← 返回
储能系统技术 储能系统 ★ 5.0

基于交互式多模型算法的变压器保护方法

Model-Based Power Transformer Protection Using Interactive Multiple Model Algorithm

作者 Saeed Nikbakhsh Jahromi · Ehsan Hajipour · Mehdi Ehsan
期刊 IEEE Transactions on Power Delivery
出版日期 2024年11月
技术分类 储能系统技术
技术标签 储能系统
相关度评分 ★★★★★ 5.0 / 5.0
关键词 集中式保护控制变电站 电力变压器保护 多线性模型 交互式多模型算法 实验测试
语言:

中文摘要

在集中式保护与控制变电站(CPC)中,所有电压和电流测量值均可供中央高性能处理器使用。本文提出一种基于模型的变压器保护算法,采用多个线性模型模拟铁芯非线性特性,利用交互式多模型(IMM)算法在不同时刻切换模型以跟踪变压器动态行为。相比传统基于模型的方法,该算法显著提高了精度并降低了计算负担,更易于在实际微处理器中实现。实验测试涵盖匝间短路和匝地故障,验证了所提算法的有效性。

English Abstract

In a centralized protection and control substation (CPC), all of the measurements of the substation, includes voltages and currents, are available to its centralized high-performance processor. This paper proposes a model-based algorithm for protection of the power transformer in a CPC-based substation. The main difficulty in the use of conventional model-based algorithms lies in modeling of the nonlinear behavior of the transformer core. To tackle this problem, in this paper, multiple linear models are employed to simulate the nonlinearity of the magnetic core. At each time instance, the dynamic behavior of the transformer will be followed by one of these linear models and, the proposed algorithm switches between these linear models based on the concept of interactive multiple model (IMM) algorithm. This way, not only the accuracy of the proposed algorithm increases but also the computational burden is significantly reduced compared to conventional model-based algorithms. Therefore, the proposed algorithm has better potential to be implemented in real-world microprocessors. Several experimental tests include turn-to-turn (TTF) and turn-to-ground faults (TGFs) have been employed to reveal the effectiveness of the proposed IMM-based protection algorithm.
S

SunView 深度解读

该交互式多模型变压器保护技术对阳光电源储能系统和光伏逆变器产品具有重要应用价值。在PowerTitan大型储能系统中,隔离变压器是关键部件,IMM算法可实时跟踪铁芯非线性动态,精准识别匝间短路和接地故障,相比传统差动保护可显著降低误动率。该方法计算负担低、易于在ST系列储能变流器的DSP/FPGA平台实现,可集成到iSolarCloud智能运维系统中,实现变压器健康状态的预测性维护。对于1500V光伏系统的升压变压器保护,该算法能有效区分励磁涌流与故障电流,提升系统可靠性。建议将多模型自适应思想拓展至功率器件热管理和电池SOC估计等领域。