Identification of Key Energy Harvesting Parameters through Monte Carlo Simulations

As the number of embedded systems increases, so do the demands placed upon them. Current algorithms are capable of creating adequate schedules under ideal conditions, but can be considered inadequate when many variables must be regarded. End users are demanding ever greater performance while minimizing failures and power consumption, meaning advanced power management must be incorporated into circuit designs, especially in multi-core environments. This paper summarizes the initial investigation into the simulation of an energy harvesting system to identify key parameters. This is done by initial multi-vary analysis to determine primary contributors, which are then refined through Design of Experiments (DoE). Finally, the reduced model is subject to control through Statistical Process Control (SPC) to confirm whether monitoring causes a statistical difference to the output reliability and if so, what parameter has the greatest effect.