Abstract: Paper briefs about electrical drives and its mode-
rn system. The basic conceptual principles of operation of vector-controlled AC drives or field oriented controllers a- re presented and the direct and in direct implementations are briefly described. The key performance capability of torque control is emphasized, and mention is made of the difficulties in achieving the desired performance arising f- rom feedback sensing and motor parameter variation. So- me applications such as helper drives, dynamometer drives and propulsion drives are cited as examples where vector controlled AC drives provide superior performance. I ndex terms-Electrical drives, coupling effect, vector contr- ol, field orientation
I. INTRODUCTION As the capabilities of AC motor drives have expanded, a new category of drives has emerged as a practical rea- lity. This type of drive is distinguished by the strategy of the control rather than differences in the power circu- it. These AC motor drives are called vector-controlled drives or field oriented controllers. The purpose of this paper is to explain what a vector-controlled AC drive is and how it works. Some applications requiring the sup- erior capability of the vector-controlled drive will be discussed. The terms "field-oriented" and "vector - controlled" arise from the behavior of the magnetic fields in the machine. The mechanism of torque production in electric machines has been well known for many years. Only recently (Blaschke, 1972) have these concepts been extended to form the basis of a strategy for control of AC machines with power conversion apparatus. It is important to recognize from the outset that the key output of any electric motor is torque, not speed, although in the majority of cases speed is the controlled variable of interest. The AC machine is essentially an energy converter in which current or voltage, frequency and speed are inputs and torque is the output. Note that the motor has influence on the speed of the load, but other factors like load torque and moment of inertia are equally important. The fundamental property of vector controlled AC drives is that they control the output torque of the machine accurately and quickly, in contrast to ordinary AC drives. II. ELECTRICAL DRIVES Drives are employed for systems that require motion control e.g. transportation system, fans, robots, pumps, machine tools, etc. Prime movers are required in drive systems to provide the movement or motion and energy that is used to provide the motion can come from various sources: diesel engines, petrol engines, hydraulic motors, electric motors etc. Drives that use electric motors as the prime movers are known as electrical drives. Types of electrical drives: 1) DC Drives: A DC drive is a device that is used to control the speed of an electrical motor, either a permanent magnet or a wound motor. 2) AC Drives: An AC drive is a device that is used to control the speed of an electrical motor, either an induction motor or a synchronous motor. A.CONTROLLING OF AC DRIVES 1. Scalar control is based on varying two parameters simultaneously (frequency and voltage). Scalar control of ac drives produces good steady state performance but poor dynamic response. This manifests itself in the deviation of air gap flux linkages from their set values. This variation occurs in both magnitude and phase. 2. Vector control (or field oriented control) offers more precise control of ac motors compared to scalar control. They are therefore used in high performance drives where oscillations in air gap flux linkages are intolerable, e.g. robotic actuators, centrifuges, servos, etc. In scalar control there is an inherent coupling effect because both torque and flux are functions of voltage or current and frequency. This is prone to instability because of 5 th order harmonics. Vector control decouples these effects. III. FIELDS IN MACHINES: In order to understand the concept of vector control, in which the torque output of the motor is controlled, we
need to review how torque is produced. In any electric machine, torque is produced by the interaction of flux and MMF. These magnetic fields are generated by current-carrying coils in the structure. These fields have vector properties: they have both a magnitude and a direction or orientation. (Both magnitude and direction are, in general, time varying.) Torque is proportional to the product of flux and MMF times the sine of the angle between them. (This can be elegantly expressed as the vector cross- product of flux and MMF vectors.) Since the torque depends on the relative field orientation, as well as magnitude, a basic problem of rotating machines is that of preserving optimum orientation when one of coils is rotating.
B. FIELD ORIENTATION IN DC MACHINES: The DC machine solves this problem with commutator, and two separate externally accessible windings. The field winding establishes a flux field with a fixed direction with respect to the stator. The armature windings are connected to the outside world via the commutator so that armature current generates a stationary MMF which is perpendicular (sin = 1) to the flux. As the machine rotates, the commutator automatically selects the optimally-oriented armature coils. Thus, one could think of the commutator as the field oriented controller of a DC machine. It controls the direction of current in the armature coils on the basis of armature position. See Fig. 4. Unfortunately, the commutator is the main drawback of the DC machine in that it is an expensive structure which requires high maintenance, as well as producing sparks in normal operation. This is conceptually what a vector controlled drive does in the case of an AC machine; that is, it controls the angle between stator MMF and flux. A further structural property of the DC machine is that changes in the current in one winding have little or no effect on the magnetic field created by the other winding. Thus flux and MMF magnitudes are separately controllable and do not interact; they are decoupled. Decoupling is another goal of the vector-controlled AC drive. We must determine how to control the stator current input in order to have independent control of the machine torque and flux. In a cage induction machine, only one set of windings is accessible, those on the stator. Thus, the electrical inputs to those windings must control MMF, flux, and their orientation. The basic property of AC machines utilized is that periodic time varying currents in polyphase windings produce rotating MMF's in the machine. So by judicious use of the power switches in the inverter to control stator current I s we can control the stator MMF amplitude and position. Now if our controller "knows", by calculation and measurement, the position and magnitude of the flux, we can decide how to control the stator MMF by applying the following principles:
1. If the flux is less than we want, increase in stator MMF component parallel to flux: Increase I s cos by increasing I s or decreasing or some combination thereof. Conversely if the flux is too large, decrease the component of stator MMF parallel to flux. See Fig5. 2. If we want to produce zero torque, make the stator MMF exactly parallel to the flux (sin= 0), and I S =I m (magnetizing current). 3. In order to produce torque, make the MMF lead or lag the flux in space. The direction of torque depends on the sign of . Increase torque by increasing both and I s . The vector control tracks the flux position and makes the stator MMF follow it around with an angular
separation which depends on how much torque is to be produced. Since the stator MMF must also maintain the flux, there is always a component of I s parallel to the flux. Generally speaking, the flux will be rotating so that the controller must make the stator MMF rotate in order to keep a constant orientation. If we assume constant flux operation, then a certain constant magnetizing current component I M will be required to maintain flux. We can say: I M = I S cos= Constant,I S sin torque This leads to the relationship between stator current, torque angle and torque for constant flux operation. I s 2 = I M 2 + kT where T = desired torque = arc tan (kT/I M ) See Fig. 6 By maintaining these relationships, the control achieves decoupling of flux and torque.
Although this is conceptually simple, the controller has to take flux amplitude measurements, convert into vector form, and perform fairly complex calculations to arrive at and I S . These must then be converted into individual currents per phase. The vector controlled drive requires hardware or software capable of doing complex calculations rather quickly.
IMPLEMENTATIONS OF VECTOR CONTROL: Because the amplitude and orientation of stator MMF is directly controlled by a current-fed inverter, these are particularly easy to equip with field oriented controllers. The amplitude of the DC link current is directly proportional to I S and the choice of output thyristor pair determines which of 6 directions in space the MMF may have. The inherent regenerative ability of current-fed inverters is a substantial advantage, because true field oriented controllers are designed to operate in all quadrants. It is possible to design vector controls for voltage fed drives, but this is more complex because three current regulators are necessary to establish precise control over stator MMF. Furthermore, in order to utilize the 4 quadrant capability, a line side converter capable of bidirectional current flow is required. This usually means another set of switches, for example, as in a dual thyristor converter. Although direct flux sensing control schemes are the most immune to complications arising from misadjustment and machine parameter change due to temperature, the practical difficulties of air-gap flux sensing are formidable. Therefore almost all vector controlled drives are of the indirect type, where flux position is inferred from rotor position or derived from highly precise calculations based on stator current and voltage. In those vector controls based on rotor position feedback, advantage is taken of the fact that the flux motion with respect to the rotor is very slow compared to the mechanical speed of the rotor, owing to the shorted windings on the rotor. The flux motion is computed as the sum of the rotor angular velocity plus a small component (slip) proportional to the product of torque and rotor resistance. Means must be provided to compensate for the change in rotor resistance with temperature. Figure 7 shows the block diagram for a vector-controlled drive. The main drawback is the need for a rotor incremental position encoder, although this is a far less difficult matter than flux sensors in the air gap. Since low cost computing power has increased dramatically with the development of the microprocessor, there have been many proposed implementations of vector control based on the calculation of flux position from stator current and voltage. These schemes are based on the analysis of terminal voltage into two components, namely stator IR drop, and speed voltage. The flux reveals its amplitude and position as a result of its velocity relative to the stator windings which induces voltages in them. All these approaches suffer from the serious drawback that, at low speeds, one cannot distinguish IR drops from speed voltage, particularly since stator resistance varies with temperature. Consequently, these controllers perform increasingly
better as speed increases, but are in serious difficulty below 0.2 P.U. speed. They are also the most dependent on correct "tuning" which is essentiall programming in the motor parameters. Although we have implied that flux is maintained constant, a field oriented controller can manipulate the amplitude of flux as well as torque, so that operation at reduced flux in extended speed is perfectly permissible.
To summarize the special capabilities of a vector- controlled drive, we can state: 1. The ability to control torque output of the machine quickly and precisely."Quickly" means step response of 1 ms to 20 ms depending on drive power rating and type of switching devices. Some corollaries of this property are: a. Ability to operate as a torque follower. b. Acceleration and deceleration at c. torque limit d. Ability to yield speed on overload
2. True 4 quadrant operation: Dynamic reversing and the ability to produce torque in either direction while maintaining either direction of rotation (Assumes power circuit can handle continuous regeneration.) 3. The ability to start into a machine spinning in either direction. 4. The ability to control flux in the machine. Regardless of the technique used, a vector controlled drive is far more versatile than an ordinary frequency-programmed drive. As we know, in a frequency programmed drive, the external speed
reference controls the output frequency directly, although there are always rate limit circuits to prevent the output frequency from changing too fast. That type of drive does not have flux position information, and depends on predictable behavior of the load. As the output frequency increases, if there is a load, the stator field gradually pulls ahead of the flux, torque is developed, and current is drawn. If the load torque drops to zero, the rotor speeds up, allowing the flux to catch up to the stator field. The motor torque then falls to zero. Torque is not controlled; it builds up if the motor encounters a load torque to retard the motion of the rotor (and consequently the flux). If the load is a difficult type, it may not be possible to start or accelerate with a frequency programmed drive. Nevertheless, most loads like centrifugal pumps and fans can be handled with frequency programmed AC drives.
APPLICATIONS: What types of applications are difficult for ordinary AC drives where a vector-controlled drive may be put to good use? One obvious category is helper drives where one desires to drive a large machine with several motors connected to different parts. This is difficult for frequency programmed drives because they tend to behave like very stiff speed regulators. Once the stator frequency is established, the speed change from zero to full load is determined by the motor. Low slip types like NEMA B designs may have only 10 to 15 rpm of rated slip. On the other hand, with a vector-controlled drive, there is no dependence on the motor for load sharing. One drive is set up as a speed regulator. (Vector-controlled AC drives usually include a PID speed regulator, but its use is optional depending on the application.)The output of its speed error amplifier, properly limited, is used as a master torque command for all drives. One can easily adjust the sharing by potentiometer control of the torque command into each drive. Figure 8 shows the block diagram of a drive system with AC helper drives of the field-oriented type.
Another good application for vector controlled AC drives is dynamometers where control of load torque on an engine or other prime mover under test is the main objective. Ability to continuously regenerate is vital in this situation. The non-sparking character of the AC motor is a real advantage in engine test cells where explosive vapors may occur. Still another application would be traction or propulsion drives. Typically, these applications require the drive to accelerate from standstill at torque limit to top speed, cruise in either direction at top speed, then regeneratively decelerate at torque limit to stop. Sometimes dynamic reversing at torque limit is necessary to go from some forward speed to a reverse speed as fast as possible without stopping at zero, and maintaining torque as the direction of rotation changes. Another class of applications which benefit from vector control AC drives is centrifuges, which need to accelerate from standstill at torque limit and decelerate at torque limit to stop. Due to the large moment of inertia, a substantial fraction of operating time is spent in torque limit. Hard to start loads like extruders and positive displacement pumps are good applications because a vector control drive can always generate at least 1 P.U. torque at standstill. (More torque can be provided by increasing the current rating of the drive.) Finally, the category of loads where moment of inertia and load torque are wildly variable or totally unpredictable, or the motor is already spinning at some speed in either direction, call for a drive which can adapt to the circumstances of the moment. By definition ordinary AC drive cannot, so the vector control drive gives one the maximum probability of success in an unknown area.
CONCLUSION: In conclusion, vector-controlled AC drives are the current state of the art in induction motor drive technology. Their superior performance characteristics have enabled them to begin displacing dual armature converter DC drives, which have been the performance standard for many years. This process will undoubtedly continue because the vector control techniques have only a modest effect on cost and complexity of the control. We shall see more and more vector-controlled AC drives applied as their merits become fully appreciated by drive users.
REFERENCES: [1] "Adjustable Speed AC Drive Systems". ed. B. K. Bose. IEEE Press 1981 [2] Blaschke, F., "The Principle of Field Orientation as Applied to the New Transvektor Closed-Loop Control System for rotating Field Machines", Siemens Review, may 1972 [3] Gabriel, R., Leonhard, W., Nordbjr, C., "Field Oriented Control of a Standard AC Motor Using Microprocessors", IEEE Transactions on Industry Applications, March / Apri1 1980 [4] Bose, B. K., "Power Electronics and ACDrives ", Prentice - Hal1 1986 [5] Wasko, C. R. ~ "500HP 120Hz Current Fed Field-Oriented Control Inverter for Fuel Pump Test Stands , IEEE-IAS Conference Record, 1986 Annual Meeting pp. 314-320 [6] C. D. Schauder, F, H. Choo, M. T. Roberts, High Performance Torque Controlled Induction Motor Drives" IEEE-IAS Conference Record, 1982 Annual meeting pp. 563-568 [7]Vector Controlled AC Drives by R. H. Osman, Engineering Manager, AC Drives, Robicon Corporation