· · · Institutional Repository

This thesis brings together, for the first time, the fields of energy harvesting and static balancing. The proposal of two new architectures for the design of mechanical oscillators is supported by an extensive review on the existing energy harvesters. For the first one, a statically balanced oscillator, an analytically study proved it to be ineffective. This pushed for the development of a statically balanced frequency up-converter, that can integrate an energy harvester capable of coping with low frequencies vibrations of broadband nature. On the static balancing ground, a new mechanism is proposed, with the balancing of the folded suspension, a traditional mechanism of precision engineering. Numerical analysis suggests that high quality balancing is achieved for a large amplitude of motion. A preliminary study is also executed, introducing bond graph modeling to the field of energy harvesting. Bond graphs are a natural representation for the cross-domain nature of energy harvesters, allowing an integrative view.

Homes, offices and vehicles are getting networked. This will enable context aware, autonomous operation of many support systems that could be controlled remotely. To achieve this there would be a large number of tiny devices -- sensors and actuators -- which are networked and they are termed generally as Internet of Things (IoT) devices. In future, they will be powered through harvested energy from the ambience to enable perennial lifetime and minimal manual maintenance. Some examples of energy sources are photovoltaic panels and piezoelectric crystals. Several challenges arise due to the nature of sources of energy. One of these challenges is that the devices (nodes) leave and re-enter networks due to fluctuating availability of harvested energy. This energy condition requires the adaptation of special means at every layer of the communication model. For example, as a result of fluctuating energy levels, the neighbor table maintained at each node changes quite often leading to complications in forming and maintaining routes. In fact initial neighbor discovery (ND) itself is a difficult task. Further, usage of directional antennas would affect the time taken to complete ND. Given the spatio-temporal variations in energy availability in harvesting environments, there are benefits of energy prediction. With the help of prediction, resource allocation within a single system and splitting of tasks between nodes in a network would be enhanced. In order to identify the various parameters that affect ND we first describe a generic analytical model of an energy harvesting device. Next, we study a network of these devices through exhaustive simulation study considering these various parameters. We demonstrate the benefits and challenges of using directional antennas for ND. We present a scheme that nodes could use to discover their neighbors during initial deployment and another scheme that could be used for subsequent discovery on re-entry into the network. We show that a dedicated ND protocol is necessary for energy harvesting networks and that directional ND is beneficial in these networks under some circumstances. Finally, we present light-weight energy prediction solutions that can be used to improve the performance of the ND process in particular.

Information and Communication Technology (ICT) is changing the way we live and has become an essential part of our life. With the advent of Internet of Things (IoT), and Wireless Sensor Networks (WSN) in particular, the number of devices that are networked is increasing exponentially over the years. With the increase in number of ICT devices ambient-energy harvesting has become a necessity. It is expected to liberate networked embedded systems from the use of batteries. Batteries provide us the necessary energy density and are able to power systems for a few months. Nowadays, rechargeable batteries can last a few years, typically 1 to 3 years, before they become unusable. On the other hand, energy harvesters such as photovoltaic panels, thermo generators, vibration harvesters etc., are permanent sources of energy, however with the handicap of having an unpredictable power output. The main challenge that is considered in this thesis is the perennial powering of embedded system with energy harvesting devices. This requires novel hardware and software mechanisms to deal with the unpredictability of such energy sources. As a first step towards our goal, the power output of energy harvesters has to be characterized under realistic conditions and thus in-situ measurements are performed for use throughout the thesis. Following this, efficient power-conversion and -management using clever circuit design techniques is investigated to make energy harvesters usable. Since the power output from the harvesters is unpredictable, a good energy prediction algorithm is required to make adjustments and circumvent low energy regimes. One has to look at replicable and generally applicable energy harvesting solutions to cater to as many environments and applications as possible. This means that energy harvesters have to be modelled, and a practical system has to be designed within a simulation environment to conduct detailed studies for understanding various situations and possibilities. This thesis has to also address the problem that in any given environment, there is a potential imbalance of the available energy across a network of nodes, since the amount of energy harvested by a node at a given location can be different compared to another node at a different location. Thus software and hardware algorithms need to be devised to ensure that the network remains unpartitioned. If data from a wireless sensor network is directed at an aggregation node or a gateway, power requirements for this node become crucial. Moreover, if data needs be transported over an Internet link, a system to collect data has to be devised and energy efficient mechanisms have to be implemented on such gateways. The gateway also requires mechanisms to send information to end-users in the most energy efficient manner. Data security under no circumstance can be compromised and at the same time, other performance objectives such as throughput and delay guarantees have to be satisfied. Thus mechanisms have to be devised to support security. Throughput and delay requirements may differ from one type of embedded system to another and depend on the availability of the system resources. Throughput maximization schemes are required to support a large variety of embedded platforms. Further, policies that enforce a throughput and delay optimum have to devised taking into account energy harvesting. Finally, real-time support with increased reliability, by exploiting spatial node-diversity, is required to address some of the important challenges in energy harvesting embedded systems wireless sensor networks. The results in this thesis address the gaps in the current state-of-art and proposes novel power management algorithms driven by empirical measurements. Let us now list a few major contributions of this thesis. (1) The unique circuit design for RF energy harvesting at -18dBm shows that the embedded system can transmit sensor data by ``accumulate and use’’ topologies. (2) In a multi-node environment, channel sensing is well known to avoid packet collisions. However, the impact of detection modes (energy and signal type) on energy consumption shows a novel way to manage harvested power in the framework of improved packet reception. (3) A perennially powered multi-node multi-hop energy harvesting testbed is setup with an advanced state-of-art energy harvesting protocol stack built over well-known MAC schemes such as BoxMAC-2 and routing algorithms such as AODV. (4) The testbed runs a multi input parameter based power management optimized distributed algorithm that boosts application performance. One such input parameter is the accurate energy prediction based on time series. (5) Suitability of well known cryptographic algorithms are studied for embedded platforms with available system resources. Key insights are gained about the best suitable scheme that does not compromise security at any cost. (6) Novel algorithms are constructed from existing ``throughput optimal’’ and ``delay optimal’’ policies for a flexible performance in WSN nodes. (7) Novel and sophisticated cooperative relaying schemes optimized for maximizing the packet reception and minimizing the delay are proposed to supersede simple relay selection schemes. (8) Delay and disruption of data due to link connectivity problems is well known. However, solution to data disruption due to energy availability advances the state-of-art and shows it is possible to reliability transfer substantially high amount of data for a given quantum of available energy. (9) A well-known modelling that adheres to energy conservation principle is applied innovatively to model energy harvesting sources and embedded systems. (10) The circuit model for a supercapacitor advances the state of art in battery-less storage of energy harvesting systems. This thesis is an initial step towards networked systems that do not require external energy sources, but instead draw their energy from their ambient environment. Such networks are referred to as Zero-Energy Networks.

Direct piezoelectric strain energy harvesting can be used to power wireless autonomous sensors in environments where low frequency, high strains are present, such as in automobile tires during operation. However, these high strains place stringent demands on the materials with respect to mechanical failure or depolarization, especially at elevated temperatures. In this work, three kinds of ceramic–polymer composite piezoelectric materials were evaluated and compared against state-of-the-art piezoelectric materials. The new composites are unstructured and structured composites containing granular lead zirconate titanate (PZT) particles or PZT fibers in a polyurethane matrix. The composites were used to build energy harvesting patches which were attached to a tire and tested under simulated rolling conditions. The energy density of the piezoelectric ceramic–polymer composite materials is initially not as high as that of the reference materials (a macro-fiber composite and a polyvinylidene fluoride polymer). However, the area normalized power output of the composites after temperature and strain cycling is comparable to that of the reference devices because the piezoelectric ceramic–polymer composites did not degrade during operation.

This thesis discusses the design and implementation of a transmitter and a receiver for an FM-UWB scheme in CMOS. The main challenge is to reach a power efficiency of better than 10 nJ/bit at a data rate of 100 kbit/s for both the transmitter and the receiver design. This thesis also discusses the design and implementation of a power management unit for an autonomous wireless system. The challenge in this design is to obtain a high efficiency DC-DC conversion using the switched capacitor topology and low voltage ripple at the output voltage.

Harvesting energy from the users’ muscular power to convert this into electricity is a relatively unknown way to power consumer products. It nevertheless offers surprising opportunities for product designers; human-powered products function independently from regular power infrastructure, are convenient and can be environmentally and economically beneficial. This work provides insight into the knowledge required to design human-powered energy systems in consumer products from a scientific perspective. It shows the developments of human-powered products from the first introduction of the BayGen Freeplay radio in 1995 till current products and provides an overview and analysis of 211 human-powered products currently on the market. Although human power is generally perceived as beneficial for the environment, this thesis shows that achieving environmental benefit is only feasible when the environmental impact of additional materials in the energy conversion system is well balanced with the energy demands of the products functionality. User testing with existing products showed a preference for speeds in the range of 70 to 190 rpm for crank lengths from 32 to 95 mm. The muscular input power varied from 5 to 21 W. The analysis of twenty graduation projects from the Faculty of Industrial Design Engineering in the field of human-powered products, offers an interesting set of additional practice based design recommendations. The knowledge based approach of human power is very powerful to support the design of human-powered products. There is substantial potential for improvements in the domains energy conversion, ergonomics and environment. This makes that human power, when applied properly, is environmentally and economically competitive over a wider range of applications than thought previously.

The piezoelectric effect, which causes a material to generate a voltage when it deforms, is very suitable for making integrated sensors, and (micro-) generators. However, conventional piezoelectric materials are either brittle ceramics or certain polymers with a low thermal stability, which limits their practical application to certain specific fields. Piezoelectric composites, which contain an active piezoelectric (ceramic) phase in a robust polymer matrix, can potentially have better properties both thermally and mechanically than the present single-phase piezomaterials, and yet exhibit sufficient piezoelectricity. The two main objectives of this thesis were to find new routes for manufacturing piezoelectric ceramic-polymer composites with adequate piezoelectric properties while retaining ease of manufacturing and explore new applications using these composites. It was found that a composite with piezoelectric particles dispersed in a high performance polymer matrix possessed improved thermomechanical properties such as thermal stability and temperature dependence of the output signal compared to conventional materials. The added flexibility of piezoelectric composites leads to a greater choice in device geometry and compatibility with polymer processing. Dielectrophoretic (electric field assisted) processing of piezocomposites results in increased permittivity and piezoelectric constants compared to conventional piezoelectric composites with randomly oriented particles, at little added processing complexity. A high degree of alignment could also be achieved using elongated piezoelectric particles, with a significant improvement in piezoelectric properties as a result. As an example, dielectrophoretically processed composites were used as energy harvesting materials in automobile tyres. Energy harvesting materials convert a small amount of energy from the environment to electricity, which can be used to power electronics such as wireless sensors. In automobile tyres, low frequency, high strains are present from which electricity can be generated using the piezoelectric effect. However, these high strains place stringent demands on the materials with respect to mechanical failure or depolarization, especially at elevated temperatures, so conventional piezomaterials are not robust enough to use. After testing inside automobile tyres the energy density of the piezoelectric ceramic-polymer composite materials is comparable to the reference materials and the output of the new piezoelectric ceramic-polymer composites did not degrade during operation. Therefore, these new materials open up new application fields for piezoelectric sensing and energy harvesting.

Integration of sensors and wireless transceivers for system networking aims at emerging applications that are highly integrated, self-powered, and low cost, relying on efficient power management schemes to prolong lifetime, thus eliminating the need for batteries as a limited primary source of energy. These applications include: health monitoring and body-area networks, environment and infrastructure monitoring, security, and building automation. Autonomous wireless transmitter sensor nodes that harvest solar energy outdoors and light energy indoors via a solar cell or photovoltaic module are developed for ultra-wideband 3-10 GHz communications. Photovoltaic (PV) antennas are used as both DC power source and radio-frequency (RF) radiator or receptor. Sharing the area consumed by the primary DC power source with the antenna reduces the size and overall cost of the transceiver, resulting in a smarter integrated wireless system. The overall package consists of a temperature sensor, a supercapacitor or a rechargeable battery, DC power management, digital signal processing, analog, RF circuitry, and a PV antenna. One transmitter node uses a single solar cell behaving as a broadband monopole antenna to generate up to 20 mW-peak power outdoors and consumes an average power of 10 µW when transmitting 1 kbps every minute. Other sensor nodes employ a PV dipole antenna for outdoor applications and a flexible PV loop antenna for in an indoor environment, transmitting data packets at 10 kbps with average power consumptions of 15 µW. These sensor nodes are light weight (< 10 g), suitable for portable and wearable applications.

Studies performed in the last decade have shown that wave energy could contribute as much as 10% in the current world electricity demand, making the ocean one of the most underrated and unexploited renewable sources so far. Wave energy offers distinctive advantages compared to other renewable technologies, providing a constant and predictable source of energy with minimal environmental impact. Although research on this topic has just been reinitiated, the devices that have been developed in these programmes typically suffer from the operation conditions in the marine environment. A new approach in this field is to use deformable materials, which should result in mechanical structures with improved reliability and survivability. An innovative new concept is to use smart materials in these deformable structures, which eliminate the need for additional generators and the typically related moving parts. This research addresses the application of Dielectric Electro Active Polymers (DEAP) in wave energy applications, and focusses on the electrical aspects involved in the energy conversion. Electrically, the EAP structure is represented by a variable capacitor, which capacitance is a function of the mechanical deformation. It is found that the energy conversion of an ideal EAP film is maximized when charging and discharging is performed infinitely fast at the maximum and minimum capacitance, respectively. An intermediate constant electric field stage results in the electromechanical conversion process. An optimized current waveform is investigated that results in the optimal energy cycle, maximizing the electrical energy output of the DEAP structure. For this purpose, the current amplitudes and the amount of charge left on the film are optimized in accordance with the material and excitation properties. It is found that even with poor material properties, reasonable energy output can be obtained by applying an optimized energy harvesting cycle. The energy cycle optimization has shown that energy harvesting in applications with low deformation ratios requires an highly efficient Power Take Off system (PTO). The specifications of this PTO converter have been derived using the parameters of a given test setup and the corresponding optimal energy cycle. In accordance with the purpose of the converter, a basic Buck and Boost topology has been selected, using a Zero Voltage Switching-Clamped Voltage (ZVS-CV) switching strategy. A single high-voltage switch configuration has been chosen for complexity and reliability reasons. Based on the PTO design procedure, it can be concluded that the high voltage and typically low current application challenge the design of the converter. To investigate the converter efficiency and characterize the losses, an accurate IGBT model has been deployed in the Saber circuit simulator. However, it has been found that the current tail effect is not modeled appropriately for the ZVS-CV conditions. Especially for high-voltage switches, the observed current tail is significantly increased under soft-switching conditions. Therefore, the design has resulted in poor efficiencies. Based on these findings, it can be concluded that topologies with a single high-voltage IGBT switch are not able to operate efficiently under the given conditions. Therefore, it is recommended to extend the research to multi-level topologies or series connected switch stacks using low voltage switches.