When sub-synchronous oscillations (SSO) occur in power grid, they can affect control performance of grid-side converter (GSC) for direct-drive permanent magnet synchronous generators (D-PMSG), thereby impacting overall stability of system. To address this issue, this paper proposes a suppression strategy for SSO based on adaptive notch filter (ANF). Firstly, the mechanism by which grid SSO affects GSC control of D-PMSG based on virtual synchronous generator (VSG) is analyzed, and disturbance propagation path of SSO component is delineated. Secondly, to overcome problem that fixed-parameter notch filters are unsuitable for time varying characteristics of SSO frequencies, an improved fast Fourier transform-based algorithm for identifying SSO parameters is proposed. Based on identification results, a strategy for updating center frequency of ANF is designed. Optimal intervention point of ANF in control system is determined. Finally, by constructing a simulation model, the effectiveness of proposed control strategy for suppressing SSO is verified.
Due to the existence of the control cycle in a motor digital control system, the voltage calculated by the current regulator cannot be applied to the motor accurately and timely, which constraints the bandwidth of the current-loop system and influences the dynamic response of the system. To solve this problem, a current reconstruction method during the PWM sampling period in current-loop of permanent magnet synchronous motor is proposed. Firstly, the timing sequence of traditional current sampling and duty cycle update is improved. In one carrier cycle, the current sampling value and the current reference value are used to reconstruct the current and update the duty cycle twice. Secondly, the increment factor is introduced, and the corresponding parameter selection rules are given. By adjusting the increment factor, the current response is improved. Finally, the control system of PMSM is built in Matlab/Simulink, to verify the effectiveness of the proposed current reconstruction method. The simulation results validate the bandwidth of the current-loop is increased with the proposed method.
电流互感器二次侧电流能否真实反映一次侧电流会受到直流分量以及谐波分量的干扰,影响电能计量公平性以及继电保护装置能否准确动作.首先从理论上阐明了电流互感器的误差影响因素,然后通过仿真与实验分别探究了谐波分量和直流分量对电流互感器比值误差和相角误差影响的程度,为电流互感器在不同工况干扰下的误差提供了参考数据.仿真和实验结果一致表明,由于直流分量可能更多地转化成励磁电流,使得铁心励磁特性变差,从而对误差造成更大的影响,比值误差差影响最高在1％左右;而谐波分量产生的交变磁通可以通过铁心耦合到二次侧,所以谐波分量对误差的影响较小,比值误差影响不超过0.1％.
高增益DC/DC变换器是将新能源发电系统中的输出电压提升至并网电压的关键环节.提出一种两相交错型高增益DC/DC变换器拓扑结构(CI-TPIBS),变换器拥有反向对称的Boost-Sepic(BS)结构和以耦合电感(CI)为升压手段的倍压单元.通过控制A、B两相交错运行进一步提高变换器的电压增益,在实现高电压增益的同时,两相交错结构(TPI)避免了开关管和二极管电压应力的逐级递升,同时也抑制了输入电流纹波,且完全的对称性结构为变换器的控制提供了方便.详细分析了变换器升压原理及各个器件的工作特性,并给出了各元器件的选择依据.最后搭建出400 W的原型样机,验证了所提出的变换器的性能.
基于梅尔频率倒谱系数(MFCC)提取的声振特征无法清晰描述变压器信号本身能量分布,应用于变压器机械松动识别时存在准确率不高的局限性,因此提出一种优先进行声振特征区分的变压器故障识别方法.首先依据XGBoost贡献指标化结合粗糙集分析区分MFCC特征显隐性:显性特征对状态识别贡献更高;其次,建立以Focal损失为目标的SMA优化模型,并按显隐性为SVM输入设置权重范围;最后,利用优化后的SVM对变压器实测样本进行训练分析.结果表明,该识别方法平均准确率达98.83%,较仅参数优化SVM的识别准确率提高2.48%,且变异波动小.相比PSO、WOA和GOA算法,SMA算法在特征全局优化和损失收敛性上更突出.此外,该方法还具有一定鲁棒性,引入5%干扰样本后准确率下降在0.3%以内,从而在变压器实际运行环境中具有抗干扰价值.
