While traditional PID control methods can achieve good steady-state accuracy, their speed, anti-interference capabilities, and robustness to parameter perturbations are not ideal. In traditional PID-controlled AC servo systems, it's difficult to achieve good dynamic and steady-state performance when tuning PID parameters. A balance must be struck, with some emphasis on one aspect over the other. Achieving good dynamic, steady-state, and disturbance rejection performance is even more challenging. Fuzzy control for AC servo systems offers great flexibility, providing a good method to improve the following and disturbance rejection performance, significantly enhancing system robustness. However, the rule base of a fuzzy controller is entirely based on fuzzy expert knowledge, which, while valuable, is a summary of past experience. If the environment or object encounters unprecedented situations, the knowledge base (database and rule base) becomes rigid and inadequate for new circumstances. Traditional control theory, after decades of development and practical testing, is a relatively complete theoretical system. However, it requires establishing a mathematical model of the object. When the mathematical model of the object is uncertain, it brings significant difficulties to the design process, or even makes design impossible. When the controlled object's model exhibits uncertainty and nonlinearity, the system's operating state and environment vary significantly, the system's dynamic and static performance requirements are high, and the system aims to achieve multiple and complex objectives, intelligent control should be employed, as traditional control theory design methods are inadequate. However, relying solely on a single intelligent control method will not yield complete functionality and desired performance. To achieve more comprehensive functionality and more ideal performance, a combination of qualitative reasoning control strategies from intelligent control theory and quantitative calculation control strategies from traditional control theory should be adopted. Therefore, applying modern control theory to servo systems aims to enhance their robustness and improve their dynamic and static performance. In recent years, the application of superior composite control in AC servo systems has shown promising prospects. Combining fuzzy control technology with traditional PID control can effectively address the steady-state error inherent in fuzzy control. Currently, the most common approach is to connect fuzzy control and PID control in series or in parallel. However, PID control with fixed parameters introduces a contradiction between dynamic and steady-state states, failing to fully realize the advantages of fuzzy control. This paper proposes a fuzzy self-tuning controller that dynamically adjusts the PID controller parameters to reflect changes in system error, thus combining the robustness of fuzzy control with the steady-state error mitigation capabilities of PID control. The fuzzy PID controller combines a fuzzy logic controller with a PID controller. The fuzzy logic controller offers high dynamic performance and disturbance rejection capabilities, while the PID controller boasts high steady-state accuracy. By leveraging the advantages of both, a fuzzy (FLCR) and PID composite control AC servo system is constructed (as shown in the figure). [align=center]Figure 1. Structure of the fuzzy PID controller[/align] This is a two-stage composite control AC servo system using an FLCR and a PID controller. The basic design idea is to use a single fixed-parameter PID algorithm for control during steady-state operation. When designing the three PID parameters, typical operating states, environmental conditions, and model parameters (which may be approximate) are selected, and the PID parameters kp, ki, and kd are determined based on the best steady-state performance (through calculation and debugging). After the three PID parameters are determined, they are stored as fixed data in the database, which is effective in ensuring steady-state performance. During dynamic processes (startup, large-scale change in given parameters, large-scale load mutation), a single fixed fuzzy rule set is used for flcr control. When designing the fuzzy rule set, the given increment Δωr*, deviation e, deviation change è, and load increment Δtl are used as input fuzzy linguistic variables. The rule set is designed and debugged. After debugging, it is stored in the fuzzy rule base as a fixed rule set, which is effective in ensuring dynamic performance. To ensure a smooth transition, the following strategy is adopted: when switching from PID control to flcr, the integral value in the PID algorithm is transferred to flcr. When switching from flcr control to PID control, the last output value of flcr is transferred to PID as the current integral value in the PID algorithm. When there is more than one typical steady-state operating condition, multiple steady-state operating points occur, and multiple fixed-parameter PID control strategies can be adopted. Similarly, when there are multiple dynamic transient processes, multiple fixed fuzzy rule sets of control strategies can be adopted. It may also be different combinations of single and multiple. 1. The operation of the fuzzy control part: The input signals of this fuzzy control are the velocity deviation e and the rate of change of velocity deviation Δe, and the output signal is the control signal. The fuzzy subset of variables is {positive large, positive medium, positive small, zero, negative small, negative medium, negative large}, and the corresponding linguistic variables are set as {pl, pm, ps, zo, ns, nm, nl}. First, the actual range of speed deviation and the rate of change of deviation are taken as the input universe of discourse, and the allowable range of output control quantity is taken as the output universe of discourse. The membership degree of an element in the universe of discourse to a certain linguistic variable value is represented by a membership function. The membership functions at both ends of the universe of discourse take a semi-triangular shape, while the rest take a triangular shape. Second, the real-time value of the input variable is taken during the fuzzification process, that is, the real-time value of speed deviation and the rate of change of deviation are calculated. The third step is to compare and combine the real-time value of the input variable with the defined membership function to obtain the corresponding fuzzy input quantity. After the above three steps, the max-min inference synthesis algorithm is used for fuzzy inference and fuzzy output results are generated. Finally, the fuzzy output results are sharpened, and output control is implemented using PWM. The duty cycle is adjusted under a certain period, and the real-time value of the duty cycle is automatically adjusted by the fuzzy control rules. 2. Parameter Correction of Fuzzy PID Controller A fuzzy self-tuning controller consists of four basic parts: a fuzzy controller, a PID controller, a parameter tuning stage, and an estimator. The fuzzy controller and PID controller control the process object based on the corrected parameters. The parameter tuning stage corrects the PID controller parameters based on the dynamic changes in system error. The operation of this fuzzy self-tuning control system involves the fuzzy self-tuning controller continuously sampling and correcting until the system reaches and maintains the desired control performance indicators. The parameter correction principle of the PID controller is as follows: In the initial stage of control, a smaller PID controller correction parameter is used; when the fuzzy control reaches basic stability, the PID controller correction parameter is gradually increased according to the system error to achieve a smooth rise and eliminate the error; when the entire system is basically stable, the PID controller correction parameter is appropriately decreased according to system requirements and error. The fuzzy PID control AC servo system consists of several components, as shown in Figure 2. The PID acts as a speed regulator, comparing the given signal with the quantized value of the encoder output pulse signal to obtain the error e. Through the PIDD algorithm with parameter self-tuning, the desired adjustment value is obtained. This value is then sent to a D/A converter to become an analog signal, which is then sent to the analog voltage input of the frequency converter to control the frequency converter frequency, thereby controlling the motor speed. [align=center] Figure 2: Fuzzy PID control AC servo system block diagram[/align] The main program block diagram of the system is shown in Figure 3. To ensure good motor stability and fast response speed, the proportional, integral, and derivative parameters of the system must be correctly selected. This system utilizes the high-speed computation and judgment capabilities of the PID, and the PIDD algorithm subroutine calculates the control quantity ur, dynamically adjusting for disturbances to minimize the steady-state error. [align=center] Figure 3 Main program flowchart of the system[/align] Simulation results and conclusions: During the debugging process of the AC servo control system, the given speed is 1500r/min, the system can quickly track the given speed and has good response performance. Under the condition of sudden load, the system can quickly and automatically adjust, and the recovery time is about 0.6ms. Its speed characteristic curve is shown in Figure 4. [align=center] Figure 4 Speed ​​curve[/align] Simulation experiment and usage results show that the above fuzzy PID controller has fast dynamic response, can achieve small overshoot or no overshoot control, and the steady-state accuracy can be comparable to PID control. This controller is simple to design, easy to modularize and has good real-time performance, and has strong robustness to adverse changes of the controlled object. References [1] Liang Zhonghua. Research on AC servo system of permanent magnet synchronous motor with intelligent control. Shenyang: Shenyang University of Technology Press, 1999 [2] Li Shiyong, Fuzzy control. Neural control and intelligent control theory. Harbin: Harbin Institute of Technology Press, 1996