Bok-Nam Ha

Empirical Modeling of Superconducting Fault Current Limiter Using Support Vector Regression

The superconductor-triggered type fault current limiter (STFCL), which was developed by Korea Electric Power Corporation (KEPCO) and LS Industrial Systems (LSIS), is under operation for the verification test at KEPCOs power testing center. The STFCL is composed of superconductor, fast switch and current limiting resistor. The fault current is suppressed after a half cycle by the method of a line commutation. In this paper, we investigated the empirical modeling of STFCL using principal components and auto-associative support vector regression (PCSVR) for the prediction and fault detection of the STFCL. Signals for the model are currents and voltages acquired from high-temperature superconductor (HTS), driving coil (DC) and current limiting resistor (CLR). After developing the empirical model we analyse the accuracy of the model. The results were compared with that of auto-associative neural networks (AANN). PCSVR showed much better performance in accuracy aspect. Moreover, this model can be used for the prognosis of STFCL system.

An optimal composition and placement of automatic switches in DAS

This paper proposes an algorithm for determining the optimal composition which means number of links and section switches in a feeder and placement of automatic switches in a Distribution Automation System (DAS). A DAS is configured by automatic switches and reclosers on a power distribution line. The composition and placement of switches affect the operational applications of a DAS. More switches lead to better DAS operation but also to increased cost and maintenance. Thus, this paper proposes an approach to determining the optimal composition and placement of automatic switches. Additionally, the proposed algorithm is developed considering various system topologies in a real field. The algorithm was tested on an example power distribution system with eight-feeders and on a real power distribution system operated by KEPCO in Young-Deung-Po and Jeju of South Korea.

Loss minimization and load balancing in a distribution network

All distribution network in South Korea are operated by Distribution Management System (DMS) of KEPCO. The DMS is configured by communication infrastructures, automatic switches, reclosers, and tie-switches on a radial system. Change of a tie-switch location determining a loss and a load capacity of each feeder can take an operational efficiency such as an economic advantage by loss minimization and load balancing in distribution automation system. Because customers load patterns in a distribution network are different by time, region, seasons, and special events. Loss minimization and load balancing are considered by each but they were not considered together because their numerical unit criteria are different. Therefore, this paper proposes an algorithm in order to consider loss minimization and load balancing together. The proposed algorithm is verified by practical distribution networks of Jeju in South Korea.

Multi-response variable optimization in sensor drift monitoring system using support vector regression

In a nuclear power plant (NPP), periodic sensor calibrations are required to assure sensors are operating correctly. However, only a few faulty sensors are found to be calibrated. For the safe operation of an NPP and the reduction of unnecessary calibration, on-line calibration monitoring is needed. Most researches have been focused on improving only the accuracy of the system although sensitivity is another important performance index. This paper presents multi-response optimization for an on-line sensor drift monitoring system to detect drift and estimate sensor signal effectively. Accuracy and sensitivity of the principal component-based auto-associative support vector regression (PCSVR) were optimized at the same time by desirability function approach. Response surface methodology (RSM) is employed to efficiently determine the optimal values of SVR hyperparameters. The proposed optimization method was confirmed with actual plant data of Kori NPP Unit 3. The results show the trade-off between the accuracy and sensitivity of the model as we expected.