Monitoring data driven life cycle management with AR based on adaptive, AI-supported corrosion prediction for reinforced concrete structures under combined impacts

利用 AR 监测数据驱动的生命周期管理，基于人工智能支持的自适应腐蚀预测，对综合影响下的钢筋混凝土结构进行腐蚀预测

• 批准号: 501798687
• 负责人: Professor Dr.-Ing. Dirk Lowke
• 金额: --
• 依托单位: Institut für Rechnergestützte Modellierung im Bauingenieurwesen
• 依托单位国家: 德国
• 项目类别: Priority Programmes
• 资助国家: 德国
• 项目状态: 未结题
• 来源: https://gepris.dfg.de/gepris/projekt/501798687?language=en
• 关键词: Monitoring; data; driven; life; cycle

The aim of the proposed research project is a monitoring data driven service life management system based on adaptive, AI-supported corrosion prediction for reinforced concrete structures exposed to chloride. Therefore, a procedure is to be researched that allows the monitoring of structures with currently available measuring principles and the provision of corresponding time-variant condition information for a link within a digital twin with a structure database. The focus is on an adaptive AI-based service life prediction for a combined impact of chloride exposure and mechanical loads. Physically Informed Neural Networks (PINN) will be used both for the interpretation of the local and possibly erroneous sensor data and for the adaptation of the parameters of the 3D prognosis model. The monitoring and forecast data will be linked in an extended building information model (digital twin). Prediction data, building geometry and the position of the steel reinforcement in the component can thus be compared in order to identify problem areas and visualise them on the real component with the help of augmented reality (AR). Defects that are not yet visible on the surface can thus be found and checked more easily during structural inspections. The service life management system thus enables optimised and locally differentiated planning of future repair measures.

拟议研究项目的目的是建立一个监测数据驱动的使用寿命管理系统，该系统基于人工智能支持的自适应腐蚀预测，用于暴露于氯化物的钢筋混凝土结构。因此，需要研究一种程序，允许使用当前可用的测量原理监测结构，并为数字孪生内与结构数据库的链接提供相应的时变条件信息。重点是针对氯化物暴露和机械负载的综合影响进行基于人工智能的自适应使用寿命预测。物理信息神经网络 (PINN) 将用于解释本地和可能错误的传感器数据以及调整 3D 预测模型的参数。监测和预测数据将在扩展的建筑信息模型（数字孪生）中链接。因此，可以比较预测数据、建筑几何形状和组件中钢筋的位置，以便识别问题区域并借助增强现实 (AR) 在实际组件上将其可视化。因此，在结构检查期间可以更轻松地发现和检查表面上尚不可见的缺陷。因此，使用寿命管理系统可以对未来的维修措施进行优化和本地差异化规划。