We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Forgetting Factor Least Square Parameter Identification Based on Tool Servo Speed Tracking of the Milling Process.
- Authors
Ke Xu; Jie Yang; Weiwei Fan; Chuansheng Tang; Tao Li
- Abstract
The tool servo system of the computer numerically controlled (CNC) machine milling process is a complex nonlinear system composed of a servomechanism, a cutting process, and a detection device. In the actual machining process, accurately establishing a system model is difficult due to the coupling of parameters and nonlinear factors. Simultaneously, system parameters change with the working environment (e.g., resistance increases with an increase in temperature), leading to a decrease in the surface quality of the work piece. To improve the online identification accuracy of milling process model parameters and effectively increase the influence of parameter changes on system performance, a tool speed model prediction adaptive tracking method based on forgetting factor least square identification was proposed in this study. First, the model was discretized in accordance with the structure and characteristics of the tool servo feed system in the machining process, and the model parameters of the system were identified using the forgetting factor least square method. Second, a model predictive tracking method based on adaptive parameter estimation was designed on the basis of the discrete model of the system. Lastly, the effectiveness of the proposed method in model parameter identification and tool speed tracking was verified via numerical simulation. Results show that when uncertain factors, such as noise exist in the system, the least square identification method based on the forgetting factor can more quickly and accurately realize the model parameter identification of a tool servo feed system in the milling process than the stochastic gradient (SG) identification method. Moreover, identification accuracy is 30 times higher than that of the SG identification method. The model predictive tracking control method based on forgetting factor least square identification can quickly track tool speed without overshooting in 0.035 s. By contrast, the traditional minimum variance predictive control is completely invalid in the actual stage. The proposed method exhibits high accuracy in tool speed tracking and strong robustness to changes in model parameters.
- Subjects
PARAMETER identification; LEAST squares; BORING &; drilling (Earth &; rocks); PARAMETER estimation; SPEED; AUTOMATION; DISCRETE systems
- Publication
Journal of Engineering Science & Technology Review, 2020, Vol 13, Issue 4, p22
- ISSN
1791-2377
- Publication type
Article
- DOI
10.25103/jestr.134.02