Modern control strategy
The traditional AC servo motor drive control strategy is mostly used under the condition that the controlled object model is determined, does not change and is linear, and the operating conditions and operating environment are determined to be constant. However, the dynamic mathematical model of AC permanent magnet synchronous motor is a nonlinear, strongly coupled, time-varying multivariable system. In the case of high performance requirements, various nonlinear effects, changes in the structure and parameters of the object, and changes in the operating environment must be considered. And time-varying and uncertain factors such as environmental disturbances. The development and application of modern control theory to some extent make up for the shortcomings of classical control theory to the time-varying nonlinear stochastic system.
(1) Direct torque control
The direct torque control theory is a high-performance AC motor control strategy proposed by Professor M. depenbrock of the German University of Ruhr and the Japanese scholar i.takahash in the 1980s. The control strategy is also based on the precise mathematics of the controlled object. The model, but unlike vector control, analyzes the mathematical model of the AC motor directly in the stator coordinate system without complex coordinate transformations. The stator field orientation is adopted, no decoupling current is needed, and the torque and flux linkage are directly controlled by the two-position 砰砰 control, which avoids decomposing the stator current into torque and excitation components, and directly controls the switching state of the inverter. Good control, focusing on the fast response of torque to achieve high dynamic performance of torque. The direct torque control field orientation uses the stator flux linkage, which is not affected by the rotor parameters. As long as the stator resistance is known, it can be observed and is not sensitive to the motor parameters.
Direct torque control technology has been successfully applied in the field of induction motor inverter control, and abb has launched a series of products. However, in the application of permanent magnet synchronous motor, there are still some problems in direct torque control. The direct torque control uses the hysteresis of the magnetic chain, and the motor torque is pulsating, which directly affects the smoothness of the motor running. Direct torque control needs to observe the flux linkage and torque. The accuracy is poor at low speeds, resulting in poor motor running performance and small motor speed range. Due to the small stator inductance of the motor, the current impact is large when the motor starts and the load changes, and the flux linkage and torque ripple are large. In addition, since the initial position of the flux linkage cannot be estimated when the motor is stationary, the motor is difficult to start. Although some scholars at home and abroad have been trying and improving the direct torque control strategy of permanent magnet synchronous motor in recent years, this control scheme is difficult 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, and its nonlinearity appears as the discontinuity of control, that is, a switching characteristic that changes the "structure" of the system. Sliding mode variable structure control does not need to know the mathematical model of the system. It only needs to understand the approximate range of system parameters and their changes, so that the variable structure control has the advantages of fast response, insensitivity to parameters and disturbance changes, and no need for online identification and design. With the function of reducing the order and decoupling, when the system enters the sliding mode state, the transfer of the system state is no longer affected by the original parameter changes and external disturbances of the system, but is forced to slide near the switch plane, with complete self Adaptability and robustness, so sliding mode control has been successfully applied in permanent magnet synchronous motor servo system. However, due to the bang-bang control, the chattering problem is inevitably caused, and the chattering problem is a major difficulty in the widespread application of sliding mode variable structure control. At present, in the AC servo motor system, by changing the sliding mode structure, such as the use of high-order sliding mode structure and filtering processing, the chattering problem caused by the sliding mode variable structure control is solved to some extent.
(3) Adaptive control
Adaptive control was proposed by Golcl-well in the early 1950s. It combines feedback control with identification theory, and proposes the influence of changes in the characteristics of the controlled object, drift and environmental disturbance on the system, or when there are not many parameters of the controlled process or these parameters are in normal operation. Changes, especially when there are slow variables, are optimized by seeking certain performance indicators to complete the adjustment of the controlled object.
The adaptive methods currently applied to control are model reference adaptive, parameter identification self-correction control and various newly developed nonlinear adaptive control. The model reference adaptive control system does not require an accurate mathematical model of the control object and does not require parameter identification. The key problem is to design an adaptive parameter adjustment law to ensure the stability of the system while making the error signal tend to zero. The main advantage is that it is easy to implement and fast. However, there are some problems in the adaptive algorithm, such as the mathematical model and the cumbersome operation, which complicates the control system. For example, parameter identification and correction take a period of time. For systems with faster parameter changes, the control performance is greatly affected by the system calculation speed. The application system hardware needs to be high in the AC servo drive, which is generally implemented by 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. The core idea is to convert a nonlinear system algebra into a (all or part of) linear system so that the skills of the linear system can be applied. The fundamental difference between it and ordinary linearization is that feedback linearization is not obtained by linear approximation of the system but by state transition and feedback. In recent years, the theoretical research results of nonlinear control systems show that nonlinear state feedback and appropriate coordinate transformation can be used to accurately linearize an affine nonlinear system under certain conditions, and this state feedback can guarantee the control system. Stability and good dynamic quality. Based on the precise feedback linearization control method, the linearized control model of the permanent magnet synchronous motor is established. After the feedback linearization control, the decoupling control of the d and q axes can be realized, the current tracking performance is good, and the torque response is fast. The speed step response can gradually converge to a given value, without static difference, small overshoot and short transition process.
(5) Intelligent control strategy
Classical or modern control strategies rely on the mathematical model of the motor and do not fundamentally address the control problems of complex and uncertain systems. The intelligent control strategy has non-linear characteristics and can solve systems with more complex control objects, environments and tasks. Intelligent control gets rid of the dependence on the controlled object model, and only controls according to the actual effect. In the control, the system uncertainty and inaccuracy can be solved.
Intelligent control strategies include fuzzy control, neural network control, expert system control, and robust control and genetic algorithm control. Fuzzy control and neural network control strategies are mature in the application of permanent magnet synchronous motor servo system.
(6) Fuzzy control
Fuzzy control is a kind of computer numerical control based on fuzzy aggregation, fuzzy linguistic variables and fuzzy logic reasoning. Fuzzy control unifies mathematics and fuzziness, and uses fuzzy sets, fuzzy linguistic variables and fuzzy reasoning as its theoretical basis, that is, using fuzzy sets to describe the ambiguity in the concepts used by people everyday, with prior knowledge and expert experience as Control rules, using machine simulation to control the system, can realistically imitate the control experience and method fuzzy control of skilled operators and experts.
Fuzzy reasoning does not depend on accurate mathematical models. According to the input and output data of the actual system, the system can be controlled in real time with reference to the operating experience of the field operators. Therefore, it is suitable for solving the control problems of nonlinear systems; Good stickiness and strong adaptability, suitable for time-varying and time-delay systems. However, the fuzzy control self-learning ability is not strong, and the design control rules depend on experience and expert knowledge, which may cause the system to be inaccurate. Simply adopting the fuzzy control strategy requires more control rules, requires a lot of experience of the staff, and the control precision is relatively low. The fuzzy control technology has been well applied in the design of AC servo motor system current regulator and speed regulator. However, in the servo system with high dynamic requirements, the technology still needs to be further improved.
(7) Neural network control
The research of neural network began in the early 1940s. In the 1980s, the neural network theory made a breakthrough and became an important branch of intelligent control.
Neural network refers to an information processing system that simulates the structure and function of human cranial nerves by engineering techniques. The neural network control embeds the calculation function in the physical network. In the calculation process, each basic operation has a corresponding connection with it. The neural network model simulates the activity process of human brain neurons, including the processing, processing, and storage of information. Each neuron stores part of the content of a variety of information, and some neuron damage and information destruction only lead to partial weakening of the network. The neural network has the advantages of information distribution storage, parallel processing, nonlinear approximation, self-learning and self-organization ability. It can fully approximate arbitrarily complex nonlinear systems, and can learn and adapt to the dynamic characteristics of severely uncertain systems. Robustness, with the ability to simulate human image thinking, is suitable for dealing with systems that are difficult to describe with models or rules. In recent years, people have begun to try to apply neural network control technology (or artificial intelligence ai) to AC motor drive control systems to solve problems that are difficult to solve by traditional methods. The use of the ai adjustment system has good noise suppression characteristics, fault tolerance and scalability, and is robust to parameters. It is an important development direction of future motor control technology.
High-performance AC servo control technology development trend
The servo system based on permanent magnet synchronous motor is the development direction of servo control. Although there are many methods for implementing AC servo control, there are still problems such as low system accuracy, poor reliability, and low speed performance.
Whether it is a traditional control strategy, a modern control strategy, or an intelligent control strategy, each control strategy has its advantages, but at the same time there are some problems. It is difficult to obtain the ideal control effect from a single control strategy. It is the development direction of high-performance AC servo control technology in the future to explore how to infiltrate and compound various control strategies to better improve the control performance of the servo system. At present, the composite control strategy mainly has two forms: one is to adopt a new control strategy based on the classic pid control strategy, such as fuzzy pid control, neural network pid control, expert pid control, etc.; second, adopt two or more new types of control Strategies such as fuzzy neural network control, adaptive fuzzy control, fuzzy direct torque control, adaptive fuzzy control, direct torque sliding mode variable structure control, etc. The various strategies complement each other to further improve the performance of the AC speed control system, and at the same time have stronger robustness. The composite control strategy has become the focus of current research and a major trend in the future development.
Conclusion
Taking the permanent magnet synchronous motor system as an example, the basic principles, advantages and disadvantages of traditional control strategy, modern control strategy and intelligent control strategy in AC servo motor system are described separately, and the control technology of high performance AC servo motor system is predicted. The development trend points out that whether it is a traditional control strategy, a modern control strategy, or an intelligent control strategy, each control strategy has its advantages, but at the same time there are some problems. It is difficult to obtain the ideal control effect from a single control strategy. It is the development direction of high-performance AC servo control technology in the future to explore how to infiltrate and compound various control strategies to better improve the control performance of the servo system.
