Identification of mechanical parameters for position-controlled servo systems using sinusoidal commands
- PDF / 2,004,585 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 57 Downloads / 185 Views
ORIGINAL ARTICLE
Identification of mechanical parameters for position‑controlled servo systems using sinusoidal commands Min‑Sik Yoo1 · Seung‑Cheol Choi1 · Sang‑Woo Park1 · Young‑Doo Yoon1 Received: 9 June 2020 / Revised: 2 August 2020 / Accepted: 8 August 2020 © The Korean Institute of Power Electronics 2020
Abstract This paper proposes an identification method for mechanical parameters based on position control. To improve motion control performance, the moment of inertia and friction components must be considered. Based on mechanical equations, the proposed method estimates the moment of inertia, viscous friction coefficient, and Coulomb friction in the off-line state. Mechanical parameters are obtained from the integral values for the products of the torque, speed, and position using the 90° phase relationship between acceleration and velocity. Simulation and experimental results demonstrate the validity and accuracy of the proposed method. Since its implementation is simple, this method can be applied easily to industry. Keywords Identification of mechanical parameters · Moment of inertia · Viscous friction coefficient · Coulomb friction · Speed reducer · Off-line identification
1 Introduction Manufacturing equipment has high productivity and accuracy requirements. In addition, factory environments must be safe for workers, since they work in close proximity to machinery [1, 2]. To achieve these goals, robots such as collaborative robots [3] are used as production equipment. These robots require high performance in terms of motion control. The motion control of a robot is achieved through speed and position control, while considering the mechanical system. However, it is difficult to accurately model friction components due to nonlinearities. Therefore, in the case of a system with a small influence from friction, the controller is often designed while ignoring the friction components [4]. However, for systems with high friction effects, friction components should be considered in the controller design. To improve motion control performance, torque feed-forward compensation [5], disturbance observers [6, 7], and anti-windup control [8] can be adopted. To implement these methods, mechanical parameters are required.
* Young‑Doo Yoon [email protected] 1
Department of Automotive Engineering, Hanyang University, Seoul, Korea
The parameters of a mechanical system change depending on the load and operating conditions [9]. Considering parameter variations, many on-line methods for estimating parameters have been published, such as state observers [10–12], adaptive controls [13–15], and Kalman filters [16]. However, some of these methods are complex to implement. In addition, unstable control characteristics can appear in controllers that use the estimated mechanical parameters in the transient state. Due to the disadvantages of on-line methods, off-line identification methods have been studied. Most off-line methods have performed motion control using predetermined motion profiles, which were based on sp
Data Loading...
