Minimization Technologies for Passive Components …...Power Electronics Lab. 1 Minimization...

1Power Electronics Lab.

Minimization Technologies for Passive Components based on Circuit Topology and Its Control

Jun-ichi ItohNagaoka University of Technology, Japan

One of the most famous firework festivals in Japan is held in Nagaoka on begin of August every year.

Phoenix (width 2000m)

Sho-sanjakuDiameter 600m, altitude 600m


Contents

1. Introduction-Importance of minimization of passive components

2. Approach I: Matrix converter-Circuit topology and features-Latest technologies: Three-phase to Single phase--Matrix converter in Market: general purpose drive/air conditioner/quick charger

3. Approach II: Active power decoupling for Single-phase converter-Principle-Integrated to PFC or Inverter-Integrated to Chopper

4. Approach III: Discontinuous Current Mode converter-Principle: Robustness for inductor -Single-phase inverter with DCM-Three-phase inverter with DCM

5. Approach IV: FRT capability of grid-tied inverter with minimized inductor-Mr. Nagai presents his projects

6. Conclusion


Road to Innovation

Innovation is born when convenience is more important than price;

New applications will be invented by size reductionSize reduction realizes cost reduction because of decrease of material cost.

such as LCD, Smart phone, LED light

The door will be opened to new world!

High power density converter has possibilities;


High efficiency = low loss =small heat think High power density

4

Toward High power density

Pow

er d

ensi

ty HEV inverter

Air conditioner inverter

Unit Pw. Sup.

Gene. Inv.

On board PS

SiC inverter

Package Power Supply

Break through technologies1)Loss reduction with SiC2)High temp. operation3)Size redc. and high temp.for passive components and heat

think4)High current density: DLB,DCB

This is not true ! Passive components can not be neglected


High power density in product –Switching Power Supply-

Passive components

Control PCB

Air flow

TDK Lambda HFE2500-48semiconductor

Passive components dominates large ratio of power system


Items for size reduction of passive components

Development of new materialMagnetic and capacitive materials are developing; however…

High frequency switching by WBG device (SiC,GaN)

Maximum frequency is limited by passive components materials

Rel

ativ

e P

erm

eabi

lity

Frequency(MHz)

Implementation of PCBwill be difficult

because parasitic capacitance and inductance can not be neglected.

Cooling method will bedifficult

Consideration of best circuit topology is important. (Two level inverter is not best way in all purpose)


Three-level output voltage waveform→Equivalent switching frequency: Double

Ref：http://www.jpubb.com/press/137261/

T-type 3-level

High frequency technique with circuit topology

Multi-level topology (series connection)

Interleave topology(parallel connection)

Loss

2-level

3-level Others

Filter

Switching

Conduction

Four-phase

K. Mino, R. Yamada, H. Kimura, Y. Matsumoto: "Power electronics equipments applying novel SiC power semiconductor modules", Proc. IPEC2014, pp. 1920-1924 (2014)

Input current is composed by triangle waveforms of four-phase current

Equivalent switching frequency:Four times

Reduction of switching loss and filter loss are achieved by circuit topology


Contents


2. Approach I: Matrix converter for reduction of capacitor-Circuit topology and features-Latest technologies: Three-phase to Single phase--Matrix converter in Market: general purpose drive/air conditioner/quick charger




6. Conclusion


Approach I: Matrix converter

Direct power conversion from line frequency AC to other AC

Motor

Input current

0

Output current

0

Input voltage

0 Matrix converter

Output voltage

0

Output voltage: composed by chopping three-phase input voltage with AC switches

Circuit type: four types

Inpu

t:

Output:

Single

Three

Single Three

o

o

o

o

The matrix converter can be applied to three-phase and single-phase input and/or output.


Feature 1: Simple AC to AC direct conversion

ＡＣＤＣＡＣＡＣＡＣ

Diode Rec. +Inverter Matrix converter

Initial charging unit, DCL, Breaking unit, and Smoothing capacitor: Not necessary for MC

Drastic size reduction will be achieved


Input /output waveforms

TH

D o

f in

put

curr

ent

(%)

Load torque (%)0 50 100

10

20

30

Inpu

t po

wer

fac

tor

(%)

Load torqure (%)0 50 100

80

90

100

Input current THD(less than 20th )

Input power factor

Input power factor over 99% at over 50% load

THD 5% at rated load

99

5

Power line200V、50Hz

Matrix converter

Motor

電源電圧

入力電流

出力電流

5ms/div

5A/div

5A/div

200V/div

0

0

0

vr

ir

iu

Feature 2: Low input current harmonics

Sinusoidal input current waveform as well as PWM rectifier

Ditto

Ditto


Realizing drastic energy saving for the applications which has regenerating power flow such as

lifts, cleans and rapid adjustable speed drives with large inertia load

Power line 2

00Ｖ、50Ｈｚ

Matrix converter Motor

Line voltage）250V/div

Input current 5A/div

Output current5A/div

Line voltage250V/div

Input current5A/div

Output current5A/div

Motoring Regenerating

Feature 3: Bidirectional power flow

Regenerated power of motor can be returned to power linewith sinusoidal current

Ditto

Ditto


Comparison among AC-AC converters

Items Diode rec. + Inv. PWM rec. + Inv. Matrix converter

Circuit configuration

Num. of arms Rec:6, INV:6, DB:1 Rec:6, INV:6 CONV:18

Circuit stricture Simple Complicate Complicate

Regeneration Impossible Possible Possible

Semiconductor loss Low Large Middle

Efficiency Middle Low High

Input current harmonics

Large Small Small

Input voltage distortion

Large Small Small

Size Small Large Middle

Cost Low High Middle?

初期充電

ブレーキユニット

DCL

初期充電

同上

同上

Performance of MC is the same to PWM rectifier although smaller size


Limitation of Matrix converter

Maximum Output voltage: 0.866 of Input voltageOutput voltage is composed by three-phase input voltage within envelope curve

Output voltage is limited to less than 0.866 of input voltageOutput voltage becomes 170V at 200V input voltageOutput voltage can be increased by over modulation technique;however input current harmonics will be increased.

0．866

Instantaneous power interruptionNot continuing the operation for instantaneous power interruptiondue to no large energy storage.

Possible to stop and automatic restart without tripSpeedy restart after power return due to no initial charging .Control power supply have to keep during interruptionwith larger capacitors in control board.

Input voltage dropMaximum output voltage is decreased in input voltage drop due to no boost up capability.

PWM rectifier can keep DC voltage even voltage drop.Possible to keep the performance for input voltage distortion, voltage imbalance

Input voltage

Output current

Interruption

Power return

Stop without tripAutomatic restart

It is important to use MC to the applications which can neglect these limitations


R

S

T

U V W

OR

RB-IGBT

IGBT + FWD

Main circuit configuration

9-swtich matrix converter Indirect Matrix converter

Three-phaseAC to AC

AC to DC to AC

Three-phase to single-phase Matrix converter topology is separated intodirect type and indirect type

Indirect matrix converter has no DC link capacitor although snubber capacitor3-1 matrix converter is used for high power application with modular structure and for medium frequency changer


Isolated AC-DC converter with matrix converter

One of disadvantage of matrix converter does not appearin this application because output voltage is adjusted by turn ratio of transformer

Circuit configuration

vr

vsvt

vrvsvt

Inductor

Output sideDC

TransformerSide AC

Step-down

Buck-boost

Inpu

t fil

ter

Output voltage

Boost-up Input sideAC

Control

Simple

Simple

Complicate

Three-types converters according to location of inductor


Step-down isolated matrix converter with DC inductor

DCMediumfrequency Line frequency AC3/1 AC1/DC

requirementSize reduction of

inductor

Current ripple increasing

Solution 1: PWM method considering large DC current ripple [4]

Solution 2: DC current ripple reduction with zero vector distribution of matrix converter [5]

Problems: current ripple reduction with small DC inductor

[5]J. Afsharian, D. D. Xu, B. Wu, B. Gong and Z. Yang, "Analysis of one phase loss operation of three-phase isolated buck matrix-type rectifierwith a boost switch," 2018 IEEE Applied Power Electronics Conference and Exposition (APEC), San Antonio, TX, 2018, pp. 30-36.

[4]S. Takuma, K. Kusaka, J. Itoh: "DC Ripple Component Cancelation Method of Isolated AC-DC Converter with Matrix Converter for Input Current Harmonics Reduction", Energy Conversion Congress Exposition 2019, pp. 6754-6762 (2019)


Dual active bridge type Isolated Matrix converter

or

However, the input voltage is fluctuated by three-phase ac voltage

Basic operating concept is the same to dual active bridge converter

Low EMI and size reduction are achieved by soft switching and increase of frequency

Suitable for high-power isolated power supply, EV charger, aircraft applications etc..


Simple duty ratio calculation method by approximation

Easy to calculate duty ratioPossible to power transfer stiffing soft switching conditionsError between ideal voltage and approximate waveform

depending on load conditions = difficult to compensation

Output voltageof inverter

Transformer current

Output voltageof MC

Equivalent circuit

Output voltage of MC is approximated to square waveform as well as DAB

Use of approximationto square waveform

M. A. Sayed, et.,al "Soft-Switching PWM Technique for Grid-Tie Isolated Bidirectional DC–AC Converter With SiC Device," in IEEE Transactions on IA, vol. 53, no. 6, pp. 5602-5614, . 2017.

Problems occur such as increase of output power error, input current distortion, etc.


Asymmetry modulation method

Output voltageof inverter

Transformer current

Output voltageof MC

Circulating

Average current during one switching period becomes constant for power transfer mode

Positive (A-B) and Negative (A’-B’) ripple

Power transfer

The current ripple in positive period is cancelled by ripple in negative period in transformer

Sinusoidal input current, high power factor

in MFT，and Soft switching operation

S. Takuma, K. Kusaka, J. Itoh, Y. Ohnuma, S. Miyawaki: "A Novel Current Ripple Cancellation PWM for Isolated Three-phase Matrix DAB AC-DC Converter", EPE2019, (2019)

Input voltage

Input current

Output current

Output voltage

Input current THD:4.1 %


Input current harmonics is suppressed to 5%

General Purpose Motor Drive: U1000 Yaskawa Electric

Power Range of U1000:From 200V, 9kVA (output current 28A) up to 400V, 770kVA (output current 930A)

Ref: http://www.e-mechatronics.com/en/product/inverter/u1000/feature/power.html

The biggest motivation to use matrix converter is

the input current harmonics reduction.


Ref: http://www.e-mechatronics.com/en/product/inverter/u1000/feature/allinone.html

Size reduction

Panel size can be reduced by 1/2

Matrix converter has big contribution to size reduction of panel in practically.

The number of components: drastically reduced.

The wiring is only 6


Air conditioners Daikin

Ref: K. Sakakibara et.al., :”Application of Indirect matrix converter for Air conditioners”,IEEJ Transaction on IA, Vol. 136, No. 7, p. 471-478, 2016

rpS

rnS

spS

snS

tpS

tnS

upS

unS

vpS

vnS

wpS

wnS

IPM

cS

preS

400V/50Hz, 5.7kW

Main circuit

Active Clamp circuit: Serial/parallel switching operation according to current

directionTo absorb reactive current at low load power factor To reduce inrush current at initial charge

Current source rectifier: unidirectional power flow


CPU →

FPGA →Clump →capacitor

Filter →capacitor

←Filterreactor

←NoiseFilter

OutputInput

NF1

NF2

NF3

Filter PCB INV PCB

OutputInput

NF1

NF2

NF3

NF4

NF5NF6

NF7,8

CPU forConv. →Inv. →

Noise←Filter

Reactor

additionalX

capacitor

ElectrolyticCapacitors→

Filter PCB INV PCB

Comparison of over view between PWM Rec. and IMC

PWM rectifier +Inverter Indirect matrix converter

The weight is reduced by 30% from 12kg to 8.2kg

Same PCB size although the number of switching device is increased.

Additional noise filters are not necessary


Vdc(200V/div)

Vs(200V/div)Is(10A/di

v)

Iu(10A/div)

Vdc:DC link voltage, Vs:Input phase voltage

Is:Input current, Iu:Output current

0

0

0

4mS/div

Operation waveforms

Performance evaluations(1)

Input current harmonics analysis

Unity power factorSinusoidal input and output current

Meeting IEC 61000-3-2 Class A


85

87

89

91

93

95

97

99

0 1000 2000 3000 4000 5000 6000 7000

Effi

cien

cy

[%]

Speed [min-1]

Diode-Rect.

PWM-Rect.

IMC

Meeting CISPR14-1 (QP)without additional noise filters

Note: PWM rec. switching frequency:15kHzIndirect matrix converter:5.9kHz

Conducted emission Efficiency

Improved by 1.5 - 2% in maximum and by 6% in light load condition(Air conditioner has long operation

time in light load)

6%

2%

Performance evaluations(2)


Volume: reduced by 50% Cost: reduced by 30%

to the conventional charger.

Rapid Charger for EV

The number of products: over 6000 units

Conventional charger Matrix converter charger

co1

1700

mm

750

640

1840

mm

380 670

15kUSD 10kUSD

Anti-series connection IGBT module


PWM rectifier inverter

Three-phase DC AC DC

Downsizing of a transformer

Sinusoidal input current

Advantages

Increasing of power loss

Bulky electrolytic capacitor and input inductor

Disadvantages

Medium frequency transformer

Three-phase Isolated AC-DC converter

50 or 60 Hz

Rectifier

BulkyBulky

LC filter

Small

MC Obtains medium frequency ACfrom line frequency directly

Three-phase AC DCAC

Conventional circuit

Advantage:No electrolytic capacitor and small AC inductors compared to that of conventional circuit

Matrix converter


Experimental verifications

Operation waveforms

(Pout: 40kW)

Over 90%

Maximum efficiency 91.4 %

EfficiencyInput voltage (250V/div)

Input Current (200A/div)

Input current THD: 2.5%

100

0

Output Current (10A/div)

The sinusoidal input current is obtained with 2.5% THD.The efficiency is over 90% in products level.


Contents






6. Conclusion

for reduction of Capacitor


Single-phase grid tied systems

Bucky electrolytic capacitors are connected to DC-linkto absorb power ripple of double frequency of power grid.

Constant power Grid powerAbsorbing power ripple

100 or 120Hz

Approach II: Active power decoupling

such as PCS for PV, On Board Charger for EV etc…

have mismatching between input and output power.


2min

2max2

1VVCE

Capacitor energy

small large

Capacitor voltage is actively oscillated instead of use of large capacitor.

Boost/buck chopper(*)

Small capacitor

(*)H. Hu, S. Harb, N. Kutkut, I. Batarseh, Z. J. Shen : “Power Decoupling Techniques for Micro-inverters in PV Systems — a Review,”Energy Conversion Congress and Exposition 2010, pp.3235-3240, pp. 12-16 (2010)

Principle of Active power decoupling

How to reduce the capacitor in DC-link

Film capacitors or ceramic capacitors will be used instead of electrolytic capacitors

Long life time will be achieved due to no electrolytic capacitor


Passive decoupling Active decoupling

ProsNo cooling system

No inductorSmall capacitance

Low rated voltage switches

ConsElectrolytic capacitors

are needed.High rated

voltage switchesLarge

capacitance

Comparison of these circuits in terms of efficiency and power density

Boost type Buck type

Circuit configuration of Basic active power decoupling


98

98.5

99

99.5

100

0 5000 10000 15000 200005 10 15 20Power density [kW/dm3]

098

98.5

99

99.5

100

SiC-MOSFET 20 kHz

1 kHz

Si-MOSFET 20 kHz

IGBT 15 kHz

SiC-MOSFET 30 kHz

fsw increases.

Pareto-front optimization: Relationship among efficiency, power density and switching frequency

Maximum parameterPower density

Passive18.8 kW/dm3

Active17.0 kW/dm3

EfficiencyPassive

99.9%Active

99.5%

Power density of conventional APD is 90% of that of passive topology.

Comparison of Basic active power decoupling circuits

Output power:6kW, Input :380VDC, Output:200Vac

Grid-tied inverter for PV

J. Itoh, T. Sakuraba, H. N. Le, K. Kusaka: "Requirements for Circuit Components of Single-Phase Inverter Applied with Power Decoupling Capability toward High Power Density", 18th European Conference on Power Electronics and Applications (EPE'16),. DS2a 0291, pp. (2016)


0

0.1

0.2

0.3

0.4

0.5

0.6

1000 2000 3000 4000 5000 60000

0.1

0.2

0.3

0.4

0.5

Vol

ume

[dm

3 ]

0.6

2 3 4 5Rated power [kW]

1 6

CSPI=3 ̊ C/dm3

(Natural cooling)

50 kHz40 kHz

50 kHz

40 kHz

40 kHz35 kHz

30 kHz

35 kHz

30 kHz25 kHz

Switching frequency : 20 kHz

Boost type vs. Buck type

Boost type

Buck type

Minimum volume point is calculated in each rated power.

Buck type < Boost type Buck type > Boost type

Best topology depends on rated power. In high power, boost type is superior


0

20

40

60

80

100

120

従来 SiC‐MOSFET

(100

%=

0.32

dm

3 )

Total volume of APD components is compared to electrolytic cap.

1.1 times larger than that of passive topology

Volume analysis

Boost inductor is dominated by 30% of total volume even though SiC device is used in APD.

Buffer capacitor becomes1/3 of Elec. cap


Buffer circuitI

V

0

Active clamp snubber in indirect matrix converter is used for APD operation

Current source rectifier Voltage source inverter

Small capacitor

SC

Points of the proposed circuit

Only requiring a small capacitor and no additional a boost inductor

The buffer switch (Sc) achieves the zero current switching The efficiency improvement and size reduction are achieved by matrix converter technology

Proposed circuit I: APD with indirect matrix converter

Y. Ohnuma and J. Itoh, "Space vector modulation for a single phase to three phase converter using an active buffer," The 2010 International Power Electronics Conference - ECCE ASIA -, Sapporo, 2010, pp. 574-580.

APD is integrated to PFC


Input current is controlled by switch Sc and zero vectors of inverter

Separating three parts

Operation principle

Load current becomes current source for DC side

Inverter becomes single-switch which is operated by zero vectors

Switch Sc controls buffer capacitor voltage

Operation condition: vc> vIN

Charging : Inverter regenerates the power from load.

Discharge: Sc is turned ON

Inverter & outputBuffer circuit

Idc

vCvIN

Input &Rectifier

Equivalent circuit

In equivalent circuit,

Inverter & outputBuffer circuit

Input &Rectifier


C:50F only

Conventional circuit

(THD 49.8%)

Operation waveforms at 1 kW

Output current can not keep sinusoidal waveform

C:50F + SW

APD circuit with IMC

Output current THD 5.1%

Input current THD 3.9%

Sinusoidaloutput current waveform although the DC capacitor is only 50uF


THD of input and output currents

The minimum value

Output current THD 6.7%

Input current THD 3.3%

Input power factor

Input power factor of over 99 %

although diode rectifier is used

99%



Feature of proposed circuit

The power pulsation is absorbed by the capacitor

Small

Capacitance value is reduced by controlling the voltage variation

Small ac filter because proposed converter is current source type.

Capacitor voltage variation is not influence to the ripple of input and output

ab

IN ac

APD circuit in ChopperCurrent source inverter

Proposed inverter achieves MPPT, grid connection and power decoupling capability

Small

Proposed circuit II: APD integrated to chopper

Additional components are minimized for APD

Ref: Y.Ohnuma, K.Orikawa, J. Itoh: "A Single-phase Current Source PV Inverter with Power Decoupling Capability using an Active Buffer ", IEEE Energy Conversion Congress and Expo 2013, pp. 3094-3101 (2013)


Mode1

Mode4

Input power is directly transferred to ac side

Short circuit mode to keep input inductor current

Proposed converter is operated as boost chopper

APD circuit current pass does not appear in these modes

Operation principle I: chopper mode


Mode3

Buffer capacitor is discharged by connecting in series to input voltage.

Mode2

Buffer capacitor is chargedby connecting in parallel to input voltage

Buffering power is controlled by mode 2 and 3

Operation principle II: Energy management mode

Buffer voltage is independently controlled.



Input voltage:70VDCOutput voltage:100Vac,50HzOutput power: 400WBuffer capacitance: 50uF

Operation waveforms

Constant input current is obtained with 9.33% of input ripple.Sinusoidal output current is obtained with 4.2% of THD.

Maximum efficiency 94.9%Output power factor 99.9%


Eff

icie

ncy

[%]

Maximum power density of 1.8 times is obtained

Volume of buffer capacitor becomes 1/10 due to use of ceramic capacitor

Comparison of power density

Vol

ume

[cm

3 ]

Volume analysis

Heat sink will be smaller if SiC device applied to proposed circuit.

Conventional circuit = boost chopper +VSI


iin vbuf

Cbuf

Lboost

1

234

vdc

Cdc

iL

Vin

Lin

Cin

DCM is used for discharge of buffer capacitor Cbuf

Cbuf：Charge mode

Cbuf：Discharge mode

Average current of iL keeps constantBecause DC input current iin should be

constant direction, constant value

※Current source：Single-phase VSI

Proposed circuit III: APD integrated to chopper with DCM

Boost inductor is shared to APD operation.

Buffer voltage becomes higher than DC-link voltage = boost type

J. Itoh, T. Sakuraba, H. N. Le, K. Kusaka: "Requirements for Circuit Components of Single-Phase Inverter Applied with Power Decoupling Capability toward High Power Density", 18th European Conference on Power Electronics and Applications (EPE'16),.DS2a 0291, pp. (2016)


Without APD operation With APD

Buffer capacitor voltage oscillates at 100 Hz.

100 Hz components in input current is suppressed by 96.8%.

output power: 600 W

Operation waveform

iin iL


1 2 3 43

0 597.5

98.5

99.5

99

98

fsw

1 kHz

DC-DC converter with power decoupling capability

However total loss becomes 1.5 times higherProposed circuit achieves 1.1 times higher power density

Volume comparison

BCM will be applied to improve the efficiency because conduction loss is dominated in total loss


Cbuf：Charging mode

Cbuf：Discharging modeCurrent source：Single-phase VSI

Proposed circuit IV: APD integrated to chopper with DCM

Boost inductor is shared to APD operation.

Buffer voltage becomes lower than DC-link voltage = buck type

-Only one switch from input voltage to DC-link-Low voltage devices applied to APD

Low conduction loss

Same concept to proposed circuit III

J. Itoh, T. Sakuraba, H. Watanabe, N. Takaoka: "DC to Single-phase AC Grid-Connected Inverter using Back Type Active Power Decoupling Circuit without additional magnetic component", Energy Conversion Congress Exposition 2017, pp. 1765-1772 (2017)


Without APD operation With APD

Buffer capacitor voltage oscillates at 100 Hz.

100 Hz components in input current is suppressed by 90.2 %.

Operation waveforms


Efficiency is larger than 94% in region over 10% of rated power

Max. efficiency is 96.0% at 600 W output

Efficiency

Maximum：96.0%


Pareto-front curve of power density and efficiency

Max. power densityConventional buck

type active buffer method4.7 kW/dm3 (98.2%)

Proposed circuit achieves;

1.26 times higher power density is obtained

Proposed DCM active buffer method7.2 kW/dm3 (98.6%)

Passive method5.7 kW/dm3 (98.8%)

Volume of capacitors is reduced by 42.7 %.

Comparison of each DC-DC converter with APD


Neutral point voltage control in T-type inverter is used as APD capability

DCM is applied to ac current control

Proposed circuit V: APD integrated to T-type inverter with DCM

A. Omomo, K. Kusaka , J. Itoh, H. N. Le, N.Takaoka: "T-type NPC Inverter with Active Power Decoupling Method using Discontinuous Current Mode for Micro-Inverter", 7th International Conference on Renewable Energy Research and Applications (ICRERA 2018), (2018)


Operation principle

Neutral point current is controlled for APD

Output current is sum of APD current and reactive current

Reactive current

g g pSum of APD and reactive current become zero

at around 0 and of grid voltage phase.

Time shearing to achieve both control at the same time

Fundamental frequency of neutral point current is the same to grid frequency


Without power decoupling

vout 100 V/div

iout 2 A/div

Iin 1 A/div

With power decoupling

fluctuated

vout 100 V/div

iout 2 A/div

Iin 1 A/div

Harmonic analysis of input current

100 Hz

Time 4 ms/div

Time 4 ms/div

0 A

0 A

0 V

0 A

0 A

0 V

THDiout

3.13%

THDiout

3.19%

Operation waveforms

2nd order harmonics in DC input current is reduced by 77%


Contents




4. Approach III: Discontinuous Current Mode converter for Reduction of indcutor-Principle: Robustness for inductor -Single-phase inverter with DCM-Three-phase inverter with DCM


6. Conclusion


Motivation to use DCM

Problems: Inductor is bulky and expensiveAdvantages of DCM are;1)Use of small inductance with

small volumePeak current becomes double average current in boundary condition Example: 1/10 of inductance, : double peak current then, amount of energy 2/5

Volume is reduced by 60%

3)Small RMS current in light load under same average current 2)Low inductor loss due to optimum design

DCM is suitable for high power

density converters


Disadvantages:-Inductance-dependent-Difficult to design cut-off frequency DCM Regeneration/-Powering transition control has not been realized.

Advantages:-Inductance-independent-Easy to design cut-off frequency-DCM Regeneration/Powering transitioncontrol is realized.

Problem and solution

Pervious duty value is used instead of induction value for nonlinear compensation.

Conventional method Proposed method

Nonlinear characteristic is compensated by inverse model of Transfer function

H. N. Le, K. Orikawa, J. Itoh: "Circuit-Parameter-Independent Nonlinearity Compensation for Boost Converter Operated in Discontinuous Current Mode",IEEE Transactions on Industrial Electronics, Vol. 64, No. 2, pp. 1157-1166 (2017)


CCM/DCM Regeneration/Powering Transition CCM/DCM Syn. Switching:

obtains maximum efficiency of 99.0% and Eff. over 98.7% at all load

Parameters


Powering and Regeneration

Current command responseSame current response is obtainedwith both CCM and DCM


Single-phase Grid-Tied Inverter

small

5ms/div

Iout

[A]

Current Command

DistortionCurrent Detection Value

Applying DCM to Single-phase inverter

Mixed PWM using both CCM andDCM is used to suppress the peak current

-Input current distortion becomes higher because loop gain from command to current is different to each modulation.-Modulation mode should be changed smoothly

H. N. Le, J. Itoh: "Inductance-Independent Nonlinearity Compensation for Single-Phase Grid-Tied Inverter Operating in Both Continuous and Discontinuous Current Mode", IEEE Transactions on Power Electronics, Vol. 34, No. 5, pp. 4904-4919 (2019)


Relationship between duties in DCM and CCM

Conventional Method-Use L to calculate IBCM

-Use IBCM to detect mode

Focusing point in conventional method

Focusing point in proposed method

DCMIICCMII

Vf

VVV

LI

BCMinBCMin

outsw

inoutinBCM

:,:

2

)(1

Proposed Method-Use relationship between DutyDCM and DutyCCM to detect mode

Mode detection between CCM and DCM

Inductance detection

Calculation of DutyDCM and DutyCCM

does not require L.Inductance-independent mode detection


Characteristics of proposed control:① Does not use inductance value to calculate DDCM&DCCM

➁ Use relationship between duties to detect mode

①

➁

Control block diagram of proposed modulation

Inductance-Independent CCM/DCM control is achieved

Pervious duty value is used instead of induction value for nonlinear compensation.


vDCDC link Voltage 350 VvgGrid Voltage 200 VrmsPnNominal Power 4 kW

TdeadtimeDead Time 500 ns25 kHz

1 kHz

fgGrid Frequency 50 Hz

fswSwitching Frequency 100 kHz

fsamp

fc

Sampling Frequency

Crossover Frequency

60oPMPhase Margin

SymbolDescription Value

LgGrid Inductance

LInverter-side Inductance

LfGrid-side Inductance

RdDamping Resistance

CFilter Capacitance

nCapacitance Ratio

LCL Filter Design Case IICase I

2000 H00160 H160 H580 H10 H10 H10 H16 F16 F4 F

1.51.50.516 16 16

Case III

2000 H160 H

10 H16 F

1.58

Experimental and simulation verification

vdcCf

Lg

vg

iout ig

SW1 SW2

SW2 SW1

L Lg

Lf

Lf

LInverter side Grid side

(260,80,60)mm

%Z=1.8% %Z=0.5%


Operation in normal grid

Conventional controller, %ZL=1.8% Conventional controller, %ZL=0.5%

Proposed controller, %ZL=1.8% Proposed controller, %ZL=0.5%

Proposed controller reduces the zero-crossing current distortion


Operation in weak grid for case II & III

3700

3800

3900

4000

4100

4200

4300

-400 -300 -200 -100 0 100 200Real Axis [s]

Imag

inat

ion

Axi

s [s

]

Design Case II

Design Case III

n = 0.5

n = 1.0

n = 1.5

n = 2.0n = 2.5

n = 3.0

Rd increases

Lg = 2000 HL = 160 HLf = 10 HC = 16 F

Proposed controller, %ZL=0.5% R=16 ohm

Proposed controller, %ZL=0.5% R=8 ohm

Pole trajectory

Filter design using pole trajectory enables

achieve desired stabilitywith lowest damping loss


Zero-crossing distortion occurs in output currentas same to single-phase inverter because DCM appears by dead-time

Inverter Output Current

Without dead time With dead time

Continuous Current Mode

Discontinuous Current Mode

Applying DCM to Three-phase inverter

Vdc

LUp Vp

Un Vn

Wp

Wn

L

L

DCM employs whole period to reduce distortionH. N. Le, J. Itoh: "Discontinuous Current Mode Control for Minimization of Three-phase Grid-Tied Inverters in Photovoltaic System", IEEJ Journal of Industry Applications, Vol. 8, No. 1, pp. 90-97 (2019)


How to apply proposed DCM to three-phase

Considering interval 0o~60o

Proposed method splits control of iu and iw

into 2 intervals:(i) D1 and D2 control iu

(ii) D3 and D4 control iw

Desired current waveform

Time sharing control

Each phase current is individually controlled.


Operation of proposed DCM control(700W prototype)

u-phase voltage and INV currents Grid voltage and grid voltage

Magnified waveform

DCM operation is achieved in whole period.


Step response & Current THD characteristic

Step response Current THD characteristic

Gir

d C

urre

nt T

HD

ig[%

]

Quick, stable step response, and Current THD below 5% are obtained at rated load Good control performance


Efficiency Charactrarits Comparison

Maximum efficiency of 97.8%

synchronous switching Compared to asynchronous switching, synchronous switching reduces loss by 33.3% at rated load

1.3% up

(Loss -33.3%)


Contents






6. Conclusion


Current overshoot suppression during Fault Ride-throughfor Three-phase Grid-tied Inverter with Low Inductance

Proposed method: Current overshoot is suppressed with the high-speed update of the inverter output voltage.

Occurrence of current overshoot with low inductance=>Update delay of inverter output at voltage drop and recovery

Abstracts

Satoshi Nagai, Jun-ichi ItohNagaoka University of Technology, Japan

The current overshoot is suppressed to less than 150% in the low inductance (%Z = 0.38%) with the proposed method.


Background

Minimization of system volume with high switching frequency

The inductor achieves small size by reducing the inductance.

Low inductance

Problem

Application

Solar cell*1

Power conditioning system

DCAC

grid

*1 https://biblion.jp/articles/ADTUP*2 S. Nagai, H. N. Le, T. Nagano, K. Orikawa, and J. Itoh, “Minimization of Interconnected Inductor for Single-Phase Inverter with High-Performance

Disturbance Observer”, the 2016 8th International Power Electronics and Motion Control Conference - ECCE Asia, No.Wb8-06 , (2016)

Downsizing of inductor

High switching frequency

Applying SiC or GaN devices to grid-tied inverter

Reduction ofdisturbance suppression performance*2


100

Voltagedrop Voltage

recovery

0 s 1.0 s0

Less than150 ms

20% of rated power/ s

20

5.0 s

FRT requirements

Grid-tied inverters stop at grid disturbanceContinuous operation is necessary without disconnection from grid

Decision of FRT / Disconnection (E. ON code) Output recovery requirements (E. ON code)

(FRT：Fault Ride Through)

Period of voltage sag less than 150 ms: FRT is necessary*3.

FRT requirements

Black out

Current overshoot rate at voltage recovery : Less than 150%*4

*3: Grid Code: High and Extra High Voltage, E.ON Netz GmbH Tech. Rep., 2006, Status:1*4: RPI-M20A, Delta Electronics, Inc., Taipei, Taiwan, 2014. [Online]. Available: http://www.deltaww.com/fileCenter/Products/Download/05/05 01/RPI-M20A_TechData_20141126.pdf


Requirement of grid-tied inverter

Continuing operationduring grid fault

FRT operation with low inductanceInverter operation stopsdue to over current protection.

Purpose

Minimization of inductorSuppression of current overshoot

Voltage sag

Overshoot occurs

Volt. drop Volt. recovery

Volt. sag

Problem for FRT with low inductance

Overshoot occurs


acvt

L

acL

vi t

L

acvt

L ac

L

vi t

L

Principal of current overshoot at voltage sag

The inductor voltage vL is increased due to an update delay of the inverter output voltage vinv at the voltage drop and recovery.

Inductor voltage vL: Large => Occurrence of inductor current overshoot

The current overshoot increases with the low inductance. The update delay should be reduced to suppress the current overshoot.

Conventional FRT: Large update delay of output voltage t


Conventional FRT control (Command change)

Reactive current control*3

=>Adjustment of reactive currenttoward voltage drop level

Command limiter*5

=>Suppression of maximumcurrent during voltage sag Reactive current injection (E.ON code)*3

Current command limiter*5*5: Y. Yang, F. Blaabjerg and H. Wang, "Low-Voltage Ride-Through of Single-Phase TransformerlessPhotovoltaic Inverters," in IEEE Transactions on Industry Applications, vol. 50, no. 3, pp. 1942-1952, 2014.

Impossible to update inverter output with high-speed at voltage drop and recovery

The current overshoot is increased by applying the low inductance.

Rea

ctiv

e cu

rren

t[%

]


Conventional FRT control (high-speed control)

Current control with 1 MHz-PLL and dead-beat control*6

*6: Y. Hanashima and T. Yokoyama, "FRT capability of 100kHz single phase utility interactive inverter with FPGA based hardware controller," 2012 15th International Power Electronics and Motion Control Conference (EPE/PEMC), Novi Sad, 2012, pp. LS8b.1-1-LS8b.1-8.

Dead-beat control:Output current follows in one sampling period.(100 kHz sampling)

1 MHz-PLL: Multi rate sampling(6.7 kHz to 1 MHz)

Achieving high-speed tracking to current command=>It is impossible to suppress the current overshoot in less than

the sampling period (10 s).


Proposed current overshoot reduction during FRT

Current overshoot reduction with high-speed update of inverter output at voltage drop and recovery

At voltage drop At voltage recovery

Current overshoot is reduced=> Counter voltage vector

(Achieving by gate-block operation)

Output vector: Same as grid voltage => Current slope becomes zero

The current overshoot is suppressed with the proposed vector at voltage drop and recovery.

Voltage vectorbefore voltage drop

=>Then, the high-speed voltage sag detection is necessary.


High-speed voltage sag detection method

Voltage sag is detected by HPF and comparator.

(x = a, b, c)

Voltage drop and recovery signals are generated in FPGA.

Threshold:

Signal for the proposed operation is generated by HPF output overshoot at grid voltage drop and recovery.

Grid voltage

MU

X2

5 times for HPF maximum output during rated operation


Inductance is designed with proposed FRT operation.

LL Ii 1_

tL

VIi ac

LL

tL

Vi acL 2_

_

acbd

L th L

VL t

I I

Inductor current overshoot rate is less than 150% for rated inductor current peak at voltage recovery.

Rated current peak Transient current

+

Inductor currentovershoot

Minimized inductance design

Inductance L is designed to0.48 mH (%Z = 0.38%)

Grid voltage peak: 163 V

Rated current peak: 4.0 A

Inductance design to meet FRT requirements

u(t): unit step function


Pout

fcry

L

tdelay

Experimental conditions

Verification in maximum-current-overshoot condition

Experimental condition

Voltage sag: down to 0 V

Using programmable power source:Voltage drop and recovery occur at voltage peak.

400 V

Line to line:200 Vrms

Occurrence of phase detection failure

Voltage drop Voltage recovery

Gridvoltage

Outputcurrent

Phaseinformation

ProgrammableAC power source


Comparison of current overshoot at voltage drop

The current overshoot at the voltage drop is reduced to 230%.

Comparison between conventional and proposed FRT method

With only reactive current control

With prop. method


0

0

Inductor current 5 A/div

Grid voltage100 V/div

a phase

b phase

b phase

a phase Maximum current4.91 A (123%)

c phase

c phase

250 s/div

Voltage recovery with prop. method (a phase peak)

With counter voltage vector (gate-block) => Current overshoot due to command tracking

With Prop. method

The current overshoot with prop. method is reduced to 46%.


Conclusion of FRT

Current overshoot suppression during Fault Ride-throughfor Three-phase Grid-tied Inverter with Low Inductance

The FRT operation was achieved with the proposed method without disconnection from grid.

Inverter side inductance: 0.38%Current overshoot rate: Less than 150%

Verification of asymmetrical voltage sag operation

Future work

Considering current overshoot suppression at voltage drop and recovery with low inductance

The proposed method achieved the high-speed update of the inverter output voltage.


Conclusion

1. Introduction-minimization of passive components is important more and more to obtain

high power density2. Matrix converter is one of candidate to eliminate DC cap.

Products in market: general purpose drive/air conditioner/quick charger3. Active power decoupling is effective for Single-phase converter

It is important to integrate APD to PFC, inverter, or chopper to realize high power density

4. Discontinuous Current Mode converter is good way to minimize inductors.Inductance-Independent CCM/DCM control is achieved for single-phase and three-phase inverters

5. FRT capability of grid-tied inverter becomes important when minimized inductoris used.High speed gate block method is one of solution to meet FRT standard.


Thank you for your kind attention

Thousand of Japanese sake bottles are waiting for youwhen you visit Japan.

Minimization Technologies for Passive Components …...Power Electronics Lab. 1 Minimization...

Documents

Transcript of Minimization Technologies for Passive Components …...Power Electronics Lab. 1 Minimization...