• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.1
• /
• pp.12-18
• /
• 2011

In this paper, we proposed Interval Type-2 polynomial Radial Basis Function Neural Networks. In the receptive filed of hidden layer, Interval Type-2 fuzzy set is used. The characteristic of Interval Type-2 fuzzy set has Footprint Of Uncertainly(FOU), which denotes a certain level of robustness in the presence of un-known information when compared with the type-1 fuzzy set. In order to improve the performance of proposed model, we used the linear polynomial function as connection weight of network. The parameters such as center values of receptive field, constant deviation, and connection weight between hidden layer and output layer are optimized by Conjugate Gradient Method(CGM) and Space Search Evolutionary Algorithm(SSEA). The proposed model is applied to gas furnace dataset and its result are compared with those reported in the previous studies.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.58 no.2
• /
• pp.399-406
• /
• 2009

In this study, Polynomial Radial Basis Function Neural Network(pRBFNN) based on Fuzzy Inference System is designed and its parameters such as learning rate, momentum coefficient, and distributed weight (width of RBF) are optimized by means of Particle Swarm Optimization. The proposed model can be expressed as three functional module that consists of condition part, conclusion part, and inference part in the viewpoint of fuzzy rule formed in 'If-then'. In the condition part of pRBFNN as a fuzzy rule, input space is partitioned by defining kernel functions (RBFs). Here, the structure of kernel functions, namely, RBF is generated from HCM clustering algorithm. We use Gaussian type and Inverse multiquadratic type as a RBF. Besides these types of RBF, Conic RBF is also proposed and used as a kernel function. Also, in order to reflect the characteristic of dataset when partitioning input space, we consider the width of RBF defined by standard deviation of dataset. In the conclusion part, the connection weights of pRBFNN are represented as a polynomial which is the extended structure of the general RBF neural network with constant as a connection weights. Finally, the output of model is decided by the fuzzy inference of the inference part of pRBFNN. In order to evaluate the proposed model, nonlinear function with 2 inputs, waster water dataset and gas furnace time series dataset are used and the results of pRBFNN are compared with some previous models. Approximation as well as generalization abilities are discussed with these results.

• Journal of Korean Neurosurgical Society
• /
• v.64 no.1
• /
• pp.60-68
• /
• 2021

Objective : Until recently, the transfemoral approach (TFA) was used as the primary method of arterial approach in acute ischemic stroke (AIS). However, TFA resulted in longer reperfusion times and worse outcomes in the mechanical thrombectomy (MT) of patients with complex aortic arches and significant carotid tortuosity. We found that the transradial approach (TRA) is a more favorable alternative approach for MT in such cases. Methods : We performed a retrospective review of our institutional database to identify 202 patients who underwent MT for AIS between February 2015 and December 2019. Patient characteristics, cause of TFA failure, procedure time, intra-procedural complications, and outcomes were recorded. Results : Eleven (5.4%) of 202 patients, who underwent MT for AIS, crossed over to TRA for recanalization, and eight (72%) of 11 achieved successful recanalization (≥modified Treatment in Cerebral Infarction 2b). The mean age (mean±standard deviation [median]) was 82.3±6.6 (76) years, and five of the 11 patients were male. The last seen normal to puncture time was 467.9±264.72 (264) minutes; baseline National Institutes of Health Stroke Scale score was 28.9±14.5 (16). Six (55%) of the 11 patients had right vertebrobasilar occlusions, and the remaining five (45%) had anterior circulation occlusive disease. The time from groin puncture to final recanalization time (overall procedural time) was 78.0±20.1 (62) minutes. The mean crossover time from TFA to TRA was 45.2±10.5 (41) minutes. The mean time from radial puncture to final recanalization was 33.8±10.5 (28) minutes. Distal thrombus migration events in previously unaffected territories occurred in 3/8 patients (37%). At 90 days, three patients (28%) had a favorable clinical outcome. Conclusion : Although rare, failure of TFA has been known to occur during MT for AIS. Our results demonstrate that TRA may be an alternative option for AIS intervention for select patients with subsequent timely revascularization. However, the incidence of distal thrombus migration was high, and the first puncture to reperfusion time was prolonged because of the time taken for the crossover to TRA after failure of TFA. This study provides some evidence that the TRA may be a viable alternative option to the TFA for MT of AIS.

• Journal of the Korean Society of Radiology
• /
• v.7 no.5
• /
• pp.313-320
• /
• 2013

For some patients with joint illnesses such as rheumarthritis or varus deformity, the total knee arthroplasty (TKA) procedures are performed. However, when inserting metal cutting guide for the procedures, due to the femoral shaft bowing, complications such as the cortex of the femoral shaft damages or secondary fractures can be caused. If the central coordinate value of the femoral shaft is known, the metal cutting guide could be inserted into the anatomical center, so such complications can be prevented. In this study, CT images of femoral shafts of 10 individuals in the experiment group who are in need of receiving the total knee arthroplasty procedures and those of 10 individuals in the control group without illness in the femoral shaft have been utilized to locate the 3-dimensional coordinate values. Then, Matlab was utilized to identify the central coordinate value in order to obtain a graph reflecting the anatomical shapes as well as to acquire the 3-dimensional radial curvature values by section. As a result, the average curvature range and standard deviation of femoral shafts of the experiment group was determined to be $758.15{\pm}206.3mm$ whereas the that of the control group was determined to be $1672.97{\pm}395.6mm$. The statistical significance of the measured results was verified through f-distribution analysis. Based on these results, it was verified that the level of curvature of the femoral shaft of the experiment group was higher. If the anatomical central points are located and analyzed using this methodology, it would be helpful in performing orthopedic operations such as the total knee arthroplasty.

• The Journal of the Acoustical Society of Korea
• /
• v.30 no.8
• /
• pp.446-460
• /
• 2011

The use of induction motors has been recently increasing with automation in aeronautical and automotive industries, and it playes a significant role. This has motivated that many researchers have studied on developing fault detection and classification systems of an induction motor in order to minimize economical damage caused by its fault. With this reason, this paper proposed feature vector extraction methods based on STE (short-time energy)+SVD (singular value decomposition) and DCT (discrete cosine transform)+SVD techniques to early detect and diagnose faults of induction motors, and classified faults of an induction motor into different types of them by using extracted features as inputs of BPNN (back propagation neural network) and multi-layer SVM (support vector machine). When BPNN and multi-lay SVM are used as classifiers for fault classification, there are many settings that affect classification performance: the number of input layers, the number of hidden layers and learning algorithms for BPNN, and standard deviation values of Gaussian radial basis function for multi-layer SVM. Therefore, this paper quantitatively simulated to find appropriate settings for those classifiers yielding higher classification performance than others.

• Journal of the Korean Wood Science and Technology
• /
• v.14 no.1
• /
• pp.45-54
• /
• 1986

A study was conducted to determine the physical properties related to drying characteristics, the seasonal air drying curves and the kiln drying schedule for taun lumber imported and utilized. This kiln drying schedule was found by oven drying and developed by pilot testing of green lumber and partially air dried lumber. The results of this study were as follows; 1. Average green specific gravity and standard deviation of heartwood lumber were 0.60${\pm}$0.03 and those of sapwood lumber were 0.64${\pm}$0.02. 2. Radial shrinkage from green to air dry and from green to oven dry were 3.05 percent and 5.96 percent respectively, and tangential shrinkage from green to air dry and to oven dry were 5.49 percent and 8.74 percent respectively. 3. Drying time for 25mm thick green lumber (50 percent moisture content) air dried to 30 percent moisture content were 14 days in springtime. 6 days in summertime, and 12 days in autumntime, whereas for 50mm thick lumber in 36 days in springtime, 18 days in summertime, 38 days in autumntime. 4. Kiln drying schedules developed by oven drying were T8-B3 for 25mm thick lumber and T5-B2 for 50mm thick lumber. 5. Kiln drying curves of green 25mm and 50mm thick lumber were similar to those of partially air dried lumber from the level of 30 percent average moisture content. Green 25mm thick lumber (55.7 percent moisture content) was dried to 9.3 percent moisture content in 101.5 hours and green 50mm thick lumber (65.6 percent moisture content) was dried to 11.5 percent moisture content in 526 hours. 6. End checking for green 25mm thick lumber occured in 49.6 percent moisture content and reached maximum amount in 27.6 percent moisture content and closed in 15.8 percent moisture content. 7. End checking for green 50mm thick lumber and partially air dried lumber developed and reached maximum amount earlier then for 25mm thick lumber. 8. Final moisture content of surface layer for 50mm thick lumber was one half of that of core, and moisture content equalized in the lumber after nine days of room conditioning. 9. Casehardening for 50mm thick lumber was slight and was conditioned after nine days of room stroage. 10. Drying defects, such as end checking and surface checking, were not observed and the quality of dry lumber was first.

• Korean Chemical Engineering Research
• /
• v.60 no.4
• /
• pp.535-543
• /
• 2022

A CPFD (Computational particle fluid dynamics) model of solar fluidized bed receiver of silicon carbide (SiC: average d_p=123 ㎛) particles was established, and the model was verified by comparing the simulation and experimental results to analyze the effect of particle behavior on the performance of the receiver. The relationship between the heat-absorbing performance and the particles behavior in the receiver was analyzed by simulating their behavior near bed surface, which is difficult to access experimentally. The CPFD simulation results showed good agreement with the experimental values on the solids holdup and its standard deviation under experimental condition in bed and freeboard regions. The local solid holdups near the bed surface, where particles primarily absorb solar heat energy and transfer it to the inside of the bed, showed a non-uniform distribution with a relatively low value at the center related with the bubble behavior in the bed. The local solid holdup increased the axial and radial non-uniformity in the freeboard region with the gas velocity, which explains well that the increase in the RSD (Relative standard deviation) of pressure drop across the freeboard region is responsible for the loss of solar energy reflected by the entrained particles in the particle receiver. The simulation results of local gas and particle velocities with gas velocity confirmed that the local particle behavior in the fluidized bed are closely related to the bubble behavior characterized by the properties of the Geldart B particles. The temperature difference of the fluidizing gas passing through the receiver per irradiance (∆T/I_DNI) was highly correlated with the RSD of the pressure drop across the bed surface and the freeboard regions. The CPFD simulation results can be used to improve the performance of the particle receiver through local particle behavior analysis.