模式识别在发电机定子接地保护中的应用

大型发电机的定子绕组对地分布电容较大,困此对定子单相接地保护灵敏度的要求较高。该文提出了一种基于模糊模式识别技术的发电机定子接地保护方案。该方案运用模糊综合评判原理,对发电机定子接地故障时的多参量进行综合判断,并通过遗传算法对模糊权系数进行了优化。通过对一台水轮发电机和一台汽轮发电机定子绕组接地故障的仿真试验证明该方案比传统方法具有更高的可靠性及灵敏度。

Application of pattern recognition in generator-stator ground fault protection

The distributed capacitance to earth is high in large-scale generator, so it calls for the alarming sensitivity on the stator single-phase-to-ground fault protection. In this paper, fuzzy-pattern-recognition-based stator ground fault protection scheme is put forward. Employing fuzzy integrated discrimination theory and genetic algorithm, the scheme judges more than one parameter integratedly on stator ground fault and optimizes subjection coefficients respectively. Simulations on stator winding ground fault of a hydro-generator and a steam generator demonstrate that the proposed scheme is superior to the conventional ones in reliability and sensitivity.