机组定子纯水冷却系统渗漏的智能检测方法探讨

Probing into Intelligence Examing Method for Pure Water Cooling System Seepage of the Unit Stator

结合三峡机组定子纯水冷却系统渗漏智能检测问题,将支持向量机方法用于纯水冷却系统的膨胀水箱液位的估计。针对纯水冷却系统本身的非线性特性,使用核函数为非线性核的最小二乘支持向量机算法完成估计模型的设计。并根据该模型得到基于统计置信区间的自适应阈值。通过现场数据分析了估计模型性能,计算了针对不同工况的自适应阈值,并将此阈值用于了实际的渗漏判定。结果表明基于多项式核函数的支持向量机估计模型可以根据机组工况估计膨胀水箱液位,且自适应阈值的引入可以有效地实现渗漏的判定。

In connection with the problem of intelligence examine seepage of pure water cooling system for Three Gorges unit stator,the support vector machine method is applied to estimate level in expansion tank for pure water cooling system.According to the non-linear characteristics of pure water cooling system itself,kernel function as non-linear nuclear least squares support vector method is applied to complete designing estimation model.And on the basis of the model the adaptive threshold in statistical confidence intervals is achieved.The estimation model performance was analyzed by the data on site and the adaptive threshold was calculated in different working conditions and used to estimate practical seepage.The results indicated that the support vector machine model based on the polynomial kernel function could estimate the liquid level in expansion tank by working conditions of the unit.In addition,introducing threshold could make effective estimation of seepage.