电机驱动技术的现状与展望

Abstract— One of the main driving force behind the industrial revolution was the invention of the electric motor more than a century ago. Its widespread use for all kinds of mechanical motion has made life simple and has ultimately aided the advancement of human kind. The advent of the inverter that facilitated speed and torque control of AC motors has propelled the use of electric motor to new realms that was inconceivable just a mere 30years ago. Advances in power semiconductors along with digital controls have enabled realization of motor drives that are robust and can control position and speed to a high degree of precision. Use of AC motor drives has also resulted in energy savings and improved system efficiency. Yaskawa Electric Corporation has been at the forefront of technology, creating reliable drives that consistently push the envelope of engineering achievement. This paper reviews Yaskawa’s role in the development and application of the inverter technology to AC motor drives and introduces some futuristic vision for the motor drive technology. The development of more efficient, more powerful electric motor drives to power the demands of the future is important for achieving energy savings, environmentally harmonious drives that do not pollute the electrical power system, and improving productivity. Yaskawa wants to be an integral part of this future and hopes to contribute significantly to achieve this. I. INTRODUCTION The electric motor and its control have advanced considerably in recent years. This can be attributed to significant progress in the field of power electronics enabled by unprecedented progress in the semiconductor technology. The benefit of improvement in the motor drive industry has touched varied applications, from heavy and large industrial equipment such as rolling mills in steel making plants, paper mills, etc. to “Mechatronics” equipment used in machine tools and semiconductor fabrication machines. The AC motor controller comprises of the induction motor controller and the permanent magnet motor controller, both of which have played a key role in the overall progress of the motor drive industry. Fig. 1 shows a current inverter (induction motor controller) and AC servo drives (permanent magnet AC motor and their controllers). The controllers shown in Fig. 1 employ the latest that industrial technology has to offer [1] in power semiconductors using the most advanced motor drive control algorithms in the form of vector control. Such controllers are ubiquitous in varied industrial and commercial applications of the present day and age. As the use of AC motor drives becomes more widespread, it is difficult to ignore an important fact - the electric power used by electro-mechanical energy conversion equipment, of which electric motors form the bulk, exceeds 70% of the total industrial electric power produced. Given the fact that future residential applications will soon be using motor drives in washing machines to HVAC applications, it is Figure 1: Typical AC motor drives. (a) 3-level Induction motor controller; (b) AC servo drives and servomotors. important to concentrate R&D efforts in achieving higher efficiency and smaller size products that use less raw material, are less toxic to the environment, have a long MTBF, and are easy to recycle. Yaskawa Electric Corporation wants to be a part of such a future. The concepts, ideas, and equipment used in the motor drives industry are easily applicable to harnessing energy from alternate sources, including Solar Energy and Wind Energy. Hence, it is not surprising to find power electronics to play an important role in these applications. The motor drives industry can thus become a key player in solving the future energy crisis and simultaneously contribute significantly to environmental preservation. II. AC MOTOR DRIVES The present day industry categorizes AC motor drives into two distinct categories – Induction Motor Drives, and Permanent Magnet AC Motor Drives. The basic difference between the two types of drives is performance and cost. Induction motor still forms the work horse of today’s industry. Applications that use induction motor may not need very high precision position and velocity control. Such applications typically use what is known in the industry as “General Purpose AC Motor Drives”. However, the machine tool industry that caters to the semiconductor manufacturing and other sophisticated industries, require highly precise and controlled motion. Permanent magnet motors are the motor of choice because of their smaller size, higher efficiency, lower inertia, and hence higher controllability. Such motors are clubbed into the Servo Motor category and are controlled by Permanent Magnet AC Motor (PMAC) Drives and are typically more expensive than their induction motor counterpart. A. General Purpose AC Motor Drives – V/f Control The power structure of the General Purpose AC Motor Drives is similar to the PMAC motor drives. Both of these PRESENT STATE AND A FUTURISTIC VISION OF MOTOR DRIVE TECHNOLOGY Mahesh Swamy Yaskawa Electric America, Inc. Waukegan, IL, U.S.A. Email: mahesh-swamy@yaskawa.com Tsuneo Kume Yaskawa Electric Corporation Kitakyushu City, Fukuoka, Japan Email: tjkume@yaskawa.co.jp drives are referred to as Voltage Source Inverters, a term which will soon be clear. Since the power topology includes a large DC bus capacitor as a filter, and since it is the voltage that is modulated to provide variable voltage, variable frequency to the AC motor, such an inverter topology is called a Voltage Source Inverter and forms the integral part of most present day AC motor Drives. A typical schematic of the present day AC motor drive is shown in Fig. 2. + - Qu1 Qu2 Qv1 Qv2 Qw1 Qw2 E 2 E 2 Soft Charge IM U V W Optional DC link Choke Figure 2: Schematic of a typical voltage source inverter based AC motor drive. The general purpose AC motor drives typically provide constant flux into the induction motor. Since the motor flux is the ratio of the voltage to the frequency (V/f) applied to the motor, this ratio is held constant to achieve constant flux operation. The motor current increases almost linearly with load. Conveyor belts and other frictional loads require such profiles. For centrifugal loads like fans and pumps, the flux in the motor can be altered to follow a square function. By doing this, the power consumed by the motor becomes a cubic function of speed (P∝f3) enabling significant energy savings. Even if the V/f is held constant in these types of applications, there is still significant energy savings compared to constant speed drives, where relatively large losses are associated with valve or damper control. Thanks to the square type torque characteristics of the load, voltage reduction at lower speed range is possible that improves efficiency further. The resulting improvement in efficiency is so significant that even the member countries that ratified the Kyoto agreement in the year 2000 agreed to convert fans and pumps from being operated directly across the line to be operated via AC motor drives to save energy and reduce the overall carbon foot print of a given plant. It is significant and important not only for those countries but for all people using centrifugal loads to convert the fixed speed fans and pumps to variable speed. B. High Performance AC Motor Drives – Vector Control Though the majority of industrial applications require unsophisticated V/f control, there are quite a few applications that require higher performance. Such applications include machine-tool spindle drives, paper making machines, winders and pinch rolls in Iron and Steel industries, elevators, top drives for oil drilling, winders/un-winders, pick and place operations, printing, rolling mills, and other applications requiring high torque at low speed. Such performance was achievable in the past using DC motors, which are now being replaced by vector controlled AC Motors. The term vector control refers to techniques where the torque component of the input current is controlled orthogonally to the magnetic field in the induction motor to result in optimal torque production. Such orientation based control is called Field Oriented Control. Similar to a DC machine, it is now possible to independently control the field flux and motor torque to achieve high performance from AC motors. The basic idea of field oriented control is to transform the input time varying current flowing into the motor from three phase to time varying two phase components called α and β components. These α and β components are then transformed into two axis (d-axis and q-axis) that rotate synchronously with the air-gap magnetic field of the motor thereby making them stationary with respect to the rotating magnetic field of the AC motor (Fig. 3(a)). By maintaining the orthogonal relationship between the d-axis and q-axis components and by controlling the q-axis component, optimal torque is produced even at standstill condition. The transformation of the motor current from 3-phase to d-q axis requires instantaneous position and speed of the rotor, which is achieved using pulse encoders mounted on the shaft of the AC motor. There are two fundamental approaches to field oriented control. They are: a. Direct Field Oriented Control, and b. Indirect Field Oriented Control [2]. In the direct field oriented control method, the position and magnitude of the air-gap flux in the AC motor is derived from measurement of motor input voltage and current. The measured flux is compared with a steady reference flux, and is fed into a flux regulator that forces the q-axis flux to go to zero to achieve complete decoupling between the two orthogonal axes. The d-axis value of the measured flux is also used to compute the measured electromechanical torque being produced by the motor, which is then compared with the reference torque. The torque regulator controls the torque producing component of the current to achieve desired torque at desired speed. The angle information from the encoder is directly used to perform the transformation from three-phase to two-axis and vice-versa. The control philosophy in the indirect field oriented control is quite different from the direct field oriented control. Air-gap flux is not explicitly calculated in the case of indirect field oriented control. The motor slip is calculated based on measured current parameters. The calculated slip is used to calculate the slip angle, which is then added to the angle information from the encoder to achieve the correct position of the air-gap flux. The newly estimated angle is used for the transformations so that the d-axis motor current is aligned correctly with the air gap flux to achieve high performance torque control even at standstill. This is clearly one significant advantage of the indirect field oriented control over the direct field oriented control. However, the calculation of the motor slip angle requires information about the rotor parameters that is sensitive to temperature and other operating conditions. This sensitivity is more pronounced in higher power motors. At higher speeds, the resolution of the encoder and the computation time available for the microprocessor to compute the slip angle are typical limitations with the indirect field oriented control method. This limitation does not exist with the direct field oriented control method and the use of both of this type of control – indirect field oriented control for standstill and low speed range and direct field oriented control for high speed range is a classical way of modern control, given the fact that the present day microprocessors are robust enough to do computations for both methods and switch over from one to the other depending on the state of a flag that is settable based on the speed of the motor. Typical control schematic for the two types of control along with the concept of coordinate transformation is shown in Fig. 3 [2]. c-axis b-axis ia ib ic a-axis q-axis θr d-axis iα iβ idr iqr ν iqs ids q-axis d-axis iqs idsν (a) VECTOR ROTATOR 2-to-3 Phase Transf PWM Volt. Mod. ω m* ω fdbk 1/p θ m ω err SPEED REG. TORQUE LIMIT T T* TORQUE REG. iq* iq CURRENT REG. Vqo Vq* Vd* Vdo id id* CURRENT REG. FLUX REG. λ r λ r* FIELD WEAK. (b) VECTOR ROTATOR 2-to-3 Phase Transf K1 PWM Volt. Mod. ω m* ω fdbk θ m ω err SPEED REG. TORQUE LIMIT iq* iq CURRENT REG. Vqo Vq* Vd* Vdoid id* CURRENT REG. λ r* FIELD WEAK. p K2 Slip Calc. 1/p ω slip θ slip + + θ sync (c) Figure 3: Schematic of typical induction motor control in modern AC drives. (a) 3-phase to 2-phase to 2-axis transformation, (b) Direct Field Oriented control and (b) Indirect Field Oriented control. C. High Performance AC Motor Drives – Sensorless Control In the control schemes discussed above and shown in Fig. 3, the encoder feedback forms an integral part. Unfortunately, in many industrial applications, there is fear that either the signal wires carrying the encoder signal could break or the encoder itself could be rendered un-operational due to hostile environment like heat and humidity at the motor. In other cases, the mounting of the encoder on the shaft may present an expense that the consumer is not prepared to bear. In either case, there is a need to achieve high performance from AC motors without the use of encoder signal. The situation described above leads to a new breed of controllers called sensorless control. Some drive manufacturers call this type of control as “open-loop control”. The advent of sophisticated microprocessors with capability of performing real time highly-intense calculations has made this field of study very interesting and challenging. Many researchers have worked on this topic and it still forms an important research and development topic at many major motor drive manufacturers. Two methods are gaining popularity. They are: a. using the motor itself as a sensor by injecting high frequency signal into the motor to reveal saliencies due to slot and teeth of the stator structure; and b. Flux observer based on a machine model that is updated as the motor temperature changes. In the latter case, operation at zero input frequency is not possible, while the exploitation of motor saliencies to identify the rotor position has been shown to be able to control the motor even at zero input frequency [3]. Practically speaking, zero shaft speed is adequate in many high performance applications like winders and top drives, where the drill bit is typically tightened and loosened when the bit needs to be changed. Hence, the flux observer, employed in Direct Torque Controlled (DTC) drives is more than adequate for such applications. Other flux observers that use standard PWM techniques are also sufficient provided the internal microprocessor used is fast enough to perform the needed calculations for the flux observer. Many researchers have worked in this area and quite a few motor drive manufacturers are coming out with sophisticated sensorless algorithms that push the boundaries consistently. III. ADVANCES IN POWER TOPOLOGY Significant progress in semiconductor technology has facilitated higher switching frequency of PWM based voltage source inverters – the workhorse of the modern day AC motor drive. Carrier or switching frequencies in the range of 10-kHz to 15-kHz is quite common. This significantly contributes to improved controllability of voltage, current, and torque. It also helps in the reduction of acoustic noise. However, high-speed switching of IGBTs increases high frequency leakage currents, bearing currents, and shaft voltage. It also contributes to voltage reflection issues that result in high voltage at the motor terminals, especially when the motor is at distances farther than 20m from the drive. Researchers and engineers in the area of power electronics and ac motor drives have long recognized this and have developed many tools that are inserted in between the drive and the motor to handle such application issues. A. Three-Level Neutral Point Clamped Inverter Instead of adding a component in between the drive and the motor, modifying the power topology to reduce the problems described above is a much prudent approach. Yaskawa Electric Corporation was the first drive manufacturer to come out with a three-level drive structure for general purpose low voltage application [4]. The three-level drive topology employed by Yaskawa is called the Neutral Point Clamped three-level inverter. The neutral point clamped (NPC) three-level inverter was introduced first by A. Nabae, I. Takahashi and H. Akagi in 1980 and published in 1981 [5]. With this circuit configuration, the voltage stress on its power switching devices is half that for the conventional two-level inverter (Fig. 2). Because of this nature, it was applied to medium and high voltage drives. Early applications included the steel industry and railroad traction areas in Europe [6][7] and Japan [8]. In addition to the capability to handle high voltage, the NPC inverter has favorable features; lower line-to-line and common-mode voltage steps, more frequent voltage steps in one carrier cycle, and lower ripple component in the output current for the same carrier frequency. These features lead to significant advantages for motor drives over the conventional two level inverters in the form of lower stresses to the motor windings and bearings, less influence of noise to the adjacent equipment, etc. Combined with a sophisticated PWM strategy, it also makes it possible to improve the dynamic performance employing the dual observer method. In order to benefit from the above features, general purpose pulse-width modulated (PWM) NPC inverters have been developed for low voltage drive applications [9], [10]. In this product, a unique technology is used to achieve balancing of the dc bus capacitor voltages [11]. Details are described in the following sections. Figure 4 shows the circuit diagram of the NPC three-level inverter [4]. Each phase has four switching devices (IGBTs) connected in series. Taking phase U as an example, the circuit behaves in the following manner. When IGBTs QU1 and QU2 are turned on, output U is connected to the positive rail (P) of the dc bus. When QU2 and QU3 are on, it is connected to the mid-point (O) of the dc bus, and when QU3 and QU4 are on, it is connected to the negative rail (N). Thus, the output can take three voltage values compared to two values for the conventional two-level topology. Relation between the switching states of IGBTs and the resulting output voltage with respect to the dc mid-point is summarized in Table 1. U V W Qu1 P N E iw O E 2 E 2 Qu2 Qu3 Qu4 Qv1 Qv2 Qv3 Qv4 Qw1 Qw2 Qw3 Qw4 iviu Figure 4: The neutral point clamped three-level inverter circuit topology. DC bus capacitors need to be connected in series to get the mid-point that provide the zero voltage at the output. This is not a drawback since series connection of the dc capacitors is a common practice in general-purpose Table 1: Relation between switching -States and output voltage Qu1 Qu2 Qu3 Qu4 Vu ON ON OFF OFF +E/2 OFF OFF ON ON -E/2 Switching State OFF ON ON OFF 0 inverters rated at 400 – 480 V range due to the unavailability of high voltage electrolytic capacitors. The current from the inverter b