• Structural Engineering and Mechanics
• /
• 제38권5호
• /
• pp.651-674
• /
• 2011

This paper presents a novel method for predicting the failure probability of structural or mechanical systems subjected to random loads and material properties involving multiple design points. The method involves Multicut High Dimensional Model Representation (Multicut-HDMR) technique in conjunction with moving least squares to approximate the original implicit limit state/performance function with an explicit function. Depending on the order chosen sometimes truncated Cut-HDMR expansion is unable to approximate the original implicit limit state/performance function when multiple design points exist on the limit state/performance function or when the problem domain is large. Multicut-HDMR addresses this problem by using multiple reference points to improve accuracy of the approximate limit state/performance function. Numerical examples show the accuracy and efficiency of the proposed approach in estimating the failure probability.

• International Journal of Contents
• /
• 제13권2호
• /
• pp.21-28
• /
• 2017

Abnormal data in the manufacturing process makes it difficult to find useful information that can be applied in data management for the manufacturing industry. It causes various problems in the daily process of production. An issue from the abnormal data can be handled by our method that uses big data and visualization. Visualization is a new technology that transforms data representation into a two-dimensional representation. Nowadays, many newly developed technologies provide data analysis, algorithm, optimization, and high efficiency, and they meet user requirements. We propose combined production of the data visualization approach that uses integrative visualization of sources of abnormal pattern analysis results. The perceived idea of the proposed approach can solve the problem as it also works for big data. It can also improve the performance and understanding by using visualization and solving issues that occur in the manufacturing process with a calendar heat map.

• 유체기계공업학회:학술대회논문집
• /
• 유체기계공업학회 2006년 제4회 한국유체공학학술대회 논문집
• /
• pp.33-36
• /
• 2006

As the computing environment is rapidly improved, the interests of CFD are gradually focused on large-scale computation over complex geometry. Keeping pace with the trend, essential computational tools to obtain solutions of complex aerospace flow analysis and design problems are examined. An accurate and efficient flow analysis and design codes for large-scale aerospace problem are presented in this work. With regard to original numerical schemes for flow analysis, high-fidelity flux schemes such as RoeM, AUSMPW+ and higher order interpolation schemes such as MLP (Multi-dimensional Limiting Process) are presented. Concerning the grid representation method, a general-purpose basis code which can handle multi-block system and overset grid system simultaneously is constructed. In respect to design optimization, the importance of turbulent sensitivity is investigated. And design tools to predict highly turbulent flows and its sensitivity accurately by fully differentiating turbulent transport equations are presented. Especially, a new sensitivity analysis treatment and geometric representation method to resolve the basic flow characteristics are presented. Exploiting these tools, the capability of the proposed approach to handle complex aerospace simulation and design problems is tested by computing several flow analysis and design problems.

• Smart Structures and Systems
• /
• 제21권5호
• /
• pp.601-609
• /
• 2018

The estimated probabilistic model of wind data based on the conventional approach may have high discrepancy compared with the true distribution because of the uncertainty caused by the instrument error and limited monitoring data. A sequential quadratic programming (SQP) algorithm-based finite mixture modeling method has been developed in the companion paper and is conducted to formulate the joint probability density function (PDF) of wind speed and direction using the wind monitoring data of the investigated bridge. The established bivariate model of wind speed and direction only represents the features of available wind monitoring data. To characterize the stochastic properties of the wind parameters with the subsequent wind monitoring data, in this study, Bayesian inference approach considering the uncertainty is proposed to update the wind parameters in the bivariate probabilistic model. The slice sampling algorithm of Markov chain Monte Carlo (MCMC) method is applied to establish the multi-dimensional and complex posterior distribution which is analytically intractable. The numerical simulation examples for univariate and bivariate models are carried out to verify the effectiveness of the proposed method. In addition, the proposed Bayesian inference approach is used to update and optimize the parameters in the bivariate model using the wind monitoring data from the investigated bridge. The results indicate that the proposed Bayesian inference approach is feasible and can be employed to predict the bivariate distribution of wind speed and direction with limited monitoring data.

• Structural Engineering and Mechanics
• /
• 제75권4호
• /
• pp.519-527
• /
• 2020

Dynamic analysis of a vehicle-track coupling system is important to structural design, damage detection and condition assessment of the structural system. Deterministic analysis of the vehicle-track coupling system has been extensively studied in the past, however, the structural parameters of the coupling system have uncertainties in engineering practices. It is essential to treat the parameters of the vehicle-track coupling system with consideration of uncertainties. In this paper, a method for predicting the bounds of the vehicle-track coupling system responses with uncertain parameters is presented. The uncertain system parameters are modeled as fuzzy variables instead of conventional random variables with known probability distributions. Then, the dynamic response functions of the coupling system are transformed into a component function based on the high dimensional representation approximation. The Lagrange interpolation method is used to approximate the component function. Finally, the bounds of the system's dynamic responses can be predicted by using Monte Carlo method for the interpolation polynomials of the Lagrange interpolation function. A numerical example is introduced to illustrate the ability of the proposed method to predict the bounds of the system's dynamic responses, and the results are compared with the direct Monte Carlo method. The results show that the proposed method is effective and efficient to predict the bounds of the system's dynamic responses with fuzzy variables.

• Structural Engineering and Mechanics
• /
• 제26권6호
• /
• pp.687-707
• /
• 2007

High-tech facilities engaged in the production of semiconductors and optical microscopes are extremely expensive, which may require time-domain analysis for seismic resistant design in consideration of the most critical directions of seismic ground motions. This paper presents a framework for generating three-dimensional critical seismic ground acceleration time histories compatible with the response spectra specified in seismic design codes. The most critical directions of seismic ground motions associated with the maximum response of a high-tech facility are first identified. A new numerical method is then proposed to derive the power spectrum density functions of ground accelerations which are compatible with the response spectra specified in seismic design codes in critical directions. The ground acceleration time histories for the high-tech facility along the structural axes are generated by applying the spectral representation method to the power spectrum density function matrix and then multiplied by envelope functions to consider nonstationarity of ground motions. The proposed framework is finally applied to a typical three-story high-tech facility, and the numerical results demonstrate the feasibility of the proposed approach.

• 한국데이터정보과학회:학술대회논문집
• /
• 한국데이터정보과학회 2005년도 춘계학술대회
• /
• pp.127-134
• /
• 2005

A suitable representation for the conceptual simplicity of the data in statistics and signal processing is essential for a subsequent analysis such as prediction, pattern recognition, and spatial analysis. Independent component analysis (ICA) is a statistical method for transforming an observed high-dimensional multivariate data into statistically independent components. ICA has been applied increasingly in wide fields of spectrum application since ICA is able to extract unknown components of a mixture from spectra. We focus on application of ICA for separating independent sources and predicting each composition using extracted components. The theory of ICA is introduced and an application to a metal surface spectra data will be described, where subsequent analysis using non-negative least square method is performed to predict composition ratio of each sample. Furthermore, some simulation experiments are performed to demonstrate the performance of the proposed approach.

• Computers and Concrete
• /
• 제30권5호
• /
• pp.339-355
• /
• 2022

The high cost of ultra-high-performance engineered cementitious composite (UHP-ECC) is currently a crucial issue, especially in terms of the polyethylene (PE) fibers use. In this paper, cheap calcium carbonate whiskers (CW) were evaluated on the feasibility of hybrid with PE fibers. Diverse combinations of PE fibers and CW were employed to investigate the multi-scale enhancement on the UHP-ECC performance. A probabilistic-based UHP-ECC tensile strain reliability analysis approach was utilized, which was in general agreement with the experimental results. Furthermore, a multi-dimensional integrated representation was conducted for the comprehensive assessment of UHP-ECC. Results illustrated that CW improved the compressive strength and energy dissipation capacity of UHP-ECC owing to the microscopic strengthening mechanism. CW and PE fiber further promoted the saturated cracking of composite by multi-scale crack arresting effect. In particular, PE1.75-CW0.5 specimen possessed the best overall performance. The ultimate cracking width of PE1.75-CW0.5 group had 98 ㎛, which was 46.18% lower compared to PE2-CW0 group, the 28d compressive strength were slightly improved, the tensile strain capacity was comparable to that of PE2-CW0 group. The results above demonstrated that combinations of PE fiber and CW could significantly enhance the comprehensive performance of UHP-ECC, which was beneficial for large-scale engineering applications.

• 한국콘텐츠학회논문지
• /
• 제9권3호
• /
• pp.1-8
• /
• 2009

본 논문에서는 퍼지 논리 기반의 유전자 알고리즘(GA)과 의미 벡터 확장 기술을 이용한 문서 클러스터링 시스템을 제안한다. GA에 관련된 여러 논문에서 이미 알려졌듯이 GA알고리즘의 성공 여부는 군체의 다양성과 수렴하는 능력에 따라 결정된다. 이러한 두 인자 사이의 영향력을 조절하기 위하여 우리는 퍼지 논리 기반의 연산자를 사용한다. 전통적인 문서 클러스터링 알고리즘에서 문서를 나타내기 위한 가장 일반적이고 직선적인 방법은 벡터 공간 모델이다. 그러나 이 방법은 다차원 특징 공간의 원인이 될 뿐만 아니라, 클러스터링의 정확성에 영향을 미칠 수 있는, 단어 간의 의미상 관계성을 무시한다. 본 논문에서는 LSA를 사용하여 문서를 관련되는 의미상의 벡터 개념으로 확장시킨다. 또한 이것은 벡터의 크기를 크게 줄일 수 있다. 본 논문에서 제안한 클러스터링 알고리즘을 테스트하기 위하여 20개의 뉴스 그룹과 로이터 데이터를 사용했다. 제안된 방법은 문서를 표현하는 다양한 환경에서 일반적인 GA보다 더 나은 결과를 보여준다.

• Smart Structures and Systems
• /
• 제29권1호
• /
• pp.181-193
• /
• 2022

Data-driven structural health monitoring (SHM) of civil infrastructure can be used to continuously assess the state of a structure, allowing preemptive safety measures to be carried out. Long-term monitoring of large-scale civil infrastructure often involves data-collection using a network of numerous sensors of various types. Malfunctioning sensors in the network are common, which can disrupt the condition assessment and even lead to false-negative indications of damage. The overwhelming size of the data collected renders manual approaches to ensure data quality intractable. The task of detecting and classifying an anomaly in the raw data is non-trivial. We propose an approach to automate this task, improving upon the previously developed technique of image-based pre-processing on one-dimensional (1D) data by enriching the features of the neural network input data with multiple channels. In particular, feature engineering is employed to convert the measured time histories into a 3-channel image comprised of (i) the time history, (ii) the spectrogram, and (iii) the probability density function representation of the signal. To demonstrate this approach, a CNN model is designed and trained on a dataset consisting of acceleration records of sensors installed on a long-span bridge, with the goal of fault detection and classification. The effect of imbalance in anomaly patterns observed is studied to better account for unseen test cases. The proposed framework achieves high overall accuracy and recall even when tested on an unseen dataset that is much larger than the samples used for training, offering a viable solution for implementation on full-scale structures where limited labeled-training data is available.