Control of electrical systems using artificial intelligence for more electric aircraft applications

Hussaini, Habibu (2023) Control of electrical systems using artificial intelligence for more electric aircraft applications. PhD thesis, University of Nottingham.

[img] PDF (Corrected thesis) (Thesis - as examined) - Repository staff only until 14 December 2025. Subsequently available to Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Available under Licence Creative Commons Attribution.
Download (15MB)


The increasing demand for fuel efficiency and environmental sustainability has driven the emergence of More Electric Aircraft (MEA) concepts, aiming to reduce weight, fuel consumption, environmental pollution, and operational costs. To achieve this, aircraft propulsion systems are transitioning towards electrical energy, resulting in a high demand for efficient power management and control. In this context, DC microgrids (MGs) have gained attention for power distribution in MEA due to their simplicity and the single-bus high voltage direct current (HVDC) electrical power system (EPS) architecture has been identified as a promising solution. Droop control is commonly employed in this architecture to achieve autonomous power management among power-generating sources without the need for a communication network. However, traditional droop control methods have limitations that require innovative and intelligent solutions. The primary goal of this thesis is to overcome the limitations of traditional droop control methods and enhance the power management, overall performance, and stability of the MEA EPS HVDC distribution network. To achieve this, the study focuses on the application of artificial intelligence (AI) techniques and introduces the following proposed approaches:

1. This thesis develops an advanced droop settings design strategy using AI-based optimisation with artificial neural networks (ANNs). This approach optimises droop coefficient settings offline, leveraging a trained ANN-based surrogate model to enhance voltage regulation and power sharing among power-generating sources.

2. A real-time tuning approach for droop control using reverse data training of an ANN-based surrogate model is presented. By directly predicting optimal droop gain settings based on specific performance requirements, the controller can adapt in real-time to reference changes, enhancing flexibility and responsiveness in dynamic conditions.

3. An innovative AI-based hierarchical control design strategy to address voltage deviations and improve overall voltage regulation under various operating conditions is proposed. The trained ANN-based surrogate model computes optimal droop and voltage compensation coefficients based on desired power-sharing ratios and bus voltage regulation requirements.

4. A new approach utilising two ANNs to address controller limitations in the presence of varying line resistance conditions is presented. One ANN optimises droop gain settings, while the second estimates real-time interconnecting cable resistances to compensate for terminal voltage deviations, leading to accurate current sharing and enhanced voltage regulation in DC MGs.

Extensive simulations and hardware-in-the-loop (HIL) experiments are conducted to validate the effectiveness and reliability of the AI-based approaches in improving droop control performance, offering promising prospects for enhancing the operation of MEA EPS.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Yang, Tao
Bozhko, Serhiy
Keywords: Artificial intelligence, Artificial neural network, Droop Control, More Electric Aircraft
Subjects: Q Science > Q Science (General)
T Technology > TJ Mechanical engineering and machinery
T Technology > TL Motor vehicles. Aeronautics. Astronautics
Faculties/Schools: UK Campuses > Faculty of Engineering
UK Campuses > Faculty of Engineering > Department of Electrical and Electronic Engineering
Item ID: 76737
Depositing User: Hussaini, Habibu
Date Deposited: 31 Jan 2024 16:16
Last Modified: 31 Jan 2024 16:16

Actions (Archive Staff Only)

Edit View Edit View