Magnetostrictive vibrational power generator for battery-free IoT application

The Internet of Things (IoT), which refers to the interconnected network of physical devices, exchanging data via electronics and sensors, is considered the next level of connectivity that can facilitate greater ease of living. One of the biggest challenges facing IoT is the power supply for wireless modules utilized in IoT applications. Recently, various low-power wireless modules such as those based on Low-power wide-area (LPWA) networks and Bluetooth Low Energy (BLE) have been developed; however, their power supplies rely on the button cell or dry cell, which forms a major limitation to the spread of IoT technology. Such cells need to be replaced over the course of time, and their management and maintenance over their lifetime is troublesome even though such cells have a long life (e.g. three years). If the number of modules is 100 or 1,000, the corresponding labor and costs cannot be ignored. Thus, for the 50 billion items covered by IoT, it is impractical to use cells for power supplies. In this regard, research has indicated that energy harvesting can solve the abovementioned problem. Energy harvesting utilizes technology that generates electricity from ambient-energy phenomena such as light-and-temperature differences and vibrations, and in particular, vibration-based power generation has been considered useful and practical for IoT. Because vibrations form a part of the working environments of production machinery, automobiles, and infrastructure such as roads and bridges in addition to the movement of people and things can be used as sources of power generation. In this regard, I have been developing vibration power generation technologies based on magnetostrictive materials. In 2010, I proposed a device based on a parallel beam structure comprising Fe–Ga alloy¹ and a magnetic material,² and demonstrated the functioning of several such devices that can yield output power of the order of milliwatts.^3,4 However, these devices suffer from the problem of lack of durability and are further not suitable for mass production. Consequently, here, I propose a novel structure based on a lamination (unimorph) of Fe–Ga alloy and magnetic material. Such a device is highly practical since it is simple to fabricate and affords robustness and high sensitivity, thereby overcoming the problems of conventional cells. Thus, this device completely overcomes the battery problem of wireless modules while providing higher power with a semi-permanent life. This paper describes the structure of the device, principles of power generation, evaluation results, and two applications of a battery-less wireless sensor based on the proposed device.

Let us first consider the conditions under which the vibration power generation device must substitute for cells. In order to exceed the performance of a button cell (CR 2032) with a lifetime of 3 years, the average power of the device must be more than 20 μW, and it must have a semi-permanent lifetime. Considering that the efficiency of a standard power conversion circuit (AC to DC) is about 10%, the output of the device is required to be at least 200 μW, which is 10 times larger than the input to the module. In addition, machines are generally made so as to not vibrate, and thus, the device (which is attached to the machine of interest) must have high sensitivity to generate an output when subjected to minute vibrations. The proposed power generation device based on Fe–Ga alloy satisfies all the above conditions and can easily substitute for cells.

Figure 1 top shows the structure of the device.⁵ The device consists of a U-Shaped magnetic frame, with one leg of the frame acting as a unimorph, wherein an Fe–Ga alloy plate (easy magnetization axis along the longitudinal direction) is bonded onto the upper part of the frame. Further, a coil is wound around the unimorph, and a permanent magnet for bias is placed inside the frame between the two legs. As shown in Fig. 1 bottom, the device is affixed to a vibrating object and is equipped with a proof mass positioned on the tip of the shorter arm. The Fe–Ga alloy is appropriately magnetized by the magnetic bias. When this device vibrates along the vertical direction, an inertial force acts on the mass and leads to bending of the unimorph, and subsequently, tensile and compressive stresses occur alternately along the longitudinal direction of the alloy. At this time, the magnetic flux increases and decreases due to the inverse magnetostrictive effect, and an electromotive force corresponding to the time-varying change in the flux linkage occurs in the coil. In general, the phenomenon of resonance is utilized to maximize power generation. When the frequency of the vibrating object and the resonance of the device are coincident, the device deforms to a large extent, and a high voltage is generated.

I focus on the features of the device. Because of the long leg of the U-shaped frame, the small inertial force exerted by the proof mass is magnified via the leverage principle and acts on the Fe-Ga alloy. Further, the large magnetic flux change caused by the stress flows through the closed magnetic circuit in the frame that acts as the yoke. This structure and magnetic circuit contribute to the high energy conversion efficiency of the device. Furthermore, sufficient winding space is available for the coil, and the output is greatly improved over the conventional parallel-beam type structure. Importantly, the device is structurally simple and suitable for mass production. The device durability is high. The device lifetime is determined by metal fatigue of the frame and peeling of the bonding. If the bond is of sufficient strength and the displacement is limited, these factors will not significantly affect the device functioning and lifetime.

A prototype device as shown in Fig. 2 was tested to verify its output characteristics and durability.⁶ In the device, the Fe–Ga plate had dimensions of 4 × 0.5 × 16 mm and was cut from a single-crystalline bulk Fe_81.4Ga_18.6 ingot produced by Fukuda X'tal Laboratory via the Czochralski process.⁷ The frame was wire-discharge machined from a 0.5-mm plate of cold-rolled Fe and subsequently bent. The Fe–Ga plate was affixed to the frame by use of an epoxy-type structural adhesive. The coil had a wire diameter of 0.05 mm with 3500 turns (520 Ω), and the magnet was a Nd–B–Fe magnet with dimensions of 4 × 3 × 2 mm. The weight of the device was 4 g and its volume was 0.5 cc.

Figure 3 shows the voltage waveform for a proof mass with a weight of 1.7 g, sinusoidal wave oscillations at a frequency of 88.7 Hz, and accelerations of 3.0, 6.0, and 9.0 m/s². The maximum voltage of 5.4 V is generated at the acceleration of 9.0 m/s². The relationship between the displacement of the tip of the frame and the change in the magnetic flux density of the plate, as calculated by the time integration of the voltage waveform, is shown in Fig. 4. A magnetic flux density change of 1.3 T, equivalent to that of the conventional parallel-beam type,^4,8 is observed to occur. Under an acceleration of 6.0 m/s², the relationship between the resistance and power (Joule loss) is measured as in Fig. 5. A maximum power of 5.1 mW and effective power of 2.0 mW are obtained at approximately 1000 Ω.

Next, the voltage waveforms for the weight of 10.2 g, frequency of 28.4 Hz, and accelerations of 0.34, 0.49, and 0.61 m/s² are shown in Fig. 6. A voltage of 1 V is generated at an acceleration of 0.49 m/s², thereby indicating that a practical working voltage can be extracted with minimal vibration. Figure 7 illustrates the relation between the resistance and electrical power at an acceleration of 0.75 m/s². A maximum electrical power of 1.05 mW and effective electrical power of 0.39 mW are obtained at a resistance of approximately 520 Ω.

Next, fatigue tests were conducted. Without a mass on the tip, the device was vibrated 0.1, 1, 10, and 100 million times at a resonance frequency of 420 Hz and acceleration of 78.4 m/s² at room temperature (25 °C). Figure 8 shows the results of the fatigue tests. We note that the frequency characteristics remain almost unchanged from the first cycle to even after 100 million cycles. Importantly, neither the separation of elements nor metal fatigue was observed after testing.

Two types of battery-free wireless sensors were fabricated using the proposed power generation device. One sensor was fabricated with the aim of realizing short distance wireless communication when subjected to small movements. In this case, the device is fixed on an acrylic plate, and the tip is flicked (resulting in a free vibration). The system as shown in Fig. 9 consists of the device, a voltage doubler rectifier circuit and a wireless communication module. The rectifier circuit contains a Schottky diode and a 33-μF electrolytic capacitor, and the wireless module is the IM 315 TX (315 MHz, Interplan Co., Ltd., maximum 50 m as per forecast) module. In the wireless module, when the specified voltage of 1.9 V to 3.6 V is provided, the signal is set to be transmitted. Figure 10 shows the terminal voltage of the device and the input voltage of the module, wherein we note that communication is achieved with a single-shot operation i.e. voltage more than 1.9 V is supplied to activate the module.

We next demonstrate a second battery-free wireless sensor powered by normal vibrations, which can transmit sensor data over long distances at fixed time intervals with the use of LPWA technology. As shown in Fig. 11, the system consists of the device, a rectifier circuit, storage module, DC/DC converter, microcomputer, temperature sensor, and wireless module. The rectifier circuit contains four tandem voltage doubler rectifier circuits, with a 10-mF aluminum electrolytic capacitor being used for storage. The step-down-type LTC 3588 (Linear Technology Corp.) is used as the DC/DC converter, the STM 32 F 303 CBT 6 (STMicroelectronics N.V.) functions as the microcomputer (MCU), and the IM 920 (920 MHz, Interplan Co., Ltd., maximum 5 km as per forecast) forms the wireless module. The transmission data of interest are the temperature and vibration frequency, wherein the temperature is detected by the LMT 86 LP, and the frequency is extracted from the generated voltage by counting cycles from the voltage wave form (using no sensor). For the vibration of 99 Hz and 2.94 m/s², wireless communication was successfully achieved in intervals of about every 3 min. Figures 12 and 13 show the device terminal voltage and capacitor voltage, respectively, wherein energy is consumed by data processing and wireless communication and then charged over 3 min for the next cycle of operation. This system works when the generated (open-circuit) voltage of the device is more than 1.2 V, and it is confirmed to operate even under a vibration of 30 Hz, 1.96 m/s².

This study proposed a simple, robust, highly sensitive vibration power generation device based on the unimorph of an Fe-Ga plate and a U-shaped frame. The prototype device was of the same size as the conventional battery, and it was confirmed that the device yields milliwatt-order electrical power when subjected to vibrations in the frequency range of 30 Hz to 90 Hz. Further, it has a semi-permanent lifetime, as evidenced by vibration tests of 100 million cycles. As a practical application, I fabricated short-distance and long-distance wireless sensors powered by the device. As regards the former sensor, I succeeded in transmitting ON signals in a single-shot operation, while sensor signals could be transmitted once every several minutes with the latter sensor. The vibration power generation can substitute for batteries and contribute to the dissemination of IoT technology while affording maintenance-free operation. The device applications include monitoring, crime prevention, disaster prevention, and predictive preservation of machinery and infrastructure. Several companies have begun to develop IoT products, all of which can utilize the proposed battery-free device.

This work was supported by JST CREST (Grant Number JPMJCR15Q1), Japan. The fatigue tests were conducted at Toyo Tire & Rubber Co., Ltd. The prototype of the wireless sensor system (long-distance type) was fabricated by Tokyo Drawing Ltd.