Thread Content
(1) Appropriate voltage level for high-power energy-saving speed control Frequency conversion speed control can save a large amount of electrical energy in medium and high-power fans and pumps, with most of these systems having power levels in the range of 0.2–2 MW. At present, most motors with a capacity of over 200KW are of medium voltage, with a voltage level of 10KV in most cases, and 6KV in a few cases. Choosing a 10KV \"direct\" frequency conversion is unreasonable from both technical and economic perspectives. All “direct” frequency converters are not truly direct frequency converters; they have transformers on their input side, so it is not necessary for the motor and the frequency converter to be at the same voltage as the grid voltage. This article discusses the appropriate voltage levels for different power ranges. (2) Vector control and direct torque control in high-performance speed control systems Vector control in high-performance speed control systems was invented in the late 1970s and commercialized in the 1980s; it is still used by many companies to this day. Direct torque was invented in the late 1980s and adopted by some companies; it was commercialized in the early 1990s and widely promoted as a next-generation technology. This article presents the author’s views on these two systems. (3) Systems with speed (position) sensors and systems without speed (position) sensors Speed (position) sensors (encoders) were required in the early stages of development for vector control and direct torque control systems. In some situations, it is difficult to install encoders; therefore, speed (position) sensor-free systems were developed. Their performance is inferior to that of the former systems, but it is better than that of V/f open-loop systems. Some current claims suggest that the low-speed starting performance of encoder-less systems has reached the level of encoder-equipped systems, but this statement is ambiguous. This article discusses when an encoder should be installed and when it can be omitted. 2 Rational voltage levels for high-power, energy-saving variable-speed drives Variable-frequency speed control can save a significant amount of electrical energy in large and medium-sized fans and pumps, with most of these systems having power ratings in the range of 200–2000 KW. Our existing AC motors of 200KW represent a boundary; below 200KW it is low voltage at 380V, while above 200KW it is medium voltage at 3KV, 6KV, and 10KV. From the perspective of reducing line losses, the power sector aims to increase the supply voltage; 3KV has been phased out, 6KV is being phased out as well, and 10KV is being promoted vigorously; in the future, it might even be raised to 20KW. Starting from the need for simplified configuration, users naturally requested that motors and transformers with a voltage of over 200 KV also use 10 KV. Unfortunately, meeting this reasonable request is technically difficult and costly from an economic perspective, for the following reasons: A. Manufacturing 10 KV motors is not difficult, but as the voltage increases, the insulation requirements rise, which in turn leads to an increase in the weight and cost of the motors. Taking the YJS series 4-pole 560 KW motor as an example: it weighs 3.6 tons at 380 V and costs 110,000 yuan; it weighs 3.9 tons at 6 KV and costs 150,000 yuan; it weighs 4.4 tons at 10 KV and costs 200,000 yuan. B. Limited by the voltage of power electronic devices and the allowable dv/dt of the motor, 10KV frequency converters must be multi-level with multiple devices connected in series. This results in complex circuits, high costs, and poor reliability. For a 10KV inverter using 1700V IGBT devices, 10 strings are required, resulting in a total of 120 devices for three phases. If 3300V devices are used, 5 strings totaling 60 devices are also required, which is a large number. On the other hand, when the current is low, the current-carrying capacity of the devices cannot be fully utilized. Taking 560 KW as an example again, the current in a 10 KV motor is only around 40 A; whereas with 1700 V IGBTs, the current can reach 2400 A, and with 3300 V devices, it reaches 1600 A. Instead of using devices capable of handling high currents, people prefer to use a large number of low-current devices connected in series, which is highly unreasonable. Even when the motor power reaches 2000 KV, the current is only around 140 A, which remains quite low. For level isolation, to improve the input current waveform and reduce harmonics, all medium-voltage \"direct conversion\" inverters today are not truly direct conversion types; they are equipped with input transformers on the input side, and this arrangement is not expected to change in the near future. Since there is a transformer on the input side, it is not necessary for the voltage of the inverter and the motor to be the same as that of the power grid; 10KV and 6KV are not mandatory. Thus, the issue of an appropriate voltage level for the inverter and the motor arises. Additionally, in the past, the 200KW threshold for low and medium voltage motors was determined with regard to direct motor starting; the starting current is 7-8 times the rated current. The capacity of 10KV/380V power transformers was set at 2000KVA, with a short-circuit impedance of around 6%. When motors start, the voltage drop on the 380V bus was limited to about 5%. If the transformer is made larger, the short-circuit current becomes too high, and the low-voltage switch cannot handle it. After speed control is achieved using an inverter, the starting current is limited to the rated value, and the boundary condition between medium and low voltage should also change accordingly. Currently, the capacity of 660V low-voltage motors has reached 1000–1200KW, which also provides a basis for discussing appropriate voltage levels. The starting point for analyzing the appropriate voltage level in this article is: A. Low-voltage frequency converters use IGBTs with a voltage of 1200V or 1700V; the rated current of these devices is less than 1800A–2400A, and the number of units connected in parallel is no more than 2. If implementing parallel connections proves too troublesome, it is better to opt for multi-level series connection with medium-voltage frequency conversion. B. There are many types of devices and voltage levels available for medium-voltage frequency conversion, and the corresponding circuit designs also vary. This article is based on the products currently available on the market, which include the Separate DC Power Multiplexing (H-bridge series) scheme (SDM) using 1700V IGBTs, as well as the Three-Level scheme (THL) utilizing IGBTs, IGCTs, or IEGTs at 3300V, 4500V, and 6000V. The literature has analyzed the appropriate voltage levels, so this will not be repeated here; only a few points are listed below: A. For variable frequency speed control systems with capacities of 800–1200 KW or less, 380V or 660V voltage levels are suitable. It has a simple circuit design, mature technology, high reliability, low dv/dt, and is inexpensive. Taking the 560KW motor as an example again, a low-voltage frequency converter with 630KW and 660V costs around 500,000, while a medium-voltage frequency converter of the same capacity but with 2300V costs around 900,000. The implementation methods include low-low, low-high, high-low, and high-low-high, among others. Since the cost of motors and transformers is much lower than that of frequency converters, it is reasonable to replace them as well. B. For speed control requirements of over 1000-1500 KW, medium-voltage frequency converters can be used. Foreign medium-voltage frequency converters are available in various voltage levels: 1.1 KV, 2.3 KV, 3 KV, 4.2 KV, 6 KV; these levels are primarily determined by the voltage ratings of the power electronic components used. The relationship between the device voltage and the inverter voltage, under the conditions of no series connection of devices in THL and no series connection of bridges in SDM, is shown in Table 1. Table 1 Relationship between device voltage and inverter voltage without series connection Device voltage (V): 1700, 3300, 4500, 6000 Inverter voltage (KV): 1.1, 2.3, 3, 4.2 Currently, the maximum device voltage is 6000 V; without a series connection, the maximum inverter voltage is 4.2 KV. 6KV frequency converters must be connected in series, resulting in complex circuits and numerous components, which affects reliability. 6KV frequency converters are rarely produced abroad, and 10KV ones are almost never made. In principle, SDMs are connected in series through H-bridge units, so the output voltage of the inverter is not limited by the voltage of the components; it can be high. However, the cost of increasing the voltage is an increase in the number of components, which reduces reliability. For inverters with the same output power, the cost of using a higher voltage with more units connected in series is greater than the cost of using a lower voltage with fewer units but higher current per unit; in other words, under the condition that the current limits of the components permit it, the lowest possible voltage level should be chosen. Many applications require a bypass function, which allows the motor to be bypassed in the event of an inverter failure and to operate at a constant speed directly connected to the power grid. To reduce the cost of inverters, a problem that needs to be addressed is how to bypass the circuit when the motor voltage falls below the grid voltage. This problem can be solved; different bypass methods are used for different frequency converters. Bypassing a frequency converter means connecting the motor directly to the power grid when the frequency converter fails, allowing it to operate at a constant speed. If the motor voltage matches the grid voltage, there is no problem with bypassing. To reduce the cost of inverters, the issue to be discussed here is how to bypass the motor when its voltage falls below that of the power grid. If low-voltage frequency conversion is used, the AC voltage at the inverter input is the same as the rated output voltage, allowing the motor to be connected directly to a low-voltage power supply of 380V or 660V, bypassing the inverter. If a THL medium-voltage inverter is used, the two sides of the input inverter can be connected in series to supply power to the motor, as shown in Figure 1. When the three toggle switches are set to “1”, the inverter operates; when the switches are set to “2”, bypass mode is activated. The voltages across the two secondary windings of the input transformer are each equal to 1.5Vm/2 (where Vm is the rated input voltage of the motor), with a difference of 300 degrees between them. When these voltages are connected in series, the resulting voltage is 1.5Vm cos150 = 1.01Vm, which is just sufficient to enable the motor to operate at a constant speed. If an SDM inverter is used, there are too many connections on the secondary side of the inverter; it is not possible to bypass the inverter by changing the wiring. Instead, only the faulty unit can be bypassed, by short-circuiting its output through contacts, which prevents the IGBTs in that unit from functioning. The operating conditions of the bypass unit have been taken into account in the design of such frequency converters. If it is necessary to bypass the frequency converter, an additional standby step-down transformer must be installed, which is reasonable when multiple frequency converters are connected to the same electrical network. When designing the bypass circuit, it is necessary to verify the starting torque of the motor when started directly. For example, if the short-circuit impedance of the transformer is 6%, its capacity is 1.1 times that of the inverter, and the motor’s starting current is 7 times the normal value, then the motor’s starting voltage will be 0.72 Vm, and its starting torque will be 0.52 times the rated starting torque; this value must be greater than the load torque. If the starting torque is insufficient, the transformer capacity must be increased or a transformer with a low short-circuit impedance should be used. 3 Vector Control and Direct Torque Control in High-Performance Speed Control Systems The task of a speed control system is to regulate speed, which is altered through torque; the performance of such a system depends on the quality of torque control. Both vector control (VC) and direct torque control (DTC) aim to achieve high-performance torque control, and their speed regulation mechanisms are identical. The torque of an asynchronous machine is equal to the vector product of the flux vector and the stator current vector. Flux linkage cannot be measured directly; it must be calculated using the stator voltage and current as well as motor parameters. Since both the stator voltage and current are alternating quantities, which makes them difficult to handle, in VC control systems, coordinate transformation is employed to convert them into direct currents in the dq coordinate system. The control values obtained through this calculation are then transformed back to the alternating-current coordinate system to generate PWM signals. To achieve good performance at both high and low speeds, two models of voltage and current must be used, which involves many motor parameters. In the DTC system, torque and flux are calculated directly using the alternating current values, and then PWM signals are generated through two Band-Band controllers for torque and flux, eliminating the need for coordinate transformation. In the early stages of developing DTC, low-speed operation conditions were not taken into account; it was based on the stator flux, with only Rs as one of the motor parameters involved. As a result, the suppliers of DTC heavily promoted its simplicity in calculation, the minimal number of required motor parameters, and high accuracy. In fact, when considering low-speed operation conditions, the DTC also must incorporate a current model and make use of the rotor flux; it involves the same number of motor parameters as VC, so its accuracy is the same as well. The DTC method does not involve coordinate transformation, and its calculation formula is simple; however, to achieve Band-Band control, it is necessary to perform numerous calculations within one switching cycle, which requires fast computation speeds. Taking ABB’s ACS600 series as an example, its calculation cycle is 25 μs. In VC, the average values of voltage and current over one switching cycle are measured, with the calculation being performed once per cycle; thus, low computational speed is required. Taking Siemens’ 6SE70 series as an example, its calculation cycle is 400 μs, which represents a difference of 16 times. Vector transformation calculations involve just 4 multiplications and 2 additions; given the capabilities of today’s processors, this is not much of a challenge. Furthermore, using the stator flux as a basis is not exclusive to DTC; some VC systems also use the stator flux as a basis. According to the product samples, the torque control response time for ACS600(DTC) is 5 ms, and that for 6SE70(VC) is also 5 ms; even mechanisms with faster response times cannot handle this. Some believe that DTC achieves an approximately circular magnetic field through Band-Band control of the flux amplitude; fluctuations in the flux amplitude lead to torque fluctuations, whereas VC is a continuous control method in which the flux amplitude remains constant, resulting in no torque fluctuations. This view is also incorrect. In DTC, thanks to the torque Band-Band control, the average torque does not fluctuate as a result of changes in flux; flux variations only affect the current waveform. In the case of VC, since the inverter operates in PWM mode, it cannot be controlled within a single switching cycle, and there is no continuous control either; this leads to current fluctuations that in turn cause torque fluctuations. The torque fluctuation in 6SE70 is 2%. In summary, the author believes that there is no essential difference between these two systems; they merely adopt different approaches to achieve torque control, and there is no issue of one being superior to the other or one replacing the other. 4 Systems with speed (position) sensors and systems without speed (position) sensors In the early stages of developing vector control and direct torque control systems, it was necessary to install encoders on the motor shaft to measure speed (position) signals. In some cases, installing encoders proved difficult; therefore, sensorless systems were developed. Speed-sensorless systems are a hot topic nowadays, with many methods available, but those actually used in industrial products are all based on the same principle—the voltage and current model method. The voltage model uses fewer motor parameters; it provides high calculation accuracy at speeds above 5-10% (high speed), while at speeds below 5-10% (low speed), the calculation errors are large due to the low voltage levels. The current model relies on many motor parameters, and is particularly sensitive to changes in rotor resistance, resulting in slightly larger calculation errors; however, these errors are independent of the speed. In systems equipped with speed sensors, a voltage model is used at high speeds, providing high control precision; a current model is used at low speeds, and although the precision is lower than that at high speeds, the system can still function properly. In a speed-sensorless system, the rotational angular velocity at high speeds is determined by identifying the results calculated using the voltage-current model; therefore, it can only achieve the performance level of a speed-sensor-equipped system at low speeds. At low speeds, due to the inaccuracies in the voltage model, a reference point is no longer available, making identification impossible. As a result, the system has to abandon vector control and operate in open-loop mode. Current speed-sensorless vector control systems available on the market are open-loop systems at low speeds, resulting in poor performance. They are only suitable for applications without long-term low-speed operation, where high precision in speed control is not required at high speeds. Some companies claim that their speed-sensorless vector control system can generate full torque even when at rest, and this is true, but there is also a element of marketing to it. Because it is known that the velocity is zero when at rest, and no identification is needed; however, once it starts moving, long-term operation at low speeds is not possible. Table 2 lists the performance of the 6SE70 series inverters with and without an encoder. 6 Conclusions 1) It is unreasonable to use 6kV or 10kV \"direct frequency conversion\" for high-power energy-saving speed control; a suitable voltage level should be selected based on the power capacity. High-power inverters use high voltages and prefer components with high current to reduce the number of components in series. 2) Vector control and direct torque control each have their advantages and disadvantages; different companies have chosen different approaches, and there is no question of one being superior to the other or of one replacing the other. 3) The speed-sensorless system is only suitable for applications without long-term low-speed operation, where performance requirements are not high at high speeds. 4) For motors in the 200kw-315kw power range, in China, the 380v voltage level accounts for a large proportion. 5) The advantages and disadvantages of vc and dtc – with so much commercial hype involved, let them argue about it. 6) Consider the rationality of selecting the variable frequency voltage level from the perspective of cost-performance; for capacities below 315 kW, 380 V can be used, for 250–800 kW, 660 V is an option, and for capacities above 500 kW, 6 kV high-high frequency conversion or 6 kV/10 kV high-low-high frequency conversion can be employed. 7) High-high product prices are high. The high-low-high type of product, due to its position within the transformer, has slightly higher power consumption, but it offers better performance and a better price.