Vibration energy harvesters are expected to become a new source of electrical power. Piezoelectric vibration energy harvesters that employ a piezoelectric transducer, a rectifier, and a storage capacitor are being used widely as electro-mechanical harvesters. Synchronized switch harvesting on inductor enhances harvesting performance due to employing a simple additional circuit and incorporating suitable switch control functionality. Switching is usually based on the displacement of a vibrating structure; hence, sensing the vibrational states is of critical importance. Conventionally, the structural displacement is measured by displacement sensors or accelerometers attached to the target vibrating structure. Although enhancement of performance through synchronized switch harvesting on inductor equipped with sensors is important, the arrangement requirements of sensors have adverse effects on the compactness and usability of the harvesters. This study aimed to eliminate the use of sensors from switch-controlled harvesters. We developed a new state estimation method that uses the piezoelectric transducer’s voltage as an observation value. Using the proposed state estimation method, the modal state values of the vibrating structure can be determined by simply measuring the voltage of the transducer. With the switch device being controlled by the estimated modal state values, no sensors are required for ensuring effective harvesting. A comparison of the harvesting performances by the proposed self-sensing state estimation method and the conventional sensor-equipped state estimation method showed that there is little difference in harvested power between the two methods over a wide range of load resistances. The proposed method is superior to the sensor-equipped method in terms of compactness and usability as it does not require any external sensors.

This study aims to increase the amount of electrical energy harvested from a piezoelectric vibration energy harvester under unloaded and high-load resistance conditions. Although increased piezoelectric charge due to the synchronized switch harvesting on inductor (SSHI) strategy damps mechanical vibrations, the mechanical vibration amplitude of a mechanical element in a harvester is assumed to be constant for most discussions regarding the active harvester with SSHI strategy. However, this assumption is not valid under excessive switching actions, in which case the performance of the harvester deteriorates. This problem is known as the vibration suppression effect. To address this problem, in this study, two switching strategies for the charge inversion circuit—namely, switching considering vibration suppression-threshold (SCVS-t) and adaptive SCVS-t (ASCVS-t)—are proposed through intermittent switching actions. During the harvesting process, intermittent switching using these strategies is performed based on the output voltage threshold, thus maintaining high mechanical vibration amplitude and excellent harvesting performance by avoiding excess switching. The ASCVS-t adopts a tuning algorithm for the time-varying threshold and can achieve appropriate intermittent switching and effective harvesting under various vibration conditions without pre-tuning. Experimental comparisons with conventional strategies confirm that the proposed strategies achieve 2.9 times and 2.0 times greater harvested energy storages than a standard harvester and conventional switching strategy, respectively.

We propose a novel semi-active vibration suppression method based on model predictive control (MPC). Semi-active vibration suppression provides excellent damping performance, energy consumption, and stability during control. As the semi-active control input is often discontinuous, it may be difficult to predict. Hence, we combine semi-active vibration suppression and MPC to determine the control input trajectory arbitrarily. The proposed method, called predictive switching based on piecewise constant input (PSPCI), assumes that the piezoelectric charge remains constant when the control circuit is in the open state. Under this assumption, the future system state can be predicted for semi-active vibration suppression while reducing the computational load. The PSPCI method predicts the future work done by the transducer and effectively suppresses vibrations. Its effectiveness and robustness are demonstrated through simulations and experiments. The proposed PSPCI method enables the prediction of the semi-active control input and diversifies the control input determination for effective semi-active vibration suppression.