PI: Prof. Wen-Jong Wu
Intel Champions: Dr. Lilly Huang
In the thriving era of internet, the internet connections are ready to connect things and machines to create intelligent life for convenience. The enormous amount of the de- vices may consume considerable extra power other than daily usages, which conflicts with the trend of power conservation. Hence the device should be supplied with an independent source with continuity.
In combination with the lower power limit requirements of the improved devices, the output of the MEMS power generators shall be collected in a sufficient way to realize applications. Hence, we should carefully design the interface circuits to eliminate the power loss and gain the best power output from the harvesting device. Due to the coupling between transducers and the interfacing circuits, it is very important to study both fields in the same time, so that the interaction between the effects can be pre- cisely studied. In this way, the energy output can be more efficiently used and also transformed.
The current micro-energy harvester has its mechanical limits of material strength, resonance frequency, and operating bandwidth. The material strength is attributed to the substrate material and the piezoelectric material characteristic. Comparing to sili- con, the stain-less steel substrate that we use has already improved the sustainability of the device. It can provide a larger displacement and hence larger voltage output. On the other hand, the piezoelectric ceramic thick film also may crack from the large displacement under low frequencies. Our device is targeted to be around 100~120 Hz, for the usage of city power equipment.
As for the interfacing circuits, synchronized switching technique are often ex- ternally powered or requires high voltage for self-powering. The external power sup- ply for the switching component often overrides the harvested power. Other than the regulating circuit, the complicated circuit within the self-powered switch also creates a remarkable component loss. Hence, the critical spec of the component loss factors should certainly be discussed to limit the power loss.
Additionally, the behavior of the ambient energy harvesting is random and the available converted power is in the range of mW, leading to the extremely low relia- bility for the power management circuits. Therefore, multiple harvesters from differ- ent sources not only enhance the reliability of the power management systems but al- so boost the available power.
When it comes to human motion of 0.5 to 1 Hz, we propose linear techniques featuring spiral or similar structures and nonlinear dynamic methods such as up con- version or bi-stable mechanisms for low frequency applications.