• Geomechanics and Engineering
• /
• 제30권2호
• /
• pp.107-121
• /
• 2022

Coal pillar assessment is of broad importance to underground engineering structure, as the pillar failure can lead to enormous disasters. Because of the highly non-linear correlation between the pillar failure and its influential attributes, conventional forecasting techniques cannot generate accurate outcomes. To approximate the complex behavior of coal pillar, this paper elucidates a new idea to forecast the underground coal pillar stability using combined unsupervised-supervised learning. In order to build a database of the study, a total of 90 patterns of pillar cases were collected from authentic engineering structures. A state-of-the art feature depletion method, t-distribution symmetric neighbor embedding (t-SNE) has been employed to reduce significance of actual data features. Consequently, an unsupervised machine learning technique K-mean clustering was followed to reassign the t-SNE dimensionality reduced data in order to compute the relative class of coal pillar cases. Following that, the reassign dataset was divided into two parts: 70 percent for training dataset and 30 percent for testing dataset, respectively. The accuracy of the predicted data was then examined using support vector classifier (SVC) model performance measures such as precision, recall, and f₁-score. As a result, the proposed model can be employed for properly predicting the pillar failure class in a variety of underground rock engineering projects.

• Journal of Electrical Engineering and Technology
• /
• 제5권2호
• /
• pp.179-190
• /
• 2010

This paper describes a new stochastic heuristic algorithm in engineering problem optimization especially in power system applications. An improved particle swarm optimization (PSO) called adaptive particle swarm optimization (APSO), mixed with simulated annealing (SA), is introduced and referred to as APSO-SA. This algorithm uses a novel PSO algorithm (APSO) to increase the convergence rate and incorporate the ability of SA to avoid being trapped in a local optimum. The APSO-SA algorithm efficiency is verified using some benchmark functions. This paper presents the application of APSO-SA to find the optimal location, type and size of flexible AC transmission system devices. Two types of FACTS devices, the thyristor controlled series capacitor (TCSC) and the static VAR compensator (SVC), are considered. The main objectives of the presented method are increasing the voltage stability index and over load factor, decreasing the cost of investment and total real power losses in the power system. In this regard, two cases are considered: single-type devices (same type of FACTS devices) and multi-type devices (combination of TCSC, SVC). Using the proposed method, the locations, type and sizes of FACTS devices are obtained to reach the optimal objective function. The APSO-SA is used to solve the above non.linear programming optimization problem for better accuracy and fast convergence and its results are compared with results of conventional PSO. The presented method expands the search space, improves performance and accelerates to the speed convergence, in comparison with the conventional PSO algorithm. The optimization results are compared with the standard PSO method. This comparison confirms the efficiency and validity of the proposed method. The proposed approach is examined and tested on IEEE 14 bus systems by MATLAB software. Numerical results demonstrate that the APSO-SA is fast and has a much lower computational cost.

• Investigative Magnetic Resonance Imaging
• /
• 제23권1호
• /
• pp.38-45
• /
• 2019

Purpose: To demonstrate the high-resolution numerical simulation of the respiration-induced dynamic $B_0$ shift in the head using generalized susceptibility voxel convolution (gSVC). Materials and Methods: Previous dynamic $B_0$ simulation research has been limited to low-resolution numerical models due to the large computational demands of conventional Fourier-based $B_0$ calculation methods. Here, we show that a recently-proposed gSVC method can simulate dynamic $B_0$ maps from a realistic breathing human body model with high spatiotemporal resolution in a time-efficient manner. For a human body model, we used the Extended Cardiac And Torso (XCAT) phantom originally developed for computed tomography. The spatial resolution (voxel size) was kept isotropic and varied from 1 to 10 mm. We calculated $B_0$ maps in the brain of the model at 10 equally spaced points in a respiration cycle and analyzed the spatial gradients of each of them. The results were compared with experimental measurements in the literature. Results: The simulation predicted a maximum temporal variation of the $B_0$ shift in the brain of about 7 Hz at 7T. The magnitudes of the respiration-induced $B_0$ gradient in the x (right/left), y (anterior/posterior), and z (head/feet) directions determined by volumetric linear fitting, were < 0.01 Hz/cm, 0.18 Hz/cm, and 0.26 Hz/cm, respectively. These compared favorably with previous reports. We found that simulation voxel sizes greater than 5 mm can produce unreliable results. Conclusion: We have presented an efficient simulation framework for respiration-induced $B_0$ variation in the head. The method can be used to predict $B_0$ shifts with high spatiotemporal resolution under different breathing conditions and aid in the design of dynamic $B_0$ compensation strategies.

• 한국산학기술학회논문지
• /
• 제21권1호
• /
• pp.169-177
• /
• 2020

과학기술 분야의 연구·개발 결과는 연구보고서 형태로 국가과학기술정보서비스(NTIS)에 제출된다. 각 연구보고서는 국가과학기술 표준 분류체계 (K-NSCC)에 따른 분류코드를 가지고 있는데, 보고서 작성자가 제출 시에 수동으로 입력하게끔 되어있다. 하지만 2000여 개가 넘는 세분류를 가지고 있기에, 분류체계에 대한 정확한 이해가 없이는 부정확한 분류코드를 선택하기 십상이다. 새로이 수집되는 연구보고서의 양과 다양성을 고려해 볼 때, 이들을 기계적으로 보다 정확하게 분류할 수 있다면 보고서 제출자의 수고를 덜어줄 수 있을 뿐만 아니라, 다른 부가 가치적인 분석 서비스들과의 연계가 수월할 것이다. 하지만, 국내에서 과학기술표준 분류체계에 기반을 둔 문서 자동 분류 연구 사례는 거의 없으며 공개된 학습데이터도 전무하다. 본 연구는 KISTI가 보유하고 있는 최근 5년간 (2013년~2017년) NTIS 연구보고서 메타정보를 활용한 최초의 시도로써, 방대한 과학기술표준 분류체계를 기반으로 하는 국내 연구보고서들을 대상으로 높은 성능을 보이는 문서 자동 분류기법을 도출하는 연구를 진행하였다. 이를 위해, 과학기술 표준분류 체계에서 과학기술 분야의 연구보고서를 분류하기에 적합한 중분류 210여 개를 선별하였으며, 연구보고서 메타 데이터의 특성을 고려한 전처리를 진행하였다. 특히, 가장 영향력 있는 필드인 과제명(제목)과 키워드만을 이용한 TK_CNN 기반의 딥러닝 기법을 제안한다. 제안 모델은 텍스트 분류에서 좋은 성능을 보이고 있는 기계학습법들 (예, Linear SVC, CNN, GRU등)과 비교하였으며, Top-3 F1점수 기준으로 1~7%에 이르는 성능 우위를 확인하였다.