Dual

INTRODUCTION Background

Objective

Thesis Outline

ABSTRACT

To fulfil the growing demand for independently supplying a number of ac loads in several

applications , needs separate inverters for each system. But, the main disadvantage of these

system is that, this may lead to undesirable increase in system cost, size and weight. The

conventional reduced switch count system consist of single phase three legs, four legs & six legs

converter respectively. Thus to overcome the above drawback our project uses reduced switch

topology three switch with respect to six-switch topology. There is a growing trend in power

electronics for reduced switch count power converters with the aim of sustaining high power

quality and enhancing the system reliability. A three switch single leg topologies are developed

which are functionally to full bridge inverter working independently though with a less number

of semiconductor switches and hence control and gate drive circuit components. The main

objective of our project is to reduce the semiconducting switches and increase the loads. In this

paper 3 semiconductor switches which is parallel connected to capacitor and using two ac loads

with equal frequency and different frequency . Keywords-Reduced switch count inverter,power

converter,full bridge,EF & DF.

INTRODUCTION

There is a growing trend in power electronics for reduced switch count power converters with

the aim of sustaining high power quality and enhancing the system reliability. The two ac outputs

are independent from each other regarding both frequency and amplitude. The nine-switch

structure is a recent reduced switch count inverter topology proposed for independently

supplying two three-phase ac loads. The usual approach to fulfilling the growing demand for

independently supplying a number of ac loads in several applications is using separate inverters

for each of them. This may lead to undesirable increase in system cost, size and weight. There is

a growing trend in power electronics for reduced switch count power converters with the aim of

sustaining high power quality and enhancing the system reliability. Dual-terminal converters

provide the researches with an extra degree of freedom in realizing switch reduction of power

converters resulting in system cost, size and weight optimization. Dual-terminal reduced switch

count topologies may be utilized as ac/ac converters for conditioning the input ac power

(regulating input voltage/current) or be employed as dual-output inverters to independently

supply two ac loads.The nine-switch structure is a recent reduced switch count inverter topology

proposed for independently supplying two three-phase ac loads.

1.1. BACKGROUND 1.1.1. Power quality

The PQ issue is defined as “any occurrence manifested in voltage, current, or frequency deviations

that results in damage, upset, failure, or disoperation of end-use equipment.” Almost all PQ issues are

closely related with PE in almost every aspect of commercial, domestic, and industrial application.

Equipment using power electronic devise are residential appliances like TVs, PCs etc. business and

office equipment like copiers, printers etc. industrial equipment like programmable logic controllers

(PLCs), adjustable speed drives (ASDs), rectifiers, inverters, CNC tools and so on. The Power

Quality (PQ) problem can be detected from one of the following several symptoms depending on the

type of issue involved.

• Lamp flicker

• Frequent blackouts

• Sensitive-equipment frequent dropouts

• Voltage to ground in unexpected

• Locations

• Communications interference

• Overheated elements and equipment.

PE are the most important cause of harmonics, inter harmonics,

notches, and neutral currents. Harmonics are produced by rectifiers, ASDs, soft starters,

electronic ballast for discharge lamps, switched-mode power supplies, and HVAC using ASDs.

Equipment affected by harmonics includes transformers, motors, cables, interrupters, and

capacitors (resonance). Notches are produced mainly by converters, and they principally affect

the electronic control devices. Neutral currents are produced by equipment using switched-mode

power supplies, such as PCs, printers, photocopiers, and any triplets generator. Neutral currents

seriously affect the neutral conductor temperature and transformer capability. Interharmonics are

produced by static frequency converters, cyclo-converters, induction motors & arcing devices.

Equipment presents different levels of sensitivity to PQ issues, depending on the type of

both the equipment and the disturbance. Furthermore, the effect on the PQ of electric power

systems, due to the presence of PE, depends on the type of PE utilized. The maximum acceptable

values of harmonic contamination are specified in IEEE standard in terms of total harmonic

distortion.

Power electronics are alive and well in useful applications to overcome distribution

system problems. Power electronics has three faces in power distribution: one that introduces

valuable industrial and domestic equipment; a second one that creates problems; and, finally, a

third one that helps to solve those problems. On one hand, power electronics and

microelectronics have become two technologies that have considerably improved the quality of

modern life, allowing the introduction of sophisticated energy-efficient controllable equipment to

industry and home. On another hand, those same sensitive technologies are conflicting with each

other and increasingly challenging the maintenance of quality of service in electric energy

delivery, while at the same time costing billions of dollars in lost customer productivity.

1.1.2. Solutions to power quality problems There are two approaches to the mitigation of power quality problems. The first approach is

called load conditioning, which ensures that the equipment is made less sensitive to power

disturbances, allowing the operation even under significant voltage distortion. The other solution is to

install line-conditioning systems that suppress or counteract the power system disturbances. Passive

filters have been most commonly used to limit the flow of harmonic currents in distribution systems.

They are usually custom designed for the application. However, their performance is limited to a few

harmonics, and they can introduce resonance in the power system. Among the different new

technical options available to improve power quality, active power filters have proved to be an

important and flexible alternative to compensate for current and voltage disturbances in power

distribution systems. The idea of active filters is relatively old, but their practical development was

made possible with the new improvements in power electronics and microcomputer control strategies

as well as with cost reduction in electronic components. Active power filters are becoming a viable

alternative to passive filters and are gaining market share speedily as their cost becomes competitive

with the passive variety. Through power electronics, the active filter introduces current or voltage

components, which cancel the harmonic components of the nonlinear loads or supply lines,

respectively. Different active power filters topologies have been introduced and many of them are

already available in the market.

1.1.3. Power filter topologies

Depending on the particular application or electrical problem to be solved, active power

filters can be implemented as shunt type, series type, or a combination of shunt and series active

filters (shunt-series type). These filters can also be combined with passive filters to create hybrid

power filters.

The shunt-connected active power filter, with a self-controlled dc bus, has a topology similar to

that of a static compensator (STATCOM) used for reactive power compensation in power

transmission systems. Shunt active power filters compensate load current harmonics by injecting

equal-but opposite harmonic compensating current. In this case the shunt active power filter

operates as a current source injecting the harmonic components generated by the load but phase-

shifted by 180°.

Series active power filters were introduced by the end of the 1980s and operate mainly as a

voltage regulator and as a harmonic isolator between the nonlinear load and the utility system.

The series-connected filter protects the consumer from an inadequate supply-voltage quality.

This type of approach is especially recommended for compensation of voltage unbalances and

voltage sags from the ac supply and for low-power applications and represents an economically

attractive alternative to UPS, since no energy storage (battery) is necessary and the overall rating

of the components is smaller. The series active filter injects a voltage component in series with

the supply voltage and therefore can be regarded as a controlled voltage source, compensating

voltage sags and swells on the load side. In many cases, series active filters work as hybrid

topologies with passive LC filters. If passive LC filters are connected in parallel to the load, the

series active power filter operates as a harmonic isolator, forcing the load current harmonics to

circulate mainly through the passive filter rather than the power distribution system. The main

advantage of this scheme is that the rated power of the series active filter is a small fraction of

the load kVA rating, typically 5%. However, the apparent power rating of the series active power

filter may increase in case of voltage compensation.

The series-shunt active filter is a combination of the series active filter and the shunt active filter.

The shunt active filter is located at the load side and can be used to compensate for the load

harmonics. On the other hand, the series portion is at the source side and can act as a harmonic

blocking filter. This topology has been called the Unified Power Quality conditioner. The series

portion compensates for supply voltage harmonics and voltage unbalances, acts as a harmonic

blocking filter, and damps power system oscillations. The shunt portion compensates load

current harmonics, reactive power, and load current unbalances. In addition, it regulates the dc

link capacitor voltage. The power supplied or absorbed by the shunt portion is the power

required by the series compensator and the power required to cover losses.

Hybrid power filters are a combination of active and passive filters. With this topology the

passive filters have dynamic low impedance for current harmonics at the load side, increasing

their bandwidth operation and improving their performance. This behavior is reached with only a

small power rating PWM inverter, which acts as an active filter in series with the passive filter.

Multilevel inverters are being investigated and recently used for active filter topologies.

Three-level inverters are becoming very popular today for most inverter applications, such as

machine drives and power factor compensators. The advantage of multilevel converters is that

they can reduce the harmonic content generated by the active filter because they can produce

more levels of voltage than conventional converters (more than two levels). This feature helps to

reduce the harmonics generated by the filter itself. Another advantage is that they can reduce the

voltage or current ratings of the semiconductors and the switching frequency requirements. The

more levels the multilevel inverter has, the better the quality of voltage generated because more

steps of voltage can be created.

1.1.4. Voltage source converters

Most of the active power filter topologies use voltage source converters, which have a voltage

source at the dc bus, usually a capacitor, as an energy storage device. This topology, shown in

Figure 1.1, converts a dc voltage into an ac voltage by appropriately gating the power

semiconductor switches. Although a single pulse for each half cycle can be applied to synthesize

an ac voltage, for most applications requiring dynamic performance, pulse width modulation

(PWM) is the most commonly used today. PWM techniques applied to a voltage source inverter

consist of chopping the dc bus voltage to produce an ac voltage of an arbitrary waveform. There

are a large number of PWM techniques available to synthesize sinusoidal patterns or any

arbitrary pattern. With PWM techniques, the ac output of the filter can be controlled as a current

or voltage source device.

Figure 1.1.Voltage source converter topology for active filters.

Figure 1.2 shows the way PWM works by means of one of the simplest and most common

techniques: the triangular carrier technique. It forces the output voltage va over a switching

cycle, defined by the carrier period of Vcar, to be equal to the average amplitude of the

modulating wave Va ref. The resulting voltages for a sinusoidal modulation wave contain a

sinusoidal fundamental component Va(1) and harmonics of unwanted components. These

unwanted components can be minimized using a frequency carrier as high as possible, but this

depends on the maximum switching frequency of the semiconductors (IGBTs, GTOs, or IGCTs).

Figure.1.2. the PWM carrier Technique (triangular carrier).

The modulation strategy shown in Figure 1.3 uses a triangular carrier, which is one of many

strategies applied today to control power inverters. Depending on the application (machine

drives, PWM rectifiers, or active power filters), some modulation strategies are more suitable

than others. The modulation techniques not only allow controlling the inverters as voltage

sources but also as current sources. Figure 1.3 shows the compensating current generated for a

shunt active power filter using three different modulation techniques for current-source inverters.

These three techniques are periodical sampling (PS), hysteresis band (HB), and triangular carrier

(TC). The PS method switches the power transistors of the active filter during the transitions of a

square wave clock of fixed frequency: the sampling frequency. The HB method switches the

transistors when the error exceeds a fixed magnitude: the hysteresis band. The TC method

compares the output current error with a fixed amplitude and fixed triangular wave: the

triangular carrier. Figure 1.3 shows that the HB method is the best for this particular waveform

and application because it follows more accurately the current reference of the filter. When

sinusoidal waves are required, the TC method has been demonstrated to be better.

Figure.1.3. Current waveforms obtained using different modulation techniques for an active

power filter: (a) PS method, (b) HB method, (c) TC method.

Voltage source converters are preferred over current source converter because it is

higher in efficiency and lower initial cost than the current source converters [3, 4, 9]. They can

be readily expanded in parallel to increase their combined rating and their switching rate can be

increased if they are carefully controlled so that their individual switching times do not coincide.

Therefore, higher-order harmonics can be eliminated by using converters without increasing

individual converter switching rates.

1.1.5. Control strategies

Most of the active filters developed are based on sensing harmonics [7,10,11] and

reactive volt-ampere requirements of the non-linear load. [4,12,17] and require complex control.

In some active filters, both phase voltages and load currents are transformed into the α-β

orthogonal quantities, from which the instantaneous real and reactive power. The compensating

currents are calculated from load currents and instantaneous powers. The harmonic components

of power are calculated using high pass filters in the calculation circuit. The control circuit of the

dc capacitor voltage regulates the average value of the voltage to the reference value [4].

Reactive power compensation is achieved without sensing and computing the reactive current

component of the load, thus simplifying the control circuit. Current control is achieved with

constant switching frequency producing a better switching pattern. An active filter based on the

instantaneous active and reactive current component in which current harmonics of positive and

negative sequence including the fundamental current of negative sequence can be compensated.

The system therefore acts as a harmonic and unbalanced current compensator. A comparison

between the instantaneous active and reactive current component - method and the instantaneous

active and reactive power method is realized [17].

A new scheme has been proposed in [10], in which the required compensating current is

generated using simple synthetic sinusoid generation technique by sensing the load current. This

scheme is further modified by sensing line currents only [8,13]. An instantaneous reactive volt-

ampere compensator and harmonic suppressor system is proposed [13] without the use of voltage

sensors but require complex hardware for current reference generator. The generated reference

current is not a pure sine wave but stepped sine wave. Also, without the use of voltage sensors,

the scheme generates balanced sine wave reference currents but do not compensate reactive

power completely (if source voltage is unbalanced/distorted) due to waveform difference

between voltage and current [14]).

Control scheme based on sensing line currents is described in [2]. The 3-phase currents/voltages

are detected using only two current/voltage sensors compared to three used in [8,16]. DC

capacitor voltage is regulated to estimate the reference current template. Selection of dc capacitor

value has been described in [4,7,13].

Conventional solutions for controller requirements were based on classical control theory or

modern control theory. Widely used classical control theory based design of PID family

controllers requires precise linear mathematical models. The PID family of controllers failed to

perform satisfactorily under parameter variation, non linearity, load disturbance, etc.[18]

During the past several years, fuzzy control has emerged as one of the most active and fruitful

areas for research in the applications of fuzzy set theory, especially in the realm of industrial

processes, which do not lend themselves to control by conventional methods because of a lack of

quantitative data regarding the input-output relations. Fuzzy control is based on fuzzy logic-a

logical system that is much closer in spirit to human thinking and natural language than

traditional logical systems. The fuzzy logic controller (FLC) based on fuzzy logic provides a

means of converting a linguistic control strategy based on expert knowledge into an automatic

control strategy[19,21]. Recently, fuzzy logic controllers (FLC’s) have generated a good deal of

interest in certain applications. The advantages of FLC’s over the conventional controllers are:

1. It does not need accurate mathematical model;

2. It can work with imprecise inputs;

3. It can handle nonlinearity, and

4. It is more robust than conventional nonlinear controllers.

1.2. OBJECTIVE

In modern electrical distribution systems there has been a sudden increase of single phase and

three-phase non-linear loads. These non-linear loads employ solid state power conversion and

draw non-sinusoidal currents from AC mains and cause harmonics and reactive power burden,

and excessive neutral currents that result in pollution of power systems. They also result in lower

efficiency and interference to nearby communication networks and other equipments. Active

power filters have been developed to overcome these problems. Shunt active filters based on

current controlled PWM converters are seen as viable solution. The techniques that are used to

generate desired compensating current are based on instantaneous extraction of compensating

commands from the distorted currents or voltage signals in time domain. Time domain

compensation has fast response, easy implementation and less computation burden compared to

frequency domain. In this work both PI and fuzzy logic controlled shunt active power filter for

the harmonics and reactive power compensation of a nonlinear load are implemented. Both

controllers performance under certain conditions and different system parameters is studied. The

advantages of fuzzy controllers over conventional controllers like PI controllers are that they do

not need accurate mathematical model, they can work with imprecise inputs, can handle non-

linearity, load disturbances etc.

1.3. THESIS OUTLINE

The body of this thesis consists of the following seven chapters including first chapter:

• In Chapter 2, a description of the structure of the shunt active power filter, the basic

compensation principle, how reference source current is estimated and role of DC side

capacitor is given.

• Chapter 3 gives the PI control scheme of shunt active power filter in which DC voltage

control loop design n how to select PI controller parameters is presented.

• Chapter 4 deals with the fuzzy logic, fuzzy login controllers and implementation of fuzzy

control scheme for shunt active power filter. In this chapter basic fuzzy algorithm and

design of control rules is also described.

• The entire active filter system is composed mainly of a three-phase source, a non-linear

load, a voltage source PWM converter, and a PI or fuzzy controller. All these

components modeling is described separately in chapter 5.

• In chapter 6, simulation results are put and discussed in detail. Both PI and fuzzy controller

performances are compared under certain conditions.

• The conclusions of the thesis and recommendations for future work are summarized in

Chapter 7.

SIX- SWITCH INVERTERBasic compensation principle

Estimation of reference current

Role of DC side capacitor

Selection of Lc and Vdc,ref

Design of DC side capacitor (Cdc)

SIX- SWITCH INVERTER It consists of two three-switch legs and two single phase loads connected to the joints between

the switches. The dual-leg structure can be considered as two full-bridge inverters which share a

row of switches. the proposed configuration reduces the number of switches. The two output

voltages can be expressed as Vox =MxVgSin(ωxt+φx) where x denotes upper or lower output

similar to single-phase full bridge inverter except that some constraints are imposed on the

modulation indices, M, frequencies, ω, and phase difference, φ of the output references

depending on the subsequently defined EF and DF operation modes.Two-switch leg which

creates four possible switching states of which two are acceptable, when the switches of the same

row are OFF both outputs are in zero state , when the upper or lower switch rows are ON, the

corresponding output is in zero state and the other is in active state and finally, when the opposite

switches of the two legs are OFF, both outputs are in active state. Figure

Fig-1 six switch inverter

THREE-SWITCH INVERTER

As Fig. 2 shows, the proposed three-switch inverter is actually developed by replacing a leg of

the six-switch inverter with a series combination of capacitors at the cost of losing zero output

voltage state so as to further economize the inverter cost by reducing the number of power

switches. The single-leg topology which uses only three semiconductor switches for

independently supplying two single phase loads is structurally comparable to two half- bridge

inverters with a common row of switch and capacitor. The output voltages of the three-switch

inverter can be expressed as Vox = (1/2)MxVgSin(ωxt+φx) where x denotes upper or lower

output similar to single phase half bridge inverter except that the -modulation indices, M,

frequencies, ω, and phase difference, φ of the output references face some constraints depending

on the inverter.

Fig- 2 Three-switch inverter

DF and EF operation modes.Three-switching states are possible for the three switch

inverter.When the lower switch is OFF, both outputs are positive, when the upper switch is OFF,

both outputs are negative and when the middle switch is OFF, the upper output is positive and

the lower output is negative.

A. Different frequency mode of operation.

DF operation enables the inverter to produce two output voltages independent of each other

from both frequency and amplitude aspects. Turning the lower switch on for producing the upper

output voltage and vice versa, the two inverter outputs can function independently. This is

achieved by portioning the modulation space out among the two voltage references via adding

appropriate offsets to them. For this purpose, upper and lower output voltage reference

waveforms should respectively be shifted up and down in the modulation space occupied by the

carrier signal. This prevents interference of modulating signals and hence provides frequency

independency of the outputs. Gate signals of the three-switch configuration are generated similar

to six-switch.

B. Equal frequency mode of operation

The inverter outputs have equal frequencies in this mode, the maximum modulation index can be

increased up to one depending on the phase difference between the references. Similar to the six-

switch inverter, should be satisfied andOff sets of can be added to the output voltage could be

used for upper, middle and lower switch gate signal generation.

2.2. Estimation Of Reference Source Current

From Figure.2.1.1, the instantaneous currents can be written as

Source voltage is given by

If a non-linear load is applied, then the load current will have a fundamental component and

harmonic components which can be represented as

The instantaneous load power can be given as

From (2.2.4), the real (fundamental) power drawn by the load is

From (2.2.6), the source current supplied by the source, after compensation is

Where Ism=I1cosΦ1.

There are also some switching losses in the PWM converter, and hence the utility must supply a

small overhead for the capacitor leakage and converter switching losses in addition to the real power

of the load. The total peak current supplied by the source is therefore.

If the active filter provides the total reactive and harmonic power, then is(t) will be in phase with the

utility voltage and purely sinusoidal. At this time, the active filter must provide the following

compensation current:

Hence, for accurate and instantaneous compensation of reactive and harmonic power it is necessary

to estimate, i.e. the fundamental component of the load current as the reference current.

PI CONTROL SCHEME Dc voltage control loop

Transfer function of PWM converter Selection of PI controller parameters

PI CONTROL SCHEME The complete schematic diagram of the shunt active power filter is shown in figure 3.1. While figure

3.2.gives the control scheme realization. The actual capacitor voltage is compared with a set

reference value.

Figure .3.1. Schematic diagram of shunt active filter.

Figure .3.2. APF Control scheme with PI controller.

The error signal is fed to PI controller. The output of PI controller has been considered as peak

value of the reference current. It is further multiplied by the unit sine vectors (usa, usb, and usc)

in phase with the source voltages to obtain the reference currents (isa*, isb

*, and isc*). These

reference currents and actual currents are given to a hysteresis based, carrierless PWM current

controller to generate switching signals of the PWM converter[2]. The difference of reference

current template and actual current decides the operation of switches. To increase current of

particular phase, the lower switch of the PWM converter of that particular phase is switched on,

while to decrease the current the upper switch of the particular phase is switched on. These

switching signals after proper isolation and amplification are given to the switching devices. Due

to these switching actions current flows through the filter inductor Lc, to compensate the

harmonic current and reactive power of the load, so that only active power drawn from the

source.

3.1. DC VOLTAGE CONTROL LOOP

The block diagram of the voltage control loop is shown in figure 3.3. Where, Gc is the gain of the PI

controller and Kc is the transfer function of the PWM converter.

Figure.3.3.Block diagram of voltage control loop.

3.2 TRANSFER FUNCTION OF PWM CONVERTER (KC) The derivation between input (ac link) and output (dc link) quantities of the PWM converter is

obtained by equating average rate of change of energy associated. Equating the average rate of

change of energy quantities of input and output side of the PWM converter

Pcap = Pconv - Pind (3.2.1)

In order to linearize the power equation a small perturbation ΔIc is applied in the input filter current f

converter Ic, about a steady state operating point Ico, the average dc link voltge will also get

perturbed by a small amount ΔVdc, about its steady state operating point Vdco (Vdcref).The transfer

function of the PWM converter for a particular operating point can be obtained from (3.1) as

3.3. SELECTION OF PI CONTROLLER PARAMETERS

A proportional-integral-derivative controller (PID controller) is control loop feed back

mechanism used in industrial control systems. In an industrial process a PID controller attempts

to correct the error between a measured process variable and a desired set point by calculating

and then outputting a corrective action that can adjust the process accordingly. The PID

controller calculation (algorithm) involves three separate modes; the Proportional mode, the

Integral mode and Derivative mode. The proportional mode determines the reaction to the

current error, the integral mode determines the reaction based on recent errors and the derivative

mode determines the reaction based on the rate by which the error has been changing. The

weighted sum of the three modes is outputted as a corrective action to a control element such as a

control valve or heating element. By adjusting constants in the PID controller algorithm the PID

can provide individualized control specific to process requirements including error

responsiveness, overshoot of set point and system oscillation. Some applications may require

only using one or two modes to provide the appropriate system control. A PID controller will be

called a PI, PD, P or I controller in the absence of respective control actions. PI controllers are

particularly common, since derivative action is very sensitive to measurement noise.

Proportional mode responds to a change in the process variable proportional to the current

measured error value. The proportional response can be adjusted by multiplying the error by a

constant Kp, called the proportional gain or proportional sensitivity. With integral mode, the

controller output is proportional to the amount and duration of the error signal. The integral

mode algorithm calculates the accumulated proportional offset over time that should have been

corrected previously (finding the offset's integral). While this will force the controller to

approach the set point quicker than a proportional controller alone and eliminate steady state

error, it also contributes to system instability as the controller will always be responding to past

values. This instability causes the process to overshoot the set point since the integral value will

continue to be added to the output value, even after the process variable has reached the desired

set point. The characteristic equation of the voltage control loop is used to obtain the constants of

PI controller in this case, can be written as [2]:

Thus a second order transfer function can be found for the closed loop system. This characteristic

equation is used to found the components of PI controller. The analysis of this characteristic

equation shows that Kp determines the voltage response and Ki defines the damping factor of the

voltage loop. The current controller has been designed on the basis of 5% overshoot, to step the

change in the amplitude of current reference.

HREE-SWITCH INVERTER

MULTILEVEL CONVERTERS OFFER GOOD SOLUTION FOR HIGH POWER APPLICATIONS

AND HAVE BECOME VERY POPULAR AND ARE REPLACING THE CONVENTIONAL TWO

LEVEL TYPE FOR MANY APPLICATIONS [1]. THE MAIN REASON OF THIS APPROACH IS THAT THEIR

OUTPUT VOLTAGE HAS SIGNIFICANTLY LOWER DV/DT AND REDUCED HARMONIC DISTORTION [2-3].

THEREFORE, PROBLEMS RELATED TO THE DRIVES WITH TWO-LEVEL INVERTERS SUCH AS COMMON

MODE VOLTAGES, STATOR WINDING INSULATION STRESS, BEARING CURRENTS, HIGH

ELECTROMAGNETIC INTERFERENCE, ETC. CAN BE SIGNIFICANTLY REDUCED WHEN USING

MULTILEVEL TECHNOLOGY [4-12]. SEVERAL TOPOLOGIES FOR MULTILEVEL INVERTERS HAVE

INVESTIGATED IN THE LITERATURE AND THEY CAN BE BROADLY CLASSIFIED AS [4];

T

Neutral point clamped or diode clamped type,

Flying or floating capacitor type, and

Cascaded H-bridge type

The neutral point topology uses diodes to clamp the voltage levels while the flying-capacitor

topology uses floating capacitors to clamp the voltage levels. The major problems associated

with these two topologies are the balancing of capacitors voltages. The balancing problem

becomes more pronounced with increasing number of levels. The cascaded H-bridge inverter is

modular in nature with each unit connected in series. The major problem with this type of

topology is the requirement of complex transformers [4].

The Flying capacitor topology has some distinct advantages over other topologies of multilevel

inverter e.g. the redundancies of voltage levels, which means that more than one switch

combination can create different output voltage. The Flying Level Capacitor inverter has phase

redundancies unlike the NPC that has only line to line redundancies [9-10]. This feature allows

charging and discharging of particular capacitors in a way that may be integrated within control

algorithm for balancing the voltages. Other advantage of this inverter is that the active and

reactive power flow can be controlled. Also, the large number of capacitors enables an operation

through short duration outages and significant voltage sags [11].

Flying capacitor multilevel topology is a good approach particularly for more than three

voltage levels. A negative aspect of using Flying Capacitor is that the output current changes the

voltage of the capacitors. In general, the capacitance must be selected high to make the changes

as low as possible. Therefore, the size of the capacitors increases as an inverse with switching

frequency. This topology is therefore not practical for low switching frequency applications [11].

Flying capacitor voltage balancing is a tedious task in Flying Capacitor topology [12].

Nemours techniques have been presented in the literature to balance the voltage across the

clamping capacitors [12-18]. These include the carrier based PWM as well as space vector PWM

using switching redundancies. Among the carrier-based techniques three methods are popular

namely, phase shifted PWM, modified carrier redistribution PWM and saw-tooth rotation PWM.

A comprehensive comparison of these methods are presented in [16] for three-phase system. It is

concluded that the modified carrier redistribution PWM offer optimum performance. This

technique is further utilized in [14,17]. Hence the same technique is employed in this paper for

three-level flying capacitor type five-phase voltage source inverter.

Multiphase multilevel inverters are quite attractive solutions for high power multiphase drive

systems. However, multilevel inverters have been mainly investigated for three-phase system

and still little work has been reported on multiphase multilevel inverters except for [19-23]. The

neutral point clamped topology of five-phase inverter is considered in these papas.

Balancing the capacitors’ voltages is a major disadvantage of multilevel inverters. Therefore, in this paper we will focus on solving the problem of balancing FLC, even when lowering the switching frequency without a need to increase the size of the capacitor. The modified carrier redistribution PWM technique is employed.

OPERATION OF THREE-LEVEL FLYING CAPACITOR INVERTER

Power circuit topology of a five-phase three-level Flying Capacitor type voltage source

inverter is shown in Fig. 1. In this topology of multi-level inverter, floating capacitors are

employed to clamp the node voltages of the series connected power switching devices. Four

power switching devices are connected in series to form one leg of the inverter, five legs are used

in the proposed topology to obtain five-phase output. The complimentary switches are indicated

in the figure ( (sn1 , sn 1' ) , (sn2 , s2

' );n∈a ,b , c ,d ,e ); the top and the bottom switches and the two middle

switches are complimentary. Two power switching devices are turned on simultaneously to

provide three different voltage levels at the output phase; 0.5Vdc, -05Vdc and 0. Since the voltage

across the flying capacitors is limited to 0.5Vdc, the same voltage stress will be borne by each

switch. As the number of level increase the required voltage blocking value reduces, thus lower

rating switches can be used or alternatively higher voltage can be achieved.

The relationship between the pole voltage and the output phase voltage remains same as that of

a two-level inverter and is given as [24];

va=(4 /5 )V A−(1/5 ) (V B+V C+V D+V E )

vb=(4 /5 )V B−(1/5 ) (V A+V C+V D+V E ) vc=( 4/5 )V C−(1/5 ) (V A+V B+V D+V E )

(1)

vd=(4 /5 ) V D−(1/5 ) (V A+V B+V C+V E )

ve=( 4/5 ) V E−(1/5 ) (V A+V B+V C+V D )

However, the relationship between the pole voltages and the switching function is given as;

V k=λ−1

2V dc , k=A ,B ,C , D , E

(2)

Where λ is the switching function defined as;

λ=2 if sn1 and sn2 are on1 if sn 1 and sn2 are complimentary0 if sn 1 and sn2 are off (3)

One cycle of operation of one leg of flying capacitor inverter is depicted in Fig. 3a to 3h. The

path of current is shown in red dotted colour line. The pole voltage is 0.5Vdc and the current

either flows to or from the load in Fig. 3a and Fig. 3b. The capacitor state remains unchanged.

Fig. 2. Power Circuit of a five-phase VSI.

Fig. 3a. Switch State 1100 with positive current flow

Fig. 3b. Switch state 1100 with negative current flow

Fig. 3c. Switch state 1010 with positive current flow (Flying capacitor charges)

Fig. 3d. Switch state 1010 with negative current flow (Flying capacitor discharges)

Fig. 3e. Switch state 0101 with positive current flow (Flying capacitor discharges)

Fig. 3f. Switch state 0101 with negative current flow (Flying capacitor charges)

Fig. 3g. Switch state 0011 with positive current flow

Fig. 3h. Switch state 0011 with negative current flow

The flying capacitor charges during the switching state of Fig. 3c and Fig. 3f. The charging

current is equal to the load current. Hence the design of flying capacitor should take into account

the maximum load current. The pole voltage is zero during this operation.

The flying capacitor discharges during the switching state of Fig. 3d and Fig. 3e, once again

the current flow through the capacitor is equal to the load current. The pole voltage remains zero

during this switching state.

Fig. 3g and 3h shows the switching state when the pole voltage attains the value of -0.5Vdc and

the flying capacitor conditions remains unchanged.

It is evident from the above discussion that the capacitor charging state changes during Fig. 3c-

3f. Thus to balance the flying capacitor voltage these switching states need special attention.

Thus the PWM technique described in the next section exploit optimally these redundant

switching states.

III. PWM TECHNIQUE TO BALANCE FLYING CAPACITOR VOLTAGE

The PWM utilized in the proposed three-level five-phase flying capacitor inverter is based on the

carrier based scheme [14, 17]. The major goal of the PWM scheme is to balance the flying

capacitor voltage and keep it constant at half of the dc link voltage. The basic idea behind

voltage balancing is equal charging and discharging time of flying capacitor.

From the switching diagram (Fig. 3) it is clear that for switch states of Fig. 3c and Fig. 3f, the

output pole voltage remains the same (zero) thus these states are redundant state. However, the

flying capacitor charges in state of Fig. 3c and Fig. 3f and discharges to the load in state of Fig.

3d and Fig. 3e. Thus states (Fig. 3c and Fig. 3f) and Fig. 3 (Fig. 3d and Fig. 3e) should persist for

the same time for equal charging and discharging of the flying capacitor. To accomplish this, the

carrier signal should be formulated as shown in Fig. 4 and Fig 5 for switch Sn1 and Sn2,

respectively. The upper carrier is used when reference voltage lies between 0.5Vdc and Vdc while

the lower carrier is used when reference voltage lies between 0 and 0.5Vdc. The gate signals are

shown in Fig. 6 and Fig. 7 for the two different values of the reference voltages. The carrier

signal for four sampling period is shown. It is evident that the redundant switching state is active

in first sample and then third sample in Fig. 6. Hence, the flying capacitor charge in the first

sampling time and discharge in the third sampling time, keeping the same charging and

discharging time in four sample time. The switching time of each power switch will be equal to

2Ts as evident from Fig. 6 and Fig. 7. The similar observation can be made from Fig. 7. Thus it is

seen that for all the values of the reference voltages, the charge and discharge time of the flying

capacitors remains same. Thus the voltage is balanced at the flying capacitor terminals in

average in four sample time. This average balancing time will be higher for higher level

numbers. Vdc

0Ts 2Ts 3Ts 4Ts

Vdc/2

Vdc/2

Carrier wave for S1

Fig. 4. Carrier wave for Sn1 and Sn1’

Vdc

0Ts 2Ts 3Ts 4Ts

Vdc/2

Vdc/2

Carrier wave for S2Fig. 5. Carrier Wave for Sn2 and Sn2’

III A Start-up Procedure

One of the important requirements of a flying capacitor inverter is their start-up procedure. This

problem still persists in a five-phase inverter similar to a three-phase inverter. At start the flying

capacitor needs to be charged equal to 0.5Vdc. For this the gating signal is provided in such a way

that the switch Sn1 and Sn1’ remains on to charge the capacitor and the Switches Sn2 and Sn2’

remains off. Once the voltage of the flying capacitor build up to the required voltage level (i.e.

0.5Vdc), all the switches are kept off so that the flying capacitor voltage floats at 0.5Vdc. Then the

normal PWM inverter operation is implemented.

Fig. 6. Gate signal generation when V dc /2≤|vref|≤V dc

Fig. 7. Gate signal generation when 0≤|vref|≤V dc/2

IV. SIMULATION RESULTS

Simulation model is developed in Matlab/Simulink using simpower system block sets to validate

the PWM technique proposed in the previous section. At first R-L load is considered followed by

a five-phase induction motor load.

IVA Results with R-L Load

DC link voltage of 100 V is kept and the switching frequency is kept at 2.5 kHz, the flying

capacitor is chosen as 100 µF and the load parameters are initially R = 5 Ω and L = 0.5 mH and

then changed to R = 10 Ω and L = 1 mH. The fundamental frequncy is kept at 50 Hz. The laod is

kept constant until 100 ms and then the load is halved to observe the effect of load change on the

flying capacitor balancing. The resulting waveforms are depicted in Fig. 8-11. The carrier is

formulated according to Fig. 4 as the amplitude of the reference is assumed equal to the dc link

voltage. The output phase current is illustrated in Fig. 8, where it has initial value of 10 A which

then reduces to half the value due to change in the load. Good quality of current waveform is

achieved. The voltage across the flying capacitor is depicted in Fig. 9. The voltage builds up to

half the dc link voltage of 50 V and stay there even if there is a change in the load, except change

in the ripple voltage. Thus perfect voltage balance is obtained. The current through the flying

capacitor is also presented in Fig. 10 which shows the load current passing through the capacitor.

0 0.05 0.1 0.15-15

-10

-5

0

5

10

15

Time (s)

Out

put p

hase

cur

rent

s(A

)

Fig. 8. Output five-phase currents.

0 0.05 0.1 0.15-10

0

10

20

30

40

50

60

Time (s)

Flyi

ng C

apac

itor V

olta

ge(V

)

Fig. 9. Voltage across Flying capacitor.

The phase ‘a’ voltage and non-adajacent line-to-line voltage is shown for initial setting of the

load in Fig 10 and Fig. 11, respectively. The frequency domain characteristics is also depcited in

the lower trace. The output phase voltage is 50 V with 2.5 kHz switching harmonics. This is the

typical value obtainable with the carrier-based scheme. The non-adajacent line-to-line voltage

magnitide is 90.1 V with very good spectrum and three level output.

IVA Results with Five-phase Induction Motor Load

Simulation is also carried out with five-phase induction motor load. The parameters of the

simulated induction machine are R s=10Ω , Rr=6 .3Ω , Lls=Llr=0 .04 H, Lm=0 . 42 H . Inertia and the

number of pole pairs are equal to 0 .03 kgm2 and 2, respectively. Rated phase current, phase-to-

neutral voltage and per-phase torque are 2.1 A, 220 V and 1.67 Nm, respectively. The dc link

voltage is kept at 600 V and the switching frequency is chosen as 2.5 kHz. The flying capacitor

value is taken as 750 µF. The aim of this section is to show the applicibility of the proposed

PWM technique in balancing the flying capacitor voltage even in regerenetion mode of this

motor. The motor is operated in open-loop mode. The motor is initially started at no-load then a

negative load is applied at t = 0.4 sec. The resulting waveforms are depcited in Fig. 13-15.

0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1-100

-50

0

50

100

Phas

e V

olta

ge (V

)

Time (s)

101 102 103 1040

20

40

60

Spec

trum

(V)

Frequency (Hz)

Fig. 10. Phase voltage time and frequency domain waveform.

0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1-100

-50

0

50

100

Non

-Adj

acen

t Lin

e V

olta

ge (V

)

Time (s)

101 102 103 1040

50

100

Spec

trum

(V)

Frequency (Hz)

Fig. 11. Line voltage time and frequency domain waveform.

The increase in the speed due to negative laod is clearly seen from Fig. 12 and thus the motor

is then running in regenerative mode. The balancing of capacitor voltage even under regeneration

is evident from Fig. 13.

Another test is carried out on the flying capacitor inverter by changing the switching frequency

and observing the capacitor voltage in R-L load. The resulting waveform is given in Fig. 14. The

switching frequency is kept at 1.5 kHz and is then reduced to 1 kHz. The capacitor voltage

shows correct balancing without the need to incerease the value of the capacitor. The ripple

voltage incerases with lowering the switching frequency which is an obvious effect of the choice

of load parameters. The important point is that lowering the switching frequency do not have

impact on the balaning of the clamped voltage. This validate the correct working of the modified

redistributed carrier wave.

We demonstrate a coherent imaging system based on a terahertz

(THz)frequency quantum cascade laser (QCL) phase-locked to a near-

infraredfs-laser comb. The phase locking enables coherent electro-optic

sampling of the continuous-wave radiation emitted by the QCLthroughthe

generation of a heterodyne beat-notesignal that carries the amplitude and

phase of the laser field. We use this beat-note signal to demonstrate raster

scan coherent imaging using a QCL emitting at 2.5THz. At this frequency, the

detection noise floor of our system is of 3pW/Hz and the long-termphase

stability is <3deg/hour, limited by the mechanical stability of the apparatus.

Significant scientific effort over recent yearshas aimed at the realizationof terahertz (THz)

frequency imaging systemsbased on quantum cascade lasers (QCLs). Even though the

continuous-wave output power of these devices can be as high as tens of mW[1-3], there are still

several technological issues that need to be addressed with the detection in order to realize a

system that is sufficiently sensitive, as well as fast and compact. For this reason, several groups

have tested different detection techniquesand configurations. Incoherent imaging systems have

been demonstrated using Golay cells, pyroelectric detectors, cryogenically cooled bolometers,

andcommercial focal plane array microbolometric cameras [4-7]. More recently,a QCL-based

imaging system was also demonstrated using an amorphous-silicon microbolometric camera that

was specifically developed for operation in the THz range [8]. Coherent imaging techniques have

also been reported in the literature, includinga pseudo-heterodyne technique based on mixing

between longitudinal modes of a multimode QCL [9], self-mixing[10],andexploiting the

heterodyne mixing between a QCL and gas laser [11]. In the latter, the THz QCL was frequency-

locked to the gas laser line in order to reduce the phase instability of the emitted radiation field

and allow the application of an inverse synthetic aperture radar technique. Our work here is

based on a coherent imaging technique that was first demonstrated by Loffler et al. [12], who

usedan harmonic of the repetition rate of a mode-locked Ti:S laser as a local oscillator, and

mixed it with a quartz-stabilized Gunn oscillatoremitting at 0.6THz. In this case,

both sources were free running since their intrinsic phase/frequency stability

was sufficiently high. However, recently,it has been demonstrated that

although THz QCLs have sub kHzquantum noise limited linewidths[13,14], up to a

Fourier frequency of ~10kHz they are affected by a ~1/f2noise component, giving rise to theline

broadenings and frequencydriftsthat have been observed in several heterodyne experiments

[15,16].As a consequence the scheme of Ref.[12] cannot be implemented with

a THz QCL without active stabilization.

Over the last few years, techniques have been developed tostabilize THz

QCLs to near-IR frequency combs produced by fs-mode-locked fiber lasers [17-

20].Theseexploit the sampling (electro-optic or photoconductive) of the radiation field emitted by the

QCL using the mode-locked pulses from the fs-laser. In the RF domain, this sampling gives rise to a

series of heterodyne beatsignalsbetween the harmonics of the fs-laser repetition rate and theTHz field.

The lowest frequency beat-notesignal can be fed into standard RF electronics and used to phase-lock the

QCL to the fs-laser repetition rate, thus eliminating the phase jitter between the two sources. The sub-Hz

beat-note linewidth obtained with this techniqueallowsthe coherent accumulation of the QCL signal over

long integration times and the achievement of high signal to noise ratios [20]. In theworkpresented here,

this harmonic sampling technique has been used to implement a coherent imaging setup using a single-

mode 2.5THz QCL and a frequency doubled fs-laser comb as local oscillator. We have achieved imaging

with a noise detection limit of 3pW/Hz, anda long-term phase stability of less than 3 °/hour.

Themaximum detection bandwidthis determined only by the sampling rateand not by the detector rise

time as found when usinga Golay cell or bolometer.

The QCL used in our experiments operates at 2.5THz and is based on a 2.5mm-long,

240m-wide, ridge-waveguide Fabry-Perot cavity that was fabricated by optical-lithography and

wet-etching (details on the waveguide and active region design can be found in Ref.[3]). The

QCL was kept at a stabilized temperature of 20K using a continuous-flow, liquid-helium

cryostat, and driven at a constant current of 1.49A with a commercial power supply. Under these

operating conditions, the emission is single-mode and the output power measured with a

calibrated THz power-meter (Thomas Keating Ltd.) is 2mW. For the near-IR laser comb, we

usea frequency doubled, mode-locked fs-fiber laser (Menlo Systems, M-fiber) operating at =

780nm and emitting a train of ~100fs-long pulses at a repetition rate of ~250MHz.

The experimental apparatusis shown in Fig. 1 and is based on two identical electro-optic (EO)

detection units, labeled EO1 and EO2, that are used respectively to (i) lock the QCL frequency to the fs-

laser comb, and (ii) detect the QCL beam after it has been reflected by the imaged target[12].For our

experiments, the QCL and near-IR comb beams areeach split into two using separate beam splitters

(labeled BS), and detection is achieved in both units by collinearly focusing the beams onto a 2mm-

thick, <1,-1,0> oriented ZnTe crystal,which is followed by /4 and /2 waveplates, and a

polarizing beam splitter. These elements form an ultrafast, near-IR electro-optic amplitude

modulator driven by the THz ac field (see Ref. [17] for details of the operating principle).

Assuming for simplicity that the QCL is single mode, this modulator generates two sideband

combs at ±2.5THz from the comb carrier centered at 780nm (385THz). Since the comb

bandwidth is approximately twice the QCL frequency, the carrier overlaps with the sideband

combs producing a series of heterodyne beat-notes, oscillating at frequencies|(QCL – n xfrep)|,

where QCL is the emission frequency of the QCL, frep is the fs-laser repetition rate (250MHz), and

n is an integer number [17]. Therefore,the lowest frequency beating, fbeat, corresponds to n

=Int(QCL/frep) ~ 104(= 2.5THz/250MHz), and has a frequency fbeat<frep/2 ~ 125MHz.This beat-note

is detected using shot-noise limited balanced detection, based on a pair of Si-photodiodes, and is

followed by a fast amplifier with a bandwidth of approximately 200MHz. Ultimately, the

detection bandwidth of the system is limited by the Nyquist criterion to half the sampling rate,

i.e. 125MHz.

To lock the QCL frequency, the beat-note generated by EO1 is compared, using a mixer, to

a signal atfRF1 ~10MHz produced by an RF synthesizer (RF1). The error signal oscillating at (fbeat-

fRF1) is then used to control the QCL current through a phase-lock loop (PLL) circuit with a

bandpass of ~2 MHz, and phase-lock QCLto the ~ 104harmonic of the fs-laser repetition rate. In

Fig. 2(a) we show an example of the spectrum of the phase-locked beat signal,fbeat, acquired with

a spectrum analyzer with a resolution bandwidth of 10Hz, and a THz power impinging on EO1

of 250W. This phase-locking is critical to allow image acquisition using a conventional lock-in

amplifier.

The second half of the QCL beam isfocused on the imagedtargetusing an f/1, gold-coated,

off-axis parabolic mirror (identical to the one used to collimate the QCL beam).Approximately

half of the THz radiation reflected from the targetis finallyfocused on EO2. As found in previous

experiments [13,17-19], the QCL is affected by optical feedback which, in this case, mainly

arises from the fraction of the beam reflected from the target that is transmitted through the beam

splitter. As shown in Ref. [13], optical feedback has a strong influence on the QCL frequency

noise. In particular, in the present case, the phase of the radiation reflected into the QCL cavity,

as well as the intensity,, changed markedly when the beam was raster scanned across the imaged

object owing to changes in the surface profile and morphology. We found experimentally that

this effect couldsuddenly drive the QCL out of lock,thus compromising the image acquisition.

To limit the effect of this feedback we inserted an optical isolator, consisting ofa wire-grid

polarizer (WGP) oriented parallel to the TM-polarised light from the QCL. This wasfollowed,

after the beam splitter, by a 3.1mm-thick, quartz quarter-waveplate with its fast-axis at 45° with

respect to the polarizer [13]. As a result, after the quarter-wave plate,the THz light emitted by the

QCL is left circularly polarised, and the polarisation is changed from left- to right-circular after

reflection from the target. Going back through the quarter-waveplate,the reflected beam recovers

its linear polarisation, now orthogonal to the wire grid polariser (and to the QCL polarisation),

thus providing the isolation. Using a power meter we measured an isolation of ~16dB. However,

this wasnot sufficient to suppress the random un-locking of fbeatcompletely during image

acquisition. Therefore an additional ~10dB of attenuation was added (~5dB attenuation per

pass).

For the measurement of the radiation reflected by the imaged object, we used a standard

lock-in amplifier (SRS model SR830), and a reference oscillator, RF1. As shown in Fig. 1, in

order to bringfRF1= 10MHz (= fbeat)within the lock-inreference frequency range (1mHz to

100kHz),the latter was down-converted to 70kHz by mixing with another synthesiser, labelled

RF2, oscillating atfRF2= 10MHz + 70kHz. The same synthesiser was mixed with fbeat generated by

EO2 to providethe lock-in input signal.It is important to note that the phase-lockingofQCLis a

crucial requirement for the realization of our coherent imaging system.Indeed, whilst in principle

the free-running fbeat generated in EO1 could be used as a reference to demodulate the output of

EO2, as in Ref [12], in practice this is not possible. In fact, withoutphase-locking, the coherence

between QCL and frepis dominated by thelow frequency noise component of QCL,

proportional to ~1/f 2, produced by mechanical vibrations and thermal and/or current

fluctuations in the device[13,14]. As can be derived from the frequency noise spectral density of

the QCL (identical to the present one) reported in Ref. [13], this generates an fbeat with an

“instantaneous” linewidth of the order of 10kHz on a1ms timescale, and is subject to drifts of

several MHz/s [15,16,21].As a consequence the “free running”fbeat is not sufficiently stable to

provide a reference signal for typical commercial DSP lock-in amplifiers including the one used

here[22].On the other hand, phase-locking of QCLensures coherence withfrep, eliminating their

mutual jitter, and thus allows use offbeatas a reference.

Before proceeding to the image acquisition, we evaluated the sensitivity of our EO

detection. Using a calibrated power meter, we measured THz powers of 250W and 60W

impinging on the ZnTe crystals of EO1 and EO2 respectively. Given the 2mW emitted power

from the THz QCL, these values are in agreement with the attenuations from the optical elements

shown in Fig.1, including the attenuator used to decrease the optical feedback and, in the case of

EO2, the reflection from the flat part of a 10 cent Euro coin that we used as a test target for our

imaging system. In Fig. 2(b) we show the power in dBmof the phase-locked fbeatmeasured at the

output of EO2 with a spectrum analyzer with a resolution bandwidth of 1Hz. The THz power

was progressively attenuated from 60W to 10pW by superimposing up to 14 A4 paper sheets.

Down to 2nW, the power was measured with a Golay cell detector that had been previously

calibrated using a THz absolute power meter, while the two points below the 300pW detection

limit of the Golay cell were obtained using calibrated attenuators. The noise floor of -140dBm is

determined by the shot noise of the photocurrent generated by the 15mW of near-IR power

incident on the balanced detection, and corresponds to the minimum detectable THz power of

~3pW (3pW/Hz noise equivalent power). This is consistent with the spectrum of Fig.2(a) at the

output of EO1 where 250W of THz power yielded a SNR of ~70dB in a 10Hz bandwidth,

which corresponds to a noise equivalent power of ~2.5pW/Hz (the SNR scales linearly with the

bandwidth).

Figs. 3(a), (b) and (c) show respectively the amplitude, power and phase image in grey-

colour scale of a 10cent Euro coin, obtained by raster scanning the object in the focal plane of

the mirror using a motorized XY stage. The image was obtained by continuous line scanning

along the Y direction (from left to right in the figure) at a speed of 2.2mm/s, with a step between

each line of 100min the X direction (the acquisition rate along the Y direction was set to obtain

the same spatial resolution of 100m). The lock-in time constant was 30ms, which allowed a

dynamic range of 60dB, as shown in Fig.3(b).The width of the vertical lines on the left side of

the coin was measured with a profilometer(top graph in Fig.3) and was found to be 160m,

showing that the system resolution is diffraction limited. Fig.3(c) displays the phase image, as

recorded from the lock-in amplifier. As expected, compared to the amplitude or power plots

where the edges of the reliefs appear very clear in black owingto scattering, here the various

reliefs are displayed in different grey colors corresponding to different heights, and hence

different phases owingto changes in the optical path. By monitoring the phase with time we

measured a phase stability, with a shift of < 3 °/hour, corresponding to an uncertainty of <1m in

the determination of the profile height. This phase stability is completely limited by the

mechanical stability of the experimental apparatus. We verified with the profilometer that the

height of most of thefeatures is largerthan /2 = 60m; however,the phase displayed by the lock-

in amplifieris limited to ±, which reduces the effective length over which the phase changes

continuously to/2. This, together with the fact thatthe coin is not perfectly parallel to the focal

plane (a ~0.3°inclination is enough to produce a phase change of 2 across the coin),partially

explains why the phase image does not display clearly the different shapes. Another difficulty is

related to small details, such as the stars and stripes on the top left corner, where the size is close

to the resolution limit of the system, resulting in a poor phase contrast. Moreover, as shown by

the height profile in Fig.3, some of the details such as the vertical lines, have a height that is

close to /2. In contrast, the amplitude image, resulting fromscattering,allows much clearer

identification of these details.Fig.3(d) displays a processed phase image where the effect of the

limited phase-range of the lock-in amplifier has been partially removed by adding or subtracting

2 to a fraction of the pixels of Fig.3(c).

The imaging speed in these experiments is solely limited by the speed of the acquisition

software. From Fig.2(a), an integration time of 10s would still allow a signal to noise ratio of

more than 30dB, which would allow acquisition of in Figs. 3 without degradation. This could be

achieved by replacement of the XY translation stage with a fast steering mirror,enabling

acquisition within a few seconds.

Acknowledgments

We acknowledge partial financial support from the AgenceNationale de

la Recherche (project HI-TEQ), the EPSRC (UK), and the European Research

Council programme ‘TOSCA’. We thank Pierre Gellie for taking the scan of

the profile shown in the top panel of Fig. 3.

Figures

Fig.1Experimental apparatus.The fs-laser comb and QCL beams are split into two parts, using

two beam splitters (BS), and are focused onto EO detection units EO1 and EO2. EO1 is used to

phase-lock fbeatto fRF1, and therefore to phase-lock QCLto~ 104 xfrep. EO2 is used to detect the THz

radiation back-reflected from the imaged target. This is achieved by mixing fbeat(= fRF1) with fRF2

= fbeat + 70kHz and monitoring the difference frequency, oscillating at 70kHz, with a lock-in

amplifier. The reference of the lock-in amplifier is obtained from the difference frequency fRF2 -

fRF1 = 70kHz. All mirrors used to collimate and focus the radiation from the QCL are 90° off-

axis, gold coated parabolic mirrors. The fs-laser comb is a frequency doubled 1550nm fiber laser

emitting a train of ~100fs pulses at a repetition rate frep = 250MHz.There is approximately 15mW

of optical power at 780nm impinging on each EO detection unit.

Fig.2 Measurement of the noise equivalent power of the imaging system. The THz power

focused on EO2 was decreased from 60W to 8pW using calibrated attenuators. The noise floor

at -140dBm is set by the shot noise of EO2 in a 1Hz bandwidth and gives a noise equivalent

power of 3pW/Hz. Inset: Spectrum of fbeatdetected on EO1 and recorded on the spectrum analyzer

with a resolution bandwidth of 10Hz. The THz power used for EO detection is 250W.

Fig.3 THz image of a 10 cent Euro coin:(a) amplitude, (b) power. (c) raw phase image, and (d)

phase image corrected for the limited phase range of the lock-in amplifier (see text). The coin

diameter is 19.75mm. The imageswere acquired with lines scans from left to right at a speed of

2.2mm/s, with a step between each line of 100min the vertical direction. The lock-in amplifier

time constant was set to 30ms. Top: Height profile obtained by scanning with a profilometer

along the red dashed line of (a).

A. Energy Efficient Motor Drive Systems

1) IntroductionSince the industrial revolution, motors have replaced humans and animals as the primary source of useful work. James Watt observed that a horse pulling 180 pounds of force made 144 trips around the circle in an hour, at an average speed of 181 feet per minute. The horse generated 33,000 ft. lbs. per minute, which Watt called one “horsepower”. At the time, generating 1 hp required a 1,000 pound, 6 foot tall horse that in today’s dollars costs about $5,000 per year to board. Today, generating 1hprequires a 32 lb motor (30x less than a horse), which is about 4 x 6 inches (12x less than a horse) and costs about $250 per year to power (20x less than a horse). Thus, motors are essential to our modern economy and can be viewed as a primary generator of wealth in modern society.

Today, motors consume about 75% of all the electricity used by industry. Their popularity is a testament to their reliability, versatility and efficiency. However, the large quantity of energy consumed by motors means that small improvements in motor drive system efficiency result in large savings. This chapter discusses fundamentals of motor drive systems and how to make them more efficient.

2) How Motors WorkMotors produce useful work by causing a shaft to rotate. By far the most common type of motor is the squirrel cage induction motor; its operation is described here. Other motor types include shaded pole, synchronous, permanent magnet, reluctance and DC motors for precise speed control.

Squirrel cage induction motors consist of a rotating shaft to which a rotor is attached. The rotor is made of solid, uninsulated metal bars connected to solid metal rings of the same material at each end. The rotor has no external electrical connections. The stator surrounds the rotor, and has two or more poles. Each pole consists of an iron bar wrapped in conductive aluminum or copper wire. The poles create

magnetic fields when electrical current passes through their wire wrappings. The magnetic fields are rotated from pole to pole around the stator, causing the rotor to rotate.

The rotor rotates at a synchronous speed given by the following equation:

rpm = frequency of applied voltage (Hz) x 60 / number of pair poles

Thus, a two pole motor with a 60 Hz power supply rotates at 3,600 rpm and a four pole motor rotates at 1,800 rpm. The difference between the synchronous speed and the actual speed is called the slip. For example, most motors with a synchronous speed of 1,800 rpm rotate at about 1,750 rpm. Slip generally decreases with motor size.

Source: Nadel et al., 1991

a) Synchronous MotorsSynchronous motors can run at low speeds, and are thus good fits for low speed applications such as reciprocating compressors. Synchronous motors are also slightly more efficient than induction motors. Finally, synchronous motors can generate or absorb reactive power, whereas induction motors can only absorb reactive power. Thus, large synchronous motors are sometimes able to correct the power factor of an entire plant.

b) Direct Current MotorsDirect current motors use direct, rather than alternating, current. In DC motors, the speed can be varied simply by varying the voltage.

3) Motor Selection a) NEMA ClassThe National Electric Manufacturers Association (NEMA) classifies motors as NEMA A, B, C, and D, depending on the relationship between torque and speed. Design B motors are by far the most common and are used for most fan, pump and compressor applications.

Source: Nadel et al., 1991

b) Motor SpeedMotors are also selected according to synchronous speed. Common speeds are 3,600, 1,800, 1,200, …, rpm. The most common motor speed is 1,800 rpm.

c) Enclosure TypeMotors come with different enclosure types for different surroundings. NEMA defines about 20 enclosure types divided into two basic groups: open and totally enclosed. The most common types are open drip proof (ODP) and totally enclosed fan cooled (TEFC).

d) Service FactorThe service factor indicates the capacity of the motor to withstand prolonged overloading. Service factor 1.0 indicates that the motor should only be operated at 100% load or less. Service factor 1.3 indicates that the motor could be operated at 130% capacity without failing, although the life of the insulation may be reduced.

e) Frame SizeThe frame defines the shape and size of the motor. For motor replacements, the frame size of the new motor must match the frame size of the older motor to avoid expensive mounting modifications. In 1952, the industry standardized U-frames, such that all 254U frames were identical. In 1964, the industry standardized T-frames, which are smaller and lighter. Most current motors are T-frames.

4) Determining Motor Input PowerThe best way to determine motor power consumption is with a power meter. Power meters simultaneously measure the current, voltage and power factor and combine these measurements to determine power. For a three-phase motor, the power consumption is:

Power (kW) = Average Current of 3 Phases (A) x Voltage (V) x √3 x PF (W/VA)

Many motors are sized to be 75% to 80% loaded. Thus, power can be estimated as:

Power (kW) =

Rated Output (hp ) x 0 . 746 kW/hp x FractionLoadedEfficiency

Example:

If a motor draws 100 A at 480 V with PF = 0.80, calculate input power:

Pin = 100 A x 480 V x 0.8 kW/kVA x 31/2 / 1,000 VA/kVA = 66 kW

Alternately, if motor power input is measured, fraction loaded can be determined as the ratio of actual output power to rated output power.

Example:

Calculate fraction loaded of a 100 hp, 95% efficient motor drawing 66 kW

Output Power = 66 kW x 0.95 / 0.75 kW/hp = 84 hp

Fraction Loaded = 84 hp / 100 hp = 84%

5) Motors and Power FactorMany types of electrical equipment, such as motors or lighting ballasts, require that more power be supplied to the equipment than is actually consumed by the equipment. The ratio of the actual power consumed by equipment (Pa) to the power supplied to equipment (Ps) is called the power factor. The reactive power (Pr) is a measure of the unusable power.

In inductive motors, the current lags the voltage, creating unusable reactive power. Because of this, the power factor of inductive motors is always less than 1.0 and declines under decreasing load.Typical power factors as functions of rated load and horsepower areshown in the table below.

Load / hp 0.25 0.50 0.75 1.00

10 0.54 0.70 0.77 0.80

20 0.58 0.72 0.79 0.81

25 0.58 0.73 0.79 0.81

30 0.59 0.73 0.79 0.81

40 0.61 0.74 0.80 0.82

50 0.63 0.76 0.81 0.82

60 0.65 0.77 0.82 0.83

75 0.67 0.78 0.83 0.84

100 0.70 0.80 0.84 0.85

Moving Motor Use to Off-Peak Shift

Some motors are used only during one shift. Moving use to an off-peak shift can reduce electrical demand charges.

Example:

Calculate the savings from moving a grinding operation to off-peak shift, if the 50-hp grinder motor is 80% loaded and 90% efficient. The cost of electrical demand is $10 /kW-month.

Annual Savings:

50-hp x 80% / 90% x 0.75 kW/hp = 33 kW

33 kW x $10 /kW-mo x 12 mo/yr = $4,000 /yr

Note that most motors operate at 900, 1,800 or 3,600 RPM.

Thus, although motors typically draw more current (power) as the motor comes up to speed, they typically come up to speed in a few seconds. Most utilities calculate peak electrical demand over a 15 or 30 minute period. Thus, a few seconds of high power draw is rarely enough time to significantly increase total use during demand period. Hence, there is negligible demand penalty from motor start up, and motors should be turned off motors when not in use.

6) Turning Off Motors When Not Equipment is IdleMotor power consumption depends on the load on the motor. In some types of manufacturing equipment, such as CNC machines, the load on the motor powering the lathe is very small when the equipment is idling. Thus, the savings from turning off CNC machines when the equipment idles are relatively small. However, motors in other equipment, such as stamping machines with flywheels and hydraulic power systems, are typically highly loaded even when the equipment is idling. Thus, turning off this this type of equipment when not in use can result in large savings.

Example

A stamping press motor is 80% loaded while stamping and 65% loaded when the press is idle. Thus, 81% of peak power is dissipated as heat due to friction. Calculate the savings from turn off a 90% efficient, 50-hp stamping press motor for 2,000 hr/yr when the press idles. The cost of electricity is $0.10 /kWh.

Annual Savings:

50 hp x .65 / 90% x .75 kW/hp x 2,000 hr/yr = 54,167 kWh/yr

54,167 kWh/yr x $0.10 /kWh = $5,417 /yr

Example

A 20-hp hydraulic system motor draws 8 kW while the machine it powers is loaded and 5 kW while the machine idles. Thus, it draws 63% of loaded power when the machine idles by forcing the unneeded hydraulic fluid through a bypass valve. Calculate the savings from turning off the hydraulic motor for 2,000 hr/yrwhen the equipment it powers idles. The cost of electricity is $0.10 /kWh.

Annual Savings:

5 kW x 2,000 hr/yr = 10,000 kWh/yr

10,000 kWh/yr x $0.10 /kWh = $1,000 /yr

7) Power TransmissionThe drivetrain, or transmission, connects the motor shaft to the load. The most common types of drivetrains are direct shaft couplings, gears, belt drives and chains.

a) Direct Shaft CouplingsDirect shaft couplings transfer virtually 100% of the power from the motor to the load and are the most energy efficient type of power transmission.

b) Gear DrivesGears are typically used for loads which must run slowly and which require high torque such that a belt may slip. Helical and bevel gears are widely used and have an efficiency of about 98% per stage. Worm gears allow a large reduction ratio in a single stage and are usually cost less than helical or bevel gears. Worm gears have efficiencies between 55% and 95%. Gear drive efficiency decreases significantly at low load and low speeds.

c) V-Belt DrivesBelt drives allow flexibility in the positioning of the motor and load. In addition, the desired rotating speed of the load can be achieved by selecting pulleys (sheaves) with the proper diameters. Because of these advantages, belt-drives are very common.

The most common type of belt is a V-belt. As V-belts travel around the pulleys, they flex and heat up. This heat is energy supplied to from the motor and reduces the efficiency of power transmission between the motor and the load.

Standard V-belts have a smooth inner surface. Cogged, or notched, V-belts work are identical to standard V-belts, but are notched on their inner circumference. The notches make them more flexible and improve the efficiency of power transmission.

Standard and cogged V-belts. Source: Dayco CPT, www.cptbelts.com

A belt manufacturer (Dayco, www.cptbelts.com) reports that:

Cogged V-belts are 98% efficient compared to 94-95% efficient standard V-belts. Life of cogged V-belts is 50% longer than standard V-belts Cost of cogged V-belts is about 50% more than standard V-belts

Another study of motor electricity consumption with variable loads showed electricity savings of 2% to 3.75% when standard V-belts were replaced with cogged V-belts. In general, the savings increased with motor loading (see Figure and reference below).

Source: MichiganManufacturingTechnologyCenter, Manufacturing Efficiency Decision Support, Case Study - Cog Belts, http://meds.mmtc.org/casestudy.asp?X=Cog%20Belts

Another source reports that cogged V-belts increase the efficiency of transmission by about 2% (Energy Tips: Replace V-Belts with Cogged or Synchronous Belt Drives, DOE/GO-102000-0972, Office of Industrial Technologies, U.S. Department of Energy).

In addition to significantly improving power-transmission efficiency, cogged V-belts also last at least 50% longer than standard V-belts. Some maintenance personnel report that cogged V-belts last four times as long as standard V-belts. This longevity reduces equipment downtime and replacement costs, which more than compensates for the 20% to 50% higher cost of cogged V-belts.

Because of these advantages, we recommend using cogged V-belts in virtually all V-belt applications except for clutching applications because of their aggressive grip (Grainger, 2001-2002) or in noise sensitive environments since cogged V-belts are slightly noisier than standard V-belts.

Example

Replace smooth with notched V-belts on 100-hp, 91% efficient motor if the end use load is 75% of rated motor output. The motor operates 6,000 hours per year and the cost of electricity is $0.10 /kWh.

Annual Savings

100 hp x 75% / 0.91 x (1/.92-1/.95) x 0.75 kW/hp = 2.12 kW

2.12 kW x 6,000 hours/yr x $0.10 /kWh = $1,273 /year

Implementation Cost

Notched belts last 50% to 400% longer than smooth belts, but cost only 30% more than smooth belts, thus implementation cost is negligible

Simple Payback

Immediate

d) Synchronous Belt DrivesSynchronous belts have teeth that engage in the teeth of the sprocket pulley. Because synchronous belts are designed for maximum flexibility and have virtually no slippage, they are about 98% efficient and last about 4 times as long as V-belts.

However, synchronous belt drives are more costly than V-belts and typically require more attention to alignment to operate properly. Thus, they are typically used on large motors with long operating hours where energy savings outweighs the other issues.

De Almeida and Greenberg estimate that synchronous belts last about 24,000 hours, and cost between $8 and $16 per horsepower, with large drives at the low end of this spectrum.

Source: Nadel et al., 1991

Example

Replace smooth V-belt with synchronous belt drive on 100 hp, 91% efficient motor if end use load is 75% of rated motor power output.The motor operates 6,000 hours per year and the cost of electricity is $0.10 /kWh.

Annual Savings

100 hp x 75% / 0.91 x (1/.92-1/.98) x 0.75 kW/hp = 4.11 kW

4.11 kW x 6,000 hours/yr = 24,680 kWh/year

24,680 kWh/year x $0.10 /kWh = $2,468 /year

Implementation Cost

$12 /hp x 100 hp = $1,200

Simple Payback

$1,200 / $2,468 = 6 months

8) Motor EfficiencyThe cost of electricity for operating a motor is typically many times greater that the cost of the motor. The figure below shows typical electricity costs for various size motors and annual operation hours.

Source: Motor Decisions Matter, www.motorsmatter.org).

In fact, the cost of energy during the first year of use alone typically exceeds the purchase cost.

Example

Compare the $1,200 purchase cost of a 20-hp, 93% efficient motor to the annual energy cost if the motor operates 6,000 hours per year, is 75% loaded, and the cost of electricity is $0.10 /kWh.

Annual energy cost:

20 hp x 75% / 93% x .75 kW/hp x 6,000 hr/yr x $0.10 /kWh = $7,258 /yr

Ratio of 1 year energy to purchase cost:

$7,258 / $1,200 = 6x !

Ratio of 10 year energy to purchase cost:

($7,258 x 10) / $1,200 = 60x !

Thus, it is usually highly cost effective to purchase the highest efficiency motor available. Besides energy savings, high-efficiency motors offer several other important benefits. First, high-efficiency motors run cooler than standard motors because of lower losses and because they operate at a higher power factor. In addition, high-efficiency motors often use heavier duty bearings. Because of these changes, high-efficiency motors typically run longer than standard motors.

In the 1980s, most motor manufacturers began offering energy-efficient motors (EEMs) with efficiencies 2% to 10% higher than standard-efficiency motors (SEMs). As of October 1997, all motors manufactured or imported into the United States had to meet new, higher efficiency standards. Today, these motors are typically called "energy-efficient" motors (EEMs). Today’s highest efficiency motors are called "premium-efficiency" motors (PEMs). The table below shows the efficiency of standard motors before 1997, minimum efficiency ratings for all motors after 1997 (energy-efficient motors), and minimum efficiency ratings to qualify as a NEMA premium-efficiency motor after 1997.

Efficiencies for Totally Enclosed Fan Cooled, 4 Pole, ~1740 RPM Motors.

Size (hp) StdEff

Before 10/1997

EngyEff

After 10/1997

PremEff

After 10/1997

1 78.5 82.5 85.5

2 84.0 84.0 86.5

5 86.5 87.5 89.5

10 88.5 89.5 91.7

15 88.5 91.0 92.4

20 90.2 91.0 93.0

25 91.0 92.4 93.6

30 91.0 92.4 93.6

40 91.7 93.0 94.1

50 93.0 93.0 94.5

60 92.4 93.6 95.0

75 93.0 94.1 95.4

100 93.0 94.5 95.4

125 93.0 94.5 95.4

150 93.6 95.0 95.8

200 95.0 96.2

Source: StdEff Before 10/1997: Grainger 1995

Source: EngyEff After 10/1997: The Impacts of the Energy Policy Act of 1992 on Industrial End Users of Electric Motor-Driven Systems, U.S. Department of Energy, http://www.oit.doe.gov/bestpractices/motors/factsheets/e-pact92.pdf.

Source: PremEff After 10/1997: NEMA Premium, Product Scope and Nominal Efficiency Levels, www.nema.org/premiummotors.

9) Replace or Repair?Most motors failures are due to electrical or mechanical failure.The primary cause of electrical failure is degradation of winding insulation. Over time, winding insulation degradesdue to heating, aging and over-voltage transients. As winding insulation degrades, the efficiency of the motor may also degrade, causing its operating temperature to increase. Another indication of winding failure is a greater than 10% difference in the amperage drawn by each leg of a three-phase motor. Eventually, winding failure can lead to shock, fire hazard and total motor failure. Sometimes motor failures are related to mechanical breakdowns, especially for motors in high-vibration environments. 75% of all mechanical failures are due to bearing failure.

a) Deciding Whether to Repair or Replace MotorsWhen motors fail, they can be replaced or repaired. The decision to replace or repair must often be made quickly, because the costs of lost production may outweigh energy savings. For this reason, we recommend establishing a motor replacement plan before motors fail. This plan may include pre-selecting energy-efficient replacement motors before key existing motors fail. It may also include working with your motor vendor to ensure that the replacement motors you want can be obtained quickly.

Efficiency and cost data to determine the economics of replacing or repairing motors is incorporated into the MotorMaster+ software. MotorMaster+ includes a database of most motors currently on the market, is available free of charge from the U.S. D.O.E.Typical costs and efficiencies from MotorMaster+ are shown in the table below.

Source: US DOE Motor Master+ 4.0 (2007)

Size (hp) Efficiency Rwd Cost RwdEfficiency Premium

MotorCost Premium

Motor

1 74.0% $213 85.6% $333

1.5 76.8% $223 86.5% $352

2 79.3% $237 87.0% $406

3 80.5% $249 89.9% $489

5 83.0% $271 90.4% $495

7.5 84.9% $319 91.8% $706

10 85.7% $377 92.0% $805

15 86.5% $463 92.8% $1,089

20 88.3% $536 93.5% $1,286

25 88.9% $614 93.9% $1,696

30 89.2% $722 94.1% $2,104

40 89.4% $858 94.4% $2,747

50 91.1% $1,033 94.9% $3,183

60 91.4% $1,171 95.1% $4,400

75 91.5% $1,365 95.5% $5,118

100 91.7% $1,716 95.5% $6,167

125 91.1% $2,028 95.3% $8,075

150 92.3% $2,420 95.6% $9,301

200 92.8% $2,985 96.2% $11,281

250 93.6% $3,571 96.3% $14,051

300 93.7% $4,211 96.1% $19,846

350 94.0% $4,734 95.8% $24,404

400 93.1% $5,168 95.8% $26,750

450 94.0% $5,722 95.8% $27,896

500 93.7% $6,212 95.8% $29,436

These data can be used to determine the payback of replacing rather than repairing failed motors. In general, the payback timefor replacing rather than repairing failed motors is short, but increases with motor size.

Example

Determine the simple payback of replace rather than rewinding a 20-hp motor. The motor operates 6,000 hours per year, is 75% loaded, and the cost of electricity is $0.10 /kWh.

Annual Savings

20 hp x 75% x (1/.883-1/.935) x 0.75 kW/hp = 0.71 kW

0.71 kW x 6,000 hours/yr = 4,251 kWh/year

4,251 kWh/year x $0.10 /kWh = $425 /year

Implementation Cost

$1286 - $536 = $750

Simple Payback

$750 / $425 = 21 months

Example

Determine the simple payback of replace rather than rewinding a 100-hp motor. The motor operates 6,000 hours per year, is 75% loaded, and the cost of electricity is $0.10 /kWh.

Annual Savings

100 hp x 75% x (1/.917-1/.955) x 0.75 kW/hp = 2.44 kW

2.44 kW x 6,000 hours/yr = 14,645 kWh/year

14,645 kWh/year x $0.10 /kWh = $1,464 /year

Implementation Cost

$6,167 - $1,716 = $4,451

Simple Payback

$4,451 / $41,464 = 36 months

10) Motor MaintenanceA good motor maintenance program can extend the lifetime of motors, reduce production downtime from unexpected motor failures and reduce motor repair costs. Key elements of a good motor maintenance program are discussed below.

a) CleaningClean motors run better. Dirt acts as an insulator and causes the motor to run hotter. Dirt can also damage lubricants, bearings and insulation.

b) LubricationMost small motors and motors with factory-sealed bearings do not require lubrication. For motors that do require lubrication, we recommend carefully following manufacturer guidelines. Do not over lubricate and be careful not to introduce contaminants. For motors used seasonally, we recommend lubrication before the season of use.

c) MountingCheck the mounting system and hold-down bolts at every maintenance interval. Loose bolts, cracks or failure of the mounting structure can cause vibration and deflection that can damage bearings.

d) Belt DrivesCheck belt drives for proper tension or wear frequently. Loose v-belts vibrate, wear rapidly and waste energy. If the belts are too tight, high lateral loading may damage motor bearings. Always replace all belts on a drive at the same time. Check belts frequently until they are broken in, usually after about 48 hours of use.

e) Electrical, Thermal and Vibration TrackingWe recommend measuring and logging the current draw, voltage, temperature and vibration of large, process-critical motors. By tracking these key indices, process downtime can be avoided and repair expenses minimized.

Measure and log motor current draw and voltage across each phase. Voltage imbalances are not usually indicative of motor problems, but can cause current imbalances. Balanced voltage with unbalanced current may indicate motor problems. Some maintenance personnel send motors in for repair when the current imbalance between legs exceeds 10%. Changes in the current draw usually indicate changes in the load. This may indicate clogged filters or other problems.

Always make temperature and vibration measurements on the same spot on the motor. Simple hand-held contact thermometers and vibration meters work well. Make measurements after the motor is warmed up and when under the same motor loading conditions. An increasing temperature trend can indicate electrical or bearing problems. An increasing vibration trend indicates bearing wear.

f) Storage and TransportMotors can have a shorter life in storage than while operating. If the motor shaft is not periodically rotated, lubricant may drain away from bearings causing metal-to-metal contact and rust. Stored motors are especially susceptible to bearing damage from vibration, since lubricant may be pushed away allowing metal-to-metal contact. Insulation on the motor windings can absorb moisture from humid air.

When storing motors, rotate the shaft or operate the motor at least once per month. Avoid storing motors near vibratory machinery. Don’t store motors in cold or damp spaces where the relative

humidity exceeds 70%. Before shipping a motor, block the shaft, drain the oil, and label the motor as needing oil.

11) Oversized MotorsMost motors are sized to operate at 75% to 80% of full load power. At this level of loading, motor efficiency and power factor remain relatively high. However, motor efficiency and power factor decrease rapidly at less than 25% loading. Thus, dramatically oversized motors should be replaced with appropriately sized motors.

Example

Right-size a 100-hp, 10% loaded motor.The motor operates 6,000 hours per year and the cost of electricity is $0.10 /kWh.

From the figure above, the efficiency of a 100-hp motor at 10% load is about 60%. At this load and efficiency, input power is:

100 hp x 10% / 60% x .75 kW/hp = 12.50 kW

Output power is:

100 hp x 10% = 10 hp.

Thus, the load could be driven with a 10-hp motor. From the figure above, the efficiency of a 10-hp motor at 100% load is about 82%. Input power of a 10-hp motor at 82% efficiency is:

10 hp / 82% x .75 kW/hp = 9.15 kW

Annual savings:

(12.5 kW - 9.15 kW) x 6,000 hr/yr = 20,100 kWh/yr

20,100 kWh x $0.10 /kWh = $2,012 /yr

Implementation Cost

A 10-hp motor costs about $500

Simple Payback

$500 / $2,012 /yr x 12 mo/yr = 3 months

12) Soft-Start ControlsWhen a motor starts under a high-torque load, it typically draws much more than full load power as it comes up to speed. The time for a motor to come up to speed varies according to the load, but is usually less than 10 seconds. Over time, full-voltage hard starts with high in-rush currents can damage the motor. The high in-rush current may also trip circuit breakers.

Source: Power Efficiency Corporation product literature, www.performancecontrol.com, 734-975-9111.

Soft-start controls limit the in-rush of current during motor startup to create smoother but longer starts. Because of this, soft-start controls may extend motor life, especially for motors that frequently cycle on and off. Soft-start controls have minimal affect on peak electrical demand because the high inrush startup current typically lasts only a few seconds and the peak demand is usually averaged over 15 minutes to one hour.

Soft-start controls can also reduce electricity consumption for under-loaded motors by modifying the voltage waveform. Soft-start controller manufacturers claim energy savings of between 20% and 50% for lightly-loaded motors. The controllers react fast enough to track variable loads without difficulty. Reduced energy consumption also results in cooler running motors. In addition, soft-start controls may also reduce current imbalance between phases and improve power factor.

Ideal applications for soft-start controls are motors that are frequently cycled on and off, and/or motors that run at reduced load most of the time. Typical prices for soft-start controls are shown below.

Motor Size Cost (including installation kit)

<= 20 hp $840

40 hp $2,100

100 hp $2,800

200 hp $5,700

Source: Power Efficiency Corporation

www.performancecontrol.com, 734-975-9111.

In addition, manufactures report that the installation time for a qualified electrician is about two hours.

13) ReferencesAmbs, L. and Frerker, M., 1997, “The Use of Variable Speed Drives to Retrofit Hydraulic Injection Molding Machines”, National Industrial Energy Technology Conference, Houston, TX, April.

Danfoss, “Application Reference Guide”, Version 0.00, Danfoss, Inc.

Machelor, J., 1999, “Root Cause Failure Analysis on AC Induction Motors”, Energy Matters, Office of Industrial Technologies, US Department of Energy, May 1999.

McCoy, G. and Douglass, J., 1997, “Energy Management for Motor-Driven Systems”, Motor Challenge Program, US Department of Energy, Washington, DC.

MotorMaster+ 3.0, U.S. Department of Energy, Office of Industrial Technologies, Washington, D.C.

Nadel, S., Shepard, M., Greenberg, S., Katz, G., and Almeida, A., 1991, “Energy Efficient Motor Systems”, American Counsel for an Energy Efficient Economy, Washington D.C.

De Almeida and Greenberg ,"Technology assessment: Energy-Efficient Belt Transmissions."

SIMULATION RESULTS

The prototypes of the proposed inverters are built using TMS320F2812 digital signal

controller. The experimental results are categorized according to the inverters’ DF or EF

operation mode to provide a basis for performance comparison. The output RL loads which are

similar in value for DF and EF operations respectively are R=14Ω, L=8.9mH, and R=25Ω,

L=7.1mH for upper and lower outputs. The level of dc bus voltages are adjusted. The loads with

the same demanded power at corresponding modulation indices. In the experimental tests, the

base value of Vdc is 50V and the switching frequency fSW is 6 KHz supply with three 2200uF

capacitor banks employed at the dc-link of the three-switch inverter. The reference signals are

presumed to be in phase and appropriate offsets are added to them for producing modulating

signals. The output waveforms of the six-switch inverter while working in EF mode Compared

to DF mode, lower level of dc bus voltage is required to supply the load with the same power.

Moreover, the output currents are less distorted which is in compliance with the conducted THD

analysis. The performance of the three-switch inverter Offsets of 0.1 and -0.1 are respectively

added to the upper and lower output reference signals. Consequently, upper and lower dc link

voltage levels are the same and equal to 40V and the middle voltage level is 20V. As a result the

upper PWM voltage is being switched between 40V and -60V and the lower PWM voltage is

being switched between 60V and -40V. The output currents are less distorted at the peaks since

first, the difference between voltage levels of output. PWM voltages has decreased and second,

smaller offsets are added to the reference signals and therefore the generated voltage pulses are

of almost equal width at the peaks of the reference signals.

Simulation wave forms

the simulations results of the proposed shunt active power filter controlled by fuzzy logic

and a conventional PI controller with MATLAB program. The parameters selected for

simulation studies are given in table 6.1. The three phase source voltages are assumed to be

balanced and sinusoidal. The source voltage waveform of the reference phase only (phase-a, in

this case) is shown in fig.6.1. A load with highly nonlinear characteristics is considered for the

load compensation. The THD in the load current is 28.05%. The phase-a load current is shown in

figure 6.2. The source current is equal to the load current when the compensator is not

connected.

Simulation wave forms

The compensator is switched ON at t=0.05s and the integral time square error (ITSE)

performance index is used for optimizing the and coefficients of the PI controller. The optimum

values (Kp and Ki) are found to be 0.2 and 9.32, respectively, which corresponds to the

minimum value of ITSE. The source currents for PI and fuzzy controllers are shown in Figs.6.3

and 6.6, respectively. Compensating currents of PI and fuzzy controllers are shown in figures 6.4

and 6.7. The DC side capacitor voltage during switch on response is shown in figures 6.5. and

6.8 of PI and fuzzy controllers

Simulation wave forms

From the wave forms it is clear that harmonic distortion is reduced after connecting compensator.

Compared to PI controller fuzzy controller fuzzy controller gives better harmonic compensation.

The system studied has also been modeled using simulink and performance of PI and Fuzzy

controllers is analyzed. The system parameters selected for simulation study are given in table 6.1

and 6.2. Figures 6.9-6.18 shows the simulation results of the implemented system with PI controller

and fuzzy controllers with simulation parameters mentioned in table 6.1. The source voltage

waveform of the reference phase only (phase-a, in this case) is shown in fig.6.9. A diode rectifier

with R-L load is taken as non-linear load. The THD of the load current is 27.88%. The optimum

values (Kp and Ki) are found to be 0.2 and 9.32 respectively.

Simulation wave formsFrom the responses it is depicted that the settling time required by the PI controller is

approximately 8 cycles whereas incase of fuzzy controller is about 6 cycles. The source current THD

is reduced form 27.88% to 2% incase of PI controller and 2.89% incase of fuzzy controller which is

below IEEE standard with both the controllers. Figures 6.19-6.28 shows the simulation results of the

implemented system with PI controller and fuzzy controllers with simulation parameters mentioned

in table 6.2. The source voltage waveform of the reference phase only (phase-a, in this case) is shown

in fig.6.19. A diode rectifier with R-L load is taken as non-linear load. The THD of the load current

is 28.34%.

Simulation wave forms

From the responses it is depicted that the settling time required by the PI controller is

approximately 10 cycles whereas incase of fuzzy controller is about 7.5 cycles. The peak

overshoot voltage incase of PI controller is 880Volts (approx) whereas incase of fuzzy controller

is 780volts (approx). The source current THD is reduced form 28.34% to 4.7% which is below

IEEE standard with both the controllers. After compensation both source voltage and current are

in phase with each other means that the harmonics are eliminated and reactive power is

compensated to make power factor close to unity. As the source current is becoming sinusoidal

after compensation power quality is improved.

CONCLUSION

Dual-leg and single-leg reduced switch count dual-output inverter topologies based on three-

switch inverter legs were proposed in this paper with the aim of reducing cost, size and weight of

low power inverters. The performance of the proposed topologies are compared with the

conventional topologies regarding the output waveform characteristics and semiconductor power

losses.

SCOPE FOR THE FUTURE WORK

Experimental investigations can be done on shunt active power filter by developing a prototype

model in the laboratory to verify the simulation results for both PI and fuzzy controllers.

REFERENCES

[1]. G. T. Kim, T. A. Lipo, “VSI-PWM Rectified/Inverter System with a Reduced Switch

Count”, IEEE Trans. Industry Applications, vol. 32, pp. 1331-1337, no. 6, 1996.

[2]. C. B. Jacobina, M. B. de R. Correa, A. M. N. Lima, E. R. C. da Silva, “AC Motor Drive

Systems with a Reduced Switch Count Converter”, IEEE Trans. Industry Applications, vol. 39,

no. 5, pp. 1333-1342, 2003.

[3]. H. W. Park, S. J. Park, J. G. Park, and C. U. Kim, “A novel high performance voltage

regulator for single-phase ac sources,” IEEE Trans.Industrial Electronics, vol. 48, no. 3, pp. 554–

562, Jun. 2001.

[4]. J.-H. Choi, J.-M. B. Kwon, J.-H. Jung, and B.-H. Kwon, “High performance online UPS

using three-leg-type converter,” IEEE Trans.Industrial Electronics, vol. 52, no. 3, pp. 889–897,

Jun. 2005

[5]. C. B. Jacobina, I. S. de Freitas, E. R. C. da Silva, A. M. N. Lima, R. L. D. A. Ribeiro,

“Reduced Switch Count DC-Link AC_AC Five-Leg Converter”, IEEE Trans. Power Electronics,

vol. 21, no. 5, pp. 1301- 1310, 2006.

[6]. C. B. Jacobina, I. S. de Freitas, E. R. C. da Silva, “Reduced-Switch- Count Six-Leg

Converters for Three-Phase-to-Three-Phase/Four-Wire Applications”, IEEE Trans.Industrial

Electronics, vol. 54, no. 2, pp. 963-973, 2007.

[7]. B. Jacobina, I. S. de Freitas, A. M. N. Lima, “DC-Link Three-Phaseto- Three-Phase Four-

Leg Converters”, IEEE Trans. Industrial Electronics, vol. 54, no. 4, pp. 1953-1961, 2007.