The vector conversion control method of asynchronous motor has developed rapidly since it was put forward in 1970s. The main idea of this theory is to simulate asynchronous motor as DC motor, and control excitation current component and torque current component respectively through coordinate transformation, so as to obtain the same good dynamic speed regulation characteristics as DC motor. This control method is now mature and commercialized, and the product quality is stable. Because this method uses coordinate transformation, it requires high performance of the controller, such as operation speed and processing ability. In recent years, scholars at home and abroad have done a lot of research on the defects of vector conversion control, such as complex system structure, nonlinearity and the influence of motor parameters on system performance.
1985, a new control method, namely the direct torque control system of asynchronous motor, was proposed by Professor Depenbrock of Germany. It is the result of the above research. It does not need coordinate transformation, nor does it need to rely on the mathematical model of the rotor, so it is very attractive in theory. Under the laboratory conditions, a prototype with quite high performance index was also made. However, there are still some unsolved problems, such as torque observer and speed fluctuation at low speed, which cannot be commercialized. At present, most systems claiming to realize direct torque control in the market combine flux orientation with direct torque control, and adopt flux orientation vector transformation control at low speed and direct torque control at high speed. Or observe the rotor flux linkage while the direct torque control system is correcting. First of all, it is difficult to determine the timing of smooth switching of this method. At present, doctors from German universities are studying this problem. Secondly, if flux oriented vector control is adopted at low speed, or the method of observing rotor flux linkage is adopted, it still depends on rotor parameters. That is to say, as long as there is a component of rotor flux linkage in it, it is still sensitive to rotor parameters. Can not reflect the advantages of direct torque control. It seems that complete torque direct control is still a long way from commercialization.
In addition, the introduction of sliding mode variable structure control technology based on modern control theory, nonlinear decoupling control using differential geometry theory and model reference adaptive control improves the system performance. However, these theories are still accurate mathematical models based on objects. Some of them need a large number of sensors and observers, and their structures are complex. Some of them still can't get rid of the influence of nonlinearity and motor parameter changes, so it is necessary to further explore ways to solve the above problems.
Throughout the development history of the motor industry, almost every big development is a theoretical breakthrough. But now as some mature modern communication systems, it is not easy to put forward epoch-making theories. Therefore, in the future development, we should combine the existing control theories for a long time, learn from each other's strengths, or introduce the theories and methods of other disciplines into motor control and take an interdisciplinary road to solve the above problems. In recent years, the research of intelligent control is very active and has been applied in many fields. Typical examples are fuzzy control, neural network control and control based on expert system. Because intelligent control does not need an accurate mathematical model of the object and has strong robustness, many scholars introduce intelligent control methods into the research of motor control system, and predict that a new era of power electronics and motion control will be ushered in the next decade. Fuzzy control is more mature. Its advantage is that it does not depend on the precise mathematical model of the controlled object, overcomes the influence of nonlinear factors, and has strong robustness to the parameter changes of the controlled object. Fuzzy control has achieved satisfactory results in AC DC speed regulation system and servo system. Its typical applications are: fuzzy controller is used for motor speed control; Application of fuzzy logic in motor model and parameter identification: optimal control of asynchronous motor efficiency based on fuzzy logic; research on intelligent inverter based on fuzzy logic, etc. In recent years, some literatures have discussed the introduction of neural network control or expert system into the direct torque control system of asynchronous motor, and it is believed that practical results will be achieved in the near future.
2 controller aspect
With a good control method, you need a controller that can realize it. High reliability and good real-time performance are the basic requirements for the control system. At first, motor control used analog circuits with discrete components. Later, with the progress of electronic technology, integrated circuits and even application-specific integrated circuits were used. Most of these circuits are analog-digital mixed circuits, which not only improve the reliability and anti-interference, but also shorten the development cycle and cost and reduce the volume, so they develop rapidly.
As an important aspect of ASIC, almost all semiconductor manufacturers in advanced industrial countries can provide their own ASIC for motor control. Therefore, there are many kinds and specifications of motor control ASIC, and the product data and application data are very rich. But at the same time, because there is no uniform standard among manufacturers, products are extremely scattered and new products appear constantly. In order to meet the needs of a design, it often takes a lot of energy and time to collect and sort out information. However, as mentioned above, the development of motor control is becoming more and more diversified and complicated. So sometimes it may not be able to meet the increasingly demanding performance requirements. At this time, we can consider developing our own special control chip for motor. Field programmable gate array (FPGA) can be used as a solution. As a development device, FPGA can be easily modified many times. To make a simple analogy, FPGA relative to ASIC is like EEPROM relative to ROM produced by mask. Because the integration of FPGA is very large. An FPGA has as few as several thousand equivalent gates and as many as tens of thousands or hundreds of thousands of equivalent gates. Therefore, an FPGA can realize very complex logic, replacing the circuit composed of multiple integrated circuits and discrete components. It uses the hardware description language (VHDL or VerilogHDL) to design the system, and abandons the traditional methods from the gate circuit to the whole system. It adopts three levels of hardware description and top-down design style (starting from system function description), and can mix and simulate the three levels of description to facilitate digital circuit design. The specific levels and their brief introduction are as follows: the first level is behavior description, mainly functional description, which can be used for functional simulation; The second layer is RTL description, mainly logical expression description and RTL level simulation; The third layer is the gate-level description, that is, the description of the basic gate circuit and the corresponding gate-level simulation. Finally, the gate-level network table is generated, and then the programming code points of FPGA are generated with special tools, so that FPGA programming can be carried out. After the trial production is successful, if mass production is needed, ASIC chips can be customized according to the design of FPGA to reduce the cost. At present, there are feasible articles in this field, and interested readers can refer to them.
The appearance of integrated circuit has a far-reaching impact on motor control. It has greatly promoted the development of motor control industry, and there is still a broad market, but it is a pity that domestic ic manufacturers cannot occupy their due share in this market. With the development of technology, especially the digital trend is widely popular today, people will not be satisfied with staying in the era of mixed analog and digital.
At present, most common frequency converters on the market are controlled by single chip microcomputer. 8096 series products are widely used. However, the processing ability of single chip microcomputer is limited, especially in the vector transformation control system. Because of the large amount of data to be processed and the high requirements of real-time and accuracy, the single chip microcomputer can no longer meet the requirements. So people naturally think of digital signal processor (DSP). In recent years, the performance of various integrated single-chip DSPs has been greatly improved, and more and more software and development tools have become better and better. However, the price has fallen sharply. At present, the low-end products are close to the price level of single chip microcomputer, and have high cost performance. Therefore, DSP devices and technologies are easier to use and the price can be accepted by users. More and more SCM users began to choose DSP devices to improve product performance, and the time is ripe for DSP devices to replace high-end SCM. Moreover, with the widespread popularity of DSP in all walks of life, the contradiction between supply and demand of professional talents will soon be solved.
Compared with single chip microcomputer, DSP device has higher integration. DSP has faster CPU, larger memory, built-in baud rate generator and FIFO buffer. Provide high speed, synchronous serial port and standard asynchronous serial port. Some on-chip analog-to-digital converters and sample-and-hold circuits can provide PWM output. More differently, DSP devices are RISC devices, and most instructions can be completed in one instruction cycle, while through parallel processing technology, multiple instructions can be completed in one instruction cycle. DSP adopts an improved Harvard structure, has independent program and data space, and allows simultaneous access to programs and data. Built-in high-speed hardware multiplier and enhanced multi-stage pipeline make DSP devices have high-speed data operation ability. Single chip microcomputer is a complex instruction system computer (CISC), and most instructions need 2 ~ 3 instruction cycles to complete. Single-chip microcomputer adopts Neumann structure, program and data are accessed in the same space, and only instructions or data can be accessed separately at the same time. ALU can only do addition, and multiplication needs software, so it takes up many instruction cycles and is slow. Therefore, the structural differences make DSP devices execute an instruction 8 ~ 10 times faster than 16-bit single chip microcomputer, and complete a multiplication and addition operation 16 ~ 30 times faster. Simply put, the calculation function of DSP device is very strong, while the transaction processing ability of single chip microcomputer is very strong. The DSP device also provides a highly specialized instruction set, which improves the operation speed of FFT and filter. In addition, DSP devices provide JTAG(JointTest Action Group) interface, which has more advanced development means and is more convenient for batch production testing.
In order to seize the market share of motor control, major DSP manufacturers have introduced their own embedded DSP motor-specific control circuits. For example, Texas Instruments, which accounts for 45% of the DSP market, has introduced a special DSP for motor controllers-TMS320C24X (TMS320F24X on-chip ROM is erasable). The new TMS 320 c 24 DSP adopts the core of TI's T320C2xLP 16 fixed-point DSP, and integrates the motor control event manager, which is characterized by electronically controlling the motor steering in the best way. The equipment uses TI's reusable DSP core technology to show TI's special ability-by integrating a DSP core and its digital and mixed signal peripherals on a single chip, DSP solutions for various applications can be manufactured. As the first digital motor controller with special DSP series, TMS320C24x can support motor steering, command generation, control algorithm processing, data exchange and system monitoring. A single-chip digital motor control scheme can be provided by adding many factors such as integrated DSP core, optimized motor controller event manager and single-chip A/D design. This series of TMS320C240 includes a 20MIPS DSP core, an event manager, two serial interfaces, a pair of 10 bit analog-to-digital converters, a 32-bit digital I/O system, a watchdog timer, a low-voltage monitor and a 16K character flash memory (TMS320F240 type). Rely on compatibility to realize system upgrade. The coding of TMS320C240 is compatible with the DSP of TI's TMS320Clx, TMX320C2x, TMS230C2xx and TMS320C5x series. It uses the fixed-point DSP software development tool of TMS320 and JTAG simulation support, so that the developers in the field of motor controller can easily upgrade from microcontroller to new DSP. American analog equipment (AD) company is not far behind. It cooperates with the famous intel company to produce ADMC3xx series special DSP for motor control, and its performance is not much different from that of TI company. It is also designed based on AD company's 16-bit fixed-point DSP core ADSP2 17 1. In addition, it also integrates a three-phase PWM generator (16 bits) and an analog-to-digital converter. Other well-known DSP manufacturers include Motorola and NEC. Another advantage of using motor ASIC based on DSP is that it can reduce the requirements for peripheral devices such as sensors. Complex algorithm can achieve the same control performance, reduce cost, and have high reliability, which is beneficial to the confidentiality of patented technology.
Sometimes the system needs more human-computer interaction, printing and other controls, and a DSP is not competent. At this time, you can use single chip microcomputer to handle transactions and DSP to handle operations. However, this not only increases the synchronization and communication burden between the two processors, but also deteriorates the real-time performance of the system and prolongs the development time of the system. In this case, Tricore is a good way to solve the problem, which concentrates the capabilities of microprocessor, microcontroller and digital signal processor on one chip, thus solving most engineering problems on one chip.