introduction
With the rapid development of power electronics technology, motor manufacturing technology, large-scale integrated circuits and microprocessor control technology, people have higher and higher requirements for the performance, function and cost performance of AC servo control products. AC servo control systems with permanent magnet synchronous motors as actuators have been widely used in CNC machine tools, robots, office automation equipment, large-scale integrated circuit manufacturing, radar and flexible manufacturing systems.
AC servo systems, as one of the important driving sources for modern industrial production equipment, are a core technology involved in the modernization of contemporary industrial technology. Developed countries such as Japan, the United States, Germany, the United Kingdom, and France possess the vast majority of core technologies in this field and impose technological blockades on my country. Among these, servo drive control technology, a key technology determining the performance of AC servo systems, is a major part of the foreign servo technology blockade. As domestic hardware technologies such as AC servo motors and drivers gradually mature, servo drive technology, which exists in software form within control chips, has become a bottleneck restricting the development of high-performance AC servo technology and products in my country.
Therefore, researching and developing digital AC servo drive control technology to create AC servo systems and motion control technologies with independent intellectual property rights, and establishing them as an industry, will provide favorable technical support for the development of my country's equipment manufacturing industry. This has significant practical implications and broad social application prospects.
Current Status of Servo Drive Technology
The development history of high-performance servo systems and servo drive technology is closely related to servo motors, and has gone through three main development stages in its nearly 60-year history:
Before the 1960s, this stage mainly relied on direct drive of power stepper motors, and most of them were open-loop position control systems. The system has advantages such as short response time and small size of drive components, and it has been widely used in EDM machine tools, dot matrix printers, automated production lines and other fields. However, it also has disadvantages such as high heat generation, low efficiency, easy environmental pollution, and difficulty in maintenance.
From the 1960s to the 1980s, DC servo motors were widely used in industry and related fields because they were easier to control and had better speed regulation performance than AC servo motors, and the relevant theories and technologies were relatively mature. The position control of servo systems also evolved from open-loop systems to closed-loop systems.
Since the 1990s, with the rapid development of microelectronics technology and the increasing integration of circuits, along with advancements in sensor technology, rare-earth permanent magnet materials, and motor control theory, AC servo control technology has made significant progress. This has led to the emergence of various new types of motors, such as brushless DC servo motors (BLDC) and AC servo motors (PMSM), which have gradually replaced DC servo systems and are widely used in many high-tech fields. The control methods of AC servo systems are rapidly evolving towards digital control, shifting from hardware servo to software servo, with intelligent software servo becoming a major trend in servo control.
AC servo motor drive control strategy
AC servo motor models, represented by permanent magnet synchronous motors, are strongly coupled, time-varying nonlinear systems with complex control strategies. Therefore, the performance of an AC servo system is directly related to its control strategy. A superior control strategy can not only compensate for shortcomings in hardware design but also further improve system performance, playing a crucial role in AC servo systems. The requirements for control strategies in high-performance AC servo systems can be summarized as follows: the system must not only have fast dynamic response and high dynamic and static accuracy but also be insensitive to parameter changes and disturbances.
Representative control strategies for permanent magnet synchronous motors include traditional control strategies such as open-loop constant voltage-frequency ratio (u/f=constant) control, classical PID control, and field-oriented control (vector control); modern control strategies such as direct torque control, sliding mode variable structure control, adaptive control, and nonlinear feedback linearization theory; and intelligent control strategies such as fuzzy control and neural network control.
Traditional control strategies
(1) Constant voltage frequency ratio control
Constant voltage-frequency ratio control with stator voltage drop compensation ensures a constant air gap flux in the synchronous motor, and adjusting the frequency setpoint synchronously changes the motor speed. This control strategy is open-loop control, controlling only the air gap flux and not adjusting torque, which can easily lead to rotor oscillation and loss of synchronization. Furthermore, since constant voltage-frequency ratio control relies on the motor's steady-state model, its dynamic control performance is not high, making it unsuitable for high-performance servo drive control applications.
To achieve good dynamic performance, a dynamic mathematical model of the motor is essential. Since the dynamic mathematical model of an AC permanent magnet synchronous motor is a nonlinear, strongly coupled, time-varying multivariable system, decoupling control of angular velocity and current, i.e., vector control technology, is required to obtain good control performance.
(2) Classic PID control
PID controllers use proportional, integral, and derivative calculations to determine the control input and thus control the controlled object based on the system error. PID controllers are currently the most widely used controllers, possessing advantages such as simple structure, high stability, reliable operation, and convenient adjustment. They have long been a key technology in industrial control and can meet the needs of most servo control applications.
However, the classic three-loop PID control method for AC servo synchronous motors still has some problems, such as complicated regulator parameter tuning and large errors, and strong dependence on system model and parameters. In some high-precision applications, it is difficult to meet system requirements.
(3) Magnetic field orientation control (id=0)
Vector control, based on an accurate mathematical model of the controlled object, allows AC motor control to extend from external macroscopic steady-state control to transient control of the internal electromagnetic processes of the motor. Vector control transforms the complex, coupled nonlinear variables within the AC motor into DC variables (current, flux linkage, voltage, etc.) that are stationary relative to the coordinate system through coordinate transformation, achieving approximately decoupled control. It then identifies constraints and obtains the optimal control strategy for a specific objective. id=0 control is a specific vector control strategy that achieves AC and DC axis current decoupling in the rotor coordinate system of the permanent magnet synchronous motor. Due to the existence of the id and iq dual current closed loop, the motor's iq current dynamically follows the system torque command (te=ktiq, where kt is the motor torque coefficient), achieving electromagnetic torque control of the motor. This control strategy gives the motor system good output torque linearity and can obtain maximum linear torque. Simultaneously, since all current is used to generate electromagnetic torque, the motor's overload capacity can be fully utilized, improving the motor's starting and braking speeds and ensuring excellent starting and braking performance.
Vector control technology has undergone more than 20 years of research and improvement, and its performance in speed control systems is excellent. Whether at low speed (constant torque control mode) or high speed (constant power control mode), its disturbance rejection characteristics, start-up and braking characteristics, and speed stability characteristics meet or exceed those of DC speed control systems. However, the vector control model and algorithm are relatively complex, and implementation requires coordinate transformations, making it difficult to ensure complete decoupling of the voltage and current of the motor system on the direct and quadrature axes. This can affect the dynamic performance and efficiency of the motor system.
Modern control strategies
Traditional AC servo motor drive control strategies are mostly used when the controlled object model is deterministic, unchanging, and linear, and when the operating conditions and environment are fixed. However, the dynamic mathematical model of an AC permanent magnet synchronous motor is a nonlinear, strongly coupled, time-varying multivariable system. In high-performance applications, it is necessary to consider various nonlinear effects, changes in the object's structure and parameters, alterations in the operating environment, and other time-varying and uncertain factors such as environmental disturbances. The development and application of modern control theory have, to some extent, compensated for the inability of classical control theory to handle time-varying nonlinear stochastic systems.
(1) Direct Torque Control
Direct torque control (DTC) is a high-performance AC motor control strategy proposed in the 1980s by Professor M. Depenbrock of Ruhr University, Germany, and I. Takahash of Japan. While based on a precise mathematical model of the controlled object, DTC differs from vector control in that it directly analyzes the AC motor's mathematical model in the stator coordinate system, eliminating the need for complex coordinate transformations. Employing stator field orientation, it avoids current decoupling. Both torque and flux linkage utilize direct feedback in a two-position bang-bang control, preventing the decomposition of stator current into torque and excitation components. This allows for optimal control of the inverter's switching states, focusing on rapid torque response for high dynamic torque performance. DTC field orientation utilizes stator flux linkage, which is unaffected by rotor parameters; it can be observed simply by knowing the stator resistance, making it insensitive to motor parameters.
Direct torque control (DTC) technology has been successfully applied in the field of frequency converter control for induction motors, with ABB of Sweden launching a series of products. However, DTC still faces some challenges in its application to permanent magnet synchronous motors (PMSMs). DTC employs flux hysteresis, resulting in torque pulsation that directly affects the smoothness of motor operation. DTC requires monitoring of flux and torque, leading to poor accuracy at low speeds, resulting in poor low-speed performance and a limited speed range. Due to the small stator inductance, the motor experiences significant current surges during startup and load changes, causing large flux and torque pulsations. Furthermore, the inability to estimate the initial flux position when the motor is stationary makes startup difficult. Although researchers both domestically and internationally have continuously attempted and improved DTC strategies for PMSMs in recent years, this control scheme currently struggles to meet the requirements of AC servo drive technology.
(2) Sliding mode variable structure control
Variable structure control belongs to the category of nonlinear control. Its nonlinearity manifests as discontinuous control, i.e., a switching characteristic that causes changes in the system's "structure." Sliding mode variable structure control does not require knowledge of the system's mathematical model, only the system parameters and their approximate range of variation. This gives it advantages such as fast response, insensitivity to parameter and disturbance changes, and no need for online identification and design. It also features order reduction and decoupling. When the system enters sliding mode, the transition of the system state is no longer affected by changes in the original system parameters and external disturbances, but is forced to slide near the switching plane, exhibiting complete adaptability and robustness. Therefore, sliding mode variable structure control has been successfully applied in permanent magnet synchronous motor servo systems. However, due to the use of bang-bang control, chattering is unavoidable, which is a major challenge for the widespread application of sliding mode variable structure control. Currently, in AC servo motor systems, the chattering problem caused by sliding mode variable structure control has been partially solved by changing the sliding mode structure, such as using a higher-order sliding mode structure and filtering.
(3) Adaptive control
Adaptive control was proposed by Caldwell in the early 1950s. It combines feedback control with identification theory and is designed to address the effects of changes, drift, and environmental disturbances on the controlled object's characteristics. It is also used when there is limited knowledge of the parameters of the controlled process or when these parameters change during normal operation, especially when there are slow-changing factors. By seeking the optimal performance indicators, it aims to regulate the controlled object.
Currently, adaptive methods applied to control include model reference adaptive control, parameter identification self-tuning control, and various newly developed nonlinear adaptive control methods. Model reference adaptive control systems do not require a precise mathematical model of the controlled object, nor do they require parameter identification. The key issue is designing adaptive parameter adjustment laws to ensure system stability while minimizing the error signal. Its main advantages are ease of implementation and fast adaptive speed. However, adaptive algorithms have some problems, such as complex mathematical models and computations, which complicate the control system. Furthermore, parameter identification and correction require time, and for systems with rapidly changing parameters, control performance is significantly affected by the system's computation speed. In AC servo drives, the system hardware requirements are relatively high, generally using a 32-bit digital signal processor (DSP) or a field-programmable gate array (FPGA).
(4) Nonlinear feedback linearization control
Feedback linearization is a nonlinear control design method. Its core idea is to algebraically transform a nonlinear system into a (fully or partially) linear system, allowing the application of linear system techniques. Its fundamental difference from ordinary linearization lies in the fact that feedback linearization is achieved not through linear approximation of the system but through state transformation and feedback. Recent research in nonlinear control system theory shows that, under certain conditions, an affine nonlinear system can be accurately linearized using nonlinear state feedback and appropriate coordinate transformation. This state feedback ensures the stability of the control system and exhibits good dynamic characteristics. Based on the accurate feedback linearization control method, a linearized control model for a permanent magnet synchronous motor is established. After adopting feedback linearization control, decoupling control of the d and q axes can be achieved, resulting in good current tracking performance, fast torque response, asymptotic convergence of the speed step response to the given value, no steady-state error, small overshoot, and a short transient response.
(5) Intelligent control strategy
Classical and modern control strategies rely on mathematical models of motors, failing to fundamentally solve control problems for complex and uncertain systems. Intelligent control strategies, with their nonlinear characteristics, can address systems with far more complex controlled objects, environments, and tasks. Intelligent control eliminates reliance on models of the controlled object, controlling only based on actual effects, thus resolving uncertainties and imprecision in system control.
Intelligent control strategies include fuzzy control, neural network control, expert system control, robust control, and genetic algorithm control, among which fuzzy control and neural network control strategies are relatively mature in the application of permanent magnet synchronous motor servo systems.
(6) Fuzzy control
Fuzzy control is a type of computer digital control based on fuzzy set theory, fuzzy linguistic variables, and fuzzy logic reasoning. It unifies mathematics and fuzziness, using fuzzy sets, fuzzy linguistic variables, and fuzzy reasoning as its theoretical foundation. Specifically, it uses fuzzy sets to characterize the fuzziness in everyday concepts, employs prior knowledge and expert experience as control rules, and uses machines to simulate human control of the system. This allows it to realistically mimic the control experience and methods of skilled operators and experts.
Fuzzy inference does not rely on precise mathematical models. Based on the input-output data of the actual system and referencing the operational experience of on-site personnel, real-time system control can be achieved, making it suitable for solving control problems of nonlinear systems. Fuzzy control exhibits good robustness and strong adaptability, making it suitable for time-varying and time-delay systems. However, fuzzy control has weak self-learning capabilities, and the control rules rely on experience and expert knowledge during design, which can easily lead to low system accuracy. Simply adopting a fuzzy control strategy requires numerous control rules and extensive experience from operators, resulting in relatively low control precision. Fuzzy control technology has found good applications in the design of current regulators and speed regulators for AC servo motor systems. However, in servo systems with high dynamic requirements, this technology still needs further improvement.
(7) Neural network control
Research on neural networks began in the early 1940s, and breakthroughs were made in neural network theory in the 1980s, making it an important branch of intelligent control.
A neural network is an information processing system that uses engineering techniques to simulate the structure and function of the human brain. Neural network control embeds computational functions into a physical network, with each basic operation having a corresponding connection during computation. The neural network model simulates the activity of neurons in the human brain, including information processing, handling, and storage. Each neuron stores partial information; damage to some neurons or information corruption only weakens part of the network's function. Neural networks possess advantages such as distributed information storage, parallel processing, nonlinear approximation, strong self-learning and self-organizing capabilities. They can fully approximate arbitrarily complex nonlinear systems, learn and adapt to the dynamic characteristics of systems with severe uncertainties, exhibit strong robustness, and possess the ability to simulate human visual thinking, making them suitable for handling systems difficult to describe using models or rules. In recent years, researchers have begun to apply neural network control technology (or artificial intelligence, AI) to AC motor drive control systems to solve problems that are difficult to address using traditional methods. AI-controlled systems exhibit excellent noise suppression characteristics, fault tolerance, and scalability, and are robust to parameters. This represents an important future development direction for motor control technology.
Development Trends of High-Performance AC Servo Control Technology
Servo systems based on permanent magnet synchronous motors represent the current development direction of servo control. Although many methods exist for AC servo control, problems such as low system accuracy, poor reliability, and poor low-speed performance still exist.
Whether traditional, modern, or intelligent control strategies, each has its advantages and disadvantages. A single control strategy is unlikely to achieve ideal control results. Exploring the integration and combination of various control strategies to better improve the control performance of servo systems is the future development direction of high-performance AC servo control technology. Currently, composite control strategies mainly take two forms: one is to adopt new control strategies based on classic PID control strategies, such as fuzzy PID control, neural network PID control, and expert PID control; the other is to adopt two or more new control strategies, such as fuzzy neural network control, adaptive fuzzy control, fuzzy direct torque control, adaptive fuzzy control, and direct torque sliding mode variable structure control. By leveraging the strengths and compensating for the weaknesses of various strategies, the performance of AC speed control systems can be further improved, while also exhibiting stronger robustness. Composite control strategies have become a current research focus and a major trend for future development.
Conclusion
This paper takes a permanent magnet synchronous motor system as an example to elaborate on the basic principles, advantages, and disadvantages of traditional, modern, and intelligent control strategies in AC servo motor systems. It also predicts the development trend of control technology for high-performance AC servo motor systems, pointing out that each control strategy—traditional, modern, and intelligent—has its advantages, but also its limitations. A single control strategy is unlikely to achieve ideal control results; therefore, exploring the integration and combination of various control strategies to better improve the control performance of servo systems represents the future development direction of high-performance AC servo control technology.
