Tian-Yau Wu

The objective of this research is to investigate the feasibility of utilizing tooling vibration signals and cutting parameters to predict workpiece surface profiles in the milling process, enabling a comprehensive evaluation of surface quality. Information-contained (IC) signals were synthesized from milling vibration measurements using principal component analysis (PCA) and empirical mode decomposition (EMD). Statistical features and sweeping-frequency features were then extracted from individual intrinsic mode functions (IMFs) of the IC signals in both time and frequency domains. Subsequently, independent long short-term memory (LSTM) networks were used to estimate the individual IMFs of the surface profiles, incorporating the milling parameters and selected signal features as inputs. Results demonstrated that mid-low frequency components of surface profiles could be estimated accurately with an average mean absolute percentage error (MAPE) of 14.72%, while low-frequency components achieved an average MAPE of 7.79%.

This research investigated the feasibility of applying hardware order-tracking (HOT) and segmentalized amplitude normalization (SAN) to enhance the diagnosis of multiple bearing defects at different levels under varying rotation speeds. The vibration of operating bearings may present an energy variation phenomenon due to different levels of bearing defects, while the fluctuation of vibration amplitude may be attributable to changes in rotation speeds. These two factors inevitably interfere with each other when diagnosing bearing defects at multiple levels and classes under varying rotation speeds. In this paper, the research focuses on conducting an in-depth analysis of signal signatures, followed by providing a physical insight into feature extraction. Consequently, it enables the application of simple machine learning methods to accurately diagnose various bearing defects, even when dealing with significantly different patterns in training and testing data due to varying rotation speeds. To verify the effectiveness of the proposed SAN method for cases involving varying rotation speeds, the training and testing sets used datasets (vibration measurements) corresponding to different rotation speed profiles. The experimental and analytical results revealed that the proposed SAN method can normalize datasets with disparate vibration patterns, and alleviate the coupling of vibration energy variation and shaft rotation speed. This enhancement resulted in approximately 18.6% increase in the accuracy of bearing diagnosis for cases involving varying rotation speeds.

The active vibration isolation of a quarter-vehicle suspension system using an electromagnetic damper was achieved by designing the feedback gain of a linear quadratic regulator (LQR) and H-∞ controllers in modal coordinates. The efficacy of vibration isolation was analyzed in the frequency domain, while the performance of vibration suppression and passengers’ comfort quality were evaluated in the time domain. The research results demonstrated that by employing modal coordinate transformation in state space, the designated system’s resonance peaks can be directly suppressed in the frequency domain. The findings also indicate that the active control in modal coordinates can achieve a higher vibration suppression percentage per unit control effort in comparison to traditional active control in physical space. In other words, the modal controller is more efficient in vibration suppression than the controller designed in physical space.

The objective of this research is to investigate the feasibility of utilizing the Elman neural network to predict the surface roughness in the milling process of Inconel 718 and then optimizing the cutting parameters through the particle swarm optimization (PSO) algorithm according to the different surface roughness requirements. The prediction of surface roughness includes the feature extraction of vibration measurements as well as the current signals, the feature selection using correlation analysis and the prediction of surface roughness through the Elman artificial neural network. Based on the prediction model of surface roughness, the cutting parameters were optimized in order to obtain the maximal feed rate according to different surface roughness constraints. The experiment results show that the surface roughness of Inconel 718 can be accurately predicted in the milling process and thereafter the optimal cutting parameter combination can be determined to accelerate the milling process.

The purpose of this research is to investigate the feasibility of evaluating the human’s balancing ability by means of the human body’s swaying acceleration measurements instead of the traditional center-of-pressure (COP) measurement. The COP measurement has been used broadly for assessing the balance ability of patients in hospitals. However, the force plate system which is employed to measure the COP signals of the human body is generally restrictive due to the very high cost as well as the bulky portability. In this study, the balancing ability of the human body was evaluated through the measurements of a capacitive accelerometer. The segmentalized principal components analysis (sPCA) was employed to reduce the influence of the gravity component in acceleration measurement projected onto the horizontal components while the accelerometer inevitably tilts. The signal relationship between the COP and the acceleration was derived, so that the swaying acceleration measurements of human body can be utilized to evaluate the human body’s balancing ability.

The article Prediction of surface roughness in milling process using vibration signal analysis and artificial neural network, written by T. Y. Wu and K. W. Lei, was originally published electronically on the publisher’s internet portal (currently SpringerLink) on 03 January 2019 without open access.