电力系统输变电设备动态诊断技术研究

Research on Dynamic Diagnosis Technology of Power Transmission and Transformation Equipment

通过分析输变电设备的不良工况、故障模式和异常征兆三者之间的因果关系,建立了三层因果网,在这个网络模型中,可以考虑不良工况的影响,与异常征兆一同作为证据信息,这种处理方法弥补了现有的诊断方法因缺少对不良工况的分析以致证据信息不完备这一不足。在三层因果网的基础上设定概率值形成贝叶斯网络模型,利用贝叶斯因果网的推理方法,求取网络的最大可能解释,推理结果包含输变电设备目前可能遭受的故障模式,其中包括各种并发故障模式和其它未检测的异常征兆的状态,能为进一步的诊断试验提供重要依据,最后利用诊断实例说明所建立模型的有效性。

A three layers causal network is established based on causal relationship analysis of undesirable service condition, failure mode and abnormal symptom. For this network model, undesirable service condition and abnormal symptom are treated as evidence in order to diagnose. However, undesirable service condition is not included in existing transformers diagnosis methods, resulting in imperfect evidence for diagnosis. The model established in this paper overcomes this deficiency. A Bayesian causal network reasoning method is used and most probable explanation of the network is obtained. The result includes all failure modes transformer currently possibly confronting, and the condition of abnormal symptoms that are not detected which can provide an important basis for the next diagnostic test performed. Finally, examples are applied to validate the advantages of this model.