The design of a microscale energy harvester from multiple sources is the main motto behind the present work. The load is a capacitor that must be charged to 1.8V where the system stops working as soon as the capacitor is charged to the desired voltage. Scope for improving the system design in various modules was described at the end of the thesis.
Energy Harvesting
Hybrid energy Harvesting
Objective of work
Literature review
The second chapter gives a detailed description of all the modules in the designed multi-source, multi-mode hybrid energy harvesting system. For a hybrid energy recovery system, more than one inverter must be connected on the input side. The work carried out mainly focuses on the design of a hybrid energy harvesting system that uses TEG and Solar as input sources.
System Architecture
Input source selection
Maximum power point tracking (MPPT)
As known from circuit theory, maximum power is supplied by a source with internal resistance to a load when the resistance of the load is equal to the internal resistance of the source. For solar, the maximum power point will shift depending on the intensity of available light, while for a TEG source, the maximum power point will vary based on the change in temperature difference between the plates of the transducer. For the current system, with the TEG input voltage range between 10 mV-700 mV, the change in the internal resistance of the transducer will not be huge, allowing us to assume that the resistance value is a constant of 3.9 ohms. MPPT for TEG source can be achieved by following the equation below.
Rteg is the internal resistance of the TEG source, fs is the switching frequency of the converter, L is the inductor value in the converter. As mentioned above, the internal impedance of the load must be equal to the impedance of the load to achieve maximum power transfer from source to load. In an energy harvesting system, the impedance of the converter, as viewed from the source, depends on the frequency at which the converter switches.
The above formula provides a relationship between the internal frequency of the transducer and the switching speed of the converter. To achieve MPP for the TEG source, the frequency of the converter must be determined based on the resistance offered by the source and the inductance value of the converter. The MPPT for piezo is performed via the Fractional Open Circuit Voltage (FOCV) method, in which the transducer is disconnected from the converter when the transducer voltage falls below half of the open circuit voltage.
An analog comparator always compares the output voltage of the piezo transducer to 0.5 times its open circuit voltage and outputs a HIGH/LOW signal which is fed to the next stages of the system.
Power converter
Inductor-based DC-DC power converters can be boost, buck, or buck-boost depending on the input and output voltage values. A buck-boost converter operates in both boost mode and buck mode, depending on the voltage available at the input and the voltages required at the output. When the high side switch is ON, the voltage across the source, in addition to the stored energy in the inductor, is applied across the load, causing the load voltage to exceed the source voltage.
The impedance that the converter provides to the source depends on the switching frequency of the converter and values of other elements in the converter. The impedance provided by the inverter depends on the switching frequency and the values of the other components in the inverter topology. Buck-boost converter performs both boost and buck operation depending on the voltage requirements of the load and the available source.
For the hybrid energy harvesting system designed, the converter is a buck boost converter working in DCM mode. The output capacitor must be charged to a voltage of 1.8V, which clearly indicates the need for a buck-boost converter. The high side switch is a PMOS because it must pull the node up to higher voltages, while the low side switch is an NMOS because it must pull the node down to 0V. The frequencies at which both switches operate vary depending on the input connected to the converter ie.
In the boost converter shown in the schematic view above, switches S1 and S2 are power switches that connect either the TEG or the piezo depending on source availability and inductor current.
Digital controller
When the temperature difference is smaller, the output voltage across each source is also less and the three sources are connected in series there by increasing the input voltage, resulting in an increase in input power. When the sources are connected in series, the impedance of the three sources is increased to 11.7 ohms, resulting in a switching frequency of 38 KHz for the DC-DC converter to track the maximum power point. Similarly, when the sources are connected in parallel, the impedance offered by the TEG array is 1.3 ohms, resulting in a switching frequency of 4 KHz for the DC-DC converter to track the maximum power point.
Clocks at these 50 percent duty cycle frequencies must be provided to the high-side and low-side switches of the boost converter when the TEG array has sufficient current to power the converter. The inductor is charged during the first half cycle of the clock signal, while it is discharged and remains at zero (called the inductor current dead time) during the second half cycle of the clock signal. The above values for the clock frequencies for the two sources TEG and PV clearly.
The idea of using the dead time of the inductor current after inductor discharges its stored energy from TEG source for piezo source is the main concept of the design of the current hybrid energy harvesting system. TEG source has high priority over piezo source due to its high energy densities. Piezo source can operate in boost or buck mode depending on its input voltage and the clock signals are generated from MOS driver logic block. If the output of the comparator is high, the capacitor has not yet reached 1.8V and so the system should work.
System operation is aborted by stopping the free-running clocks for TEG (38 KHz/4 KHz for series/parallel) and Piezo (200 KHz).
Load
Experiment to verify the startup mode, it can be seen that on startup, the system starts at 10ms, whereas the system is unable to start with ramp VDD without startup. The table below provides an overview of various conditions under which the system must operate. The simulation results mentioned below have the waveforms of the inductor current and the mode in which the system operates, ie.
The graph above is the results for the system operating when the TEG source has minimum voltage and Photo Voltaic has no current to supply the load. From the graph above it appears that the TEG voltage across a single source is 20 mV and the system is working. That is, when the output voltage reaches 1.8 V, the input voltage reaches the no-load voltage, because the system does not draw any current from the source.
The diagram above is the results for a system operating when the TEG source has a maximum voltage of 200 mV and the Photo Voltaic has no power supply to power the load. From the graph below, it can be concluded that without TEG power and with minimal PV power, the output voltage is charged to 1.8 V. It can also be seen that the PV source is at MPP when the system is operating. It can also be seen that the PV source is at MPP when the system is operating.
It can be clearly seen that the PV is at MPP until the time it supplies power to the system. The load charging to 1.8V can also be seen from the graph. The efficiency of the system for different cases has been calculated and the information has been mentioned in the form of tables. From the above tables, it can be deduced that the system has a maximum efficiency of 76.65% when both TEG and PV are operating under maximum current conditions.
Zero Current Switching Comparator
LDO for Multiple VDD
TEG having minimum voltage (20mV) while PV doesn’t have power 45
TEG doesn’t supply any power while PV supplying Minimum power 47
From the graph below, it can be concluded that without TEG power and with maximum PV power, the output voltage is charged to a voltage of 1.8V. Average input power 23.68uW Average output voltage 1.8V Average output current 6.3uA Average output power 11, 34uW. Lower system efficiency in other cases is due to power converter losses and high energy consumption in the digital controller block.
Efficiency can be improved by optimizing the power converter design for switching and conduction losses, calculating comparators and other methods. Comparisons from previous works on hybrid power systems with the designed system can be seen from the comparison table. The single PV and TEG sources can be replaced by a multi-source PV and TEG array in order to increase the input power.
Kuo Three-Switch Single-Inductor Dual-Input Recycled DC-DC Converter with 93% Peak Conversion Efficiency and 0.5mm2 Active Area for Light Energy Harvesting IEEE International Solid-State Circuits Conference - (ISSCC) Digest of Technical Papers, San Francisco, CA, 2015, p. 8] Claude Vankecke et al., “A Multi-Source, Battery-Free Power Harvesting Architecture for Aeronautical Applications”, IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 9] Yen Kheng Tan et al., “Energy Harvesting from Hybrid Indoor Light and Thermal Energy Sources for Enhanced Performance of Wireless Sensor Nodes,” IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL.
10] Michele Dini et al.,” A Nanocurrent Power Management IC for Multiple Heterogeneous Energy Harvesting Sources”, IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL.
