摘要
船舶推进变频器在故障状态时运行时间很短,故障样本数据很少.为了得到理想的预测效果,采用支持向量机技术,实现推进变频器的运行数据预测和工作模式识别.先对能反映船舶推进变频器运行状态的关键运行参数进行分析,提取正常运行和故障状态下的样本数据,然后采用支持向量机建立数据预测模型和模式识别模型,对变频器未来时刻的运行参数和运行模式完成预测.得到的预测数据具有较高的准确度,预测得到的变频器工作模式也与实际运行情况较好吻合,证明基于支持向量机技术的变频器状态预测方法可以用于变频器状态评估和故障诊断.
The time of the marine propulsion converter running in the fault status is short,so the fault sample data are rare.The key operating parameters of marine propulsion converter are analyzed and the sample data under normal condition and fault conditions are extracted.The support vector machine(SVM)method is used to establish the data prediction model and pattern recognition model in order to predict the operation parameters and operation mode of the converter.The result shows that the forecast data have a high accuracy and the forecasted operating mode accords with the real operation condition well,which proves converter status prediction method based on support vector machine is feasible to be used for converter status evaluation and fault diagnosis.
出处
《武汉理工大学学报(交通科学与工程版)》
2015年第2期423-426,共4页
Journal of Wuhan University of Technology(Transportation Science & Engineering)
基金
教育部博士点基金项目(批准号:20100143110004)
武汉理工大学研究生自主创新基金项目(批准号:135205003)资助
关键词
支持向量机
分类
模式识别
预测
评估
support vector machine(SVM)
classification
pattern recognition
prediction
assessment