NearLimits

The NearLimits project contributes to the development of kinetic energy harvesting by proposing a technique called "Near-Limits Energy Harvesting", which is based on the active control of the electromechanical dynamics of the system. This is a break with conventional kinetic energy harvester design approaches. Our technique is particularly suited to power biomedical devices (portable or implanted) by recovering energy from human body movements. These movements are characterized by irregular vibrations and/or very low frequencies (< 10Hz). Conventional kinetic energy harvesting systems, based on linear or non-linear resonators, are not very efficient in this context. Indeed, such systems are often optimized for canonical vibrations, such as harmonic or white Gaussian noise. The approach proposed in this project targets a gain of one to two orders of magnitude in terms of the amount of power converted by targeting a power density of 100 µW/cm3 for typical vibration profiles. The associated design method is firmly oriented towards the specificities of the vibrations of each application. This project aims at proving the feasibility of this "Near-Limits" approach for electromechanical energy conversion from low frequency and/or irregular ambient vibrations. It intends to show that an active synthesis of the trajectory of the proof mass, as required by the "near limit" principle, can be achieved by an adiabatic control circuit, which should itself be appropriately actuated. The physical implementation of this active synthesis is based on the control of the dynamics of the moving mass by a differential electrostatic transducer fabricated in MEMS (microelectromechanical systems) silicon technology, which also integrates the sensing functions needed by the control law. In addition, the design of the control circuit + transducer system implementing the "near limit" approach is made in close connection with the vibration characteristics of the targeted application. Thus, the vibration characteristics that determine the performance of the "near limit" harvesters will be identified, as well as their impact on the sizing of the system. This will allow to restrict the class of targeted vibrations to those for which the "near limit" concept is fundamentally superior to conventional approaches. There are three main scientific and technical challenges: - The search for a conditioning circuit whose dynamics result in generating the “near-limits” control law on the moving mass, while ensuring adiabatic energy transfer. This is a challenge because the problem is not generally posed in these terms in classical control, where the control law is sought relatively independently of its future hardware implementation; - The control circuits will have to withstand the high voltages inherent to the physics of miniature electrostatic transducers (typically of the order of 100 V), while consuming little power. This aspect is an additional constraint in the choice of the circuit and thus of the control law; - The architecture of the MEMS device must simultaneously include a differential transduction structure and a position measurement block, resulting in complex structures. Our objective is to build two prototypes over the course of the project: one based on a 3D printed macroscopic mechanical device, the other on a MEMS device realized in a thick-layer silicon technology. The device and its control electronics will be sized to target a pre-identified realistic type of vibration (e.g., human body motion). The complete system will then be tested to experimentally prove the advantage of the "near limit" approach in this case.