Detailed parameters for transformer model [22].

Power transformers are essential elements in power systems and thus their protection schemes have critical importance. In this paper, a scheme is proposed for accurate discrimination and location of internal faults in power transformers using conventional measuring devices attached to the transformer. Different types of internal winding faults are intensely considered: partial discharge, inter-disk faults, series and shunt short circuit faults and axial displacement. Depending on the transformer measured output voltage, input voltage and the input current, the construction of a locus diagram (ΔV-Iin) serves as an indicator for any physical modification to the winding. Using five suggested features extracted from the developed locus, an artificial neural network (ANN) technique is applied to accurately distinguish any deviation from the transformer healthy condition. The exact location of each fault inside the windings of power transformer is then determined. The obtained results validate the usefulness of the proposed scheme for different internal faults. The superiority of the proposed scheme is extensively examined by comparing its results with some published schemes.

电力变压器是电力系统中的核心元件，因此其保护方案具有至关重要的意义。本文提出一种基于变压器外接常规测量装置的方案，可精准识别并定位电力变压器的内部故障。本文重点考量了多种绕组内部故障类型：局部放电、盘间短路故障、串联与并联短路故障以及轴向位移故障。基于变压器实测的输出电压、输入电压与输入电流，构建ΔV-Iin轨迹图，以此作为绕组发生任何物理变化的判定依据。从所构建的轨迹图中提取五项预设特征后，采用人工神经网络（Artificial Neural Network, ANN）技术，可精准识别变压器健康状态的任何偏离情况，进而确定电力变压器绕组内部各类故障的具体位置。所得结果验证了所提方案对各类内部故障的有效性。通过将本方案的结果与部分已发表的同类保护方案进行对比，本文全面验证了所提方案的优越性。