摘要
为了保证变压器的安全运行从而提高其经济效益,对变压器进行剩余寿命预测并采取相应对策显得尤为重要。研究表明,通过对油浸式电力变压器油中溶解的糠醛体积分数以及气体测定可以进行变压器固体绝缘寿命的监督。为此,笔者利用BP神经网络建立起油中CO、CO2、糠醛含量以及运行年限来预测绝缘纸老化程度的关系模型。实例结果表明,通过此方法能够提供比较准确的变压器寿命预测。
The oil-filled transformer＇s solid insulating life can be measured by detecting the content of furfural and the gases both dissolved in oil. According to this principle, the relational model, with the help of BP neural networks, is to predict the aging degree of insulating paper by operated length and the content of furfural dissolved in oil . The result of examples show that this method can predicts exactly the life of transformer.
出处
《高压电器》
CAS
CSCD
北大核心
2010年第4期84-87,共4页
High Voltage Apparatus
关键词
油浸式电力变压器
糠醛
BP神经网络
可靠性
oil-immersed power transformer
furfural
BP neural networks
reliability
