Design and development of safety systems for high frequency inductive power transfer
File(s)
Author(s)
Lan, Lingxin (Steve)
Type
Thesis
Abstract
As wireless charging is gaining its popularity among consumer electronics, e.g., phones, smart
wearables, electric toothbrushes, etc., there has been a trend of expanding this technology into
a wider range of applications e.g. drones, robots, electric vehicles etc.. To achieve this, both the
charging power and range need to be increased. This thesis discusses the limitations of widely
used kHz inductive power transfer systems and emphasises the challenge of deploying into a
wider range of applications. High-frequency inductive power transfer (HF-IPT) systems are
then discussed with two real-world applications presented to showcase HF-IPT’s potential over
kHz IPT systems. Some of the benefits of the HF-IPT, e.g., the large air gap and tolerance to
misalignment, could increase the chances for live or other unintended objects to be coupled into
the wireless charging system, which could cause safety hazards if the system was not designed
carefully. This thesis, therefore, focuses on the safety systems design and development for
HF-IPT systems.
A number of existing and potential foreign and live object detection methods (FOD/LOD)
including a new FOD/LOD method based on reflected impedance are introduced. The proposed
method can operate without additional sensors, and without a communication link between
IPT transmitter and receiver. A detection accuracy of 95% is achieved by implementing such
FOD/LOD method. In addition, a FOD/LOD technique based on a mmWave radar sensor
is also introduced. Differing from typical radar applications, the proposed method leverages
machine learning techniques to perform object recognition to reduce the false detection rate.
The developed FOD/LOD system could classify six different charging scenarios with an average
accuracy of 96%. For applications that do not involve any live or unintended objects, this thesis
also introduces a localisation technique based on the IPT system to help guide a drone or robot
to a specific location e.g. a wireless charging point. Such a system was designed to reduce the
risk of charging by minimising human’s involvement.
wearables, electric toothbrushes, etc., there has been a trend of expanding this technology into
a wider range of applications e.g. drones, robots, electric vehicles etc.. To achieve this, both the
charging power and range need to be increased. This thesis discusses the limitations of widely
used kHz inductive power transfer systems and emphasises the challenge of deploying into a
wider range of applications. High-frequency inductive power transfer (HF-IPT) systems are
then discussed with two real-world applications presented to showcase HF-IPT’s potential over
kHz IPT systems. Some of the benefits of the HF-IPT, e.g., the large air gap and tolerance to
misalignment, could increase the chances for live or other unintended objects to be coupled into
the wireless charging system, which could cause safety hazards if the system was not designed
carefully. This thesis, therefore, focuses on the safety systems design and development for
HF-IPT systems.
A number of existing and potential foreign and live object detection methods (FOD/LOD)
including a new FOD/LOD method based on reflected impedance are introduced. The proposed
method can operate without additional sensors, and without a communication link between
IPT transmitter and receiver. A detection accuracy of 95% is achieved by implementing such
FOD/LOD method. In addition, a FOD/LOD technique based on a mmWave radar sensor
is also introduced. Differing from typical radar applications, the proposed method leverages
machine learning techniques to perform object recognition to reduce the false detection rate.
The developed FOD/LOD system could classify six different charging scenarios with an average
accuracy of 96%. For applications that do not involve any live or unintended objects, this thesis
also introduces a localisation technique based on the IPT system to help guide a drone or robot
to a specific location e.g. a wireless charging point. Such a system was designed to reduce the
risk of charging by minimising human’s involvement.
Version
Open Access
Date Issued
2021-11
Date Awarded
2022-03
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Mitcheson, Paul
Publisher Department
Electrical and Electronic Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)