Fundamentals of Near Field IPT

# Near Field IPT 
### History and the Fundamentals

.TitleAuthor[Duleepa J Thrimawithana & Grant A Covic]

---

.logo-slide[]
.footer[[D J Thrimawithana](https://www.linkedin.com/in/duleepajt) & [G A Covic](https://www.linkedin.com/in/grant-covic-179546a/), IEEE Wireless Week, Wireless Power Transfer School (June 2021)]

---

# The University of Auckland

- Highest ranked New Zealand university and 85th in the QS World University Ranking
- Over 5,000 staff members and 40,000 students
- Nine faculties including Medical & Health Sciences, Engineering, Business & Economics and Science

---

# Dept. of Electrical, Computer & Software Eng.

.left-column[
- One of the 5 departments in the Faculty of Engineering
- Offers 3 undergraduate degree programs
 - Electrical & Electronics, Computer Systems and Software
 - Project based teaching
- 35+ full-time academic staff members and 15+ post-doctoral research fellows
- 150+ postgraduate students and 700+ undergraduate students
- Regular visiting research scholars and research students
- Research groups include Power Electronics, Power Systems, Signal Processing, Robotics, Embedded Systems, Parallel Computing, Telecommunications and Control Systems
]

---

# Dept. of Electrical, Computer & Software Eng.

---

# Power Electronics Research Group

---

# Part I - Wireless Power Transfer
### A Review of Current Status

---

---

# Wireless Power Transfer

.left-column[
- Can be divided in to two broad categories
- Radiative transmission through far-field principles
 - Examples include power transfer through radio-frequency, microwave, optical and ultrasonic technologies
 - At an experimental stage of development
 - AirFuel Alliance is developing standards
- Non-radiative transmission through near-field principles
 - Resonant inductive power transfer (IPT) and capacitive power transfer (CPT) technologies are utilised
 - IPT, also referred to as highly resonant wireless power transfer or magnetic resonance coupling,  is the most mature technology 
 - WPC, SAE, etc. standards are being developed
]

---

# Early Developments

- 1890s to 1920s: Tesla (US), Hutin (FR) and LeBlanc (FR) demonstrated wireless power transfer 
- 1960s: Shuder (US) and his team worked on high power IPT systems for biomedical applications
- 1970s: Otto (NZ) and Bolger (US) proposed roadway power
- 1980s: Santa Barbara electric bus project (US) 
- Early 1990s: Boys (NZ), Green (NZ) and Covic (NZ) developed IPT systems for dynamic industrial systems & stationary people movers which were adopted and commercialized by Difuku and Wamplfer in materials handling applications
- Late 1990s: Meins (DE) developed multiphase IPT systems for industrial applications

---

# Recent Developments

- Early 2000s: Boys (NZ) and Covic (NZ) developed IPT systems with Conductix-Wamplfer and Daifuku for automated guided vehicles and busses
- Mid 2000s: Hui (HK), Hu (NZ) and Soljacic (US) developed low power IPT systems for powering/charging biomedical implants, sensors and consumer devices
- Late 2000s: Hu's (NZ) work lead to forming Power by Proxi, Soljacic's (US) work lead to forming WiTricity  
- Early 2010s: Boys's (NZ) and Covic's (NZ) work lead to forming HaloIPT, which lead to Qualcomm Halo and Qi standard was released 
- Late 2010s: WiTricity holds the largest patent portfolio as they acquired Qualcomm Halo, while Apple acquires Power by Proxi, and number of new research groups as well as companies starts to emerge
- Early 2020s: Standards for EV stationary charging completed

---
name: S5

# Current Status of IPT

- Power ratings of tens of kWs with uni or bi-directional power flow
 - Magnetic designs include circular coil, solenoidal coil, polarised coil, multi coil structures as well as track based systems 
 - Ferrites and/or reflection coils are often used to shape the magnetic fields generated
 - Transmission range of over 300 mm with over ± 200 mm XY tolerance easily achievable 
 - Efficiency typically over 85% but can be as high as 97%
 - Operating frequency typically ranges from tens of kHz to tens of MHz

.center[<img src="img/fund/LongDistance.jpg" height="175px">]
.center[.credits[ [Dipole coil resonant system developed by KAIST](https://phys.org/news/2014-04-wireless-power-five-meter-distance.html)]]

---
name: S6

# IPT Applications

- Factory automation and materials handling
 - Includes hoists, AGVs and clean room systems  
- Consumer electronics and appliances
 - Charging phones, laptops, etc. as well as powering appliances 
- Transportation electrification 
 - Charging stationary and in-motion electric vehicles, charging electric ferries, etc.
- Biomedical
 - Charging and powering medical devices as well as implants
- Lighting  
 - Wireless power and communication to tunnel lighting, stage lighting and mines 
- Industrial applications 
 - Industrial machinery, robots, electric fense energizers, etc.

---

# Part II - IPT Fundamentals
### Modelling the Coils and Power Transfer

---

---
name: S7

# Components of an IPT System

- Primary consists of a DC-AC switched-mode converter, a compensation network and a transmitter coil(s) 
 - A lumped transmitter coil is often referred to as a primary pad/coil/coupler 
 - An elongated coil is often called a primary track
 - Can be directly fed by a DC source or through a grid-connected AC-DC converter
- Secondary/pick-up consists of a receiver coil(s), a compensation network and an AC-DC converter
 - The reciever coil is is often referred to as a secondary/pick-up pad/coil/coupler
- Primary and secondary coils are magnetically coupled but coupling coefficient is typically 40% or lower
- The compensation networks help improve efficiency by minimizing the VA requirements of converters

---
name: S8

# Modelling the Primary and Secondary Coils

- The behaviour of the coils can be modelled using a T-equivelent transformer model
 - `\(L_{pt}\)` and `\(L_{st}\)` are the self-inductances of primary and secondary coils, respectively
 - `\(\phi_{m}\)` represents flux linkage while `\(\phi_{lk_p}\)` and `\(\phi_{lk_s}\)` represent leakage flux
 - The coupling coefficient, k, between the coils is less than 0.4 in a typical IPT system
- The mutual inductance, M, between the two coils is given by 
\\[M = k\sqrt{L\_{pt}L\_{st}}\\]
- If the mean-turn lengths of the coils do not change, then `\(n = \frac{N_{pt}}{N_{st}} = \sqrt{\frac{L_{pt}}{L_{st}}}\)`, where `\(n\)` is the turns ratio

---
name: S9

# Coupled Inductor Model

- Although a T equivalent transformer model can be used to model the coils, the coupled inductor model is widely used when analysing an IPT system in the phasor-domain
- In the coupled inductor model, `\(L_{pt}\)` and `\(L_{st}\)` represent the self-inductances of primary and secondary coils
- `\(V_{sr}\)` represents voltage induced across `\(L_{st}\)` due to current `\(I_{pt}\angle 0\)` flowing through `\(L_{pt}\)` and is given by
\\[V\_{sr} = \omega MI\_{pt}e^{j\pi/2}\\]
- Similarly, `\(V_{pr}\)` represents voltage induced across `\(L_{pt}\)` due to `\(I_{st}\angle\theta\)` and is given by
\\[V\_{pr} = \omega MI\_{st}e^{j(\theta+\pi/2)}\\]

???

If the current through primary coil is `\(I_{pt} \sin (\omega t)\)` then,
\\[\phi\_{m} \propto I\_{pt} \sin (\omega t) \\]
The voltage induced on the secondary coil is therefore,
\\[v\_{sr} \propto \frac{\mathrm{d} \phi\_{m} }{\mathrm{d} t} \\]
\\[v\_{sr} \propto \omega I\_{pt} \cos (\omega t) \\]
Since `\(M\)` represent the coupling between the two coils,
\\[v\_{sr} = \omega M I\_{pt} \cos (\omega t) \\]
In the phasor-domain,
\\[V\_{sr} = \omega M I\_{pt} e^{j\pi/2} \\]

---
name: S10

# Open-Circuit Voltage & Short-Circuit Current

.left-column[
- The voltage measured across an open-circuited secondary coil, `\(V_{oc}\)`, is the same as `\(V_{sr}\)` and therefore
\\[V\_{oc} = V\_{sr} = \omega MI\_{pt}e^{j\pi/2}\\]
- The current flowing through a short-circuited secondary coil, `\(I_{sc}\)`, is 
\\[I\_{sc} = \frac{V\_{sr}}{j\omega L\_{st}} = I\_{pt}\frac{M}{L\_{st}}\\]
- The uncompensated volt-ampere (VA) capacity of the secondary coil can be given by
\\[S\_{u} = \left | V\_{oc} \right | \left | I\_{sc} \right | = I\_{pt}^2\frac{\omega M^2}{L\_{st}}\\]
]

---
name: S11

# Power Transferred Across Airgap

.left-column[
- Assume that the `\(V_{sr}\)` induced by `\(I_{pt}\angle 0\)` causes a current `\(I_{st}\angle\theta\)` to flow through a load, `\(Z_{sc}\)`, attached across the secondary coil 
- Under these conditions, power transferred is
\\[P\_{o} = \Re \left \\{ V\_{sr}I\_{st}e^{-j(\theta)} \right \\} = \left | V\_{sr} \right | \left | I\_{st} \right | \cos(\pi/2-\theta)\\]
- The VA rating of the coils required to achieve this is
\\[\mathit{VA}\_{pt} = \omega L\_{pt} \left | I\_{pt} \right |^2 \quad \text{and} \quad \mathit{VA}\_{st} = \omega L\_{st} \left | I\_{st} \right |^2 \\]
- Thus `\(P_{o}\)` is related to the VA in the coils as given by
\\[P\_{o} = k \sqrt{\mathit{VA}\_{pt}\mathit{VA}\_{st}} \sin(\theta) \\]
 - Typically, `\(\sin(\theta) \approx 1\)`, and thus `\( P_o \approx k \sqrt{\mathit{VA}_{pt}\mathit{VA}_{st}} \)`
]

???

Note that 
\\[P\_{o} = \Re \left \\{ V\_{sr} \bar {I\_{st}} \right \\} \\]
Substituting `\( \left| V_{sr} \right | = \omega M \left| I_{pt} \right | \)`, 
\\[P\_{o} = \omega M \left | I\_{pt} \right | \left | I\_{st} \right | \cos(\pi/2-\theta)\\]
\\[P\_{o} = k\sqrt{ (\omega L\_{pt} \left | I\_{pt} \right |^2 ) (\omega L\_{st} \left | I\_{st} \right |^2) } \sin(\theta)\\]
\\[P\_{o} = k \sqrt{\mathit{VA}\_{pt}\mathit{VA}\_{st}} \sin(\theta) \\]

---
name: S12

# Coil Q, Losses & Maximum Efficiency

- Assume that the losses in `\(L_{pt}\)` and `\(L_{st}\)` are modelled using equivelent series resistances (ESRs) `\(r_{pt}\)` and `\(r_{st}\)`
- The quality factors, `\(Q_{pt}\)` and `\(Q_{st}\)`, of the primary and secondary coils are
\\[Q\_{pt} =  {\omega L\_{pt}}/{r\_{pt}} \quad \text{and} \quad Q\_{st} = {\omega L\_{st}} / {r\_{st}} \\]
- The total losses in the two coils can be expressed as
\\[P\_{loss(c)} = \left | I\_{pt}^2 \right | r\_{pt} + \left | I\_{st}^2 \right | r\_{st} = \frac{\mathit{VA}\_{pt}}{Q\_{pt}} + \frac{\mathit{VA}\_{st}}{Q\_{st}}\\]
- The maximum efficiency is achieved when losses in the coils are matched (i.e. `\(\mathit{VA}_{pt}/Q_{pt} = \mathit{VA}_{st}/Q_{st}\)` )
\\[ \eta\_{c\_{max}} = \frac{ k \sin(\theta) -  \frac {1} { \sqrt {Q\_{pt} Q\_{st}} }} {k \sin(\theta) +  \frac {1} {\sqrt {Q\_{pt} Q\_{st}} }} \\]

---
name: S12b

# Impact of k and Coil Q on Performance

- Typically, k is in the range 0.1 to 0.3 and coil Q is in the range 400 to 600 for EV charging
- Design of a 10 kW IPT system 
 - Assume both `\(Q_{pt}\)` and `\(Q_{st}\)` are 400
 - Note that `\( P_o \approx k \sqrt{\mathit{VA}_{pt}\mathit{VA}_{st}} \)` needs to be always met

---

# Part III - Compensating the Coils
### Behavior of the Compensated Coils

---

---
name: S19

# An Uncompensated Secondary

- Assume load resistor, `\(R_{L}\)`, is connected directly across a secondary coil, which is coupled to a primary coil 
- The primary coil is driven by the primary converter generating a current `\(I_{pt}\angle 0\)` at an angular frequency `\(\omega = 2\pi f\)`
- The current flowing through secondary coil as well as the load will be
\\[ I\_{st} = \frac {-V\_{sr}} {R\_{L} + j\omega L\_{st}} = \frac {-\omega MI\_{pt}e^{j\pi/2}} {R\_{L} + j\omega L\_{st}}  \\]
- The impedance of `\(L_{st}\)` limits the ability to supply current to `\(R_{L}\)`

---
name: S20

# Power from an Uncompensated Secondary

.left-column[
- The power delivered to the load is then
\\[ P\_{o} =  \left | I\_{st} \right |^2 R\_{L} = \left ( \omega MI\_{pt} \right )^2  \frac {R\_{L}} {R\_{L}^2 + \omega^2 L\_{st}^2 } = k^2 \mathit{VA}\_{pt} \frac {\omega L\_{st} / R\_{L}} { 1 + \omega^2 L\_{st}^2 / R\_{L}^2 } \\]
- At a load `\(R_L = \omega L_{st}\)`, maximum power is delivered 
\\[ P\_{o\_{(max)}} =  \frac {\left ( \omega MI\_{pt} \right )^2} {2 \omega L\_{st} } = 0.5 V\_{oc}I\_{sc}  = 0.5k^2\mathit{VA}\_{pt} \\]
- As an example, when `\(k = 0.1\)`, to deliver 5 mW to the load, the primary coil should be driven at least at 1 VA
 - The impedance of `\(L_{st}\)` limits the power transfer
]

---
name: S21

# What Should an 'Ideal' Secondary Look Like?

- The load `\(R_{L}\)` 'ideally' needs to be able to take any amount of power from secondary
- This could be achieved, for example if the secondary coil is made to to look like an ideal voltage source
 - The impedance of `\(L_{st}\)` need to be made zero 
 - In this case we would have an ideal voltage source, `\(V_{sr}\)`, supplying the load, `\(R_{L}\)` 
- Alternatively, the secondary coil can be made to look like an ideal current source, `\(I_{sc}\)`
- Resonance can be used to achieve this 'ideal voltage source' or 'ideal current source' type behavior
 - A technique discovered by Hertz in the 1880's

---
name: S22

# Reflected Impedance on to Primary

- Secondary can be modelled as `\(V_{sr}\)` driving into `\(Z_s\)`, which represents secondary compensation and load
- Current flowing through the secondary coil is therefore 
\\[ I\_{st} = -V\_{sr} / Z\_s = -j \omega M I\_{pt} / Z\_s \\]
- `\(I_{st}\)` induces a voltage, `\(V_{pr}\)`, across the primary coil, which can be given by
\\[ V\_{pr} = j \omega M I\_{st} = \omega^2 M^2 I\_{pt} / Z\_s \\]
- `\(V_{pr}\)` can also be represented as a reflected impedance where `\(Z_{pr} = \omega^2 M^2 / Z_s \)`

---
name: S23

# Primary Compensation Networks

- Directly driving the highly inductive load consisting of `\(L_{pt}\)` and `\(Z_{pr}\)` with a power converter can be very inefficient and costly
 - For example the VA rating of the converter, which is required to deliver `\(1W\)` to the secondary load, can be in excess of `\(1/(k^2Q_{ss/sp_{(op)}})\)`
 - If `\(k = 0.1\)` and `\(Q_{ss/sp_{(op)}}=5\)`, which are typical for an IPT system, the primary converter will need to be supply 20VA to deliver 1W to the load
- Resonance is also used to reduce the VA demand of the primary converter
 - Many different resonant networks have been developed to-date to cater for different applications

---
name: S24

# Compensation Methods

- The two most common resonant/compensation networks are the parallel and series types
 - Parallel resonance is not common on primary in modern systems as it needs to be driven by a current sourced inverter
- Primary side commonly employ LCL compensation networks 
- More complex compensation networks include
 - LCCL and LCLC
 - Hybrid compensation
 - Impedance compression networks

---
name: S25

# Compensation Networks - Summary

---

# Part IV - Power Converters in IPT Systems 
### An Overview

---

---
name: S53

# Voltage Source Primary Converters

.left-column[
- A full-bridge converter is typically used to drive a series/LCL compensated primary coil
- Half-bridge converters may be used in low power applications
- The primary converter can be fed from the grid through a PFC stage or a grid-tied inverter or directly from a DC source
- The frequency of the fundamental component of `\(v_{pi}\)` generated by the converter is at or near the tuned frequency of the compensation network (10's of kHz)
- Traditional PWM schemes cannot be used to regulate, `\(V_{pi(1)}\)` (same as `\(V_{pi}\)`), the phasor magnitude of the fundamental component of `\(v_{pi}\)`
- Typically, voltage cancellation schemes are used, for example using phase-shift modulation to regulate the magnitude of `\(V_{pi}\)`
]

---
name: S54

# Phase-Shift Modulation

- The switches are operated at a frequency `\(f_{s}\)`, where `\(f_{s}\)` is equal to or close to the tunned frequency of the compensation network
- Each leg of the full-bridge operated at 50% duty-cycle, generating the square-waves `\(v_{pa}\)` and `\(v_{pb}\)`
- The phase-shift modulation, `\(\phi_{p}\)`, applied between `\(v_{pa}\)` and `\(v_{pb}\)`, regulates the magnitude of the fundamental component of `\(v_{pi}\)`, since `\(v_{pi}=v_{pa}-v_{pb}\)`
  - When `\(\phi_{p} = \pi\)`, `\(V_{pi}\)` is a maximum

---
name: S56
  
# Modelling the Converter

.left-column[
- The fourier expansion of `\(v_{pi}\)` is
\\[ v\_{pi} = V\_{in} \frac {4} {\pi} \sum\_{n=1,3,..}^{\infty} \frac {\sin \left(  n \phi\_p/2 \right)} {n} \cos \left(  n \omega\_s t \right)  \quad \text{where} \quad \omega\_s = 2 \pi f\_s \\]
- RMS of `\(v_{pi}\)` is therefore
\\[ V\_{pi\_{rms}}^2 = V\_{pi(1)}^2 + \sum\_{n=3,5,..}^{\infty} V\_{pi(n)}^2 \quad \text{where} \quad V\_{pi(1)} = V\_{in} \frac {4} {\pi} \sin \left(  \phi\_p/2 \right) \cos \left( \omega\_s t \right) \\]
- System behavior can be accurately predicted by only considering the fundamental
 - High order resonant circuit formed by the compensation network filters the harmonics
]

---
name: S58
 
# Hard Switching

.left-column[
- Consider a half-bridge leg and assume that `\(Q_2\)` has been turned-on creating a low-voltage across `\(v_{pa}\)`
- To create a high-voltage, at `\(t = 0s\)`, `\(Q_2\)` starts turning-off and after the dead-time, `\(t_{d}\)`, `\(Q_1\)` starts turning-on  
 - Assume that `\(i_{pi}\)` is positive during this time (out of the leg)
- As `\(Q_2\)` turns-off, `\(i_{pi}\)` shifts to `\(D_2\)`, but the voltage across `\(Q_2\)` as well as `\(v_{pa}\)`, is held low by `\(D_2\)`
- Once `\(Q_2\)` is turned-off, until the end of `\(t_{d}\)`, `\(D_2\)` conducts `\(i_{pi}\)`
- At `\(t_{d}\)`, `\(Q_1\)` starts turning-on and `\(i_{pi}\)` shifts to `\(Q_1\)`
- Then, `\(D_2\)` starts recovering and the voltage across `\(Q_1\)` starts to drop
- At `\(t_{d}+t_{on}\)`, `\(v_{pa}\)` has completed the transition from low-high
- `\(Q_1\)` was subjected to high-stresses leading to high turn-on losses
- Capacitive loading and/or modulation lead to hard-switching

]

---
name: S60
 
# Soft Switching

.left-column[
- Consider a half-bridge leg and assume that `\(Q_2\)` has been turned-on creating a low-voltage across `\(v_{pa}\)`
- To create a high-voltage, at `\(t = 0s\)`, `\(Q_2\)` starts turning-off and after the dead-time, `\(t_{d}\)`, `\(Q_1\)` starts turning-on  
 - Assume that `\(i_{pi}\)` is negative during this time (in to the leg)
- As `\(Q_2\)` turns-off, `\(i_{pi}\)` shifts to `\(C_2\)` and `\(C_1\)` causing `\(v_{pa}\)` to increase 
- Once `\(Q_2\)` is turned-off and `\(V_{pa}\)` has completed the transition from low-high, until the end of `\(t_{d}\)`, `\(D_1\)` conducts `\(i_{pi}\)`
- At `\(t_{d}\)`, `\(Q_1\)` starts turning-on while voltage across it is low and `\(i_{pi}\)` shifts from `\(D_1\)` to `\(Q_1\)`
- Both `\(Q_1\)` and `\(Q_2\)` had very low-stresses leading to significantly lower switching losses
- Inductive loading and/or improved topologies help soft-switching 
]

---
name: S65

# Current Source Converters

- Current source push-pull converters are widely used in low power IPT applications
 - Drives a parallel compensated primary coil at resonance in autonomous operating mode or near resonance in fixed-frequency operating mode
- Not widely used in high-power applications 
 - Require a large DC inductor
 - Sensitive to changes in operating conditions
- Current source full-bridge converters may be used in higher power applications

---
name: S68

# Secondary Converters

<div style="height:1px;font-size:1px;"> </div>
- In most applications a rectifier is used to derive a DC output
- In some applications, current doubler rectifiers are used to increase the power transfer capability of a parallel compensated secondary 
- Following the rectifier, a DC-DC converter is often used to regulate the output 
- Alternatively, an active rectifier may be used for rectification and regulation while also providing additional functionalities such as bi-directional power flow or improved misalignment tolerance
 - However, the active rectifier needs to be synchronised with the primary converter 
- Direct AC-AC processing secondary converters may be used if an AC output is required

---
name: S74

# Short-Circuit Regulator

- Following the rectifier, a short-circuit (boost) regulator, can be used to regulate the output of a parallel compensated IPT secondary
- The duty-cycle, `\(D_{scr}\)`, of `\(Q_{1}\)` is controlled to regulate `\(I_{out}\)`, and therefore `\(V_{out}\)`
- `\(Q_{1}\)` can be operated either at a 'low' switching frequency using a hysteritic controller or at a 'high' switching frequency using for example a PI controller
- `\(L_{dc}\)` should be sufficiently large to maintain `\(I_{Ldc}\)` continuous

---
name: S79

# Active Rectifiers

- A typical active rectifier employs a full-bridge converter that is controlled to generate a `\(V_{si}\)` that leads or lags `\(V_{pi}\)` by a phase-angle `\(\theta\)`
- `\(\theta\)` can be controlled together with the amplitude of `\(V_{si}\)` to generate a specific AC impedance, `\(Z_{L(ac)}\)`, across secondary compensation network
- The reactive part of `\(Z_{L(ac)}\)` can be used to improve power transfer and/or reduce losses, especially when operated under misaligned conditions
- In applications requiring only uni-directional power flow, two of the switches in the full-bridge can be replaced with diodes

---

# References
#### These slides were derived from a presentation prepared by the author on 
#### "Fundametals of Inductive Power Transfer: A Workshop with Example Designs"
#### @wirelesspower.github.io/SeminarsWorkshops/FundamentalsofIPTDJT/presentation.html

---

# Thank You