Abstract: For the health diagnosis of the tower crane boom structure, a dual-input single-output ARX tower crane boom structure damage diagnosis model based on time series data is proposed. Take the dynamic displacement data of the middle node of the three adjacent nodes on the boom as the output, and use the dynamic displacement data of the two adjacent nodes of the middle node as the input to establish a diagnostic model. The least squares estimation of the model parameters is derived, and in order to further reduce the deviation of the parameter estimation of the least squares method, the flower pollination algorithm is used to optimize the model parameters. The tower crane is modeled by finite element and the dynamic excitation load is applied to the model when the tower crane is hoisted to obtain the dynamic displacement data of the intact and damaged conditions of the relevant nodes on the boom. The diagnosis model is used to process the acquired dynamic displacement data to obtain the damage characteristic parameters, and to define the damage degree parameters, so as to make a more accurate diagnosis of the health status, damage location and damage degree of the boom structure.

Key words: tower crane boom, structural health diagnosis, dual-output single-output ARX model

摘要： 针对塔机起重臂结构的健康诊断，提出了一种基于时序数据的双输入单输出ARX的塔机起重臂结构损伤诊断模型，即以起重臂上相邻三个节点中的中间节点的动态位移数据作为输出，以中间节点的相邻两节点的动态位移数据作为输入建立诊断模型。推导了模型参数的最小二乘估计，并且为了进一步减小最小二乘法的参数估计产生的偏差，采用花朵授粉算法（Flower Pollination Algorithm，FPA）对模型参数进行优化。通过有限元对塔机建模并对模型施加塔机起升时的动态激励载荷，获取起重臂上相关节点完好工况与损伤工况的动态位移数据。用诊断模型对所获取的动态位移数据处理得到损伤特征参数，并定义损伤程度参数，以此对起重臂结构健康状态、损伤位置及损伤程度进行较为准确的诊断。

关键词: 塔机起重臂, 结构健康诊断, 双输出单输出ARX模型