摘要
从频域角度对用于故障波形重现的逆变器进行建模和分析。通过引入导通时间偏移量和时间–电压转换系数的概念提出一种改进模型,实现了占空比到输出电压的线性变换;克服了近似模型由于线性化造成的频带受限。引用能消除固有跟踪误差的准PID控制器对改进模型进行控制,为消除参数漂移等对其性能的影响,提出了基于"有感知器监督的Hebb学习规则"的单神经元自适应准PID控制器,很好地改善了控制性能,满足了故障波形重现的精度要求;且该算法运算量小,适合于实时实现和现有软件的更新升级。
An inverter for the electrical fault waveform reconstruction is modeled and analyzed in the frequency domain. By introducing the concepts of the biased on-time and timevoltage transfer coefficient, a modified model （MM） is proposed to realize the linear transition from the duration ratio to the output voltage and break the frequency limitation of the approximate models due to linearization. A quasi-PID controller that can eliminate the inherent tracking error is assigned to control the MM. To inhibit the parameters drifting impacts on the controller, a neuron adaptive quasi-PID controller based on the perceptron supervised Hebb learning rule is proposed, ameliorating the control performance excellently and making the strategy fulfill the accuracy requirements of fault waveform reconstruction; and the algorithm＇s computational burden is so low that it is suitable for real time implementation and software update.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2009年第34期63-70,共8页
Proceedings of the CSEE
基金
国家自然科学基金项目(50677046)~~
