Development of Optimal Lighting System to Improve Crop Productivity

Abstract

This PhD thesis explores the development of an optimal lighting system aimed at enhancingcrop productivity. The primary technical contribution of this thesis is divided into two parts. Part I centres on constructing a mathematical model of the plant circadian system to generate an optimal lighting protocol, while Part II focuses on developing an efficient power conversion system to implement the lighting protocol generated in Part I.
In Part I, a comprehensive model for the PCS is developed, considering various light qualities as inputs and providing various growth indicators as outputs. Firstly, an existing compact plant circadian model is modified and validated through laboratory-scale experiment to make it suitable for multiple light qualit-ies, in contrast to the existing version designed for single light qualities. Subsequently, further modifications are proposed to account for different plant growth indicators. The developed com-prehensive model, validated through experimental data from the existing literature, quantifies the relationship between multiple light qualities and two vital indicators of plant productiv-ity: hypocotyl growth and days to flowering. Utilising this model, an optimal lighting protocol is proposed through numerical simulation for the model plant Amb·idops'is thal·iana, ensuring sustainable plant productivity (i.e. reducing the time to grow the plant) with lower energy consumption compared to similar results found in the literature.
In Part II, an efficient power conversion control system is developed for the LED driver that powers the LED lights to im-plement the developed lighting protocol. The LED driver control system comprises two parts: the power grid detection scheme and the converter control system, the latter relying on the grid detection scheme. Through a comprehensive performance comparison, a suitable grid detection scheme is initially identified and later refined to provide fast and accurate grid detection, even in the face of various unforeseen disturbances. Subsequently, an efficient power converter control scheme is proposed, employing the nonlinear PI method to alleviate challenges such as the dy-namic performance vs. disturbance rejection trade-off faced by its linear counterpart. Numerical simulation results are provided to demonstrate the advantages of the proposed controller.

Date of Award	Oct 2024
Original language	English
Awarding Institution	Coventry University
Supervisor	Jesper Christensen (Supervisor), Fei He (Supervisor), Laura Roden (Supervisor), Hafiz Ahmed (Supervisor) & Mathias Foo (Supervisor)