Study On The Strategies Of Self-tuning And Anti-disturbance For PMSLM Servo System

Permanent magnet synchronous linear motor (PMSLM) has been widely used in many fields, such as precision engineering and industrial automation, because of the advantages of high speed, high accuracy, and big thrust force. Nevertheless, compared with a rotational motor, PMSLM is more sensitive to various disturbances because of the reduction of gears. Moreover, the servo system of PMSLM generally uses a traditional proportional integral (PI) controller with fixed control parameters to achieve the control process, it is hard to obtain the satisfactory control performance according to the variations in operating conditions. To solve the problems, in this paper, the controlled model of the servo system of PMSLM is built, then, the strategies of self-tuning and anti-disturbance are proposed. Finally, experiments are made to verify the performance of the servo system of PMSLM.This paper, aiming at obtaining high-performance the servo system of PMSLM, combined with the National Science and Technology major projects "Product development and integrated validation of open advanced numeric control system> servo drive and motor " and National Natural Science Fund "Study on control parameters self-tuning strategy of Feature-based CNC machine tool motion ", mainly completes the following studies:The mathematical model of the servo system of PMSLM is built, and the vector control of the servo system of PMSLM is introduced; Meanwhile, the applied disturbances of PMSLM were analyzed, and the theoretical basis for the strategies of self-tuning and anti-disturbance is obtained.The on-line identification strategy of controlled object in the linear servo system is proposed:In the running process of the servo system of PMSLM, a k-Vector Nearest Neighbors (k-VNN) is proposed to adopt to different conditions; In order to apply the Just-in-time Learning (JITL) technique into the linear servo system, the database update strategy of the traditional JITL technique is improved for enhancing the computation efficiency and keeping the adaptive performance.The self-tuning strategies of the linear servo system speed loop are proposed:In off-line auto-tuning process, the performance evaluation of the transition process is chosen as the objective function, and the identification results of the controlled object of speed loop were used by a least squares algorithm, then, the off-line auto-tuning of the speed loop is completed by optimizing the identification model. In on-line self-tuning process, the quadratic performance evaluation is chosen as the objective function, the controlled object model of the speed loop was identified on-line, using the identification model to predict the performance of the speed loop, and the speed-loop control parameter were tuned according to the predictions.The anti-disturbance strategies of the linear servo system position loop are proposed: To suppress the force ripple, when PMSLM keeps a low and constant velocity, the dominant frequency of the force ripple can be firstly obtained by applying a Fast Fourier Transform (FFT) to analyze the force current; then, the coefficients of the force ripple will be identified by using an recursive least squares algorithm and an JITL algorithm; finally, the identification result is using for compensating the force current. To suppress the load force, a model reference adaptive identification algorithm is used to estimate the total mass of the mover, on the basis of the identification result, a disturbance observer is built to obtain the desired force current item.The linear servo system platform is constructed, and the proposed strategies of on-line identification、self-tuning and anti-disturbance are carried out. The experimental results demonstrate the accuracy, the stability and the suitability.