供新的CPS设计理论的核心线索，但需要与其他方面进行重要的相互联系。例如，计算机辅助设计模型在网络世界中对某些产品来说是足够的，但一般来说，系统是多功能和多学科的，需要一系列的建模方法来提供必要的设计评估数据，例如全寿命成本。同样，虽然可以与制造（例如CNC机床）进行通信，但智能数据直接反馈到实时设计中尚未完成，并且设计系统和组织都需要新的方法来实现这种能力。随着系统和底层方法的发展，将进化算法叠加到这些算法上，必然要求所有元素都要适应和改变。因此，这些性质类似的过程和一系列替代方法（例如分形，基于代理的系统，响应面方法等）将被探索。

The fundamental goal of this proposal is to Re-Imagine Design Engineering so that new ideas and concepts are generated rapidly, and where both the product and its associated manufacturing system (including its supply chain and people) are designed concurrently and fully tailored to each other. By doing this the >70% of lifecycle and supply chain costs that are "locked in" at the concept design stage can be understood, minimised and verified.This programme will target the transformation of Design Engineering via Interoperable Cyber-Physical-Social (CPS) Services in which: (i) engineering competences and multiscale physics are integrated by innovative digital capabilities, (ii) advanced analytics support capture of knowledge, enhance resilience and predict compliance by interoperable 'smart testing' and fully simulated lifecycle analyses to validate model-centric designs, (iii) novel business/supply chain models provide a transparent value stream from digital design through to manufacturing and pathways to ensure the UK develops the next generation of digital engineering talent. Our vision of the future where manufacturing systems are self-organising, self-aware and distributed, brings a radically different manufacturing industry than exists today.This leads naturally to identifying four major research challenges to this programme:1. Interoperability - CPS Design Theory: How can we generate ideas and concepts rapidly such that artefacts are designed concurrently with manufacturing systems to create resilient extended enterprises with open communication throughout the whole system?2. The Cyber World - CPS Modelling Design & Manufacture: How can we represent concepts virtually such that key design characteristics driving intended behaviour are understood, coded and realised via robust, intelligently manufactured product variants?3. The Physical World - CPS Concept to Reality: What verification and validation concepts are needed to find the shortest and most beneficial pathway to physical realisation aided by a cyber-physical-socio manufacturing ecosystem?4. The Socio World - CPS The Extended Manufacturing Enterprise: How can we translate and exploit concepts in new organisational structures within a cyber-physical-socio ecosystem to accelerate evolution of design solutions across extended enterprises?The four technical challenges are integrated and pose interdependent challenges. They form the four threads which are to be woven together in this programme. A range of approaches for modelling, evaluation and prediction are needed for the whole programme, and dealing with such diverse system entities from simulation models to individual human and business organisations necessitates a diversity of technical approaches.The concept of 'cyber-genes' and 'cyber-seeds' that can be used in an evolutionary approach form the core thread to provide a new CPS design theory but requires significant interlinkage with the other aspects. For example, CAD models in the cyber world are sufficient for some products, but in general systems are multi-functional and multi-disciplinary and will require a range of modelling methods to provide the necessary design evaluation data, such as with whole life costing. Similarly, although possible to communicate with manufacturing (e.g. CNC machines), feedback of intelligent data directly into a live design is not yet done, and new methods are needed in both design systems and the organisation to allow this capability. Overlaying evolutionary algorithms to these will necessarily require all elements to be adapted and changed, as both the system and underlying methods evolve. Therefore, these nature analogous processes and a range of alternative approaches (e.g. fractals, agent-based systems, response surface methodologies etc.) will be explored.