基于置信规则库的油浸式变压器故障诊断

Fault Diagnosis of Oil-immersed Transformer Based on Belief Rule Base

针对油浸式变压器的故障诊断问题,提出了基于置信规则库(belief rule base,BRB)的有效诊断模型,通过模型结构简化和模型参数优化来提高BRB建模的效率和精度.首先,合理约减故障气体类型和减少训练模型参数以实现对BRB模型结构的简化.其次,提出一种具有自适应更新策略的人群搜索算法(seeker optimization algorithm with adaptive update strategy,AUS-SOA),对简化BRB模型的参数进行优化.然后,根据简化模型和优化参数建立AUS-SOA-BRB诊断模型.实验结果表明,AUS-SOA-BRB模型具有较高的诊断精度,也验证了所提建模方法的有效性.

Aiming at the fault diagnosis of oil-immersed transformers,a method of establishing an effective diagnostic model was proposed based on belief rule base(BRB).The efficiency and accuracy of BRB model was improved by simplifying the model structure and optimizing the model parameters.First,the structure of BRB model was simplified by reasonably reducing the types of faulty gas and reducing the parameters of the training model.Second,a seeker optimization algorithm with adaptive update strategy(AUS-SOA)was proposed to optimize the parameters of the simplified BRB model.Third,an AUS-SOA-BRB diagnostic model was established based on the simplified model and the optimized parameters.Results show that the AUS-SOA-BRB model has higher diagnostic accuracy,and verifies the effectiveness of the proposed modeling method.