基于ANSYS优化分析的导管架损伤识别

采用有限元软件ANSYS自带的APDL语言和可以考虑水影响的PIPE59单元，建立了管架平台的参数化有限元模型，通过改变有限元模型中的单元几何参数和物理参数的方法，描述导管架在作业中常见的腐蚀、弯曲等损伤情况。选择和测取导管架上部远离损伤位置节点处的模态位移，创建目标函数；利用ANSYS的优化功能寻找某种对原完好导管架有限元模型进行修改后的模型，使其在测试节点的数据与预先建立的存在损伤的模型对应节点的数据趋于一致，从而实现识别导管架的损伤类型和损伤位置。计算结果表明当计算步数达到20000时，对于单处损伤的情况，损伤位置以及损伤程度识别准确率平均可达到90%以上；对于两处损伤同时存在的情况，损伤位置以及损伤程度识别准确率平均可达到70%以上。

Jacket damage identification based on ANSYS optimization analysis

Element parametric finite element models of several common damage events such as corrosion and bending have been analyzed by changing physical and geometric parameters. ANSYS parametric design language (APDL) was employed and PIPE 59, which can reflect the influence of water, was used to establish the finite element model. Node data far from the damage location were selected and collected in the upper region of the jacket to form the objective function. ANSYS design optimization was used to find a model which was modified with an intact model of the jacket. The data from this model at test nodes were in good agreement with the pre-established damage model at corresponding nodes, so that damage location and severity can be identified. The results show that when the calculation step in the optimization module was set as 20000, the accuracy rates of location and severity identification were all over 90% on average when there was only one damage event, and the accuracy rates of location and severity identification were all over 70% on average when there were two damage events.