【作者】 左英飞；

【导师】 鞠平；

【作者基本信息】 河海大学 ， 电力系统及其自动化， 2004， 硕士

【摘要】 本文首先阐述了变压器故障诊断的发展和现状，指出目前变压器故障诊断中存在的不足：一方面具有滞后性，即要到变压器故障已经发生时才能够判断出来；另一方面具有不确切性，即能实现对故障的定性分析但不能实现故障的定位。针对这些问题提出了从电气量方面入手，对变压器进行健康诊断的思路。 通过对变压器绕组建立“灰箱”模型，利用最小二乘辨识法对模型参数进行辨识，根据辨识得到的电阻参数来对变压器绕组进行诊断。对现场试验数据的应用结果表明，该方案能判断出变压器是否发生轻微故障以及故障发生在哪一相。 为了有效地提高变压器诊断的准确性，本文提出了利用辨识结果建立动态的、开放的、基于规则的电力变压器的诊断专家系统(ES)，并结合目前研究已经成熟的油气量对变压器进行综合诊断的思路。更多还原

【Abstract】 The development and present of fault diagnosis of power transformer are surveyed in detail, and the matter is pointed out. The faults cannot be found in time and fault location cannot be confirmed. This paper proposes a concept of diagnosis of transformer based on electric quantities.A " grey " winding model is developed. Then based on the identification of these parameters by least-square technique, a new diagnosis method is presented. The results of test indicate that it is possible to judge whether occurring fault on winding and even which phase by parameters’ change.To improve the veracity of diagnosis this paper presents an idea. Based on a dynamic, open, rule-based diagnosis expert system a general method is developed that combines the oil-gas and electric quantities to monitor the transformer.更多还原