lecture power electronics interactions between electrical machine and power electronics technische...

LecturePower Electronics

InteractionsBetween Electrical Machine and Power

Electronics

Technische Universität MünchenLehrstuhl für Elektrische Antriebssysteme

und LeistungselektronikProf. Dr.-Ing. Ralph Kennel

ralph.kennel@tum.de

Additional Losses

Additional LossesCurrent Harmonics

Quelle : Prof. A. Binder, Technische Universität Darmstadt

with increasingswitching frequency

the current harmonicscaused by the inverter

decrease

Iron Losses under Inverter Supply

Quelle : PTB

f / Hz (fundamental oscillation)

Additional LossesInfluence of Switching Frequency

Quelle : Prof. A. Binder, Technische Universität Darmstadt

mains supply inverter supply

2 pole squirrel cage induction machine 3 kW, 380 V, Y connectionrated frequency 50 Hz, slip 4.5 %, torque 10 Nmvoltage source inverter 8.3 kVA, 400 V

at frequency 9.6 kHz motor efficiency is high ( less temperature rise)the overall efficiency, however, is the same as at Frequency 4.8 kHzAt frequency 19.2 kHz motor current harmonics are low, but switching losses increase

EMC

Electro Magnetic Compatibility

E

M

C

even

more

confusion

Cabinet Designwith Modern Servo Drives

signalelectronics

powerelectronics

Shielding and Grounding

following these requirements and advicethere is „conductive EMI“ only

when using power electronics inverters(… usually no „radiation EMI“)

Shielding and Grounding

Reactive Power

Reactive Power

+

-

+

-

+

-

Motor(e. g. induction machine)

U0

induction machines needreactive power

for magnetization

… in this cablereactive power

can be measured !

this is a DC link there is

no reactive powerby definition

… in this cable reactive power

cannot be measured !

Reactive Power

+

-

+

-

+

-

Motor(e. g. induction machine)

U0

induction machines needreactive power

for magnetizationthis is a DC link there is

no reactive powerby definition

… in this cable reactive power

cannot be measured !

… where doesreactive powercome from ???

… as the sumof reactive power

in all 3 phases is zero („0“) !

Reactive Power

+

-

+

-

+

-

Motor(e. g. induction machine)

U0

induction machines needreactive power

for magnetization… where does

reactive powercome from ???

… it is no problem for the inverter to provide it

… as the sumof reactive power

in all 3 phases is zero („0“) !

Reactive Power

+

-

+

-

+

-

Motor(e. g. induction machine)

U0

… it is no problem for the inverter to provide it

… as the sumof reactive power

in all 3 phases is zero („0“) !

… with regard to reactive power the inverter is like a marshalling yard (switching station) for trains !

Reactive Power

+

-

+

-

+

-

Motor(e. g. induction machine)

U0

… with regard to reactive power the inverter is like a marshalling yard (switching station) for trains !

… therefore inverters can be used easily for compensating reactive power in grids !

… especially in regenerative energy applicationslike wind power farms or solar power arrays !

Reactive Power

voltage

current

reactivepower

t

t

t

Reactive Power

voltage

current

reactivepower

t

t

t

Active Power

activepower

voltage

current

t

t

t

Active Power

activepower

…, of course, this can be split mathematicallyinto a constant term (which is active power)and a fluctuating term(which can be considered reactive power)

t

Active Power

activepower

…, of course, this can be split mathematicallyinto a constant term (which is active power)and a fluctuating term(which can be considered reactive power)

… this is, however,a mathematical operation only

… because there is no momentwith power flowing in backward direction

… in fact this is pulsating active power only !!!

… from physical perspectivethere is not really reactive power

t

Active Power

activepower

… in fact this is pulsating active power only !!!

Pulsating Active Power ≠ Reactive Power

PFC

=PowerFactorControl

Power Factor=

PW (active power)

PS (apparent power)

PFC

… when using line commutated (Thyristor-)convertersthis was an issue indeed

voltaget

currentt

cos = 10,60,2- 0,3

PFC

… when using line commutated (Thyristor-)convertersthis was an issue indeed

voltaget

currentt

cos =

firing angle control takes carefor a phase shift between current wave and voltage wave !

10,6

… for that reason the „inductive“ reactive power had to be compensated by „kapacitive“reactive power

PFC

… in case of diode rectifiers

the power factor is usually cos = 1

voltaget

currentt

cos = 1

and in case of fully controlled rectifier bridges

… PFC is meant to include harmonics as well !

– in spite of harmonics having nothing to do with „PFC“ „PFC“ is often meant

to compensate harmonics

Rectifiers

not allowed in the public grid(with respect to impact

to the grid)

RectifiersB4 bridge with capacitive load

current shape

… today‘s discussion is dealing with harmoncs !

– in spite of harmonics having nothing to do with „PFC“ „PFC“ is often meant

to compensate harmonics

… filtering effort would be significant!

0

20

40

60

80

100

120

H1 H3 H5 H7 H9 H11 H13 H15 H17 H19 H21

without filtering

harmonics spectrum

RectifiersB4 bridge with capacitive load

not allowed in the public grid(with respect to impact

to the grid)

… different (better) solution :fully controlled front end rectifier

current shape harmonics spectrum

RectifiersB4 bridge with capacitive load

… filtering effort would be significant!

Quelle : Prof. A. Binder, Technische Universität Darmstadt

Additional LossesCurrent Harmonics

with increasingswitching frequency

the current harmonicscaused by the inverter

decrease

line voltage

line current

motoring

regeneration

… different (better) solution :fully controlled front end rectifier

+

-

+

-

+

-

Netz

U0 ≈

line voltage

line current

motoring

regeneration

current shape

Rectifiers

+

-

+

-

Netz

U0 ≈

harmonics spectrum

0

20

40

60

80

100

120

H1 H3 H5 H7 H9 H11 H13 H15 H17 H19 H21

with optimized filtering

Rectifiers

current shape harmonics spectrum

fully controlled front end rectifier

• … nevertheless !!! … even with a good 1phase „PFC“ …• either the load has to be charged by pulses (law of energy conservation !)• or an energy storage device must be implemented

… no tricky control schemecan change that !!!

… or – after all – can it?

Rectifiers

PFC

… how must the current shape look liketo provide a constant power flow ?

voltaget

currentt

power

t… of course, at u = 0 and/or i = 0 no power can be transmitted (law of energy conservation !)The time to be bypassed by the energy storage device (e. g. capacitance),

(= energy), however, is significantly smaller !

PFC

… is this current shape allowed ???

voltaget

currentt

power

t… please calculate the harmonics spectrum …it is surprising,

how close one can get to this current shapewithout exceeding the standard limits of grid harmonics

PFC

… is this current shape allowed ???

voltaget

currentt

power

t… please calculate the harmonics spectrum …some companies make use of this effect,

to reduce the size of the DC link capacitance, … but, of course, they do not tell that in public.

Travelling Waves

-800

-600

-400

-200

0

200

400

600

800

0 0.002 0.004 0.006 0.008 0.01

vo

lta

ge

in

V

time in s

typical „voltage pattern“at the output of a PWM voltage source inverter

Travelling Waves

-100

0

100

200

300

400

500

600

700

0 4e-06 8e-06 1.2e-05 1.6e-05 2e-05

volta

ge

in

V

time in s

Travelling Waves

typical „voltage step“at the output of a PWM voltage source inverter

M

time

term

inal

vol

tage

Travelling Waves

M

… what now ?

… which case is it ?

„fix“ end

„loose“ end

adaptation… „it depends“ … on what ?

… whether we consider currents or voltages !!!

… in our case : voltages

… for voltages the motor is a „loose“ end

Travelling Waves

time

term

inal

vol

tage

M

… what now ?

… which case is it ?

„fix“ end

„loose“ end

adaptation

… for voltages the inverter is a „fix“ end

… if the inverter output voltagedid not change meanwhile,

the wave is inverted and travels back again

Travelling Waves

time

term

inal

vol

tage

M

„loose“ end

… on the inverter side the voltage remains constant(fix end !)

… on the motor side voltage oscillations occurrup to the double value of DC link voltage

(loose end !)

Travelling Waves

time

term

inal

vol

tage

0

200

400

600

800

1000

1200

0 4e-06 8e-06 1.2e-05 1.6e-05 2e-05vo

ltage

in

V

time in s

-100

0

100

200

300

400

500

600

700

0 4e-06 8e-06 1.2e-05 1.6e-05 2e-05

volta

ge

in

V

time in s

… so far … so good… the matter, however, is getting much worse,

as soon as the inverterswitches simultaneously „into“ the back travelling voltage wave

Travelling Waves

… on the inverter side the voltage remains constant(fix end !)

… on the motor side voltage oscillations occurrup to the double value of DC link voltage

(loose end !)

M

Travelling Waves

time

term

inal

vol

tage

M

„loose“ end

… until here everything is like before …

Travelling Waves

time

term

inal

vol

tage

M

… what now ?

„fix“ end

… in case the inverter has switchedthe voltage at its output meanwhile,

the wave travels back with amplification

Travelling Waves

time

term

inal

vol

tage

M

„loose“ end

Travelling Waves

time

term

inal

vol

tage

… on the inverter sidethe voltage is „impressed“

(fix end !)

… on the motor sidevoltage oscillations occurr

up to 2,7 timesthe DC link voltage

(loose end !)

… what is so critical ?

Travelling Waves

Voltage Flashover within Winding

… what is so critical ?

„Horror“ Picture

… the danger is real …- behind such pictures, however,

there is acommercial interest !!

… what is so critical ?

Extract from IEC paper IEC 2 (CD) 5661991 (in Germany) : appendix to IEC 34

as long as you supply standard induction motorsby inverters with• voltage peaks below 1000 V• voltage rise times below 500 V/µsyou should not expect any danger for the motor

Compatibility between Inverter and Motor

… these are realistic valuesfor modern inverters !!!

Insulation of WireReasoning

• the critical voltage resulting in a flashoverdoes not depend at all on the diameter of the wire

• doubling the thickness of wire insulationincreases the critical voltage by 15 %

(the must significant effect results fromcovering faults of the first layer by the second layer)

that is „state of the art“ today !!!

• increasing the operation temperature to 155 °Clowers the critical voltage by 15 %

Voltage Stress on Partial Coils

Voltage Stress on (Partial) Coils

0

200

400

600

800

1000

1200

0 4e-06 8e-06 1.2e-05 1.6e-05 2e-05

volta

ge

in

V

time in s

0

200

400

600

800

1000

1200

0 4e-06 8e-06 1.2e-05 1.6e-05 2e-05

volta

ge

in

V

time in s

-100

0

100

200

300

400

500

600

700

800

0 4e-06 8e-06 1.2e-05 1.6e-05 2e-05

volta

ge

in

V

time in s

Voltage Stress on (Partial) Coils

0

200

400

600

800

1000

0 2e-06 4e-06 6e-06 8e-06 1e-05

volta

ge

in

V

time in s

Voltage Stress on (Partial) Coils

… to explain this effect,the representationas a simple equivalent circuit

containing a serial connectionof concentrated inductances

is not sufficient!!!

… in this casethe motor winding has to be represented

– like an electric cable – by a serial network

of two-ports

Voltage Stress on (Partial) Coils

… voltage “waves“ spread out within ther motor windings– as in an electrical cable –

according to the laws of cable equation

Voltage Stress on (Partial) Coils

… the motor windings only has inductive behaviour,if the rising time of the voltage edge

is significantly larger than the group delay of the complete motor winding

if the rising time of the voltage edge is smallerthan the group delay of the complete motor winding,

the capacitive behaviour is predominant !!!

Voltage Stress on (Partial) Coils

0

200

400

600

800

1000

0 2e-06 4e-06 6e-06 8e-06 1e-05

volta

ge

in

V

time in s

Voltage Stress on (Partial) Coils

0

100

200

300

400

500

600

700

0 2e-06 4e-06 6e-06 8e-06 1e-05

volta

ge

in

V

time in s

Voltage Stress on (Partial) Coils

-200

0

200

400

600

800

1000

0 4e-06 8e-06 1.2e-05 1.6e-05 2e-05

volta

ge

in

V

time in s

… that is alarming !!!the first voltage pulse

appears nearly completely at the entrance coil

Voltage Stress on (Partial) Coils

… using cost effective winding processesthe single wires are distributed randomly in the slot !!

… therefore the insulation of the single wiremust be designed with respect to

the full voltage stress !!!

Voltage Stress on (Partial) Coils

… remember:

… on the motor side entstehenvoltage oscillations occurr

up to 2,7 timesthe DC link voltage

Voltage Stress on (Partial) Coils… using cost effective winding processes

the single wires are distributed randomly in the slot !!

… therefore the insulation of the single wiremust be designed with respect to

the full voltage stress !!!