为了解决三比值法在变压器故障诊断应用中存在的问题,提出一种基于B样条理论改进的方法。该方法在对三比值法进行空间几何分析的基础上,采用B样条曲面表示故障特征区域间的分界面,通过柔性自适应地调整分界面的形状,可将误诊样本对应的故障特征纳入到正确的故障特征区域内,从而提高诊断精度。方法突破了原三比值法固定区域边界的限制,为故障诊断问题求解提供了一种新的解决思路。实例分析表明,改进后的方法既保留了三比值法对大多数样本能够准确识别的特点,又使三比值法具备了学习能力,其学习过程具有可视性,且学习效果是可控的。
To solve diagnostic problems in the three-ratio method for transformer fault diagnosis, this paper proposed an improved three-ratio method based on the B-spline theory. The fault characteristic distribution of three-ratio codes in the spatial geometry was analyzed. On the basis of this, B-spline surfaces were constructed to indicate the failure interfaces. By flexibly adjusting curved shape of the interface, a diagnostic error sample produced in the vicinity of the interface can be incorporated within a new correct area of fault characteristic distribution, thereby diagnostic accuracy was improved. The method avoids the limit of the original three-ratio method with fixed boundary and is a new idea for solving fault diagnosis problems. Case studies indicated that the improved method can maintain the feature of identifying the majority of the samples, and can make the three-ratio method have learning ability. The learning process is visible and the learning effect is controllable.