At that time, only thyristor power semiconductors were available for ... major concern in the direct torque control of switched reluctance motor drives is the torque ...
CHAPTER 1 INTRODUCTION
GENERAL The concept of the Switched Reluctance Machine (SRM) is actually
very old, going back to the 19th century which was the forerunners of modern stepper motors. At that time, only thyristor power semiconductors were available for the relatively high-current, high-voltage type of control needed for SRM. These years, power semiconductor devices like GTOs, IGBTs have been developed in the power ranges required for SRM control. Simple construction is a prime feature of this motor. SRM eliminates Permanent Magnets (PMs), brushes, commutators and hence the excellent overall performance of SRM makes it an efficient competitor for AC drives. It finds wider application in automotive industries, direct drive machine tools, aerospace and ship propulsion.
The control of SRM is the recent trend of research as they offer a viable option for variable speed electrical drives, particularly for applications with high temperature and hostile operating environments such as small automotive motor drives, acting as a replacement for DC motors and brushless DC motors. Progress in the motor design and drive converter topologies has made SRM drives comparable in cost and performance to induction machines and brushless DC motors in adjustable speed applications like pumps, fans.
But SRM has complications implemented due to mutual coupling of the motor phase and parameter variation of inductance characteristics. Previous control schemes involve using of linear or nonlinear models. Though linear systems were simple, they were highly inaccurate as torque and flux are both nonlinear functions. So some schemes developed non linear characteristics of SRM. The nonlinear model implementations were complex and could not be implemented in real time.
They were expensive and
affected by variations in saturation. To overcome all these problems, the Direct Torque Control (DTC) was proposed which provided simple solution to control the motor torque and speed. Direct torque control has many promising features and advantages such as absence of speed and position sensors, absence of coordinate transformation, reduced number of controllers and minimal torque response time. In addition, there are many limitations that need to be investigated. The major concern in the direct torque control of switched reluctance motor drives is the torque and flux ripples. This is because none of the switching vectors is able to generate the exact stator voltage required to produce the desired changes in torque and flux. So the torque and flux ripple minimization is the major concern in this work. 1.2
OBJECTIVE OF THE THESIS The minimization of the torque ripple is essential in high
performance servo applications, which require smooth operation with minimum torque pulsations. The excellent positive features of an SRM can be utilized in a servo system by developing techniques to reduce the torque ripple. These types of drives have extensive applications in automotive industries, direct drive machine tools. So, the objective of the work is to minimize the torque ripple in direct torque control of switched reluctance
motor drive using multiple voltage space vectors. In this thesis, two methodologies have been introduced to increase the voltage space vectors. i)
Increasing the number of phases of SRM(Multi-phase SRM)
Introducing discrete space vector modulation techniques to SRM.
The multi-phase drives divide the controlled power on more inverter legs. The increased phase number reduces the current stress on each switch, lowers amplitude and increases the frequency of torque pulsations. It further reduces the rotor harmonic currents and increases the voltage per phase thus increasing the reliability of motor drive. Also, the increase in phase increases the torque per RMS ampere ratio for the machine of same volume. Due to increased switching frequency, increased phase number leads to increased commutation torque ripple frequency, thus making its filter easier. The second technique, Discrete Space Vector Modulation (DSVM) is based on the concept of synthesizing a higher number of voltage space vectors than which is used in classical DTC technique. This can be made possible by dividing the sampling time into N equal intervals and applying various voltage vectors in each of the sampling time. By doing this, many new equivalent voltage vectors can be synthesized. The more the voltage vectors, the more convenient it is to select the voltage vectors according to various speeds, which reduces the ripples of the torque and flux. 1.3
LITERATURE REVIEW Switched reluctance motors possess several advantages over other
motors, including high power density, high torque per ampere, high efficiency, simple stator and rotor structures, low cost and high reliability.
Their rotors do not generate significant heat, resulting in easy cooling. Their unidirectional flux and current yield low core losses and possibly simple converter design. Their independent torque generation of each phase, together with the series connection of the stator winding with the power semiconductor devices, further improves the reliability of SRM drives. In addition, they allow protection and speed control functions to be integrated in their control systems. Therefore, they offer a viable option for variable speed electrical drives, particularly for applications with high temperature and hostile operating environments. The invention of vector control in the beginning of 1970’s and the demonstration that an induction motor can be controlled like separately excited DC motor, brought a renaissance in the high performance control of AC drives. Researchers were attracted towards implementation of this vector control to both induction and synchronous drive motors and a huge number of papers were published on vector control and DTC. Later to overcome the major drawback of DTC technique, which is torque ripple tuning of the DTC to minimize the ripples, were done. With the development of variable reluctance machine and power electronic components, these techniques were implemented to the special machines. One such special machine, the switched reluctance motor attracted researchers as they were competitor replacement for induction machine and had many advantages of other machines like Permanent Magnet Synchronous Motor (PMSM) and Brushless DC motors (BLDC). The literature review in Figure1.1 clearly shows the research works carried in this area and for SRM. It can be seen that the control technique for SRM and torque ripple minimization are still to be concentrated to put the machine in market for wide spread application.
1. Vector control of induction motor drive 2. Direct torque control of induction motor drive 3. Torque ripple minimization in DTC for induction motor drive 4. Direct torque control for SRM drive 5. Torque ripple minimization for SRM drive 6. Torque ripple minimization in DTC for SRM drive
140 120 100
80 60 38
Figure 1.1 Literature Review 1.3.1
Control Techniques for SRM The switched reluctance motor has certain drawbacks due to the
motor’s doubly salient structure as well as the highly nonuniform torque output and magnetization characteristics. The double salient structure leads to the inability of exciting the motor using conventional AC motor waveforms and thus the well established AC motor rotating field theory cannot be applied to the motor. Furthermore, due to the motor’s nonuniform torque output characteristics, a high torque ripple is inherent in the motor unless a torque ripple reduction strategy is employed. Additionally, the highly nonlinear magnetization characteristics of the motor entail that the control of the motor is complex. This is further complicated by the interaction due to mutual coupling of different motor phases and parameter variation of the inductance characteristics.
Hence, the problem of controlling the SRM has been an ongoing area of research. The previous control techniques have fallen into two main categories: those which used a linear motor model and those which used a nonlinear model which takes into account the motor saturation. In general, most techniques generate a voltage or current command profile in order to control the motor torque, speed, position and/or minimize the torque ripple. Control schemes that used linear motor models include the method proposed by Taylor (1991). This scheme used an adaptive feedback linearising controller for the switched reluctance motor which assumed a linear magnetic circuit. In another scheme (Nagel and Lorenz 1999), an analytical solution for production of motor voltages to provide a smooth torque was proposed. This solution was based on the linear torque characteristics of the motor.
Matsui et al (1991) has assumed that the
inductance profile of the motor can be expressed as a simple sinusoidal function that only varies with position. Although these linear schemes have the advantage of a simplified and real-time tractable control scheme, disadvantages arise because of the highly simplified models of the motor which are inaccurate in most practical motor drives. This is due to the fact that the motor has highly nonlinear magnetization characteristics under normal operating conditions and the flux linkage (and hence inductance) is a nonlinear functions of both current and position. Other developed schemes have taken into account the nonlinear characteristics of the SRM. Ilic’-Spong et al (1987) proposed a feedback linearising control scheme that provides compensation of the magnetic nonlinearities of the motor. This method has the advantage of decoupling the effect of the phase currents in the torque production. However the method
requires exact knowledge of the motor parameters and is thus quite impractical. In addition, other nonlinear methods include those which generate current profiles for the torque sharing between two motor phases in order to produce a smooth torque (Stankovic et al 1999). However these schemes limit the excitation to two phases at a time and also require complex waveform optimization computations which cannot be implemented in realtime. Also, Vedagarbha et al (1997) proposed sophisticated nonlinear adaptive control schemes for SRM. The major disadvantages of these methods are that they are complex and computationally expensive, which makes real-time implementation difficult. 1.3.2
Direct Torque Control In order to overcome the above problems, methods of SRM control
have been proposed which used the philosophy of direct torque control of conventional AC machines. DTC is a well established control principle for AC motor drives which was proposed by Takahashi and Noguchi (1997). It has been shown to provide a simple solution to control the torque and speed of the motor and to minimize torque ripple. Previously DTC has been almost exclusively applied to AC motors which have linear characteristics and three phase balanced sinusoidal excitation. The SRM however, has a nonlinear model and nonsinusoidal excitation. Furthermore, the excitation is normally not balanced between phases, as the phases are independently excited. Therefore, the SRM has not been seen as conducive to the conventional AC machine DTC theory. But, an application of DTC to the SRM has recently been described by Jinupun and Luk (1998). This scheme used the concept of a short flux pattern that links two separate poles of the stator (Michaelides and Pollock 1993). In order to achieve this flux pattern, a new type of winding
configuration was proposed. This new winding technique limits the length of two individual short flux loops that connects two stator poles. The scheme produces a rotating magnetic flux field by continuously changing the proportion of currents between the phase windings in sequence. This produced a rotating field similar to that seen in the conventional AC motor. The major disadvantage of this scheme is that a new motor winding topology is required. Altering the motor winding configuration is both expensive and inconvenient. Although the DTC scheme explained by Jinupun and Luk (1998) may be used theoretically with a conventional motor phase winding, but to achieve this practically bipolar currents (as opposed to the normal unipolar currents seen in SRM drives) are required. This would entail more complex motor drive hardware and software with double the power switches and control signals per phase. Hence, applying the proposed scheme to a conventional SRM, leads to the added expense and inconvenience of a bipolar current drive. Adrian David Cheok and Yusuke Fukuda (2002) proposed a novel DTC methodology for the SRM which is derived from the analysis of nonuniform torque characteristics of the motor. The new method does not involve short flux patterns, a change of the motor winding configuration, or the use of a bipolar current drive. Thus, the scheme can be conveniently implemented on any normal type of SRM drive. This scheme provides the advantages of the DTC method to the SRM drive. In the scheme, torque is directly controlled through the control of the magnitude of the flux linkage and the change in speed (acceleration or deceleration) of the stator flux vector. Furthermore no model calculation is required and thus the scheme is not dependent on the accuracy of the estimated model parameters. Hence, this overcomes the disadvantages and difficulties faced by conventional linear or
nonlinear controllers of the SRM mentioned above, due to the highly nonlinear characteristics of the motor. Panda et al (2005) proposed Sliding Mode Controller (SMC) for DTC. Normally it is difficult to develop an accurate non-linear model for SRM as the torque controller should be robust to model inaccuracies. It is well known that SMC provides robustness to such model uncertainty. By exploiting this property of SMC, a simplified trapezoidal phase inductance profile is used to obtain equivalent control. For robustness, a saturated switching control variable structure controller is added to the equivalent control. The resulting controller keeps torque tracking error within a narrow band near zero. Direct torque control avoids the conversion of demanded motor torque into equivalent phase current references; which is non-trivial due to the coupled and non-linear torque, current and rotor position relationship. Panda et al (2004) proposed a controller using a sliding mode controller combined with Iterative Learning Controller (ILC) for accurate torque tracking. ILC is effective in minimizing torque ripples during steady state operations. Due to the finite time required for ILC learning, there will be performance degradation during transient periods. Manabu Mitani et al (2006) proposed position sensorless direct torque control. As in DTC, expensive position sensor such as rotary encoder is needed, which causes some serious problems in terms of reliability, size, and cost. Thus, it is very important in the motors to accomplish position sensorless driving objective. Jeong et al (2006) proposed that the sensor-less control is based on direct torque control with Direct Flux Compensation (DFC) for 4 phase 8/6 switched reluctance motor drive.
Torque Ripple Minimization in SRM Drive The primary disadvantage of an SRM is the higher torque ripple
when compared to conventional machines, which contributes to acoustic noise and vibration. The origin of torque pulsations in an SRM is due to the highly nonlinear and discrete nature of torque production mechanism. The total torque in an SRM is the sum of torques generated by each of the stator phases, which are controlled independently. Torque pulsations are the most significant at the commutation instants when torque production mechanism is being transferred from one active phase to another. The minimization of the torque ripple is essential in high performance servo applications, which require smooth operation with minimum torque pulsations. The excellent positive features of an SRM can be utilized in a servo system by developing techniques of reducing the torque ripple. These types of drives have extensive applications in automotive industries, direct drive machine tools. There are essentially two primary approaches for reducing the torque pulsations: One method is to improve the magnetic design of the motor, while the other is to use sophisticated electronic control. Machine designers are able to reduce the torque pulsations by changing the stator and rotor pole structures, but only at the expense of the performance of the motor. The electronic approach is based on selecting an optimum combination of the operating parameters, which include the supply voltage, turn on, and turn-off angles, current level and the shaft load. It must be noted that the minimization of torque ripple does lead to a reduction of the average torque, since the capabilities of the motor are not being fully utilized at every rotor position. In general, it can be stated that torque maximization and ripple minimization cannot be achieved simultaneously by electronic control.
Praveen Vijayraghavan and Krishnan (1999) described the different sources of acoustic noise. The source of noise is mainly classified into four categories that are magnetic, mechanical, aerodynamic and electronic. Also acoustic noise mitigation and cancellation are mentioned. A number of electronic noise-reduction methods for SRM’s have been summarized in (Krishnan and Vijayraghavan 1998). Cameron et al. (1989) demonstrated that the resonant vibrations of the stator are the dominant source of acoustic noise in an SRM. These vibrations are caused by radial magnetic force, which act to decrease the gap separation between the rotor and stator as their poles approach alignment. The torque ripple minimization method introduced initially (Wu and Pollock 1993) is based on the optimum profiling of the phase currents during an extended overlapping conduction period of two phases. This method of control is on an instantaneous basis instead of the conventional time-averaged torque control. The instantaneous control reduces the time of response in addition to minimizing the torque ripple. Several efforts to reduce or eliminate the torque ripple of switched reluctance motor were presented in (Lechenadec et al 1994).Some modulated phase current shape to counteract the torque ripple. But this technique requires a special motor geometry and pole shape design. A balanced commutator which works for accurate current tracking to reduce torque pulsations was also reported by Wallace et al 1992. 1.3.4
Torque Ripple Minimization in Direct Torque Control Even though DTC is getting more popular, it also has some
drawbacks, such as the torque and flux ripple. Many researchers have paid attention to this problem by now and proposed solutions for it.
However, the DTC strategy using a switching table has some drawbacks. First, switching frequency varies according to the motor speed and the hysteresis bands of torque and flux. Second, a large torque ripple is generated. So a new DTC technique is proposed by Jun-Koo Kang and Seung-Ki Sul (1999) which minimizes torque ripple by keeping constant switching frequency. Output voltage vector Vs is selected using the conventional DTC switching table, but the pulse duration of Vs is determined by the torque-ripple minimum condition. The instantaneous torque variation of the motor can be expressed as a function of the applied voltage vector Vs. Then, the rms torque-ripple equation during one switching period can be obtained from the instantaneous torque variation equations. The optimal switching instant ts (i.e., the on duration Vs) are determined by taking a partial derivative of the rms torque ripple equation with respect to ts. Joon Hyoung et al (2003) proposed a unified flux and torque control to reduce flux and torque ripple in which a voltage space vector is calculated for a deadbeat action and a minimum-distance vector selection scheme replaces the drops. The calculation algorithm is greatly simplified to generate the deadbeat control voltage in a very simple form. The minimum-distance vector selection makes the flux and torque error nearly minimum over a fixed sampling period without additional modulation. Direct torque control, selection of each voltage vector is based on the measured parameters in the beginning of the sampling period. Due to the fast dynamic of torque, this delay causes an extra torque ripple. So Kaboli et al (2004) proposed a new predictive controller for compensating this delay and reducing the torque ripple. The stator current is measured in the beginning of each sampling period and its expected value in the end of period is predicted according to an extrapolation algorithm. This algorithm is performed by calculating the slope of stator current variation according to the
stator flux variations. Thus, there is no need to solve the higher order equations of machine. The calculation of torque is performed using the predicted value of stator current. Therefore, the selection of voltage vector is more realistic and it prevents extra torque ripple. To increase the number of vectors to be applied to the machine, Casadei et al (2000) proposed the discrete space vector modulation. Takahashi et al (1989) proposed a double three phase inverter and Mei (1999) had used variable switching sectors to minimize the torque and flux ripple. However, the common problem of these methods is that they cannot work at zero-error state, i.e. these DTC algorithm cannot work properly if the torque error or flux error is zero. Zero state is not a steady state under the basic DTC. Therefore, if we can reduce the steady state error to zero, the steady state performance should be improved. To solve this problem Lixin Tang and Rahman (2001) proposed a new technique with Space Vector Modulation (SVM). Normally the inverter voltage vectors in classical DTC are determined by the errors in the electric torque and stator flux. Large and small errors are not distinguished i.e. the switching vectors chosen for large errors are the same as the switching vectors chosen for normal operation, when these errors are small. In order to overcome this problem, SVM-DTC and DSVMDTC methodologies were proposed (Casadei et al 1996). However, the switching frequency still changes. In the SVM DTC scheme a predictive flux controller is used with the SVM. It can work well with low sampling frequency and only one torque controller is used, which is better than the previous schemes. However, it seems that the torque response under this scheme is slower than its counterparts.
Pengcheng Zhu et al (2003) proposed duty ratio modulation technique to reduce the torque and flux ripple. We all know that the output voltage in DTC contains much harmonics for applying hysteresis band in this kind of system. And it is found that the torque ripple can be divided into two parts where one is related only to the motor parameter and the other is related to not only the applied voltage but also to the rotor angular velocity. Based on the analysis, a duty ratio modulation method is proposed that only sends the voltage vector in part of the sampling period, while in the rest time of the sampling period zero voltage vectors is sent. The duty ratio modulation algorithm, which is induced from the torque ripple analysis, allocates the sampling period. Hassan Halleh et al (2008) presented an improved Direct Torque Control (DTC) based on fuzzy logic technique. The major problem that is usually associated with DTC drive is the high torque ripple. To overcome this problem a torque hysteresis band with variable amplitude is proposed based on fuzzy logic. The proposed fuzzy controller is shown to be able to reduce the torque and flux ripples. It also improves the performance of DTC especially at low speed.
Grzesiak et al (2007) devoted the field of artificial intelligence for drive control. Learning of the neural controller was set on-line. Starting from a random configuration of the speed controller, the network adapts its weights according to an error criterion. Although the use of such specialized controller allows potential adaptive and robust control skills, tuning of an ANN for online learning control is a long iterative procedure. Also, optimization of the neural controller induces determination of ten parameters acting critically on the control dynamics. He used Genetic Algorithm (GA) which is inspired by genetic processes leading human race toward optimal individuals capable of controlling their environment.
Xiying Ding (2007) presents
voltage space vector modulation
technology is used for fuzzy direct torque control of AC motor with fuzzy stator resistance estimator, which can weaken the ripple of torque and flux of fuzzy direct torque control system, therefore the control system get swifter response velocity, stronger robustness and higher precision of velocity control. Bo Zhou and Xiao Fei Jing (2008) investigates a new Direct Torque Control (DTC) Strategy for induction machine based on Particle Swarm Optimization(PSO).The PSO was used to adjust the parameters of PID controller. The scheme improves the adjective capability of PID controller. The new direct torque controlled strategy is considered in detail and it is shown that the use of PSO-PID controller decrease the electromagnetic torque ripple and increase system rapidity and stability. 1.4
ORGANIZATION OF THE THESIS This thesis is composed of seven chapters. The overall organization
of rest of the chapters is as follows Chapter 2 describes the construction, working principle of switched reluctance motor and different types of converter topology like two transistors per phase, C dump converter topology. The design and simulation of a new converter topology and the corresponding results is also presented in this chapter. Chapter 3 deals with classical control techniques for AC drives like scalar control, vector control and direct torque control. It also outlines the advantages and disadvantages of those controller techniques. Chapter 4 exposes the different control techniques used in switched reluctance motor drives. This chapter explains the current control and
implementation of direct torque control to SRM drive. The simulation diagram and results are incorporated in this chapter. Chapter 5 evaluates the different types of torque ripple minimization techniques like three level inverter, unified torque and flux control, duty cycle control, space vector modulation and discrete space modulation techniques. The concept of duty cycle control, space vector modulation and discrete space vector modulation are implemented and results are produced in this chapter. Chapter 6 deals with the implementation of direct torque control to the five phase switched reluctance motor drive for the reduction of torque ripple. The simulation results are produced to validate the concept of torque and flux ripples minimization. Chapter 7 Concludes the main work of torque ripple reduction using multiple space vectors with the validation of results. It also explains the application of SRM and gives the future scope of the research in this area. 1.5
SUMMARY A detailed literature review is carried out for problem identification.
It also gives the details of thesis organization. The next chapter will give the details about SRM drives.