← 返回
储能系统技术 储能系统 多物理场耦合 ★ 5.0

基于自适应线性神经元增强ADRC的双三相PMSM谐波电流抑制方法:一种优化策略

Adaptive Linear Neuron-Augmented ADRC for Harmonic Current Suppression in Dual Three-Phase PMSMs: An Optimization Strategy

语言:

中文摘要

双三相永磁同步电机(DTP-PMSM)因反电动势非正弦、逆变器非线性及绕组不对称等因素,易受周期性与非周期性谐波干扰,影响系统性能。传统自抗扰控制(ADRC)对周期性谐波抑制能力有限,本文提出一种自适应线性神经元增强型ADRC(ALNA-ADRC)策略。通过在ADRC框架中引入自适应线性神经元(ALN),该方法在谐波平面内精准抑制5次、7次等周期性谐波,同时保持对非周期扰动的强鲁棒性。解耦观测器带宽与控制增益,简化了参数整定。实验结果表明,该方法显著降低电流总谐波畸变率(THD),稳态下最低可达6.55%,并在参数失配、变转速及突加负载等复杂工况下表现出优异的谐波抑制能力与稳定性。

English Abstract

Dual three-phase permanent magnet synchronous motors (DTP-PMSMs) are susceptible to both non-periodic and periodic harmonic disturbances due to factors such as non-sinusoidal back electromotive force (EMF), inverter nonlinearity, and winding asymmetry, which degrade system performance. To address the insufficient suppression of periodic harmonics in traditional active disturbance rejection control (ADRC), this paper proposes an adaptive linear neuron-augmented ADRC (ALNA-ADRC) strategy. By embedding an adaptive linear neuron (ALN) into the ADRC framework, the proposed method precisely suppresses periodic harmonics, such as the 5th and 7th harmonics, in the harmonic plane while maintaining ADRC’s robustness against non-periodic disturbances. Furthermore, by decoupling the observer bandwidth from the control law gain, the parameter tuning process of ADRC is simplified. Experimental results demonstrate that the proposed method significantly reduces the total harmonic distortion (THD) of the current, achieving a minimum THD of 6.55% under steady-state conditions. Additionally, it exhibits superior harmonic suppression capability and stability under complex operating conditions, including parameter mismatch, variable speed, and sudden load changes.
S

SunView 深度解读

该ALNA-ADRC谐波抑制技术对阳光电源储能与电驱产品具有重要应用价值。在ST系列储能变流器中,可显著降低双三相PMSM飞轮储能系统的电流THD至6.55%以下,提升能量转换效率与电网友好性。对新能源汽车电机驱动系统,该方法可有效抑制5次、7次谐波引起的转矩脉动,改善NVH性能。解耦观测器带宽与控制增益的设计思路,可简化PowerTitan大型储能系统多电机协同控制的参数整定复杂度。建议将ALN自适应算法与现有ADRC框架融合,增强储能变流器在电网谐波污染、参数失配等复杂工况下的鲁棒性,并可扩展至充电桩双向充放电控制中的谐波治理场景。