Jo Dec 24, 2024
Formal Concept Analysis (FCA) is one of the major formats for knowledge extraction, reduction, representation and analysis. FCA is formulated based on a formal context represented by a binary relation between a set of objects and a set of attributes. Formal concept lattice is the core of the mathematical theory of FCA.
Applying FCA methods to large formal contexts could bring many challenges, because the algorithm to extract formal concepts can have an exponential complexity in the worst case. Hence, the determination problem has been widely discussed in many scientific works ever since, which leads to the development of many efficient algorithms. In general, these algorithms can be divided into two categories, i.e. batch algorithms and incremental algorithms. Batch algorithms usually construct the lattice in a bottom-up (or top-down) approach, while incremental algorithms compute the lattice by adding objects (or attributes) to a given context one by one.
Among all types of lattice construction algorithms, incremental algorithms have a unique advantage. The input formal context may not be fixed in a realistic application of FCA, which means the present lattice has to be updated or a new lattice should be computed from scratch. Obviously, computing the corresponding changes only and updating the current lattice should be a better choice, which can be handled by incremental algorithms.
Generating all extents of the generalized one-sided concept lattice by the incremental algorithm is usually based on the operation of Galois connection, and the operation needs to scan the formal context. This becomes a significant computing overhead, especially when the formal context is large.
With respect to the description problem, the researcher who formulated FCA mentioned that the most communicative description is given by Hasse diagrams. However, it is difficult to autonomously understand the hierarchy of the concepts from Hasse diagrams for any information retrieval software. Thus, Pak Chol Hong, a researcher at the Faculty of Applied Mathematics, has proposed methods of describing the hierarchy of a concept lattice by using a hierarchy-matrix, estimating the connectivity of concepts by the hierarchy-matrix and describing the Hasse diagram via a hierarchy-matrix.
If there is a method of determining the generalized one-sided concept lattice and describing its Hierarchy-matrix sequentially, the more useful software with autonomous intelligence, which is based on FCA from the generalized one-sided formal context, can be built. On this basis, he has presented a new efficient incremental algorithm that performs the operation of Galois connection for determining the generalized one-sided concept lattice only once and describes its hierarchy-matrix sequentially. In addition, he has evaluated the feature and effectiveness of the proposed algorithms.
...
Jo Dec 22, 2024
Mg-Al alloys are widely used in various fields such as architecture and aviation due to their low density, high specific strength, good corrosion resistance and ease of processing. The components applied to special vehicles need to have high temperature creep resistance. The addition of a small amount of rare earth elements in the Mg-Al alloy can meet this requirement.
The research group led by Jang Pok Nam, an institute head at the Faculty of Metal Engineering, has already worked on the preparation of Mg-Al-La master alloy by chloride-fluoride molten salt electrolysis in MgCl2-KCl-AlF3-La2O3 system. In this method, AlF3 exists in a solid state in molten salts and it is consumed through chemical reactions going with electrolysis. However, the separation of this solid fluoride and the resulting master alloy is not easy and evaporation loss of chlorides is big due to relatively high electrolytic temperature.
Therefore, the research group has proposed a method of preparing Mg-Al-La master alloy by chloride electrolysis to reduce electrolytic temperature and facilitate the separation of electrolytes and alloys.
No studies have been reported on the preparation of Mg-Al-La master alloys by molten salt electrolysis in MgCl2-KCl-AlCl3 molten salts.
After preparing the Mg-Al-La master alloy by molten salt electrolysis in MgCl2-KCl-AlCl3 ternary molten salt, they have investigated the factors affecting the element contents and current efficiencies in the production of Mg-Al-La master alloy, such as electrolysis temperature, cathode current density and electrolysis time.
He has found that increasing electrolytic temperature favors the increase in the contents of Al and La in the alloy, and the optimum electrolytic temperature is 600-650℃, and that as cathodic current density and electrolytic time increase, the content of Mg in the alloy also increases steadily but the content of La decreases.
...
Jo Dec 20, 2024
Research for new deposits using the satellite image analysis method has already been carried out by many previous researchers. Studies to predict mineral prospecting sites by satellite image analysis have been mostly conducted to identify mineralized alteration zones and to detect linear structures.
Satellite image analysis generally begins with remote sensing. The integrated data is prepared for the user. In this stage, the integrated data is interpreted by various methods. In data interpretation, images are interpreted to identify various phenomena with regard to their importance.
It is widely accepted that hydrothermal alteration plays an important role in ore formation. The desirable distribution of alteration zones around the limited borders of an ore body facilitates outdoor exploration activities. Generally, hydrothermal alteration processes lead to emanation of clay and other silicate minerals, including kaolinite, sericite, quartz, chlorite, epidote, and calcite. The application of both spectral and statistical analyses of remotely sensed data can identify the above-mentioned minerals and explore the buried deposits. Hydrothermal alteration zones can be recognized from the spectral characteristics of main mineral assemblages.
Based on the spectral characteristics of main mineral assemblages, Ra Chung Yol, a section head at the Faculty of Earth Science and Technology, has determined the distribution of the hydrothermal alteration zone in the study area, using band-ratio analysis, principal component analysis, and colorimetric synthesis methods.
He has concluded that the hydrothermal alteration zone in the study area was widely distributed over an area of 2 000m×200m above the surface under the control of the ore-bringer structure, especially at the contact with the alkaline rock mass.
...
Jo Dec 19, 2024
A large amount of energy is consumed for buildings around the world, and therefore, energy saving is being studied extensively.
Mixing ventilation system (MVS) is one of the mechanical ventilation systems that makes the entire indoor air mixed completely using the jet flow made from blast air. Displacement ventilation system (DVS) is one of the mechanical ventilation systems, which blows air with lower temperature than indoor air into the lower part of a room and causes upward flow of indoor air by buoyant force to exhaust it at the ceiling.
The basic difference between DVS and MVS is that DVS mainly relies on the buoyancy effect, while MVS uses mechanical ventilation force. As DVS is aimed at only meeting the requirements of the workplace, there is a thermal and concentration stratification between the lower and upper parts. DVS is characterized by good air quality, efficient emission of polluted air and significant energy saving effect. In the design of DVS, the height of thermodynamic stratification has to be set to be higher than the height of the workplace, but not too high, because it might cause waste and increase in the amount of air blast. So, the determination of the height of thermodynamic stratification is one of the vital problems arising in the application of DVS. In DVS, ventilation volume for keeping the temperature below the required level is too much when the cooling load is heavy. For DVS, experiments, experiences and engineering design methods have already been presented and mathematical analysis methods have been studied. However, there is little data about DVS for high-ceiling rooms.
Paek Myong Chol, a section head at the Faculty of Heat Engineering, has investigated the temperature change at the workplace when there is heat source in MVS and HDVS with the vertical supply duct by using CFD and presented its effectiveness. In addition, he has verified the existence of the effective length of the vertical supply duct in HDVS with supply ducts from the view of energy saving.
He has found that under the conditions of providing equal temperature and velocity of air inflow, the temperature of the air in the working zone depends on the length of a vertical supply duct, and that the temperature of the air in the working zone increases when the vertical supply duct is longer than the limited value.
...
Jo Dec 17, 2024
In general, in the electromagnetic surveying, frequency dispersion of conductivity is neglected. In contrast, in the common excited polarization method, current is indirectly supplied to the underground, so the secondary voltage originated by ramp turn off in the current flow is actually frequency dispersion of conductivity. Hence, electromagnetic (EM) effect and induced polarization (IP) effect come into being at the same time, which leads to distortion, in other words, sign reversal of EM response. For this reason, many researchers deeply studied the mechanism of the sign reversal and the forward modeling in consideration of this phenomenon.
Many researchers simulated the distorted transient response by conductivity related to frequency. Most of the research results concluded that Cole-Cole model is well suited to considering IP effect observed in TDEM.
Transient response of polarizable whole space is surely different from both polarizable half space response and non-polarizable whole space response.
Jon Kwang Bok, a researcher at the Faculty of Earth Science and Engineering, has approximated the whole space electromagnetic transient field originated by rectangular loop in the whole layered space which includes polarizable layers by using the Cole-Cole model and linear digital filtering method.
As a result, he has drawn two conclusions.
First, induced voltage measured has no relationship with whether the polarizable conductive layers are above or below the measurement spot. Second, polarization must be considered in the whole space inversion as in half space because the response computed by common whole space modeling method is a lot different from that by polarizable whole space modeling method.
...
Jo Dec 16, 2024
Unmanned Surface Vessel (USV) is gaining great popularity in various fields including marine environmental monitoring and coastal waters protection and it is becoming world-focused marine equipment.
USV can reduce direct human intervention and perform safe operations in hazardous environments such as marine exploration, personnel rescue, environmental monitoring and survey, etc.
When the USV is sailing, it will deviate from the given course due to the disturbances of wave, current, wind and other marine environment.
Although several technical proposals have been proposed to solve this problem, ship control still remains a very challenging problem.
Ship course is mostly controlled by the conventional PID control law. Conventional controllers are still regarded as a good choice for an auto-steering system when they have typical SISOs, but not suitable for a complex and strong nonlinear MIMO system like USV. Indeed, this method is easy and simple, but its parameters are set by manual experience and control is possible in the limited range, which leads to low adaptability.
Choe Kum Song, a section head at the Faculty of Shipbuilding and Ocean Engineering, has obtained USV’s hydrodynamic parameters from its sailing test data and built a linear mathematical model of 3-DOF for USV planar motion. On this basis, he has proposed a course control algorithm for USV based on fuzzy PD and compared it with the conventional PID law.
The simulation results show that the designed fuzzy PD law is more effective than the conventional PID law.
...
