Efficient Physical Modeling of MEMS Energy Harvesting Devices With VHDL-AMS

Abstract

International audienceIn this paper, we propose a VHDL-AMS implementation of a physical model of a microelectromechanical systems (MEMS) piezoelectric microgenerator. Such an executable model acts as a bridge between specifications and fabricated devices. Usually, physical and geometrical parameters of electromechanical parts of a system are only considered in lower levels of the design flow, typically using finite-element tools, which, despite their accuracy, do not allow efficient optimization of the structure properties and dimensions. Thus, it would be very interesting to have a model of the entire harvesting system (the MEMS piezoelectric microgenerator cascaded with the electronic circuit) to perform efficient optimization. Some features like damping effects and process fluctuations have considerable impact on the performance of MEMS, especially the resonant structures. We propose a method of integrating such features early in the design flow, while keeping the simulation time reasonable. The resulting model is reusable, predictive (comparable to experimental results) and respects Kirchhoff laws. Consequently, it can be integrated in global simulation of multidomain and mixed signal systems like wireless sensor nodes

Automotive VHDL-AMS Electro-Mechanics Simulations

“…A further up-to-date example is in F. Gao & Moudni (2010), where a multidomain dynamic proton-exchange-membrane fuel-cell stack model is described and simulated using VHDL-AMS. In (H. Boussetta & Soudani, 2010) a physical model of a microelectromechanical system piezoelectric microgenerator and its controlling circuit is presented and their interaction analyzed. This capability is in fact particularly suitable to accurately represent automotive electro-mechanic components.…”

DoE-based Performance Optimization of Energy Management in Sensor Nodes Powered by Tunable Energy-Harvesters

“…The complexity leads to prohibitive simulation times. Models based on hardware description languages (HDL), such as VHDL-AMS and SystemC-A have recently been proposed [2,3] to accelerate simulation times. However, these solutions have not fully solved the main challenge of the traditional analogue simulation approach based on Newton-Raphson iterations which are the main cause of the long CPU times.…”

Modeling of Wireless Sensor Nodes Powered by Tunable Energy Harvesters: HDL-Based Approach

“…Pending the development of such a solver, in this paper we have used SystemC-A [14], a proprietary extended SystemC which has recently been equipped with an analog solver that uses a linearised state-space equation formulation [15], capable of simulating efficiently intricate operational scenarios of an energy harvester system. Recently, Boussetta et al [16] reported a VHDL-AMS model of a MEMS (microelectromechanical system) piezoelectric microgenerator. Their model incorporates the physical and geometric parameters of the microgenerator, is reusable, reflects experimental results well and can be integrated in global simulation of multidomain and mixed signal systems such as complete wireless sensor nodes.…”