作　　者： ; ; ; ;

机构地区： 清华大学机械工程学院先进成形制造教育部重点实验室

出　　处： 《中国电机工程学报》 2006年第8期72-76,共5页

摘　　要： 通过空化声信号判断水轮机中的空化程度是水轮机空化检测的方向和难题,为此进行了轴流转桨式水轮机模型转轮空化试验:选择了4个常用桨叶角度,固定桨叶角度和水头,调整尾水位改变空化状态,试验中用频率范围达1MHz的超声传感器测量空化声信号,并对转轮室空化进行录像。通过观察试验闪频录像,把空化严重程度划分为4个典型的级别,以此作为验证声信号辨识算法的判据。在分析试验数据的基础上,提出空化声信号特征矢量分类法,即把采集到的声信号按频段划分,每个频段的方差作为一维,组成空化特征矢量、通过对特征矢量的空间分类来判别轴流转桨式水轮机空化严重程度。用该方法处理所有试验数据,给出与录像观测进行比较的统计结果,表明该方法是可行和有效的。 Acoustic method is a main orientation and difficultness in detecting cavitation of turbine. The kaplan turbine model experiments were taken that fixed heads and tuned draft tube pressure to achieve various cavitation state at four blade rotating angles. In the experiments, acoustic signals were sampled by transducers with 1MHz frequency range and the visual detection of cavitation in runner chamber was recorded. The criterion to the acoustic signal cavitation discriminating method is established using the visual cavitation classification which is divided into four cavitation level. Based on analysis of the acoustic signal spectrum changing in different cavitation, the cavitation identification method is proposed which convert the acoustic signal character to vectors using different frequency band variance as dimension and classify these vectors to identify cavitation degree. The identification statistic result to the observable cavitation classification show it is an efficient method.

关 键 词： 轴流转桨式水轮机 模型试验 空化 可视监测 声信号 矢量分类

领　　域： [交通运输工程] [动力工程及工程热物理] [金属学及工艺] [金属学及工艺]