DSP tabanlı bir sistem ile anahtarlamalı relüktans motorunun kontrolu
DSP based control of switched reluctance motor
- Tez No: 39798
- Danışmanlar: PROF.DR. R. NEJAT TUNCAY
- Tez Türü: Yüksek Lisans
- Konular: Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 1994
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 121
Özet
independently. Parallel operation solves the problem from performing operation in different units in the same time. These features makes the computational speed much faster and improves architecture for higher throughput. Accuracy issues is the main part of a processing unit. So to provide better accuracy, the architecture can be designed for avoid numerical errors. The shared memory for both instructions and data and single instruction data bus is significant bottleneck and will cause errors in real time appplications. Architecture, accuracy and speed interact, so an efficient and optimized control solution can only be find with taken care of this issues. The fundemental primitive operation in numerical computation is multiply accumulate step and it is desired to do this in a single clock cycle. DSP can do it with pipelining and parallel operation speed processing by doing more in fewer cycles. The desirable features of a DSP is bus width, expanded arithmetic, data-address generation, instruction sequencing and scaling.To minimize the effect of numerical operations the units must have additional guard bits, bus width must large enough to operate correctly with double precision option. The basic units of DSP system is as follows : Multiplier-accumulator (MAC), Arithmetic logic unit (ALU), Barrel shifter (BS), Data address generator (DAG), Instruction sequencer (SEQ). These units perform rapid execution of arithmetic and logic operations and provides protection against errors with diffirent solutions. All these units can operate in parallel and connect to buses which gives the result in a single cycle. The main disadvantage of DSP is the memory area. For providing speed only limited memory area is put on chip. Thus, large amount of programs can not been stored into memory. There are several differences between microcomputer and DSP systems. Both system has its own advantages and disadvantages. The designer must learn the options of controlled system to make the right decision between DSP and microcomputer. Speed, memory area, software support, peripheral devices and cost will be the main points to select which processor must be chosen. In applications, especially using a sophisticated control algorithm, math operations makes longer the execution time. Microcomputers use look up table to decrease the execution time. But DSP is produced to opareta at this conditions, so it is better to use numerical methods to solve a function. However, there is no enough memory to create look up tables and store program. This is just a sample how DSP and microcomputer solve main problems. Switched reluctance motor (SRM) is one of the alternative motors to asynchronous and DC motors. SRM is a motor which both stator and rotor is salient pole. There is only winding on stator to create flux, no winding IX
Özet (Çeviri)
SUMMARY DSP BASED CONTROL OF SVVITCHED RELUCTANCE MOTOR in recent years, the revolution on the electronics, brings the new devices into the area of control. Opportunities have grown rapidly över past years to solve the problems of control applications. Ali the control systems of a few years were designed by a general purpose microcomputer. But, VLSI design technology brings DSP processor to digital signal processing area with a powerful architectural design. High performance of DSP make it possible to implement a wide range of digital control algorithms for sophisicated applications. A control system is built by a controller and the circuits which provide to connect these parts together. The control system must achieve a desired output to command a process. in the control applications of power electronics and motors, the system is preferred to be a closed loop. The closed loop system takes the reference value and feedback value which came from the sensors and actuators. The controller must compute a desired response from the entire system. The controller can be either analog ör digital. The function is to ensure the system to be stable in ali conditions and respond as quick as possible. Analog controller can be produced with operational amplifiers and passive circuit elements. Analog contollers continiously process a signal and can be used for very high bandvvidth systems. The disadvantages of this controller is hard wired design, suffering from temperature drift and component aging. The system is flexible only for simple design approachs. With the new control algorithms like adaptive ör fuzzy control, it is impossible to design a controller using analog devices. The high performance and reliability of microcomputers put them directly to the control applications. The programmable solution and flexibilty is the main advantages of the digital controller. it is easy to make upgrade and maintain design investment. To control nonlinear systems with sophisticated techniques from modem control algorithm is possible. Digital controller is not affected by enviromental conditions and shows precise behavior. The main problem is that, the processes in nature is analog. So to control these processes with a digital controller, the signals must convert to a kind that both parts understand each other. The digital controller require A/D (analog to digital) and D/A (digital to analog) converters to obtain control of real world data. The accuracy of the viapproximation determines the performance of the digital controller. The conversion betvveen two system can provide by sampling the signal at dicrete time intervals. This limits the bandwidth that can be handled by the controller. There will be always an error because of the loss of information at the conversion cycle. To avoid this error the resolution ör word lentgh of the processor must be increased. Digital controller is built onto a general purpose processor and peripheral devices to make the system flexible for usage in different applications. The choice of the processor is the main problem to determine the behavior of control system. Some tradeoffs have to be made in selecting a processor. The poor performance of digital system is caused by the wrong selection of the processor. in practical area, speed of execution, accuracy and cost determine the choice of processor. Architecture, performance and peripheral integration is the factors which must be considered. Processing signals in real time raises numerical problems, and to ensure the performance of the controller acceptable, they must resolve these numerical problems correctly. The architecture of the processor must provide ali the requirements of the controlled system. Because of the inadequate processor architecture, quantization, truncation, limit cycles and overflow handling errors appear, and several techniques developed to preset these errors. Ûuantization effect is result by representing a signal in discrete time. The analog signals are continious so to compute these signals quantized magnitude levels must be used. The system must support a large word lentgh and scaling shifters to provide the resolution and dynamic range needed. If the gain coefficients and signals are time varying and have large dynamic range, floating point processors must be used. Scaling can be used to minimize the effects of quantization. The arithmetic operations alvvays produce an error because of the finite word length. The processor takes the value, compute the operation and then throw away bits when necessary. Because of the finite amount of storage length, the processor can not produce a result using ali bits, so it made an approach to make the result understandable to computational units. Rounding and truncation is two ways to do this. If the processor compute 16*16 bit multiplication, the result will be 32 bit length. To store and send the result, needs 32 bit bus length and registers. If only 16 bit is avaliable, least meaningfull 16 bits will be thrown away. Truncation will throw away before storage, and rounding will round and then throw away before store the result. Rounding makes the result closerto the real value, so it is better than truncation. There are several ways to minimize the effects of rounding and truncation. Limit cycles can produced from viiquantization and truncation. Insufficent resolution of the output causes to oscillate around the actual value without being able to reach it. The control system can use fractional arithmetic or integer arithmetic. The binary arithmetic is used for represantion of numbers. Two s complement is the better alternate which is used by digital signal processing elements. Fraction numbers is better when the algorithm is multiplication based. If the highest number is 0.99, it is impossible to occur an overflow. But addition calculation can cause overflow. The registers which are effected by overflow, wrap around and change magnitude from most positive to most negative. The processor must take care of with these registers to avoid wrong control behaviour. The saturation mode is provided to prevent the content of registers to wrap around and change signal when an overflow occurs. This is done automaticallly in DSP. The overflow can be avoided with increasing the register size, but it is better solution to put some guard bits to protect the result from overflow. To optimize the dynamic range, data arrays are kept scaled so that the largest value fill all the bits of the data word. Precision and dynamic range of numbers must be balanced to find an optimum solution. The processor must finish processing the signal before the arrival of the next sample or the information will be lost. Also peripheral integration is important from a system cost, ease of design and interface. DSP is the digital signal processor which provide new features to control system applications. DSP is designed by using Harvard architecture, while the microprocessor is designed by Von Neumann architecture. The most important part, the units are independent so can compute same time in parallel. The second main difference is, the memory of data and program is independent from each other so to compute and to execute can be done in single cycle. All functions are performed internally in hard wired logic so that it takes a single cycle to execute most of the mathematical operations. The hardware multiplier can multiply the contents in a single instruction cycle thus the speed is excellent comparing with microprocessors. The multiple bus architecture provides better computing speed performance. High sampling rates can be provided so the precision improve better. DSP gives the real time application flexibility and performance which is predictable, accurate, repeatable and adaptable. Computational speed matched from unit to unit so no bottlenecks develop. To remove bottlenecks pipelining and parallel operation can be used. Bottleneck is a computational delay produced because of the data can not use the buses and units efficiently. In pipelining, the operations broken into subunits and these subunits compute part of operation VIIIindependently. Parallel operation solves the problem from performing operation in different units in the same time. These features makes the computational speed much faster and improves architecture for higher throughput. Accuracy issues is the main part of a processing unit. So to provide better accuracy, the architecture can be designed for avoid numerical errors. The shared memory for both instructions and data and single instruction data bus is significant bottleneck and will cause errors in real time appplications. Architecture, accuracy and speed interact, so an efficient and optimized control solution can only be find with taken care of this issues. The fundemental primitive operation in numerical computation is multiply accumulate step and it is desired to do this in a single clock cycle. DSP can do it with pipelining and parallel operation speed processing by doing more in fewer cycles. The desirable features of a DSP is bus width, expanded arithmetic, data-address generation, instruction sequencing and scaling.To minimize the effect of numerical operations the units must have additional guard bits, bus width must large enough to operate correctly with double precision option. The basic units of DSP system is as follows : Multiplier-accumulator (MAC), Arithmetic logic unit (ALU), Barrel shifter (BS), Data address generator (DAG), Instruction sequencer (SEQ). These units perform rapid execution of arithmetic and logic operations and provides protection against errors with diffirent solutions. All these units can operate in parallel and connect to buses which gives the result in a single cycle. The main disadvantage of DSP is the memory area. For providing speed only limited memory area is put on chip. Thus, large amount of programs can not been stored into memory. There are several differences between microcomputer and DSP systems. Both system has its own advantages and disadvantages. The designer must learn the options of controlled system to make the right decision between DSP and microcomputer. Speed, memory area, software support, peripheral devices and cost will be the main points to select which processor must be chosen. In applications, especially using a sophisticated control algorithm, math operations makes longer the execution time. Microcomputers use look up table to decrease the execution time. But DSP is produced to opareta at this conditions, so it is better to use numerical methods to solve a function. However, there is no enough memory to create look up tables and store program. This is just a sample how DSP and microcomputer solve main problems. Switched reluctance motor (SRM) is one of the alternative motors to asynchronous and DC motors. SRM is a motor which both stator and rotor is salient pole. There is only winding on stator to create flux, no winding IXon rotor. The system must operate normally in closed loop control, with shaft position sensors and encoders. it is necessary to know which stator phase will excited in a time period. The commutation of the phase currents with rotor position must be in synchronize. The motor can be 8/6, 6/4, 12/8 ör another stator rotor pole numbers with different phase numbers. The torque is produced by moving rotor to position where the inductance of excited winding is maximized. The aligned position of motor defined as, stator and rotor poles is aligned exactly, and airgap is minimum. The magnetic reluctance of flux path is in lowest, so inductance is maximum. The control of the phase currents gives the best performance. The requirements of control is to produce smooth torque with minimum ripples, good stability and smooth reversing capability. Different kind of power electronic circuits is defined to take maximum performance. The control of system must be in real time, because torque, current, speed and firing angles reacts as nonlinear function and vary with load conditions. in this thesis, the survey of digital controller design has been made. The microcomputer and DSP systems are introduced by practical point of view, architecture, main units, differences and applications of the systems are explained. The switched reluctance motor control design is explained with power electronic circuit, gate drive circuit and controller circuit realized with microcomputer and DSP, respectively.
Benzer Tezler
- Sliding mode yaklaşımı ile anahtarlamalı relüktans motorun hız kontrolü
Speed control of switched reluctance motor by using sliding mode approach
İBRAHİM SEFA
Doktora
Türkçe
1997
Elektrik ve Elektronik MühendisliğiErciyes ÜniversitesiElektronik Mühendisliği Ana Bilim Dalı
YRD. DOÇ. DR. ŞABAN ÖZER
- Anahtarlamalı relüktans motorun bulanık mantık tabanlı modellenmesi ve kontrolu
Modelling and control of a switched reluctance motor based on fuzzy logic
ÖMER FARUK BAY
- Anahtarlamalı relüktans motorun genetik bulanık mantık denetleyici ile nonlineer hız ve konum denetimi
Nonlinear speed and position control of switched reluctance motor with genetic fuzzy logic controller
İLKER YILDIZ
Doktora
Türkçe
2013
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve KontrolMarmara ÜniversitesiElektronik-Bilgisayar Eğitimi Ana Bilim Dalı
PROF. DR. AHMET FEVZİ BABA
DOÇ. DR. OĞUZ ÜSTÜN
- 3 fazlı aa servo motorun akıllı yöntemlerle gerçek zamanlı yörünge kontrolü
Real time trajectory control of 3 phase aa servo motor based on intelligent methods
KÜBRA BALTACI
Doktora
Türkçe
2024
Elektrik ve Elektronik MühendisliğiBursa Teknik ÜniversitesiElektrik-Elektronik Mühendisliği Ana Bilim Dalı
DOÇ. DR. DAVUT ERTEKİN
DOÇ. DR. GÖKAY BAYRAK
- Akım sensörsüz DSP tabanlı PFC devre tasarımı ve gerçekleştirilmesi
Design and implementation of a current sensorless DSP based PFC circuit
NAZLIGÜL SERTLER
Doktora
Türkçe
2008
Elektrik ve Elektronik MühendisliğiGazi ÜniversitesiElektrik Eğitimi Ana Bilim Dalı
YRD. DOÇ. DR. İBRAHİM SEFA