Solar Energy Based Self-Sustainable Power Management System

I also want to thank the researcher Rizwan sir for his guidance in the layout. Frequent battery replacement is a major challenge faced in wireless sensor nodes (WSNs) operating in remote locations due to high maintenance cost and inaccessibility. The battery is charged by solar power during daylight hours and when solar power is not available, the system is powered by the charge stored in the battery.

A self-sustaining high-power battery management system has been developed where the system operates in two modes: the first is the main converter mode and the second is the boot mode. Most of them are used in remote places and it is difficult to access the system and replace the battery regularly.

Energy Harvesting

Solar Energy Harvesting

Despite being available during most periods, solar energy collection is carried out on both a micro and macro scale. Other major challenges with these systems are the design of the start-up circuit and the MPPT technology to extract maximum power from the solar cell. Overwhelming research has been done in this area, allowing the system to operate at a point commonly referred to as the maximum power point (MPP) of the solar cell.

Figure 1.1: Current-Voltage characteristics of solar cell MP3-37 at 100% solar irradiation of [3].

Objective of Work

Henceforth technique for extracting maximum power is referred to as MPPT (Maximum Power Point Tracking) and the various prevailing techniques used for the same are discussed. The hill-climbing method calculates the MPP directly by measuring the gradient of the output power as a function of the output voltage. Disturbance and observation constantly disturb the system to detect the MPP and the panel operating voltage fluctuates around it.

The increasing conductance method is based on the fact that the slope of the solar cell power curve is zero at MPP, positive to the left of MPP, and negative to the right. The fractional open circuit voltage (FOCV) technique exploits the existing quasi-linear relationship between the MPP voltage (VM P P) and the open circuit voltage (VOC) under different irradiance and temperature levels, in which the input voltage of the solar cell is adjusted to a reference voltage which turns out to be 0.7×VOC, to obtain maximum power. Since VOC changes with radiation, we have to do periodic sampling to update the new Voc value for which we have to turn off the converter which is again a waste of energy.

Thesis Organisation

Section 1.1.1 explains how, among the available energy sources, solar is the best choice to meet the increased energy demand of WSN, Chapter 2 discusses the need of DCM mode of operation and also the selection of components.

Continuous and Discontinuous Conduction Modes

In DCM, however, all the energy in the inductor is transferred to the load during this second interval. The third interval begins when the energy in the inductor is depleted and continues until the inductor begins to charge again in the next switching period. Since all the energy in the inductor is discharged in each switching cycle during the breaker OFF time, for driving the same load, the peak inductor current (IL−pk) must be higher in DCM than in CCM.

To keep the inductor current stable at a value, CCM requires high frequency switching compared to DCM. The load is pulsating load, i.e. for most of the time it will be in sleep mode and draw less current in the order of mA. If the converter is operated in CCM for less current, very high switching frequency is required.

Therefore, by operating the system in DCM and selecting the correct switches, all losses can be minimized.

Figure 2.2: Representative CCM waveforms of [6].

Component Selection

Inductor Considerations

Output Filter Capacitance

Power Switches

Introduction

It is not possible to run the load directly from the combine, as the input power is non-deterministic in nature. So a battery is required to collect the maximum power from the combine at any given time. This harvested power with the regulator is supplied to the load as per the load requirement.

Since these storage devices (battery, solid state cell, supercapacitor) have certain under and overvoltage limits to ensure proper lifetime operation, undervoltage and overvoltage protection is also present in the system which is collectively referred to as battery management . Also, one of the essential blocks is the cold start which is required when the battery is completely discharged. Since the entire system is powered by the power itself stored by the storage device, so in a scenario where the storage device has no load, the cold start will become functional as it is directly supplied from the harvester input and charges the storage device to some level from where the main converter will skip the charging operation.

The main converter includes all circuits except the starting circuit and is usually very efficient as a starting circuit.

Energy Harvesting System Architecture

Boost Converter
Regulating Solar Cell Potential at Maximum Power Point
Current Sensor
Limiting the inductor current between 0 A and 1 A
The Requirement of the Free Running Clock When V SOLAR > V C ST OR
Digital Controller
Battery Management
Auxiliary Circuits

In the DCM mode of operation, zero-current switching is one of the critical issues or bottlenecks that must be handled carefully, as negative inductor current, as mentioned above, leads to loss of harvested energy and thus degrades system efficiency to a much greater extent. The methodology used to prevent negative inductor current in this part is explained later in this section. The coil current can rise up to a preset value, denoted ISAT, after which M2 is OFF and M3 is ON.

Now the charged inductor current is discharged to the storage capacitor CST OR through the PMOS switch M3 until the inductor current drops to zero. The charge slope for inductor current is given by VSOLAR/L and the discharge slope is given by (VSOLAR−CST OR)/L (negative since for boost converter VOU T > VSOLAR) and in this case since VSOLAR is regulated to VM P P which will be discussed later in this section, and the CST OR charges from 2.4 V to 5 V. The charge slope is relatively constant compared to the discharge slope, which increases gradually as the CST OR is charged. The inductor current is limited to a fixed peak value ISAT (in our case fixed at 1 A), determined based on the inductor used (47µH in our case) and design specification for maximum power (500 mW in our case).

As the current decreases and crosses zero, if the PMOS switch is not off, then to maintain the continuity of the current, the inductor current flows negative which means that current flows from the CST OR to the inductor which the timing switch node will be at voltage less than CST node voltage OR. By monitoring both switch nodes and the CST OR voltages, it can be deduced whether the inductor current is positive or negative. Since the load requires a maximum power of 0.5 W and with a DC 5 V supply, the solar panel must have high energy density for the system to operate even at low irradiance. The typical value of VOC of the considered solar cell is about 3 V. A condition arises where VSOLAR and VCST OR become equal due to which the inductor current slope becomes zero when CST OR is charging.

The general operation is that if VSOLAR > VM P P and VCST OR < VOV (overvoltage) then the boost converter should work in the same way as mentioned above ie. the inductor current must be limited by ISAT to zero. To avoid this situation, care must be taken that even if VM P P−SIG is low, PMOS M3 must be turned off only when the inductor current reaches zero. To ensure that CLKP is not high before the inductor current reaches zero, even if VM P P−SIG goes low, VM P P−SIG must be masked so that it does not affect the clock output, which is done using a D flip flop, via whose VM P P−SIG is transmitted only if a ZERO is received and the select signal which is the output of the D flip flop is used to select between the clock signals for the D flip flop.

Figure 3.1: Complete architecture of solar energy harvesting system of [9].

V CSTOR

Full System at Higher Solar Irradiation
Full system at Lower Solar Irradiation
Sub-Blocks Explanation

Current Sensor Waveforms
I SAT and I ZERO Waveforms
The Digital Controller Waveforms
Battery Management Waveforms
Simulation Results

Summary
Future Work

Until VSOLAR−CST OR goes low, the system is operated in open loop using a freewheeling clock. The switch node potential is matched to the Drain potential of the VSEN SE−N ODE MOSFET and the current is scaled by a factor of 116666 times IL. Resistance in current sensor is adjusted so that VIREF equal to 1.2 V corresponds to 1 A inductor current.

CST OFen switching node potentials are monitored during the discharge phase of the inductor. Once switching node potential goes below CST OR potential ISEROSIG switches on. ISEROCLK acts as an enable signal for the comparator. If VM P P−SIG goes low before inductor reaches Ipeak, high priority is given to VM P P−SIG and CLKN turns off and inductor starts discharging as CLKP immediately goes low. The table 4.1 gives the input power (PIN), output power (POU T), power dissipated (PD) (across the switches (PD−SW), inductor series resistance (PD−RL), main converter (PD−CKT)) and efficiency(η) of the energy harvesting system for different solar conditions with the number of solar cells in parallel(NP) as 4 and 8.

This work mainly deals with the modeling and design of an efficient solar energy harvesting system as a powerful aid to power IoT. A generic solution for a wider input voltage/power range harvester (1 V-3 V) is provided and the output can be regulated to the required voltage (3 V-5 V) by varying the resistance of resistive steps in under-voltage and over-voltage protection in battery management. The performance of the system is verified with the simulation results. A robust energy harvesting system is proposed which will work even with large variations in input voltage (1V-3V) and temperature (−400 C to 1200 C) that can deliver a power of 0.5 W to the load.

Small modifications can be made in the circuit to make it harvest energy from the multiple sources like TEG, RF, Vibration energy. Peak inductor current can be made variable depending on the load requirements, so that conduction losses can be further minimized. Van Hoof, 5W - 10mW Input Power Range Inductive Boost Converter for Indoor Photovoltaic Energy Harvesting with Integrated Maximum Power Point Tracking Algorithm ISSCC Dig.