## Introduction

## 1.1 A Look Back on Development of Vector Control Induction Machine Drives

In the past decennaries, DC machines have been used extensively in many applications due to its simpleness and offers high public presentation torsion control. Their mechanical commuting action forms the torsion and flux constituents are ever extraneous, which allows high public presentation control achieved under optimum conditions. Despite their first-class public presentation and simple control construction, their building utilizing the mechanical commutator and coppice cause some restrictions. For illustrations, they require regular care and can non be operated in the dirty or explosive environment, their velocity is limited and they are expensive. Unlike the DC machines, the initiation machines are robust, necessitate less care, cheaper and able to run at higher velocity operations. The initiation machines became popular and easy replacing DC machine thrusts in many industrial applications as the Field Oriented Control ( FOC ) introduced by F. Blachke in 1970 ‘s can bring forth comparable public presentation to that obtained in DC machines [ 1 ] . Furthermore, their popularity is besides assisted by the rapid development in power semiconducting material devices and the outgrowth of high-speed microprocessor and digital signal processors [ 2 ] . FOC of initiation machines can supply a decoupled control of torsion and flux utilizing their several bring forthing current constituents, which is similar to the DC machine control method. This is possible by sing the stator current vector in the revolving mention frame so that the extraneous constituents of the stator current ( represent as torsion and flow bring forthing current constituents ) appear as DC quantities. The revolving mention frame refers to the revolving infinite vector measure which can be either rotor flux or stator flux. The one which is based on rotor flux ( referred as Rotor Flux Oriented Control ( RFOC ) ) is sensitive to the parametric quantity fluctuations, and besides it requires the cognition of rotor velocity. The betterments of the FOC public presentation were reported in many proficient documents ; these include the countries of robust control [ 3-5 ] , improved flux calculator [ 6-9 ] and sensorless velocity control [ 10-12 ] .

## 1.2 Direct Torque Control of Induction Machines

Over the past old ages, Direct Torque Control ( DTC ) strategy for initiation motor thrusts [ 13 ] has received tremendous attending in industrial motor thrust applications. It was besides used as the chief platform for the ABB inverter engineering, and the first DTC thrust was marketed by ABB in 1996 [ 14-15 ] . Its popularity additions since it offers a faster instantaneous dynamic control and its control construction is simpler compared to the field-oriented control ( FOC ) strategy [ 13, 16-17 ] . In DTC, the torsion and flux are controlled independently, in which their demands are satisfied at the same time utilizing optimal electromotive force vectors. Unlike the FOC thrusts, where the torsion and flux are controlled corresponded to their bring forthing current constituents ( i.e. vitamin D and q axis constituents of stator current ) which requires frame transformer, cognition of machine parametric quantities and current regulated Pulse Width Modulation ( PWM ) .

## The Control Structure of Basic DTC

The simple control construction of DTC proposed by Takahashi [ 13 ] is shown in Figure 1.1. It contains a brace of hysteresis comparators, torsion and flux calculators, exchanging tabular array for electromotive force vectors choice and a 3-phaseVoltage Beginning Inverter ( VSI ) . A decoupled control construction is provided in the strategy, wherein the electromagnetic torsion and the stator flux can be controlled independently utilizing 3-level and 2-level hysteresis comparators, severally. To fulfill the torsion and flux demands, the appropriate electromotive force vector for a peculiar flux sector should be chosen, either to increase torsion or to cut down torsion and either to increase stator flux or to diminish stator flux ; based on torsion mistake position, flux mistake position and flux orientation. In fact, high public presentation control in DTC thrust with proper choice of electromotive force vectors can be established if the flux every bit good as torsion is estimated accurately. A combination of the electromotive force theoretical account and the current theoretical account utilizing a first-order slowdown web was implemented by the [ 13 ] ( which was introduced the DTC ) to obtain a proper flux calculator for full velocity scope operations. The usage of the current theoretical account is superior in gauging the flux at really low velocity operations, but it requires the cognition of motor velocity, which in bends requires an perceiver or a velocity detector from the motor shaft. On the other manus, the execution of the electromotive force theoretical account provides better flux appraisal for high velocity operations and requires merely stator opposition and terminal measures ( i.e. stator electromotive forces and stator currents ) , which leads to the robust control of DTC. However, the digital execution of this theoretical account using microprocessor or digital signal processor introduces some jobs, i.e. the integrating impetus and initial status jobs [ 18-20 ] .

In hysteresis-based DTC, the switching frequence of VSI is wholly contributed by the shift in the hysteresis comparators. It was highlighted in [ 21-23 ] that the inclines of torsion and flux, which comparatively affect the shift in their hysteresis comparators, vary with the operating conditions ( i.e. rotor velocity, stator and rotor fluxes and DC link electromotive force ) . This causes the switching frequence of VSI besides varies with the operating conditions. The fluctuation in the switching frequence as a consequence produces unpredictable harmonic current flow. For this ground, the exchanging devices can non be to the full utilised to its maximal frequence capableness for most operating conditions, even the choice of hysteresis set ‘s breadth is based on the worst status.

Figure 1.1 Control construction of basic DTC-hysteresis based initiation machine

Switch overing tabular array

## +

IM

Sa, Sb, Sc

sector

## –

Te, ref

## +

## –

?s, ref

Tellurium

?s

Voltage beginning inverter ( VSI )

Stator flux and electromagnetic torsion calculators ( i.e. merely utilizing the electromotive force theoretical account )

Iowa, ib, Intelligence Community

Virginia, vb, vc

## Improvements of DTC Performance

Since DTC was foremost introduced in 1986 [ 13 ] , several fluctuations to its original construction ( which referred to as DTC hysteresis-based ) were proposed to better the public presentation of DTC of initiation machines. Perceptibly, most researches in last decennaries aimed to get the better of the built-in disadvantages of hysteresis-based DTC strategies, such as variable exchanging frequence and high torsion rippling. In add-on, some recent researches interested to ease the DTC to execute under overmodulation manner. Using overmodulation scheme, it will better the dynamic public presentation and the power end product of DTC of initiation machines. To associate the survey of this thesis, some reappraisals on betterments of DTC are concentrated into three following research countries:

Torque Ripple Reduction or/and Constant Switching Frequency

Basically, the end product torque rippling can be reduced by take downing the set breadth of hysteresis comparator to the appropriate value. The choice of the appropriate set breadth is based on the worst operating conditions [ 24 ] . This will guarantee the switching frequence of exchanging devices do non transcend its bound ( or thermal limitation ) . It is besides desirable to utilize a high velocity processor to let the rippling ever maintain within the set, in such a manner the distinct hysteresis accountant will execute like the parallel one.

Alternatively of take downing the hysteresis set and utilizing a high velocity processor, the rippling can be minimized by shooting high-frequency triangular wave forms to the mistakes of torsion and flux [ 25 ] . This method is called dithering technique, which is simple and effectual to minimise the torsion rippling even executing DTC at limited trying frequence. However, it still produces unpredictable exchanging frequence since the torsion and flux inclines which are related to the shift frequence, vary on operating conditions [ 23, 26 ] .

Several methods had been proposed to supply a changeless shift frequence every bit good as reduced end product torsion rippling [ 22, 27-38 ] . In [ 27 ] , a changeless shift frequence is established in the hysteresis-based DTC by seting the set breadth of the hysteresis comparators harmonizing to the alterations in operating conditions. The adjustable of set breadth is based on a PI accountant and a pulse counter, for each comparator. This, accordingly, increases the complexness of the DTC thrust. Furthermore, this technique does non vouch decrease of torque rippling as it is inevitable in the execution of distinct hysteresis accountants. Another attack to minimise the jobs is to command the DTC exchanging based on optimum exchanging blink of an eye that satisfies the minimum-torque ripple status [ 22, 29-30 ] . In this manner, the term so called a responsibility ratio is determined so that the appropriate active province is switched for some part of a switching period, and the nothing vector is selected for the remainder of the period. It was shown in [ 24 ] that the shift frequence is strongly affected by the torsion hysteresis set. Therefore, it makes sense that the about changeless shift frequence every bit good as reduced torque rippling can be achieved if the control of torsion mistake is performed at a higher and a changeless frequence. As proposed by [ 32 ] , the chief operation of torsion accountant is substituted from the hysteresis-based to the triangular carrier-based. The betterment is achieved and at the same clip retains the simple control construction of DTC. The most common attack to work out the jobs is utilizing the infinite vector transition [ 28, 31, 33-35 ] . In this attack, the switching period is subdivided into three or more provinces, to synthesise the coveted electromotive force vector in order to bring forth the minimal torsion rippling.

Recently, prognostic control scheme to DTC has gain considerable sum of attending peculiarly due to its ability to cut down the torsion rippling and every bit good as exchanging frequence [ 36-39 ] . In peculiar theoretical account prognostic control ( MPC ) which was applied in [ 38 ] and [ 36 ] uses the hysteresis comparators but with the exchanging tabular array replaced with on-line optimisation algorithm.

Fast Torque Dynamic Control through Dynamic Overmodulation Strategy

Without the demand to utilize any excess hardware, the torque dynamic public presentation can be improved by to the full using the available DC nexus electromotive force through overmodulation. To execute the DTC under the overmodulation manner, it is preferred to utilize the Space Vector Modulation ( SVM ) [ 28, 40-44 ] instead than other techniques ( for illustration, [ 45-46 ] use the triangular bearer based ) . This is because, the SVM is more flexible to be adapted in advanced motor control where merely a individual mention electromotive force vector V, ref is employed to specify the manner of overmodulation.

During big torque demand, it is inevitable that the mention electromotive force V, ref exceeds the electromotive force vector bounds enclosed by the hexangular boundary. Under this status, the SVM has to be operated in what is termed as dynamic overmodulation manner. The electromotive force mention vector V, ref has to be modified such that it will lie on the hexangular boundary.

Several methods [ 28, 43, 47-49 ] have been proposed and to some extent, these methods have managed to minimise the electromotive force vector mistake every bit good as obtained a fast torsion response. Figure 1.2 compares some modified electromotive force mentions, v [ I ] ( e.g. when one =1, proposed in [ 1 ] ) with regard to the original electromotive force mention vector, V, ref, which is beyond the hexangular boundary of the electromotive force vectors. Note that, electromotive force vector constituents are non drawn to scale. It can be seen that ( from Figure 1.2 ) , [ 42 ] and [ 50 ] switched merely individual electromotive force vector which is vk+2 during dynamic overmodulation. This individual choice of vector shows the happening of a six-step operation that produces the fastest dynamic torsion control as will be discussed subsequently in the Chapter VI. While the other methods result in slower dynamic torsion response since two active provinces are alternately switched during the dynamic status. For illustration, [ 28 ] used two active provinces using dead-beat control in order to keep the magnitude of stator flux under control for any status. Later, [ 43 ] was proposed to simplify the complexness control construction in [ 28 ] , ( where it does non supply a dead round control of the magnitude flux as a transient torsion encountered ) and hence consequences in a faster dynamic torsion control. In this manner, the modified electromotive force vector, V [ 43 ] has the same angle, ? as the original mention electromotive force, V, ref but with a modified magnitude. In [ 47 ] , the mention electromotive force, V, ref was modified to v [ 47 ] such that the mistake between the magnitude of V [ 47 ] and V, ref is minimized. This means the modified electromotive force vector, V [ 47 ] should be closest to the original mention vector V, ref by guaranting that a line fall ining the V, ref and V [ 47 ] is extraneous to the hexagon boundary.

K

## ?s

k+1

k+2

k+3

k+4

k-1

## vk+1

## vk+2

## V [ 28 ]

## V [ 40 ]

## V, ref

## V [ 41 ]

?

## ??s ? V [ 28 ] .?t

Hexagonal boundary

Locus of flux mention

*Note:

V [ I ] proposed in [ I ]

## V [ 44 ] , V [ 45 ]

Figure 1.2 Variations of modified mention electromotive force vectors which is applied during dynamic overmodulation manner.

Enhanced Torque Capability utilizing Flux Weakening or/and Overmodulation

Scheme

A wide-speed high torsion capableness is a really of import characteristic in many electric motor thrust applications. In automotive applications, it is normally referred to as a high ‘Constant Power Speed Range ‘ ( CPSR ) . The handiness of broad scope of velocity operations with the maximal capableness of torsion is the chief concern component, particularly for the route electric vehicles where multiple cogwheels have to be avoided. This is the most likely hereafter electric vehicle system demand.

In pattern, a flux weakening scheme is usually used to widen the motor velocity operations beyond the base velocity and to heighten the capableness of torsion. Several documents were published [ 42, 50-63 ] suggesting the solution of accomplishing maximal torque capableness in field weakening part. The common attack adopted is to gauge the optimum flux degree of the motor based on the maximal values of inverter electromotive force and inverter current. Typically, the algorithms require frame transformer, cognition of machine parametric quantities and space-vector modulator. For illustrations, [ 53 ] used Field Oriented Control-Space Vector Modulation ( FOC-SVM ) while [ 63 ] used Direct Torque Control-Space Vector Modulation ( DTC-SVM ) , and they consider electromotive force and current bound conditions to calculate the governable currents ( in stator flux mention frame ) in accomplishing the appropriate flux degree in field weakening part. Besides that, some documents had been reported to supply a robust field weakening scheme so that any fluctuations of machine parametric quantities used in ciphering the optimum flux can be compensated [ 55, 59, 61-62 ] .

In order to accomplish the fastest torque dynamic response every bit good as high torsion capableness in a flux weakening part, the DTC needs to hold the ability to execute the stator electromotive force to the six-step manner. However, merely a few strategies have the ability to execute this ; i.e. [ 42 ] used DTC-SVM with the predictive-deadbeat control and [ 50 ] used Direct Self Control with hexangular flux operation.

## The Popularity Versus The Complexity of DTC-SVM based initiation machines

The most popular fluctuation of DTC of initiation motor thrusts is the 1 that is based on infinite vector transition ( SVM ) , which usually referred to as DTC-SVM [ 28, 33, 40, 42, 64-66 ] . As mentioned antecedently, the advantages provided by this strategy non merely work out the built-in jobs in hysteresis-based DTC but besides facilitate the DTC performs under overmodulation part. Furthermore, the survey of the SVM itself in the country of overmodulation scheme betterment for 3-phase VSI was reported by many research workers [ 41, 46-47, 49, 66-73 ] . All this grounds shows that the execution of SVM to run under over a transition part is good established, and no admiration it has to be the first option to be adapted for advanced motor control.

The major difference between DTC hysteresis-based and DTC-SVM is the manner the stator electromotive force is generated. In DTC hysteresis-based the applied stator electromotive force depends on electromotive force vectors, which are selected from a look-up tabular array. The choices are based on the demand of the torsion and flux demand obtained from the hysteresis comparators. On the other manus, in DTC-SVM, a stator electromotive force mention is calculated or generated within a sampling period, which is so synthesized utilizing the infinite vector modulator. The stator electromotive force mention vector is calculated based on the demand of torsion and flux demands. Due to the regular sampling in SVM, the DTC-SVM produces changeless exchanging frequence as opposed to the variable shift frequence in hysteresis-based DTC nevertheless, at the disbursal of more complex execution.

The coevals of mention electromotive force V, ref frequently involves complex computation. For illustrations ; [ 28 ] used dead-beat control with several complicated equations ( i.e. quadratic equations ) to bring forth the mention electromotive force in real-time and [ 64 ] utilised prognostic control of stator flux mistake vector to gauge the mention electromotive force and needed excess computation on the synchronal angular speed. While the others include the usage of proportional-integral current accountant [ 40 ] , stator flux vector mistake [ 65, 74-75 ] , proportional-integral torsion and flux accountants [ 33, 76-77 ] , prognostic and dead-beat accountants [ 42-43 ] . Furthermore, the execution of DTC-SVM becomes complicated as the mention electromotive force needs to modify whenever it passes outside the hexagon ( due to the physical restraint of inverter as the V, ref defines manner of overmodulation ) , peculiarly during a big torque demand. Some modified mention electromotive forces as mentioned antecedently are shown in Figure 1.2. Figure 1.3 shows a general block diagram of coevals of exchanging in DTC-SVM, which contains three chief constituents and the functional of the two of them are as mentioned above.

Ultimately, all of the proposed methods [ 28, 33, 40, 42-43, 64-65, 74-77 ] complicate the basic control construction of DTC drive systems as originally proposed in [ 13 ] .

Figure 1.3 A general block diagram of coevals of exchanging in DTC-SVM

SVM

modulator

## +

## –

Te, ref

?s, ref

Tellurium

?s

## +

## –

Reference electromotive force coevals

??s

Reference electromotive force alteration

vsd, ref

vsq, ref

vsd, mod

vsq, mod

Sa, Sb, Sc

## Thesis Aims and Contributions

The aim of this thesis is to research, implement and better the public presentation of the DTC of initiation machines. Despite the betterments, the proposed methods besides aim to continue the simple control construction of DTC thrust. The thesis proposes a simple, yet important, method of overmodulation scheme employed in the DTC-hysteresis based initiation machine to better torque dynamic control and to heighten torque capableness for a broad velocity scope operation. The thesis besides shows the proposed methods can be employed in [ 32 ] ( which has a similar control construction with the hysteresis-based DTC ) , in order to give excess advantages of supplying a changeless shift frequence and decreased torsion rippling. While making the research, the thesis makes the undermentioned parts:

It shows the decrease of torsion rippling in [ 32 ] can be farther reduced by enlarging the changeless frequence of triangular bearers. This is achieved without the demand of cut downing the sampling period which is utile for the DSP or microprocessor with a limited sampling frequence.

It analyses the consequence of different electromotive force vectors choice, angular flux speed and operating conditions on the public presentation of torque dynamic control in DTC initiation machine thrust. Although the hysteresis-based DTC is good established to give a high public presentation torsion control, there is still room to better farther the public presentation based on the observation of the analysis.

It introduces a block of alteration of flux mistake position in the hysteresis-based DTC thrust which is responsible to execute dynamic overmodulation. This technique is really simple, yet it gives the fastest torque dynamic control. The introduced block can besides be implemented in [ 32 ] to verify the betterments.

It justifies the extension of changeless torsion part and hence betterment of torsion capableness can be obtained by commanding the venue of flux into hexangular form. With this method, the stator electromotive force in the DTC thrusts ( i.e. in [ 32 ] and hysteresis-based DTC ) can be transformed with a smooth passage from PWM to the six-step manner.

It describes the alteration of flux mistake position to set up the control of hexangular flux venue, peculiarly during dynamic conditions ( i.e. big torque demand or motor acceleration ) . The intercrossed DTC is proposed to run the control of flux into double manner, i.e. round flux venue as performed in DTC during the steady province ( that produces lower current ‘s THD ) and hexangular flux venue as performed in Direct Self Control ( DSC ) when dynamic status encountered ( that gives superior dynamic public presentation ) . The feasibleness of the proposed method is verified utilizing the hysteresis-based DTC and the strategy proposed by [ 32 ] .

It improves dynamic public presentations in flux weakening part. With the proposed overmodulation scheme, the fastest torque dynamic control and higher torsion capableness are achieved with the complete six-step electromotive force.

It introduces a measure decrease of mention flux to guarantee proper torsion control when a motor velocity reaches its mark in a flux weakening part. This to avoid an inappropriate degree of mention flux ( i.e. excessively high flux degree ) as the hexangular flux alterations to the handbill.

It develops a simulation and experimental set-up to verify the proposed betterments on the DTC thrust. The simulation is based on Matlab and Simulink plan from Mathworks inc.. The experimental set-up consists of a DS1102 DSP accountant board as the chief processor and APEX20KE Altera FPGA. The combined accountant board is implemented so that the sampling period of the DSP can be minimized by administering some of the undertakings to the FPGA.

## Methodology of Research

A freshness of the overmodulation scheme was developed by look intoing assorted overmodulation schemes, peculiarly associated with the SVM attack. The exchanging form of the electromotive force vectors in overmodulation part was translated or mapped onto the stator flux plane since the hysteresis-based DTC there is no mention electromotive force vector. A thorough probe on the assorted flux weakening schemes, changeless shift frequence and decreased torsion rippling scheme was besides carried out.

The development of the overmodulation scheme for the hysteresis-based DTC is based on the alteration of flux mistake position on the stator flux plane. To accomplish the fastest torque dynamic control, the most optimized electromotive force vector is switched and held. The choice of the electromotive force is determined by the modified flux mistake position. To heighten torque capableness, the proposed overmodulation and flux weakening schemes are utilized in the intercrossed DTC strategy. This strategy performs the control of flux in double manner operations. This is established utilizing the introduced block named as ‘block of Modification of Flux Error Status ‘ . The betterments utilizing the proposed methods besides can be achieved with the excess advantages, i.e. reduced torque rippling and changeless shift frequence utilizing the strategy proposed by [ 32 ] .

Once the algorithm for the overmodulation and field weakening was developed, it had to be applied to the DTC thrust and simulated utilizing big signal theoretical account. The Matlab/SIMULINK simulation bundle was used for this intent.

The proposed overmodulation and field weakening schemes using a intercrossed DTC strategy had been verified and evaluated for its feasibleness and effectivity through hardware execution. A state-of-the-art digital signal processor and field programmable gate array devices ( DS1102 controller board and APEX20KE Altera FPGA ) were used to implement the DTC accountants including the overmodulation and field weakening schemes. A standard initiation machine with suited tonss and IGBT-based VSI had been used for this intent.

## Administration of the Thesis

The remainder of the thesis is organised as follows:

Chapter II describes the mathematical mold of initiation machine and the basic rule of DTC of initiation machines. The built-in jobs in distinct execution of hysteresis-based DTC such as variable shift frequence, high torsion rippling and the demand of high velocity processor are besides discussed.

Chapter III suggests to utilize the simple method proposed by [ 32 ] to execute a changeless shift frequence and to cut down end product torsion rippling, alternatively of utilizing the common attack that is SVM. A speedy usher to plan or to obtain a proper accountant of the Constant Frequency Torque Controller ( CFTC ) is presented. Further betterment in cut downing the torsion rippling can be achieved by widening the triangular bearer frequence to its upper limit ( which is equal to one-fourth of the maximal sampling frequence achieved by the DSP ) .

Chapter IV proposes a simple dynamic overmodulation to accomplish the fastest torque dynamic control. The consequence of choosing two possible electromotive force vectors on the torque dynamic public presentation is analyzed. Based on the observation, the most optimized electromotive force vector is identified, and it is used to bring forth the fastest torque dynamic response. Simulation and experimental consequences are presented to demo the effectivity of the proposed method. Experimental consequences on the proposed method implemented in [ 32 ] is besides presented.

Chapter V proposes a simple overmodulation and flux weakening scheme using intercrossed DTC strategy. The intercrossed DTC strategy is constructed to ease the DTC thrust to run similar to that of DSC, which based on hexangular flux venue. By commanding the venue flux into hexangular form, the bound of stator electromotive force can be reached up to the six-step manner as the motor operates in flux weakening part. As a consequence, the intercrossed DTC performs at higher torsion capableness and hence gives a lesser motor acceleration clip. Simulation and experimentation every bit good as comparing with the conventional DTC strategy to verify the feasibleness of the proposed method, are presented. Experimental consequences on the proposed method implemented in [ 32 ] is besides presented.

Chapter VI describes the experimental set-up used in the thesis. Details information of each hardware constituent and the execution of the undertakings are given.

Chapter VII gives the decision of the thesis and possible waies of farther research.