JSSC 2012第9期Power Management0.35μm CMOSEnergy HarvestingNeural Network Accelerator

Platform Architecture for Solar Thermal and Vibration Energy Combining With MPPT

提出一种结合太阳能、热能和振动能量的平台架构，采用双路径能量收集，峰值效率提升11%-13%。

输入电压20mV至5V，光伏升压峰值效率83%，热电升压58%，压电升降压79%，MPPT跟踪效率96%

能量收集最大功率点跟踪光伏热电压电

▸双路径能量收集架构：采用双路径架构替代传统两级方法，实现了11%–13%的峰值效率提升，显著提高了能量收集的整体效率。

▸可重构多输入多输出开关矩阵：通过可重构开关矩阵同时整合光伏、热电和压电三种能量源，利用单一电感实现多源能量收集，系统可处理20 mV至5 V的宽输入电压范围。

▸基于时间的功率监测器实现最大功率点跟踪：提出一种时间基功率监测器，用于光伏采集器的最大功率点跟踪，跟踪效率峰值达96%，优化了能量提取效率。

▸单电感多源能量提取：通过共享电感设计，实现了光伏升压、热电升压和压电升降压转换器的峰值效率分别为83%、58%和79%，减少了硬件复杂度和成本。

Abstract

A platform architecture combining energy from solar, thermal, and vibration sources is presented. A dual-path architecture for energy harvesting is employed that has a peak ef- ﬁciency improvement of 11%–13% over the traditional two-stage approach. The system implemented consists of a recon ﬁgurable multi-input, multi-output swit ch matrix that combines energy from three distinct energy-harvesting sources–photovoltaic, thermoelectric, and piezoelectric. The system can handle input voltages from