基于H-K聚类逻辑回归的贯流式机组水导轴承磨损性能评估研究

贯流式机组水导轴承性能对机组振动特性和稳定运行有很大影响,对此本文提出了一种基于H-K聚类逻辑回归模型用于实现贯流式机组水导轴承磨损性能评估。以黄河河口水电站3#机组振动、摆度幅值和工况参数等作为自变量,水导轴承运行状态作为因变量,同时为了增强模型泛化能力,引入H-K聚类方法对自变量进行离散化处理,通过建立变量之间的逻辑回归模型实现对机组水导轴承磨损性能评估。研究结果表明:机组轴系摆度信号和机组轴系振动信号可以更好地解释水导轴承性能变化,同时通过模型对水导轴承性能显著影响的特征信号频谱分析推断,机组水导轴承磨损的主要原因是机组轴线偏移和不平衡电磁拉力影响所致。

Research on wear properties assessment of tubular turbine guide bearing based on H-K Clustering-Logistic Regression Model

The performance of turbine guide bearing has a great influence on vibration characteristics and stability of tubular turbine, thus a H-K clustering-logistic regression model is proposed in this paper to evaluate the turbine guide bearing wear property. Taking amplitudes of vibration and throw and working parameters of 3# unit in the HeKou hydropower station as independent variables, and the operating condition of turbine guide bearing as dependent variable. In order to enhance model generalization, the method of H-K clustering is introduced to discretize variables, then the wear properties assessment of tubular turbine guide bearing is realized based on logistic regression model between independent and dependent variables. The research results show that:the throw and vibration amplitudes of tubular turbine shaft system have a better explanation to the performance variation of turbine guide bearing. In addition, according to frequency spectrum analysis of characteristic signals, it can be inferred that the main reasons for turbine guide bearing wear property degradation are axis deviation and unbalanced electromagnetic tension of tubular turbine.