Earticle

산사태 발생시점 및 지점을 예측하기 위한 연구는 서로 다른 목적과 범위로 구성되었지만 상호보완 적인 역할을 한다. 따라서 이 연구에서는 산림청에서 구축한 산사태 예보체계와 산사태위험지도의 활용성을 높이고자 정적 산사태위험도를 동적 산사태위험도로 변환하고, 이를 평가하기 위해 시공 간적 위험도의 매트릭스 결합을 수행하였다. 2017년 산사태 발생지인 충청남도 천안시의 시나리오 및 실제 산사태 발생시점을 대상으로 다양한 매트릭스 평가모형을 적용한 결과, 간단한 행렬 조합만 으로 동적 산사태위험도의 정량적 평가를 가능케 하였으며, 정확도 또한 기존의 산사태위험지도에 비해 높게 나타났다. 또한 산림청에서 제공하는 1시간 이후의 산사태 예측정보와 결합을 통해 단기 적인 미래의 동적 산사태위험도를 평가할 수 있는 점에 의의가 있다. 이 연구결과는 기구축된 국가 산사태 예보체계와 산사태위험지도의 활용성을 극대화함과 동시에 우기 시의 산사태위험지 관리 효율성 향상에 기여할 수 있을 것이다.

The models to predict when and where a landslide occurs were constructed with different purposes and scopes, but tended to be complementary to each other. The objective of this study was to convert the static landslide hazard into the dynamic landslide hazard and combine a spatiotemporal hazard matrix to assess it, in order to enhance the usability of the early warning system for landslide and the landslide hazard map developed by the Korea Forest Service. We applied various matrix evaluation models to develop scenarios and assess the 2017 landslide occurred in Cheonan, Korea. The results of this study showed that a simple combination of matrices facilitated quantitative evaluation of the dynamic landslide hazard and its accuracy was higher than the static landslide hazard map. In addition, the dynamic landslide hazard can be assessed for the near future by integrating the 1-hour prediction data provided by the Korea Forest Service. The results of this study can help maximize the utilization of the existing national landslide forecast system and landslide hazard map, which should enhance the efficiency of managing landslide-prone zones during the rainy season.

This paper presents experimental study of the static and dynamic fracture toughness behavior of a A1-6061 aluminum alloy reinforced alumina( .delta. -A1$_{2}$0$_{3}$) whiskers with 5%, 10%, 15% volume fraction. The static fracture tests using three-point bending specimen were performed by UTM25T. And drop weight impact tester performing dynamic fracture tests was used to measure dynamic locads applied to a fatigue-precracked specimes. The oneset of crack initiation was detected uwing a strain gage bonded near a crack tip. The value of static fracture toughness $K_{IC}$ and dynamic fracture toughness $K_{ID}$ were decided on the basis of linear elastic fracture mechanics. The effects of fiber volume fraction and loading on fracture toughness were investigated. The distribution of whiskers, bonding state and fracture interfaces involved in void, fiber pull-out state were investigated by optical microscopy(OM) and scanning electron microscopy(SEM)

보행행위는 건축공간에서 다양한 이용특성의 출현을 가능하게 하는 기본적인 요소이다. 이용자의 공간인지 범죄자의 이동, 희생자의 탈출 및 방재안전 관련 피난은 이용자 행태를 반영하고 구축된 공간의 조건에 반응하는 보행경로를 발생시킨다. 축선도 분석이나 VGA 기법은 선행 연구에서 분석적인 모델과 함께 제안되었고 시험되어 왔다. 그러나 이러한 기법들과 모델들은 단지 구축된 공간의 정적인 조건만을 고려하고 보행경로의 개별적이고 국지적인 변화를 탐색하지 않는다. 본 연구의 목적은 보행 에이전트의 적용과 함께 동적 변환행렬을 이용하는 보행경로 모델의 개발에 있다. 또한 보행경로 모델의 구현은 구축된 조건의 고려, 공간 구성, 보행자의 시지각 및 다른 보행자의 위치를 정량적 모델링 과정의 변수로 포함하였다. 다수의 보행자를 고려하며 개발된 모델의 적용은 개별적이고 국지적인 보행경로를 제시하였다. 그 적용은 또한 제안된 변수요소와 종합된 모델의 분석적 가능성을 제시하였다.

Pedestrian behavior is the basic element enabling the appearance of various usage characteristics in architectural space. User’s spatial recognition, criminal’s movement, victim’s escape and the evacuation process in fire life safety generate walking paths reflecting user behavior and responding condition of built environment. Axial map analysis and VGA method have been suggested and tested with analytical models in precedent researches. But, these methods and models only consider static conditions of built environment, and do not consider dynamic pedestrian behavior. In addition, these studies produce analysis result for visit frequency, and do not examine individual and local changes of walking paths. The purpose of this study was the development of agent-based pedestrian path model using dynamic transition matrix with application of pedestrian agents. The implementation of pedestrian path model also included the consideration of built conditions, spatial composition, pedestrians` visual cognition and other pedestrians' locations as the variables in quantitative modeling process. With considering multiple pedestrians, the application of developed models showed individual and local pedestrian paths. It also provided the analytical feasibility of the suggested variable components and the combined model.

The random dynamic forces experienced by structure due to wind are caused primarily by the turbulence in the approaching flow and motion of the structure. In the paper, fluctuating wind velocity for time history analysis are simulated by multi-dimensional random processes using Karman spectrum about along-wind direction and Liang spectrum about across-wind direction. This paper is analyzed and obtained the dynamic response for structures modeled as a plane frame. This study uses the stiffness method and static condensation method to determine the condensed stiffness matrix that corresponds to the lateral displacement coordinates at the various levels of the building. The numerical results, obtained by the Newmark-β method, shows the time- history responses and trends of maximum accelerations and displacements by applying various load conditions. In the paper, the values of dynamic reponse to the structure modeled as a plane frame will be discussed with the values of dynamic reponse to the structures modeled as a shear building. For the building that is modeled as a plane frame, lateral displacements and accelerations are greater than the values obtained in modeled as a shear building.

최근 통합교통서비스(Mobility-as-a-Service)의 개념을 도입하여 이용자들의 이동성과 접근성 을 향상시키고자 하는 연구가 진행되고 있다. 특히 카셰어링, 택시 등 에 대해 수요와 공급에 따라 지역을 구분하여 가격을 책정하는 유동적인 가격 책정 전략을 도입하여 단일 요금제가 가지는 서비스 기피 등의 문제를 해결함과 동시에 기업과 운전자들의 수익성에 긍정적인 영향 을 줄 수 있을 것으로 기대되고 있다. 본 연구에서는 승객과 운전자간의 배차거리, 승객의 운행 거리, 승객의 목적지에 대한 HDBSCAN 알고리즘을 통해서 정밀하게 인식된 수요 밀집지역, 승객과 운전자가 생각하는 선호가격을 고려하여 승객과 운전자의 입장에서 Matching Matrix를 생성한다. 이를 조합하고 보상에 반영하여, 강화학습이 더욱더 현실적인 유동적인 가격 책정 전략을 도출할 수 있는 새로운 방법론을 제안한다.

Research interest in the Mobility-as-a-Service (MaaS) concept for enhancing users’ mobility experience is increasing. In particular, dynamic pricing techniques based on reinforcement learning have emerged since adjusting prices based on the demand is expected to help mobility services, such as taxi and car-sharing services, to gain more profit. This paper provides a simulation framework that considers more practical factors, such as demand density per location, preferred prices, the distance between users and drivers, and distance to the destination that critically affect the probability of matching between the users and the mobility service providers (e.g., drivers). The aforementioned new practical features are reflected on a data structure referred to as the Matching Matrix. Using an efficient algorithm of computing the probability of matching between the users and drivers and given a set of precisely identified high-demand locations using HDBSCAN, this study developed a better reward function that can gear the reinforcement learning process towards finding more realistic dynamic pricing policies.

In order to lower the classification cost and improve the performance of the classifier, this paper proposes the approach of the dynamic cost-sensitive ensemble classification based on extreme learning machine for imbalanced massive data streams (DCECIMDS). Firstly, this paper gives the method of concept drifts detection by extracting the attributive characters of imbalanced massive data streams. If the change of attributive characters exceeds threshold value, the concept drift occurs. Secondly, we give Cost-sensitive extreme learning machine algorithm, and the optimal cost function is defined by the dynamic cost matrix. Build the cost-sensitive classifiers model for imbalanced massive data streams under MapReduce, and the data streams are processed in parallel. At last, the weighted cost-sensitive ensemble classifier is constructed, and the dynamic cost-sensitive ensemble classification based on extreme learning machine classification is given. The experiments demonstrate that the proposed ensemble classifier under the MapReduce framework can reduce the average misclassification cost and can make the classification results more reliable. DCECIMDS has high performance by comparing to the other classification algorithms for imbalanced data streams and can effectively deal with the concept drift.

Genetic algorithms can solve complex problems, including the problems of cryptography. What problems often occur on the Hill Cipher is the waste of time to determine the numbers that are used in the encryption process. In the encryption process, it is not a problem if the key is derived from any number. However, the problem is ciphertext cannot be returned to the original message. The key that is used must have the determinant is 1. To find the value of it is something that takes time if it must be done manually. Due to the entered value to the Hill Cipher is random, Genetic algorithms can be used to optimize the search time. By using this algorithm, the determinant calculation will be more accurate and faster. The result achieved is the program can specify some combination of numbers that can be used as the encryption key Hill Cipher and it can reject the unnecessary numbers.

The robust 𝐻∞ control design for bilinear systems with multi inputs is presented in this paper. First, the bilinear system is represented as a dynamic Takagi-Sugeno (TS) fuzzy system by using sector nonlinearity approach. The dynamic TS fuzzy system is a convex combination of local linear systems. The local robust 𝐻∞ controller is designed for each local linear system. The controller synthesis for the local linear systems is then formulated in the bilinear matrix inequalities (BMIs) problem. After that, the BMIs problem is reduced to an equivalent parameter of linear matrix inequalities (LMIs) problem which has a feasible solution. The robust 𝐻∞ controller for bilinear systems as a convex combination of the local robust 𝐻∞ controllers is obtained by using defuzzyfication. The existence condition of the robust 𝐻∞ controller for the bilinear systems is also presented. The simulation results are given to clarify the proposed method for the robust 𝐻∞ control design of the bilinear systems.

The Dynamic Matrix Control(DMC) technique was applied to nonlinear and nonminimum phase system. System model was identified by using Least Square method. Desired output trajectory was prespecified and input suppression parameter was also introduced. It was shown that DMC technique worked with great success in solving both nonminimum phase system and nonlinear system.

터빈 blade에 대해 기존의 stiffness matrix법과 finite element법에 의한 방식을 개선하여 수치계산을 행한 결과 다음의 결론을 얻었다. 1)stiffness matrix법을 적용하기 위해 th.2 order로 방정식을 유도하였으며, 회전시의 원심력의 영향 및 blade의 기하학적 형상이 수식에 고려되었다. 이 방법으로 blade의 여러 parameter의 영향을 간단히 계산할 수 있다. 계산결과는 다른 논문의 결과와 잘 일치함을 보였다. 또한, 원심력의 영향에 있어서는 th.2 order의 계산결과가 기존의 변형된 th.1 order의 결과보다 더 정확한 결과를 얻을 수 있다. 2) FEM 이용시 계산시간을 단축하면서 정확한 결과를 얻기 위해 blade를 간단히 modeling하여, 기하학적인 형상과 회전시의 영향을 고려한 식을 유도하였다. 본 방법의 결과는 타 문헌과 일치하였다.

A 160 MW drum-type boiler-turbine system is developed in the present research through a multi-multivariable dynamic matrix control (DMC) scheme and a multi-multivariable model approach. A novel intelligence-based decision mechanism (IBDM) is realized to support both model approach and control scheme. In such case, the responsibility of the proposed IBDM is to identify the best multivariable model of the system and the corresponding multivariable DMC scheme to cope with the system at each instant of time in an appropriate manner.

In this paper, we present the implementation method of Dynamic Matrix Control(DMC) block for use with a Rapid Control Prototyping(RCP) system and consider the speed control of a DC motor using the developed DMC block. Firstly, we briefly introduce a lab-built RCP system. Secondly, we present a method for implementing a DMC block using C-language, which enables the DMC algorithm to be represented in a library block that can be used in a Simulink environment. Finally, we use the developed DMC block for the speed control of a DC motor, through which we show that the DMC-based control system can be easily implemented and applied to the real-time control of systems with relatively fast dynamics.

In this paper a hybrid control strategy is presented based on Dynamic Matrix Control (DMC) and feedback linearization methods for designing a predictive controller of five bar linkage manipulator as a MIMO system (two inputs and two outputs). Analyzing the internal dynamic of robot shows the open loop system is unstable and non-minimum phase, so in order to apply the predictive controller, special modifications are needed. These modifications on non-minimum phase behavior are performed using feedback linearization procedure based on state space realization. The design objective is to track a desirable set point as well as time varying trajectories as a command references with globally asymptotical stabilization. The proposed controller is applied to nonlinear fully coupled model of the typical five bar linkage manipulator with non-minimum phase behavior. Simulation results show that the proposed controller has good efficiency. The step responses of system with and without feedback linearization process illustrated that the mentioned modification for stabilizing is performed properly. After applying the proposed predictive controller, the joint angle of robot tracks the reference input while another input acts as the disturbance and vice versa.

동역학행렬제어(Dynamic Matrix Control) 기법은 각종 산업 현장에서 가장 활발하게 적용되고 있는 고급 제어 기법으로, 최근에는 공정제어의 표준 기법으로 인식되고 있다. 일반적으로 동역학행렬제어에서는 대상 플랜트의 거동을 묘사하기 위하여 계단 응답 모델을 이용한다. 본 논문에서는 화력발전의 보일러-터빈 시스템에 동역학행렬제어 기법을 적용하기 위하여, 두 가지 방식으로 선정된 계단응답모델의 제어 성능을 비교하였다. 먼저 주어진 대상 플랜트의 정확한 비선형 모델을 알고 있다고 가정하여 이의 선형화를 통한 계단응답 모델과 대상시스템의 계단응답 실험을 통하여 얻어진 두 가지 계단응답 모델의 유효성을 비교 분석하였다. 분석 결과 실험적인 계단 응답 모델을 사용한 경우가 우월한 제어 성능을 나타냄을 확인하였다.

This paper presents comparison results of Step Response Model of Dynamic Matrix Control(DMC) for a drum-type boiler-turbine system of a fossil power plant. Two possible kinds of step response models are investigated in designing the DMC, one is developed with the linearization of theoretical model and the other is developed with the process step-test data. Then, the control performances of each model-based DMC are simulated and evaluated. It is observed that the simulation results with the step-response model based on the test data show satisfactory results, while the linearized model is not suitable for the control of boiler-turbine system.

Multi-input multi-output (MIMO) dynamic matrix control (DMC) technique is applied to control steam temperatures in a large-scale ultrasupercritical once-through boiler-turbine system. Specifically, four output variables (i.e., outlet temperatures of platen superheater, finish superheater, primary reheater, and finish reheater) are controlled using four input variables (i.e., two spray valves, bypass valve, and damper). The step-response matrix for the MIMO DMC is constructed using the four input and the four output variables. Online optimization is performed for the MIMO DMC using the model predictive control technique. The MIMO DMC controller is implemented in a full-scope power plant simulator with satisfactory performance.

This paper proposes supplementary control of conventional coordinated control of a power plant which directly affects network frequency. The supplementary control with dynamic matrix control is applied for 1000 MW power plant with ultra-supercritical (USC) once-through boiler. The supplementary control signal is added to the boiler feedforward signal in the existing coordinated control logic. Therefore, it is a very practical structure that can maintain the existing multi-loop control system. This supplementary controller uses the step response model for the power plant system, and on-line optimization is performed at every sampling step. The simulation results demonstrate the effectiveness of the proposed supplementary control in a wide operating range of a practical 1000 MW USC power plant simulator. These results can contribute the stable operation of power system frequency.

In this paper the dynamic stiffness matrix method was used for the free vibration analysis of axially moving micro beam with constant velocity. The extended Hamilton's principle was employed to derive the governing differential equation of the problem using the modified couple stress theory. The dynamic stiffness matrix of the moving micro beam was evaluated using appropriate expressions of the shear force and bending moment according to the Euler-Bernoulli beam theory. The effects of the beam size and axial velocity on the dynamic characteristic of the moving beam were investigated. The natural frequencies and critical velocity of the axially moving micro beam were also computed for two different end conditions.

We propose a dynamic incidence matrix (DIM) for reflecting states and time conditions of a timed Petri net (TPN) explicitly. Since a DIM consists of a conventional incidence matrix, two time-related vectors and two state-related vectors, we can get the advantages inherent in the conventional incidence matrix of describing a static structure of a system as well as another advantage of expressing time dependent state transitions. We introduce an algorithm providing the DIM with a state transition mechanism. Because the algorithm is, in fact, an algorithmic model for discrete event simulation of TPN models, we provide a theoretical basis of model transformation of a TPN model into a DEVS(Discrete Event system Specification) model. By executing the algorithm we can carry out performance analysis of computer communication protocols which are represented TPN models.