Natural frequency of superstructure's overall longitudinal vibration is predicted with Support Vector Machine(SVM),which possesses many characteristics such as small sample learning, global optimization and strong generalization. A nonlinear regression model of natural frequency of superstructure's overall longitudinal vibration is put forward based on the analysis of overall vibration characteristics of superstructure, with principal dimensions, layer numbers and types of superstructures, length, width and height of each layer of 23 ships served as input data, and measured natural frequencies of superstructures' overall longitudinal vibration as output data of the SVM. Natural frequencies of superstructures' overall longitudinal vibration for five ships are predicted by the nonlinear regression model, and the obtained results are coincident with the measured values. The proposed method in this paper is proved to be accurate and feasible, and provides a new idea to the prediction of natural frequency of superstructure's overall longitudinal vibration.%根据支持向量机具有小样本学习、全局寻优和泛化能力强等特点,应用支持向量机对船舶上层建筑整体纵向振动固有频率进行预测.通过对船舶上层建筑整体振动特性的分析,以23条船的船舶主尺度、上层建筑的层数、类型以及上层建筑各层的长度、宽度和高度作为支持向量机的输入数据,上层建筑整体纵向振动固有频率实测值作为支持向量机的输出数据,建立了船舶上层建筑整体纵向振动固有频率的非线性回归模型.应用此模型对5条船的上层建筑整体纵向振动固有频率进行预测,预测值与实测值接近.这证明本文方法是可行的,为船舶上层建筑整体纵向振动固有频率预报提供了一种新思路.
展开▼
