Novel Design of Energy Control Algorithm used in Solar Powered Batteryless Energy Harvesting System to power Wireless Sensor Node

Date created: 
Super sensor systems
Wireless sensor nodes
Solar powered batteryless energy harvesting system
Energy control algorithm
Model predictive control

'Internet of Things' (IoT) technology is becoming one of the most important driving forces in human productivity in recent years.  New generation of 'super sensors', in the form of wireless sensor nodes (WSN) are the most important components in IoT. Powering these devices using traditional batteries creates a tough battery longevity problem, where the rigorous demands on the batteries require them to be replaced once every few years. This is further worsened by the huge number of devices in a typical IoT application, with their demand in power becoming a serious issue.   It is commonly considered that one of the best ways to power these wireless sensor nodes is to use energy harvesters with solar energy harvesting. Due to the unpredictable nature of solar irradiation, a problem to be solved is how a wireless sensor node powered by a solar energy harvester can have continual operation while simultaneously deliver the highest possible service duty.   This thesis presents a new energy control algorithm that addresses this bottleneck problem. Firstly, the analysis of past research using PID Control, Fuzzy Logic, and Adaptive Dynamic algorithm is provided, which reveals significant shortfalls.  The use of a solar irradiation prediction model by one group of researchers results in significant system shutdown (“dead time”) when actual solar irradiation deviates from the prediction model.  Another group of researchers maintain the terminal voltage of the supercapacitor at a certain set point but this approach is not able to avoid system shut down, and it demands an unacceptable operating condition in which certain amount of light must be present for the system to operate. After analysis of these past projects, the design deficits and imprecise design objectives in these researches are elaborated. Secondly, a proposal of a new energy control algorithm with the use of a precise two branch equivalent model is presented, with the employment of Model Predictive Control (MPC) theories to compute important control parameters. An augmented MPC control algorithm is designed based on three new principles, in order to handle the two mingled system input variables of system operating current and system sleep mode current of the WSN. Thirdly, the resulting new energy control algorithm is implemented in a self designed wireless sensor node embedded system. The purpose of this self designed system is to conduct comprehensive field tests to validate the performance and the robustness of the energy control algorithm. Finally, detail results with analysis of the four field tests is presented. The four field tests include the first test with normal operating condition, the second test as a stress test with an obstructed solar panel, the third as an additional stress test with a defective supercapacitor, and the fourth field test under abnormally adverse operating conditions. Except for the third field test which exhibits some time duration (2.8% of the total testing duration) with non-maximized WSN operation, all other field tests demonstrate full fulfillment of the new energy control algorithm’s design objectives. The last part of this thesis summarizes the conclusion of the research. And the research contribution in the field of IoT as well as in other numerous application areas are interpreted.

Document type: 
This thesis may be printed or downloaded for non-commercial research and scholarly purposes. Copyright remains with the author.
Zoë Druick
Applied Sciences: School of Mechatronic Systems Engineering
Thesis type: 
(Thesis) M.A.Sc.