考虑铁心深度饱和的单相双绕组变压器改进π模型
An Improved π Model for Single-Phase Two Winding Transformers Considering Deep Saturation of the Iron Core
-
摘要: 考虑铁心深度饱和的变压器模型是准确模拟变压器低频电磁暂态的关键。现有的T模型和Γ模型均存在物理缺陷:认为铁心深度饱和时其各部分饱和程度一致,这与实际情况不符;虽然现有的π模型能够解决此问题,但尚未见能够表征铁心从开始饱和到深度饱和渐变物理过程的π模型。该文以经典π模型为基础,基于直流电源激励铁心进入深度饱和的原理,利用耦合交流小信号测量铁心励磁曲线的饱和段数据,构建整条励磁曲线。提出将励磁曲线分配到π模型2个励磁支路上的方法,由此对π模型励磁支路的非线性特性进行准确表达,进而建立一种考虑铁心深度饱和的单相双绕组变压器改进π模型。对2台不同铁心材料的单相双绕组变压器进行开路、短路和深度饱和试验,提取建模参数,在ATP/EMTP中搭建改进π模型,同时以现有T、Γ和π模型为基础,分别搭建考虑和不考虑铁心饱和过程的6种变压器模型。利用励磁涌流和直流偏磁试验对所建立的模型进行验证,结果表明,该文提出的改进π模型误差小于5%,精度明显高于其他几种模型,说明其对以励磁涌流为代表的变压器低频电磁暂态具有较高的模拟精度,能够为EMTP类电磁暂态软件提供基础变压器模型。Abstract: A transformer model considering deep saturation of the iron core is essential to accurately simulate lowfrequency transients of transformers. The defect of conventional T and Γ models is: saturation levels of each part are same when the core goes into deep saturation, which does not conform to the actual situation. Although this problem is resolved in the conventional π model, no π model that characterizes physical processes of the core from saturation to deep saturation has been presented. Based on the principle that the core is excited into deep saturation by a DC source, an AC coupling signal is used to measure the data of the magnetization curve in the saturation region, and then the entire magnetization curve is obtained. A method to distribute the data of the curve to two branches of the π model was proposed. The non-linear behaviors of magnetizing inductances were described precisely, then an improved π model for single-phase two winding transformers considering deep saturation was established. Two transformers with different core materials were selected for research. Parameters were obtained by open-and short-circuit tests plus deep saturation tests. The improved π model and six models based on T, Γ and π models were established on the ATP/EMTP platform. Inrush currents and DC bias currents were used to validate the improved π model. Results indicate that the maximum error of this model is less than 5%, far lower than other models, which demonstrates its accuracy for simulating low-frequency transients such as inrush currents. And this model can provide a fundamental transformer model for EMTP-type electromagnetic transient programs.