面向船体结构动冰载荷监测的线性形函数识别方法研究

Identification method using the combination of linear shape functions for the monitoring of dynamic ice load on ship structures

船体结构冰载荷监测系统中的载荷识别算法是对船舶在冰区安全运行进行实时评估的关键环节,而常用的影响系数矩阵法尚未考虑冰载荷的动力效应。在时域内将动冰载荷离散为诸多时间元,在每个时间元内利用线性形函数的组合形式逼近构造动冰载荷;通过结构的动力响应分析得到形函数的响应矩阵,并以此建立船体结构冰载荷识别的正问题。利用共轭梯度最小二乘迭代型正则化算法和终止迭代准则对冰载荷识别问题中的不适定性进行控制。该数值算例和试验验证均表明,该冰载荷识别模型具有良好的识别精度、求解速度及稳定性。

A load identification algorithm is the key to the real-time evaluation of operation safety of ship structures in the ice load monitoring system.However,the dynamic effect of ice load has not been considered yet in the conventional influence coefficient matrix method,which is typically adopted in the ice load identification.The dynamic ice load was discretized into multiple time elements in time domain and the combination form of linear shape functions was used to construct the dynamic ice load approximately in each time element.The response matrix of the shape functions was calculated by the structural dynamic response analysis.On this basis,the forward problem of ice load identification was established.The iterative regularization method using conjugate gradient least squares and stopping criteria for iteration selection were adopted to control the ill-posed condition in the identification problem.Both numerical examples and experimental verification prove that the ice load identification model proposed owe the advantages in aspect of identification precision,solution speed,and stability.