RESEARCH INITIATION AWARD: Markov Chain Control of Randomized Switching in Power Converters

一般来说，电力电子的使用，特别是开关电源变换器的使用，是普遍的，广泛的，对许多现代技术的成功至关重要。目前电源变换器的标准操作是基于周期性开关函数的生成。周期性开关策略是节能的，但可能需要广泛的滤波，并且由于与机械结构的谐波相互作用，在电机应用中可能产生过多的噪声。该项目的目的是通过开发电源转换器中随机开关波形的统一设计程序来减轻这些影响和其他相关影响。该建议的出发点是一组最近提出的随机交换方案，其中一个具有适当选择的统计性质的随机信号被添加到参考交换函数中。随机化在不影响正常操作的情况下改变频率内容。由马尔可夫链控制的开关设计程序的拟议开发将通过在时域和频域实现时域自适应和波形优化，进一步增强这种方法。这种新方法对制造业来说也是实用而有趣的，因为它不需要对现有设备进行任何重大的硬件更改。＊＊＊

9410354 Stankovic The use of power electronics in general, and of switching power converters in particular, is pervasive, widespread and critical to the success of much of modern technology. Current standard practice in power converters operation is based on the generation of periodic switching functions. Periodic switching policies are energy efficient, but can require extensive filtering, and can produce excessive acoustic noise in motor applications, due to interactions of harmonics with mechanical structures. The purpose of this project is to mitigate these and other related effects through the development of a unified design procedure for randomized switching waveforms in power converters. The starting point for the proposal is a set of recently proposed randomized switching schemes, where a stochastic signal with appropriately chosen statistical properties is added to the reference switching function. Randomization then alters the frequency content without affecting the proper operation. The proposed development of design procedures for switching governed by Markov chains will further enhance this approach by enabling time-domain adaptation and wave-form optimization in both time and frequency domains. The new approach is also practical and interesting for manufacturing, as it does not require any major hardware changes in existing equipment. ***