Real-time Material Measurements and Process Control in Automated Fibre Placement Composites Manufacture

先进复合材料和结构是英国未来经济成功的重要技术。复合材料领导力论坛预测，到2030年，英国有潜力将目前23亿英镑的年市场增长到120亿英镑。为了利用这一行业的潜在增长，需要以高速度每次正确制造零件的工艺。该项目的目的是将材料数据实时连接到过程模型，以提供高度自动化和强大的复合材料制造过程。与自动驾驶仪保持飞机在湍流中飞行的方式相同，该资助计划为先进复合材料制造工艺开发一种等效的“自动驾驶仪”，以检测材料干扰并在缺陷发生之前做出反应。这一想法将通过设计和建造一个新的，控制良好的测试平台，并使用它来展示主动过程控制软件。传感器将被选择并安装在试验台上，以测量材料沉积之前、期间和之后的关键材料和工艺变量。将被监控的变量包括复合材料的厚度和宽度、加工速度、温度和压力。传感器数据将通过数学模型进行实时分析，以确定传入的材料是否会制造出设计的结构。新的软件将被编写和测试，以与测试台硬件接口，克服专有工业设备遇到的电流传感器集成障碍。如果数学模拟偏离预期设计，则将调整工艺以克服任何不必要的缺陷。传感技术的发展，工艺模型和控制软件将通过使用理想和不完美材料的制造试验进行验证。该项目的产出将是一个高度仪表化，精细控制的测试台，超过目前的工业复合材料沉积机。在这项资助中收集的高分辨率数据将支持现有数值模型的验证或确定过程模型的新要求。过程模型和控制算法将用非专有编程语言编写，使其他研究人员能够在此资助下完成工作。流程模型和控制代码将在Github（基于Web的软件开发平台）上进行管理，并在达到适当成熟度后开放源代码。未来的研究将调查新材料的加工特性，为其他工艺添加新算法，或将机器学习方法应用于该项目中创建的数字平台。工业项目合作伙伴Coriolis Composites、Hexcel Composites、National Composites Centre和Rolls-Royce将带来实际的制造利益，并从复合材料结构的加速设计-制造-测试周期中受益。与慕尼黑工业大学的学术合作将加强实验特性和测试，以及复合材料制造过程的数值模拟。

Advanced composite materials and structures are an important technology for future UK economic success. The Composites Leadership forum predicts the UK has the potential to grow its current £2.3bn annual market to £12bn by 2030. Processes that manufacture parts right-every-time at high-rate are needed to take advantage of the potential growth in this sector. The aim of this project is to connect material data to process models in real-time, in order to deliver highly automated and robust composites manufacturing processes. In the same way that autopilots keep aircraft flying through turbulence, this grant proposes to develop an equivalent 'autopilot' for advanced composite manufacturing process to detect material disturbances and respond before defects occur. This idea will be delivered by designing and building a new, well-controlled testbench, and using it to demonstrate active process control software. Sensors will be selected and installed on the testbench to measure key material and process variables before, during, and after the material is deposited. Variables that will be monitored include composite material thickness and width, process speed, temperature, and pressure. The sensor data will be analysed in real-time by mathematical models to determine whether the incoming material will make the structures that was designed. New software will be written and tested to interface with the testbench hardware, overcoming current sensor integration barriers encountered with proprietary industrial equipment. If the mathematical simulation deviates from the intended design, the process will be adjusted to overcome any unwanted deficiencies. The sensing developments, process models, and control software will be validated by manufacturing trials using ideal and imperfect materials. The outputs of the project will be a highly-instrumented, finely controlled testbench that exceeds current industrial composite deposition machines. The high-resolution data collected in this grant will support validation of existing numerical models or identify new requirements for process models. The process models and control algorithms will be written in non-proprietary programming languages, enabling other researchers to build-upon the work done in this grant. The process models and control codes will be managed on Github (a web-based software development platform), and made open source once reaching the appropriate maturity. Future research will investigate the processing characteristics of new materials, add new algorithms for other processes, or apply machine learning approaches to the digital platform created in this project. Industrial project partners Coriolis Composites, Hexcel Composites, the National Composites Centre, and Rolls-Royce will bring practical manufacturing interests and benefit from an accelerated design-make-test cycle for composite structures. An academic collaboration with the Technical University of Munich will strengthen experimental characterisation and testing, and numerical modelling of composite manufacturing processes.