• 동력기계공학회지
• /
• 제21권3호
• /
• pp.60-65
• /
• 2017

A GA is one of the best method to find optimal value in searching area. A GA is driven by probabilistic selection that based on the survival of the fittest. So this algorithm need a huge solving time even if it can be used lots of optimizing problem such as structural design, machine learning, system's identification and so on. This GA's characteristic constrain the program to drive offline. Some studies try to use this algorithm on online or reduce the GA's running time with parallel GA or micro GA. Unfortunately these studies still didn't reduce amount of fitness solving. If the chromosome was imported to the system, it affected system's stability. And when the control system uses online GA, it also doesn't have enough learning time. In this study, try to find stability criterion to reduce the chromosome's affection and find the characteristic of the number of population and generation when GA was driven into the wide searching area.

• Smart Structures and Systems
• /
• 제17권2호
• /
• pp.167-194
• /
• 2016

This paper investigates the Synergetics based Damage Detection Method (SDDM) for frame structures by using surface-bonded PZT (Lead Zirconate Titanate) patches. After analyzing the mechanism of pattern recognition from Synergetics, the operating framework with cooperation-competition-update process of SDDM was proposed. First, the dynamic identification equation of structural conditions was established and the adjoint vector (AV) set of original vector (OV) set was obtained by Generalized Inverse Matrix (GIM).Then, the order parameter equation and its evolution process were deduced through the strict mathematics ratiocination. Moreover, in order to complete online structural condition update feature, the iterative update algorithm was presented. Subsequently, the pathway in which SDDM was realized through the modified Synergetic Neural Network (SNN) was introduced and its assessment indices were confirmed. Finally, the experimental platform with a two-story frame structure was set up. The performances of the proposed methodology were tested for damage identifications by loosening various screw nuts group scenarios. The experiments were conducted in different damage degrees, the disturbance environment and the noisy environment, respectively. The results show the feasibility of SDDM using piezoceramic sensors and actuators, and demonstrate a strong ability of anti-disturbance and anti-noise in frame structure applications. This proposed approach can be extended to the similar structures for damage identification.

• 전력전자학회논문지
• /
• 제6권1호
• /
• pp.98-106
• /
• 2001

본 논문에서는 영구자석 동기 전동기의 정밀 속도 제어의 방법으로 외란 관측기를 이용한 외란 보상방법과 파라미터 추정에 의해 보상기의 이득을 조절하도록 함으로서, 외란이 없는 등가 지표시스템의 응답 특성을 추정하도록 제안하였다. 외란 관측기에 의한 보상방법은 잘 알려진 데드비트 외란 관측기를 이용하였으며 잡음에 약한 데드비트 관측기의 단점을 보완하기 위하여 후단필터로서 MA처리를 통하여 잡음에 대한 영향을 줄이도록 하였다. 또한 관측기의 단점을 보완하기 위하여 후단필터로서 MA처리를 통하여 잡음에 대한 영향을 줄이도록 하였다. 또한 관측기의 파라미터와 실제 시스템의 파라미터의 차이로 발생하는 외란 추정 오차를 줄이고자 실제 시스템과 파라미터 보상기로 구성된 등가 시스템이 지표 시스템이 되도록 구성하였다. 시스템에 사용된 RLS파라미터 추정기는 외란에 의하여 편향된 추정 특성을 가진다. 이러한 파라미터 추정문제에 대하여 파라미터 추정기가 높은 성능을 갖는 데드비트 외란 관측기를 포함하도록 함으로서 외란에 의한 문제를 해결하였다. 이와 같이 제안된 제어기는 외란 및 파라미터 변화를 갖는 시스템에서 강인한 고정밀 제어를 할 수 있으며, 이의 안정성과 효용성을 컴퓨터를 이용한 모의 실험과 TMS320C31이 내장된 DS1102 DSP 보드를 이용하여 실험으로써 보였다.

• Journal of Power Electronics
• /
• 제10권5호
• /
• pp.505-517
• /
• 2010

In this paper, an intelligent sliding-mode speed controller for achieving favorable decoupling control and high precision speed tracking performance of permanent-magnet synchronous motor (PMSM) drives is proposed. The intelligent controller consists of a sliding-mode controller (SMC) in the speed feed-back loop in addition to an on-line trained wavelet-neural-network controller (WNNC) connected in parallel with the SMC to construct a robust wavelet-neural-network controller (RWNNC). The RWNNC combines the merits of a SMC with the robust characteristics and a WNNC, which combines artificial neural networks for their online learning ability and wavelet decomposition for its identification ability. Theoretical analyses of both SMC and WNNC speed controllers are developed. The WNN is utilized to predict the uncertain system dynamics to relax the requirement of uncertainty bound in the design of a SMC. A computer simulation is developed to demonstrate the effectiveness of the proposed intelligent sliding mode speed controller. An experimental system is established to verify the effectiveness of the proposed control system. All of the control algorithms are implemented on a TMS320C31 DSP-based control computer. The simulated and experimental results confirm that the proposed RWNNC grants robust performance and precise response regardless of load disturbances and PMSM parameter uncertainties.

• Smart Structures and Systems
• /
• 제25권5호
• /
• pp.569-580
• /
• 2020

The identification of delays and delay compensation are critical problems in real-time hybrid simulations (RTHS). Conventional delay compensation methods are mostly based on the assumption of a constant delay. However, the system delay may vary during tests owing to the nonlinearity of the loading system and/or the behavioral variations of the specimen. To address this issue, this study presents an adaptive delay compensation method based on a discrete model of the loading system. In particular, the parameters of this discrete model are identified and updated online with the least-squares method to represent a servo hydraulic loading system. Furthermore, based on this model, the system delays are compensated for by generating system commands using the desired displacements, achieved displacements, and previous displacement commands. This method is more general than the existing compensation methods because it can predict commands based on multiple displacement categories. Moreover, this method is straightforward and suitable for implementation on digital signal processing boards because it relies solely on the displacements rather than on velocity and/or acceleration data. The virtual and real RTHS results show that the studied method exhibits satisfactory estimation smoothness and compensation accuracy. Furthermore, considering the measurement noise, the low-order parameter models of this method are more favorable than that the high-order parameter models.