169. flexible power electronic transformer
TRANSCRIPT
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
1. INTRODUCTION
POWER electronic transformers (PETs) are proposed to replace conventional
transformers and perform voltage regulation and power exchange between generation and
consumption by electrical conversion. The previous researches show that PETs have a great
capacity to receive much more attention due to their merits such as high-frequency link
transformation and flexible regulation of the voltage and power. Although many studies have
been conducted on application and control of PET in power systems, less attention is paid to the
areas of the circuit topologies. The topology of PET can be developed in such a way to achieve
multiport electrical system that converts variable input waveform to the desired output
waveform. In addition, for higher voltage applications or three phase systems, the topology is
expandable as it is modular.
In this paper, a new PET topology named flexible power electronic transformer (FPET)
is proposed. As shown in Fig. 1, it is constructed based on modules and a common dc link,
which is used to transfer energy between ports and isolate all ports from each other. In this
bidirectional topology, each port can be considered as an input or output. Each module consists
of three main parts, including modulator, demodulator, and high frequency isolation transformer
(HFIT). The modulator is a dc– ac converter and the demodulator is an ac–ac converter; both
with bidirectional power flow capability. Each module operates independently and can transfer
power between ports. These ports can have many different characteristics, such as voltage level,
frequency, phase angle, and waveform. As a result, FPET can satisfy almost any kind of
application, which are desired in power electronic conversion systems and meet future needs of
electricity networks. Considering this point, it is named flexible. The simulation results of high-
voltage application are given to clarify the advantages of the proposed FPET over the recently
developed PETs. To show the flexibility of the proposed PET, a prototype is built and tested.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
2. POWER ELECTRONIC TRANSFORMER
An isolated high-frequency link AC/AC converter is often termed as a power electronic
transformer (PET). Because of high frequency operation of the magnetic core they have size and
cost advantages over conventional transformers. These transformers achieve high frequency AC
power transformation without any DC capacitor link. The transformer provides isolation and
voltage transformation while the power converters provide with high frequency operation. Also
termed as high frequency transformers, they have been extensively researched for various
applications, on account of many advantages over conventional line frequency transformers.
The PET has a wide range of applications including electrical distribution systems, wind power
generation etc. In the advantages of PET over conventional transformers are discussed.
2.1 Introduction
In electric power distribution system, transformers perform several functions, such as
voltage transformation, isolation, noise decoupling etc. The conventional distribution
transformers operate at low frequencies (60 Hz) making them bulky and expensive. A power
electronic transformer (PET) operates at much higher frequencies, of the range of several kHz.
The transformer size, which is inversely proportional to the frequency and saturation flux
density, could be reduced under high frequency operation conditions. The PET utilizes power
electronic converters along with a high-frequency transformer to obtain overall size and cost
advantages.
The PET substitutes the conventional 60 Hz transformer at the PCC of a micro-grid,
connecting the later with the utility. This results in an enhanced power management for the
micro-grid as well as decentralized control of the DERs and controllable loads within the micro-
grid. A dynamic control of active and reactive power flow from the utility is possible. It also
allows a bi-directional flow of active power between the utility and the micro-grid. The high
frequency AC power transformation is achieved without a DC link.
Also a smooth transition from grid-connected to isolated mode of micro-grid is possible.
To better acknowledge the claims of the PET, it is essential to understand its operation and
topology, described in the following sections.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
2.2 PET topology
The PET consists of a high frequency transformer with three-phase to single-phase
matrix converters on its primary and secondary as briefly described in chapter 1. Figure.1 shows
a more detailed schematic of the PET. The proposed topology consists of two matrix converters
with high frequency AC link. The primary side of the electronic transformer is supplied by
utility AC source and the secondary side has the equivalent AC source of a micro-grid. The
three phase input AC voltage at the line frequency (60Hz) is first converted into high frequency
(10 kHz) single phase voltage by the input side matrix converter. The output side converter is
also a three phase to single phase matrix converter.
This yields high frequency pulsating single phase AC voltage at the primary and
secondary of the high frequency transformer with leakage reactance Llk . As seen in figure 3-1
the utilities define the limit for the reference power Pref , for a particular micro-grid it serves at
the PCC. The average active power as measured at the output AC source can thus be restricted.
By calculating the corresponding modulation signal for the secondary side matrix converter, the
equivalent phase shift can be achieved. The active power flow between the input side and output
side of the PET is regulated similar to that in a dual active bridge. The active power flow is
controlled by regulating the phase shift between the primary and secondary voltages at the
transformer.
The regulated phase shift angle corresponds to the desired active power limit set by the
utilities. This is achieved by a PI controller designed for the control signal of the PWM strategy
applied for matrix converter modulation. The proportional and integral gain parameters KP and
KI are designed to provide a faster response while eliminating any steady state error.
The PWM strategy also allows a control over the phase shift between the voltage and
current at input as well as output and hence can control the reactive power. Voltage regulation
can thus be achieved by controlling the reactive power at the front-end converter.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Figure.1 PET schematic
2.3 ISOLATION TRANSFORMER
An isolation transformer is a transformer used to transfer electrical power from a source
of alternating current (AC) power to some equipment or device while isolating the powered
device from the power source, usually for safety. Isolation transformers provide galvanic
isolation and are used to protect against electric shock, to suppress electrical noise in sensitive
devices, or to transfer power between two circuits which must not be connected together.
Suitably designed isolation transformers block interference caused by ground loops. Isolation
transformers with electrostatic shields are used for power supplies for sensitive equipment such
as computers or laboratory instruments.
Strictly speaking any true transformer, whether used to transfer signals or power, is
isolating, as the primary and secondary are not connected by conductors but only by induction.
However, only transformers whose primary purpose is to isolate circuits (opposed to the more
common transformer function of voltage conversion), are routinely described as isolation
transformers. Given this function, a transformer sold for isolation is often built with special
insulation between primary and secondary, and is tested, specified, and marked to withstand a
high voltage between windings, typically in the 1000 to 5000 volt range.
Sometimes the term is exceptionally used to clarify that some transformer, although not
primarily intended for isolation, is a true transformer rather than an autotransformer (whose
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
primary and secondary are not isolated from each other).[1] Even step-down power transformers
required, amongst other things, to protect low-voltage equipment from mains voltage by
isolating the secondary and primary such as are used in older "wall warts", are not usually
described specifically as "isolation transformers".
Some very small transformers—e.g. 4 transformers in one tiny dual in-line (DIP) chip
package—used to isolate high-frequency low-voltage (logic) pulse circuits (e.g., 500V RMS
primary–secondary for one second), are described as isolation transformers.
Isolation transformers are commonly designed with careful attention to capacitive
coupling between the two windings. The capacitance between primary and secondary windings
would also couple AC current from the primary to the secondary. A grounded Faraday shield
between the primary and the secondary greatly reduces the coupling of common-mode noise.
This may be another winding or a metal strip surrounding a winding.
2.3.1 Applications
A simple 1. Isolation transformer with an extra dielectric barrier and an electrostatic shield between primary and secondary. The grounded shield prevents capacitive coupling between primary and
secondary windings.
In electronics testing and servicing an isolation transformer is a 1:1 (under load) power
transformer used for safety. Without it, exposed live metal in a device under test is at a
hazardous voltage relative to grounded objects such as a heating radiator or oscilloscope ground
lead (a particular hazard with some old vacuum-tube equipment with live chassis). With the
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
transformer, as there is no conductive connection between transformer secondary and earth,
there is no danger in touching a live part of the circuit while another part of the body is earthed.
Electrical isolation is considered to be particularly important on medical equipment, and special
standards apply. Often the system must additionally be designed so that fault conditions do not
interrupt power, but generate a warning.
Isolation transformers are also used for the power supply of devices not at ground
potential. An example is the Austin transformer for the power supply of air-traffic obstacle
warning lamps on radio antenna masts. Without the isolation transformer, the lighting circuits
on the mast would conduct radio-frequency energy to ground through the power supply.
Metal boats are subject to corrosion if they use earthed power from shore when moored,
due to galvanic currents that flow through the water between shore earth and the hull. This can
be avoided by using an isolation transformer with the primary and case connected to shore
earth, and the secondary "floating".
A metal safety screen between primary and secondary is connected to shore earth; in the
event of a fault current in the primary (due, e.g., to insulation breakdown) it will cause the fault
current to return and trip a shore-based circuit breaker rather than making the hull live.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
3. INSULATED GATE BIPOLAR TRANSISTOR (IGBT)
IGBT has been developed by combining into it the best qualities of both BJT and
PMOSFET. Thus an IGBT possesses high input impedance like a PMOSFET and has low on-
state power loss as in a BJT. Further, IGBT is free from second breakdown problem present in
BJT. All these merits have made IGBT very popular amongst power-electronics engineers.
IGBT is also known as metal oxide insulated gate transistor (MOSIGT), conductively-
modulated field effect transistor (COMFET) or gain-modulated FET (GEMFET). It was also
initially called insulated gate transistor (IGT).
The insulated-gate bipolar transistor or IGBT is a three-terminal power
semiconductor device, noted for high efficiency and fast switching. It switches electric power in
many modern appliances: electric cars, variable speed refrigerators, air-conditioners, and even
stereo systems with digital amplifiers. Since it is designed to rapidly turn on and off, amplifiers
that use it often synthesize complex waveforms with pulse width modulation and low-pass
filters.
The IGBT combines the simple gate-drive characteristics of the MOSFETs with the
high-current and low–saturation-voltage capability of bipolar transistors by combining an
isolated-gate FET for the control input, and a bipolar power transistor as a switch, in a single
device. The IGBT is used in medium- to high-power applications such as switched-mode power
supply, traction motor control and induction heating. Large IGBT modules typically consist of
many devices in parallel and can have very high current handling capabilities in the order of
hundreds of amps with blocking voltages of 6,000 V.
The IGBT is a fairly recent invention. The first-generation devices of the 1980s
and early 1990s were relatively slow in switching, and prone to failure through such modes as
latch up and secondary breakdown. Second-generation devices were much improved, and the
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
current third-generation ones are even better, with speed rivaling MOSFETs, and excellent
ruggedness and tolerance of over loads.
3.1 BASIC STRUCTURE
Fig illustrates the basic structure of an IGBT. It is constructed virtually in the same
manner as a power MOSFET. There is however , a major difference in the substrate. The n+
layer substrate at the drain in a PMOSFET is now substituted in the IGBT by a p+ layer
substrate called collector C. Like a power MOSFET, an IGBT has also thousands of basic
structure cell connected approximately on a single chip of silicon.
In IGBT, p+ substrate is called injection layer because it injects holes into n -
layer. The n- layer is called drift region. As in other semiconductor devices, thickness of n - layer
determines the voltage blocking capability of IGBT. The p layer is called body of IGBT.The n -
layer in between p+ and p regions serves to accommodate the depletion layer of pn- junction ,
i.e. junction J2.
N-Channel IGBT Cross Section
3.1.1 Equivalent Circuit
An examination of reveals that if we move vertically up from collector to
emitter. We come across p+, n- , p layer s. Thus, IGBT can be thought of as the combination of
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
MOSFET and p+ n- p layer s. Thus, IGBT can be thought of as the combination of MOSFET
and p+ n- p transistor Q1 .Here Rd is resistance offered by n – drift region. Approximate
equivalent circuit of an IGBT.
Exact equivalent circuit
The existence of another path from collector to emitter, this path is collector, p +, n-, p (n-
channel), n+ and emitter. There is, thus, another inherent transistor Q2 as n- pn+ in the structure
of IGBT. The interconnection between two transistors Q1 and Q2.This gives the complete
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
equivalent circuit of an IGBT. Here Rby is the existence offered by p region to flow of hole
current Ih .
The two transistor equivalent circuit illustrates that an IGBT structure has a
parasitic thyristor in it. Parasitic thyristor is shown in line.
3.2 WORKING
When collector is made positive with respect to emitter, IGBT gets forward biased. With
no voltage between gate and emitter, two junctions between n- region and p region (i.e. junction
J2) are reversed biased; so no current flows from collector to emitter
When gate is made positive with respect to emitter by voltage VG, with gate-
emitter voltage more than the threshold voltage VGET of IGBT, an n-channel or inversion layer,
is formed in the upper part of p region just beneath the gate, as in PMOSFET . This n- channel
short circuits the n- region with n+ emitter regions. Electrons from the n+ emitter begin to flow
to n- drift region through n-channel. As IGBT is forward biased with collector positive and
emitter negative, p+ collector region injects holes into n- drift region .In short; n-drift region is
flooded with electrons from p-body region and holes from p+ collector region. With this, the
injection carrier density in n- drift region increases considerably and as a result, conductivity of
n- region enhances significantly. Therefore, IGBT gets turned on and begins to conducts
forward current IC.
Current Ic , or Ie of two current components:
1. Holes current Ih due to injected holes flowing from collector ,p+ n- p transistor Q1, p-
body region resistance Rby and emitter .
2. Electronic current Ie due to injected electrons flowing from collector, or load, current
IC=emitter current Ie=Ih+Ie.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Major component of collector current is electronic current Ie, i.e. main current path for
collector, or load, current is through p+, n -, drift resistance Rd and n-channel resistance Rch.
Therefore, the voltage drop in IGBT in its on-state is
Vc e . o n = I c . R c h + I c . Rd + V j i
=voltage drop [in n - channel] + across drift in n- region + across forward
biased p+ n- junction J1.
Here Vji is usually 0.7 to 1v as in a p-n diode. The voltage drop Ic. Rch is due to n-channel
resistance, almost the same as in a PMOSFET. The voltage drop Vdf = Ic.Rd in UGBT is much
less than that in PMOSFET. It is due to substantial increase in the conductivity caused by
injection of electrons and holes in n- drift region. The conductivity increase is the main reason
for the low on-state voltage drop in IGBT than that it is in PMOSFET.
3.3 LATCH-UP IN IGBT
From the above that IGBT structure has two inherent transistors Q1 and Q2,
which constitute a parasitic thyristor. When IGBT is on, the hole current flows through
transistor p+ n- p and p- body resistance Rby. If load current Ic is large, hole component of
current Ih would also be large. This large current would increase the voltage drop Ih. Rby which
may forward bias the base p- emitter n+ junction of transistor Q2. As a consequence, parasitic
transistor Q2 gets turned on which further facilitates in the turn-on of parasitic transistor p+ n- p
labeled Q1. The parasitic thyristor, consisting of Q1 and Q2, eventually latches on through
regenerative action, when sum of their current gains α1+α2 reaches unity as in a conventional
thyristor .With parasitic thyristor on, IGBT latches up and after this, collector emitter current is
no longer under the control of gate terminal. The only way now to turn-off the latched up IGBT
is by forced commutation of current as is done in a conventional thyristor .If this latch up is not
aborted quickly, excessive power dissipation may destroy the IGBT. The latch up discussed
here occurs when the collector current Ice exceeds a certain critical value .the device
manufactures always specify the maximum permissible value of load current Ice that IGBT can
handle without latch up.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
At present, several modifications in the fabrication techniques are listed in the
literatures which are used to avoid latch-up in IGBTs. As such, latch up free IGBTs are
available.
3.4 IGBT Characteristics
The circuit shows the various parameters pertaining to IGBT characteristics.
Static I-V or output characteristics of an IGBT (n-channel type) show the plot of collector
current Ic versus collector-emitter voltage Vce for various values of gate-emitter voltages
VGE1, VGE2 etc .These characteristics are shown below .In the forward direction, the shape of
the output characteristics is similar to that of BJT . But here the controlling parameter is gate-
emitter voltage VGE because IGBT is a voltage controlled device. When the device is off,
junctionJ2 blocks forward voltage and in case reverse voltage appears across collector and
emitter, junction J1 blocks it. Vrm is the maximum reverse breakdown voltage.
The transfer characteristic of an IGBT is a plot of collector current Ic versus gate-
emitter voltage VGE .This characteristics is identical to that of power MOSFET. When VGE is
less than the threshold voltage VGET, IGBT is in the off state.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Static V-I characteristics
3.5 SWITCHING CHARACTERISTICS
Switching characteristics of an IGBT during turn-on and turn-off are
sketched. The turn-on time is defined as the time between by instance of forward blocking to
forward on-state. Turn-on time is composed of delay time tdn and rise time tr ,i.e. ton=tdn+tr.
The delay time is defined as the time for the collector-emitter voltage to fall from Vce to 0.9
Vce. Here Vce is the initial collector-emitter voltage. Time tdn may also be defined as the time
for the collector current to rise from its initial leakage current Ice to 0.1 Ic. Here Ic is the final
value of the collector current .
The rise time tr is the time during which collector-emitter falls from 0.9VCE to
0.1VCE. IT is also defined as the time for the collector current to rise from 0.1Ic to its final
value Ic. After time ton, the collector current Ic and the collector-emitter voltage falls to small
value called conduction drop=VCES where subscript s denotes saturated value.
The turn-off time is somewhat complex. It consists of three intervals
1. Delay time tdf
2. Initial fall time tf1
3. Final time tf2
I.e. toff=tdf+tf1+tf2
The delay time is the time during which gate voltage falls from VGE to threshold voltage
VGET.As VGE falls to VGET during tdf, the collector current falls from Ic to 0.9 Ic. At the end
of the tdf, collector-emitter voltage begins to rise. The first fall time Tf1 is defined as the time
during which collector current falls from 90 to 20 % of its initial value Ic, or the time during
which collector-emitter voltage rises from Vces to 0.1 Vce.
The final fall time tf2 is the time during which collector current falls from 20 to
10% of Ic, or the time during which collector-emitter voltage rises from 0.1 VCE to final value
VCE.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
3.6 APPLICATIONS OF IGBTS
IGBTs are widely used in medium power applications such as AC and DC
motor drives, UPS systems, power supplies and drives for solenoids, relays and contactors.
Though IGBTs are somewhat more expensive than BJTs, yet they are becoming popular
because of lower gate-drive requirement, lower switching losses and smaller snubber circuit
requirements. IGBT converter are more efficient with less size as well as cost, as compared to
converters based on BJTs. Recently, IGBT inverter induction-motor drives using 15-20KHZ.
Switching frequency favor where audio-noise is objectionable. In most applications, IGBTs will
eventually push out BJTs. At present , the state of the art IGBTs of 1200vots, 500 Amps ratings,
0.25-20 µs turn off time with operating frequency are available.
Comparison of IGBT with MOSFET
Relative merits and demerits of IGBT over PMOSFET are enumerated below.
1. In PMOSFET, the three terminals are called gate, source, drain where as the
corresponding terminal for the IGBTs are gate, emitter and collector.
2. Both IGBT and PMOSFET posses high input impedance.
3. Both are voltage control devices.
4. With rising temperature, increase in on-state resistance in PMOSFET is much
pronounced than in IGBT. So on state voltage drop and losses rise rapidly in
PMOSFET than IGBT, with rising temperature.
5. With rising
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
4. PULSE WIDTH MODULATION (PWM)
Pulse Width Modulation (PWM) is the most effective means to achieve constant voltage
battery charging by switching the solar system controller’s power devices. When in PWM
regulation, the current from the solar array tapers according to the battery’s condition and
recharging needs consider a waveform such as this: it is a voltage switching between 0v and
12v. It is fairly obvious that, since the voltage is at 12v for exactly as long as it is at 0v, then a
'suitable device' connected to its output will see the average voltage and think it is being fed 6v -
exactly half of 12v. So by varying the width of the positive pulse - we can vary the 'average'
voltage.
Similarly, if the switches keep the voltage at 12 for 3 times as long as at 0v, the average
will be 3/4 of 12v - or 9v, as shown below and if the output pulse of 12v lasts only 25% of the
overall time, then the average is
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
By varying - or 'modulating' - the time that the output is at 12v (i.e. the width of the
positive pulse) we can alter the average voltage. So we are doing 'pulse width modulation'. I
said earlier that the output had to feed 'a suitable device'. A radio would not work from this: the
radio would see 12v then 0v, and would probably not work properly. However a device such as
a motor will respond to the average, so PWM is a natural for motor control.
4.1 PULSE WIDTH MODULATOR
So, how do we generate a PWM waveform? It's actually very easy, there are circuits
available in the TEC site. First you generate a triangle waveform as shown in the diagram
below. You compare this with a d.c voltage, which you adjust to control the ratio of on to off
time that you require. When the triangle is above the 'demand' voltage, the output goes high.
When the triangle is below the demand voltage, the
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
When the demand speed it in the middle (A) you get a 50:50 output, as in black. Half the
time the output is high and half the time it is low. Fortunately, there is an IC (Integrated circuit)
called a comparator: these come usually 4 sections in a single package. One can be used as the
oscillator to produce the triangular waveform and another to do the comparing, so a complete
oscillator and modulator can be done with half an IC and maybe 7 other bits.
The triangle waveform, which has approximately equal rise and fall slopes, is one of the
commonest used, but you can use a saw tooth (where the voltage falls quickly and rinses
slowly). You could use other waveforms and the exact linearity (how good the rise and fall are)
is not too important.
Traditional solenoid driver electronics rely on linear control, which is the application of
a constant voltage across a resistance to produce an output current that is directly proportional to
the voltage. Feedback can be used to achieve an output that matches exactly the control signal.
However, this scheme dissipates a lot of power as heat, and it is therefore very inefficient.
A more efficient technique employs pulse width modulation (PWM) to produce the
constant current through the coil. A PWM signal is not constant. Rather, the signal is on for part
of its period, and off for the rest. The duty cycle, D, refers to the percentage of the period for
which the signal is on. The duty cycle can be anywhere from 0, the signal is always off, to 1,
where the signal is constantly on. A 50% D results in a perfect square wave. (Figure 1)
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
A solenoid is a length of wire wound in a coil. Because of this configuration, the
solenoid has, in addition to its resistance, R, a certain inductance, L. When a voltage, V, is
applied across an inductive element, the current, I, produced in that element does not jump up to
its constant value, but gradually rises to its maximum over a period of time called the rise time
(Figure 2). Conversely, I does not disappear instantaneously, even if V is removed abruptly, but
decreases back to zero in the same amount of time as the rise time.
Therefore, when a low frequency PWM voltage is applied across a solenoid, the current
through it will be increasing and decreasing as V turns on and off. If D is shorter than the rise
time, I will never achieve its maximum value, and will be discontinuous since it will go back to
zero during V’s off period (Figure 3).* In contrast, if D is larger than the rise time, I will never
fall back to zero, so it will be continuous, and have a DC average value. The current will not be
constant, however, but will have a ripple.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
At high frequencies, V turns on and off very quickly, regardless of D, such that the
current does not have time to decrease very far before the voltage is turned back on. The
resulting current through the solenoid is therefore considered to be constant. By adjusting the D,
the amount of output current can be controlled. With a small D, the current will not have much
time to rise before the high frequency PWM voltage takes effect and the current stays constant.
With a large D, the current will be able to rise higher before it becomes constant.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
4.2 WHY THE PWM FREQUENCY IS IMPORTANT
The PWM is a large amplitude digital signal that swings from one voltage extreme to the
other. And, this wide voltage swing takes a lot of filtering to smooth out. When the PWM
frequency is close to the frequency of the waveform that you are generating, then any PWM
filter will also smooth out your generated waveform and drastically reduce its amplitude. So, a
good rule of thumb is to keep the PWM frequency much higher than the frequency of any
waveform you generate.
Finally, filtering pulses is not just about the pulse frequency but about the duty cycle
and how much energy is in the pulse. The same filter will do better on a low or high duty cycle
pulse compared to a 50% duty cycle pulse. Because the wider pulse has more time to integrate
to a stable filter voltage and the smaller pulse has less time to disturb it the inspiration was a
request to control the speed of a large positive displacement fuel pump. The pump was sized to
allow full power of a boosted engine in excess of 600 Hp.
At idle or highway cruise, this same engine needs far less fuel yet the pump still
normally supplies the same amount of fuel. As a result the fuel gets recycled back to the fuel
tank, unnecessarily heating the fuel. This PWM controller circuit is intended to run the pump at
a low speed setting during low power and allow full pump speed when needed at high engine
power levels.
4.3 MOTOR SPEED CONTROL (POWER CONTROL)
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Typically when most of us think about controlling the speed of a DC motor we think of
varying the voltage to the motor. This is normally done with a variable resistor and provides a
limited useful range of operation. The operational range is limited for most applications
primarily because torque drops off faster than the voltage drops.
Most DC motors cannot effectively operate with a very low voltage. This method also
causes overheating of the coils and eventual failure of the motor if operated too slowly. Of
course, DC motors have had speed controllers based on varying voltage for years, but the range
of low speed operation had to stay above the failure zone described above.
Additionally, the controlling resistors are large and dissipate a large percentage of
energy in the form of heat. With the advent of solid state electronics in the 1950’s and 1960’s
and this technology becoming very affordable in the 1970’s & 80’s the use of pulse width
modulation (PWM) became much more practical. The basic concept is to keep the voltage at the
full value and simply vary the amount of time the voltage is applied to the motor windings.
Most PWM circuits use large transistors to simply allow power On & Off, like a very fast
switch.
This sends a steady frequency of pulses into the motor windings. When full power is
needed one pulse ends just as the next pulse begins, 100% modulation. At lower power settings
the pulses are of shorter duration. When the pulse is On as long as it is Off, the motor is
operating at 50% modulation. Several advantages of PWM are efficiency, wider operational
range and longer lived motors. All of these advantages result from keeping the voltage at full
scale resulting in current being limited to a safe limit for the windings.
PWM allows a very linear response in motor torque even down to low PWM% without
causing damage to the motor. Most motor manufacturers recommend PWM control rather than
the older voltage control method. PWM controllers can be operated at a wide range of
frequencies. In theory very high frequencies (greater than 20 kHz) will be less efficient than
lower frequencies (as low as 100 Hz) because of switching losses.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
The large transistors used for this On/Off activity have resistance when flowing current,
a loss that exists at any frequency. These transistors also have a loss every time they “turn on”
and every time they “turn off”. So at very high frequencies, the “turn on/off” losses become
much more significant. For our purposes the circuit as designed is running at 526 Hz. Somewhat
of an arbitrary frequency, it works fine.
Depending on the motor used, there can be a hum from the motor at lower PWM%. If
objectionable the frequency can be changed to a much higher frequency above our normal
hearing level (>20,000Hz).
4.4 PWM CONTROLLER FEATURES
This controller offers a basic “Hi Speed” and “Low Speed” setting and has the option to
use a “Progressive” increase between Low and Hi speed. Low Speed is set with a trim pot inside
the controller box. Normally when installing the controller, this speed will be set depending on
the minimum speed/load needed for the motor. Normally the controller keeps the motor at this
Lo Speed except when Progressive is used and when Hi Speed is commanded (see below). Low
Speed can vary anywhere from 0% PWM to 100%.
Progressive control is commanded by a 0-5 volt input signal. This starts to increase
PWM% from the low speed setting as the 0-5 volt signal climbs. This signal can be generated
from a throttle position sensor, a Mass Air Flow sensor, a Manifold Absolute Pressure sensor or
any other way the user wants to create a 0-5 volt signal. This function could be set to increase
fuel pump power as turbo boost starts to climb (MAP sensor). Or, if controlling a water
injection pump, Low Speed could be set at zero PWM% and as the TPS signal climbs it could
increase PWM%, effectively increasing water flow to the engine as engine load increases.
This controller could even be used as a secondary injector driver (several injectors could
be driven in a batch mode, hi impedance only), with Progressive control (0-100%) you could
control their output for fuel or water with the 0-5 volt signal.
Progressive control adds enormous flexibility to the use of this controller. Hi Speed is
that same as hard wiring the motor to a steady 12 volt DC source. The controller is providing
100% PWM, steady 12 volt DC power. Hi Speed is selected three different ways on this
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
controller: 1) Hi Speed is automatically selected for about one second when power goes on.
This gives the motor full torque at the start. If needed this time can be increased ( the value of
C1 would need to be increased). 2) High Speed can also be selected by applying 12 volts to the
High Speed signal wire. This gives Hi Speed regardless of the Progressive signal.
When the Progressive signal gets to approximately 4.5 volts, the circuit achieves 100%
PWM – Hi Speed.
4.5 HOW DOES THIS TECHNOLOGY HELP
The benefits noted above are technology driven. The more important question is how the
PWM technology jumping from a 1970’s technology into the new millennium offers:
• LONGER BATTERY LIFE
– reducing the costs of the solar system
– reducing battery disposal problems
• MORE BATTERY RESERVE CAPACITY
– increasing the reliability of the solar system
– reducing load disconnects
– Opportunity to reduce battery size to lower the system cost
• GREATER USER SATISFACTION:
– get more power when you need it for less money
5. MODELLING OF CASE STUDY
5.1 PROPOSED POWER CIRCUIT OF FPET
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
The proposed circuit is shown in Fig. 2. It should be mentioned that the proposed
topology can be expanded by connecting modules in series or parallel to obtain higher voltage
or current ratings, and to form star/delta connections for three phase applications.
As shown in Fig. 2(a), each port is composed of a full bridge dc-link inverter (FBDCI),
HFIT, and a Cyclo-converter. This topology consists of independent and similar modules and
each port can work independently. Thus, the analysis of one port is sufficient to introduce whole
topology. The FBDCI (modulator) can operate as an inverter when it converts the dc-link
voltage to an ac waveform at the HFIT side. It can operate as an active rectifier when it converts
the ac waveform of the HFIT to the dc-link voltage. The FBDCI is used to achieve zero-voltage
level, adjustable pulse width, and symmetrical switching. In addition, the number of switches
can be reduced to obtain simpler circuit than the latter, shown in Fig. 2(b). In this case, one of
the half-bridge circuits can be considered as the reference or master leg. Once gate pulses for
the master leg (i.e., switches and ) are provided, the gate pulses of the other legs (slave legs)
have a phase shift respect to the master leg. Using this control strategy, the number of switches
can be reduced to half.
Fig2.Proposed circuit of the FPET (a) Basic topology and (b) reduced switch topology
The modulator can be described as follows:
1) Bi directional power flow capability;
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
2) Adjustable switching frequency that feet voltage pulses frequency into the pass band of
HFIT;
3) Stored energy in the dc link (if the modulator is in active rectifier mode). For
cycloconverters, several circuit topologies can be proposed using unidirectional or bidirectional
switches.
In this paper, a typical cycloconverter with two bidirectional switches operates as the
demodulator. The demodulator converts high frequency voltage (i.e.) to low frequency voltage
(i.e., Vpr1 ) and vice versa. The specifications of the demodulator are listed as follows:
1) Bidirectional power flow capability; and
2) Providing zero voltage switching by turning the switches of cycloconverter ON/OFF, while
voltage of HFIT riches to zero.
5.2 MODULATION AND DEMODULATION OPERATION PRINCIPLES
The well-known phase shift modulation (PSM) method is shown in Fig. 3. The
definition of parameters is given in Table I.
TABLE IDEFINITION OF PARAMETERS
The voltage regulation is performed by the FBDCI using PSM method. The cycloconverter
chooses the PSM pulses in such a way to provide positive or negative voltage polarity at the
output. In this figure, the cycloconverter provides positive output voltage polarity as an
example. On one hand, the switches of cycloconverter turn ON/OFF with a time delay (Tcd )
respect to those of FBDCI, so they operate under zero voltage condition.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
On the other hand, the switches have a small overlapping time to provide a path for Lf
current to avoid high stresses at switching instants. Thus, the switches operate at soft switching
condition. The leakage inductance of HFIT should be minimized as much as possible. In
practice, snubber circuits must be used to damp the stored energy in the leakage inductance of
HFIT
Fig.3. Principle of PSM method
According to Fig. 3, the duty cycle of FBDCI is defined as follows:
The modulated voltage at the secondary side for one duty cycle is expressed by (2)
The modulated voltage at the output of cycloconverter (Vc) is determined as follows:
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Where sign (tk ) function determines the polarity of Vc that can be positive or negative
according to the desired output voltage and presented by (4), as shown at the bottom of the
page.
A. PSM Control Circuit
Fig. 4. Schematic presentation of PSM controller.
The control circuit is responsible for providing pulse gate of dc link switches and the
cycloconverter. The implementation of PSM is shown in Fig. 4. The input data address consists
of four lines. The first line is polarity of output voltage signi. The second line is switch-enabled
of cycloconverter (EnableCi ). The third line is switch-enabled of dc link (EnableSi ). The
fourth line provides the duty cycle data of the ith port. The enabled lines are provided by the
startup and protection circuits.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
B. Utilization of Ports as a Voltage Source
As an example when a port (assuming ith port) is designed to operate as a voltage
source, it can provide a constant voltage regardless of the active or reactive power that is
exchange between the port and the grid. So, a controllable voltage at the output of
cycloconverter can be obtained and it is given by
Where (t) is the reference voltage. According to (4), one may obtain the following
approximation:
Where the asterisk symbols show the next stage values Therefore, the duty cycle and the
sign function are achieved as follows:
Because of high switching frequency, it is expected to assume is constant over time
period of kTs < t < (k + 1)Ts. The duty cycle is a function of dc-link voltage (Vd (kTs)) and the
turn winding of the HFIT at the ith port. The block diagram of controller is shown in Fig. 5.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
5.3 ENERGY BALANCE IN FPET
In every system, there is a balance among losses, input energy and output energy. This
balance for FPET is presented as follows
Fig.5. Control circuit of a typical port that operates as a voltage source
TABLE II
DESCRIPTION OF PARAMETERS PREARRANGED IN (10)
Where Wi, WCd, and Wloss are the input/output energy, stored energy at dc link and
losses, respectively. Neglecting the power losses, (8) can be approximated by
To achieve power equilibrium in Cd and have constant dclink voltage, some of the ports
should absorb and inject desired active power. The algorithm for regulation of dc-link voltage is
as follows:
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Step 1: At the start-up instant, following two methods can be used to charge the dc-link
capacitor to the desired value.
1) The dc-link capacitor can be charged by an extra dc source. As the desired dc-link voltage
achieved, the dc source should be disconnected.
2) The cycloconverter can provide a high frequency voltage across HFIT. When the voltage
passes through HFIT, it changes to a dc voltage across dc-link capacitor by the body diodes of
FBDCI switches. The dc voltage can charge the capacitor considering the winding ratio of
HFIT. The startup current is limited by Lf .
Step 2: dc-link voltage checking.
1. If then there is no need for adjustments. The
ΔVd, Ref is a fraction of Vd, Ref that is required to provide Hystersis band.
2. If then voltage should be
regulated and the port powers should be adjusted.
Step 3: Return to the second step.
A. Balancing Ports:
For another solution to regulate voltage of dc link, some ports are considered as
“balancing ports” that provide energy to balance dc-link voltage in FPET. One of the main
objectives of these kinds of ports is to control voltage level in the dc-link voltage, particularly
when over voltage or voltage drop occurs in the dc link. Assuming the ith port is chosen as the
balancing port, the main component of the cycloconverter voltage, and output of the port are
given as follows:
Fig.6. Simplified diagram of FPET
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
The definition of the parameters is given in the Table II. Therefore, neglecting the
resistance of output filter inductance, the active power of the port is obtained as follows:
Applying the differences between Vd and Vd,Ref as an error signal to a typical PI
controller, the value of required Pi can be estimated. According to (6) and (7), the duty cycles
are achieved.
5.4 DESIGN PROCEDURE
A. DC-Link Capacitor
Fig. 6 shows the voltage and currents of all ports and the dc link capacitor. The
following equation presents the instantaneous power balance of the losses in FPET.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Fig.7. Proposed HV FPET
The voltage and current of ports can have different polarity and directions. If the
currents and voltages of ports have sinusoidal waveforms, then (12) can be rewritten as follows:
Now, the input power of dc link can be expressed as follows:
(14)
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
This input power consists of two components. The first component is the pulsation
power with angular frequency of 2ωi and the second one is the dc power . Assuming
Vd,ref as the voltage of capacitor and Id as the average current, (14) can be rewritten as follows:
The ripple voltage of the dc-link capacitor (ΔVd ) can be approximated as follows:
Thus, the minimum value of Cd can be calculated for the maximum voltage ripple.
B. Reference Voltage of DC Link and Winding Ratio of HFIT
From practical point of view, lower dc-link voltage results in lower voltage stress of
switches. But according to (17), as Vd, Ref decreases, the voltage ripple increases. In addition,
the decrease of the dc-link voltage increases the current of dc link switches. Consequently, by
selecting an appropriate dc link reference voltage (Vd,Ref ) and the maximum ripple voltage,
the minimum dc-link voltage (Vd,min) can be determined. In the worst condition, the lowest dc-
link voltage (Vd,min), maximum duty cycle (D = Dmax) and the maximum magnitude of
desired.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
TABLE III
PARAMETERS OF PETS
Fig.8. Port voltage and current of HV FPET
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Fig.9. Load voltage and current of three-phase output
Voltage (Vi,max) can determine the winding ratio as follows:
C. Matching Inductance Lf
Matching inductance Lf should limit the output current to its maximum acceptable value
(Ii,max) during the switching period (Ts). For the ith port, the following assumptions can be
considered:
Where Δifi is the variation of the cycloconverter current for one switching period Based
on these assumptions, Lf is determined by
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
5.5 COMPARISON STUDY
A comparison study is given to clarify the advantageous and disadvantageous of the
FPET. A three-phase system, contains six ports, is compared to the similar PETs. First, some of
the pros and cons of bidirectional FPET in comparison to the unidirectional topologies should
discuss. In the unidirectional systems, input power factor is not controllable but in bidirectional
structures input or even output power factor can be adjusted. This means that the reactive and
active power of each port can be regulated. Also for DG systems like wind turbine, bidirectional
capability is indispensable. Energy management for energy efficient systems is another
application of this feature. A detail comparison study (e.g., cost, efficiency, quality, etc.) is
given in Table V to clarify the pros and cons of FPET and the existing topologies proposed.
As can be seen from Table V, conversion efficiency of FPET is relatively low in
comparison to the similar circuits topologies proposed. The main reason is the usage of power
snubber, and voltage clamp circuits, which damp absorbed energy in leakage inductors of HFIT.
To reduce the size of protection circuits in FPET, a PSM approach is utilized, so the
cycloconverter switches just select the PSM pulses and can commutate naturally. Therefore, the
switches communicate at almost zero voltage. In addition, because of overlap technique the
voltage surge is reduced over the switches and the continuous current flow in the output filter
(Lf) is not interrupted.
In addition, Table V shows some of the most noticeable applications of FPET. Dynamic
voltage restorer (DVR) and active filter (AF) applications can be satisfied by the FPET, because
it can connect to the grid in series or/and in parallel. Desired voltage and current can provide by
the flexibility of FPET in providing various waveforms (see Section VI). FPET can provide
desired waveform in each phase (or port) independently, so this can be used in universal power
quality conditioner (UPQC). FPET can transfer active and reactive power from one port or
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
phase to another port or one phase. This in power distribution system is very useful for interline
power flow controller (IPFC).
Additionally, FPET can provide symmetrical three-phase voltage from an asymmetrical ac
source in the form of an uninterruptible power supply application (UPS). FPET can play a role
in providing useful power from variable low-voltage dc sources. That is suitable for renewable
energy applications such as photovoltaic and fuel cell. Design simplicity and expandability (to
achieve higher ratings) are other advantageous of FPET.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
6. CASE STUDY AND RESULTS
Flexible Transformer
Fig 6.1 flexible power electronic transformer circuit diagram 1
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Fig 6.2 Output voltage at port 8
Fig 6.3 dc link voltage
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Fig 6.4 load voltages of three phase output at ports 6,7,8
Fig 6.5 load voltage at port 1
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Fig 6.6 load current at port 1
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Flexible Transformer 1
Fig 6.7 flexible power electronic transformer circuit 1
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Fig 6.8 load voltage at port 8
Fig 6.9 dc link voltage
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
Fig 6.10 load current of three phase output at ports 6,7,8
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
CONCLUSION
Based on the requirement of a flexible power conversion system, FPET is proposed to
facilitate many requirements that are expected in power electronic and distribution systems. The
proposed topology is flexible enough to provide bidirectional power flow and has as many ports
as it is required. For low-voltage application, FPET can correct power factor and can adjust the
waveform and frequency of the output voltage. The proposed topology can be expanded for
high voltage and high current applications. The dc link plays a significant role to provide energy
balance, power management in the circuit and independent operation of ports. The measurement
results verify the basic theoretical concepts of this paper. The advantages of the FPET are:
bidirectional power flow capability of ports, module-based topology, which can be used in
different forms, independent operation of ports, flexibility in power amount and direction in all
ports, and double galvanic isolation between each port, as well as using only one storage
element.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
REFERENCES
[1] S. H. Hosseini, M. B. Sharifian, M. Sabahi, A. Yazdanpanah, and G. H. Gharehpetian,
“Bidirectional power electronic transformer for induction heating systems,” in Proc. Can. Conf.
Electr. Comput. Eng., May 4–7, 2008, pp. 347–350.
[2] S. H. Hosseini, M. Sabahi, and A. Y. Goharrizi, “Multi-function zero voltage and zero-
current switching phase shift modulation converter using a Cycloconverters with bidirectional
switches,” IET Power Electron. JNL, vol. 1, no. 2, pp. 275–286, Jun. 2008.
[3] M. Sabahi, S. H. Hosseini, M. B. Sharifian, A. Yazdanpanah, and G. H. Gharehpetian, “A
three-phase dimmable lighting system using a bidirectional power electronic transformer,” IEEE
Trans. Power Electron., vol. 24, no. 3, pp. 830–837, Mar. 2009.
[4] D.Wang, M. Changing, L. Jiming, S. Fan, and C. Luonan, “The research on characteristics
of electronic power transformer for distribution system,” in Proc. IEEE Transmits. Distrib.
Conf. Exhib. Asia Pacific, 2005, pp. 1–5.
[5] M. Huasheng, Z. Bo, Z. Jianchao, and L. Xuechao, “Dynamic characteristics analysis and
instantaneous value control design for buck-type power electronic transformer (PET),” in Proc.
IEEE Annu. Conf. Ind. Electron. Soc. IECON, Nov. 2005, pp. 1043–1047.
[6] H. Wrede, V. Staudt, and A. Steimel, “Design of an electronic power transformer,” in Proc.
IEEE 28th Annu. Conf. Ind. Electron. Soc., 2002, vol. 2, pp. 1380–1385.
[7] J. Aijuan, L. Hangtian, and L. Shaolong, “A new high-frequency AC link three-phase four-
wire power electronic transformer,” in Proc. IEEE Conf. Ind. Electron. Appl., May 2006, pp. 1–
6.
[8] H. Krishnaswami and V. Ramanarayanan, “Control of high-frequency AC link electronic
transformer,” IEE Proc. Elect. Power Appl., May 2005, vol. 152, no 3, pp. 509–516.
DEPT OF EEE SRTIST
43
FLEXIBLE POWER ELECTRONIC TRANSFORMER
[9] S. Farhangi, H. Iman-Eini, J. L. Schanen, and J. Aime, “Design of power electronic
transformer based on cascaded H-bridge multilevel converter,” in Proc. IEEE Int. Symp. Ind.
Electron., Jun. 2007, pp. 877–882.
[10] D. Chen and J. Liu, “The uni-polarity phase-shifted controlled voltage mode AC-AC
converters with high frequency AC link,” IEEE Trans. Power Electron., vol. 21, no. 4, pp. 899–
905, Jul. 2006.
DEPT OF EEE SRTIST