Papers by Author: Ki Hong Shin

Abstract: The Location Template Matching (LTM) method is known to be simple and straightforward to use for source localization problem. We have found that the method is not as simple as one might think, but there are many practical aspects that must be carefully considered. The most important parameter is the spatial resolution of the pre-defined reference points. The other parameters are the sampling rate, the time alignment between the signals, the frequency bandwidth of a signal, and the length of data. It is also shown that the accuracy of the LTM method can be improved if a couple of more sensors are used. This paper discusses these practical considerations in detail for the purpose of providing a proper guidance of using the LTM method.

Abstract: In-plane characteristics of a single degree of freedom frictional system with a compressed spring are considered. The compressed spring is vertically installed to the mass moving horizontally along the friction interface. The compressed spring can introduce a nonlinear negative stiffness into the in-plane motion. The resulting system has a multiple equilibrium points; an unstable point at the center and two stable points on either side. It is shown that two stable equilibrium points can be separated far apart by increasing the compression ratio and the stiffness of the spring. The friction system is often characterized by stick and slip motions that cause unfavorable effects such as wear, noise, and chatter etc. It is demonstrated that increasing the compression ratio and the stiffness of the spring results in decreasing the size of the stick regions.

Abstract: In general, dynamic friction coefficient in a dry surface condition is dependent on the relative velocity between two sliding bodies. A conventional method of estimating the velocity dependent dynamic friction coefficient is time consuming and requires a special jig to measure the friction forces. In this paper, a new technique called the state space mapping method is proposed based on the nonlinear dynamics of a 1-DOF friction oscillator. One body is constructed as a single degree of freedom system and another body is formed as a moving base that may be built as a rotating disk or a horizontally oscillating plate. The resulting friction induced vibration system is used to estimate the dynamic friction coefficient. The measured vibration signals are utilized to construct a three dimensional plot of acceleration versus state variables. Then, the velocity dependent dynamic friction coefficient can be estimated from the plot. The natural frequency can also be estimated from the plot, which can be used to verify the quality of the results.

Abstract: This paper presents a theoretical and experimental study of a non-collocated pair of piezopolymer PVDF sensor and piezoceramic PZT actuator, which are bonded on a cantilever beam, in order to suppress unwanted vibration at the tip of the beam. The PZT actuator patch was bonded near the clamped part and the PVDF sensor, which was triangularly shaped, was bonded on the other part of the beam. This is because the triangular PVDF sensor is known that it can detect the tip velocity of a cantilever beam. Because the arrangement of the sensor and actuator pair is not collocated and overlapped each other, the pair can avoid so called "in-plane coupling", which can be found at a matched piezoelectric sensor and actuator pair and restricts the stability and performance of direct velocity feedback control. The test beam is made of aluminum with the dimension of 200 × 20 × 2 mm. Before control, the sensor-actuator frequency response function is confirmed to have a nice phase response without accumulation in a reasonable frequency range of up to 5000 Hz. The feedback control attenuates the magnitude of the first two resonances in the error spectrum of about 6 -7 dB.