变压器的故障征兆和故障机制间,存在着很大的随机性和模糊性,而单项诊断方法由于考虑角度的片面性以及知识库的不完善性,往往很难满足变压器故障诊断的实际要求。因此鉴于单一诊断方法的不足,提出将IEC-60599三比值法、模糊故障诊断法、蚁群算法和灰色系统理论(grey system,GS)诊断法这4种方法,根据各自的诊断效果构成最优组合模型,诊断变压器故障。诊断实例结果表明,该组合模型诊断的思想可以克服单项诊断方法信息单一、诊断片面性等不足,并具有更高的故障诊断准确率和更好的鲁棒性。
There are strong randomness and fuzziness between the symptom and the mechanism of power transformer failures, and due to the one-sideness in the consideration and imperfection of knowledge base it is often hard to meet practical requirement of power transformer fault diagnosis by the single diagnosis method. For this reason, based on respective diagnosis results by IEC-60599 three-ratio method, fuzzy diagnosis method, ant colony optimization (ACO) and grey system theory (GST) algorithm an optimal combined model is constructed to diagnose power transformer fault. Practical fault diagnosis examples show that the proposed fault diagnosis idea can remedy such insufficiencies of not all-inclusive information and one-sideness of diagnosis in single diagnosis method. The proposed diagnosis method possesses better robustness and higher fault diagnosis accuracy can be achieved by it.