http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
계절별 기상조건에 기반한 태양광 발전량 예측에 관한 연구
김백천(Baekcheon Kim),정승환(Seunghwan Jung),김민석(Minseok Kim),김종근(Jonggeun Kim),김성신(Sungshin Kim) 한국지능시스템학회 2021 한국지능시스템학회논문지 Vol.31 No.2
태양광 발전량 예측은 태양광 발전 시스템을 연계하여 운용되는 마이크로그리드의 전력공급 안정성을 향상시키기 위해 필요하다. 하지만 태양광 발전은 계절별 특성 및 기상조건에 따라 발전량이 다르기 때문에 예측하기 어렵다. 따라서 본 논문에서는 기상 및 발전량 데이터를 계절 및 날씨에 따라 분류하고, 분류된 데이터에 각각 Long Short-Term Memory(LSTM)를 이용하여 태양광 발전량 예측 방법을 제안한다. 제안된 방법의 성능을 비교하기 위해 계절만 고려한 모델, 날씨별 모델과 비교하였다. 실험결과, 계절 및 날씨에 따라 분류된 데이터를 이용하여 모델을 구현하였을 때, 기존 방법들보다 예측성능이 더 우수한 것을 확인할 수 있었다. The solar power generation forecasting is necessary to improve the power supply stability of microgrids operated in connection with the solar power generation system. However, it is difficult to forecast solar power due to anomalous seasonal characteristics and weather conditions. Therefore, the proposed method classifies meteorological and power generation data by season and weather, and then predicts solar power generation by Long Short-Term Memory (LSTM) using the classified data. Weather conditional model and seasonal model are used to verify the performance of the proposed method. Experimental results show that model using classified data by season and weather is outperforms than conventional models.
kNN-LSTM을 이용한 시간별 일사량 예측 성능 개선에 관한 연구
김민석(Minseok Kim),정승환(Seunghwan Jung),김백천(Baekcheon Kim),김진용(Jin Yong Kim),김성신(Sungshin Kim) 한국지능시스템학회 2022 한국지능시스템학회논문지 Vol.32 No.3
세계 에너지시장 동향 및 재생에너지 관련 제도 개선으로 태양광발전 보급이 증가하고 있다. 태양광발전은 탄소를 발생시키지 않으면서 발전할 수 있다는 점에서 유망한 대체 에너지원이다. 하지만, 이러한 재생에너지를 이용한 전력원은 자연에너지를 이용하는 특성상 불안전한 발전출력으로 인해 발전량 예측, 유지보수 및 관리 시스템이 필요하다. 태양광 발전량을 예측하기 위해서는 먼저 태양광 발전에 가장 많은 영향을 미치는 일사량을 예측해야 하므로 이와 관련된 연구가 지속적으로 수행되고 있다. 따라서, 본 논문에서는 kNN-LSTM (k-Nearest Neighbor and Long Short-Term Memory) 기반 일사량 예측 방법을 제안한다. 제안된 방법은 kNN을 이용하여 전 날 일사량 패턴과 유사한 과거 데이터를 탐색한 다음 LSTM에 입력변수로 사용하여 다음 날의 시간별 일사량을 예측한다. 실험 결과, 제안된 방법이 서포트 벡터 회귀, 인공신경망 및 LSTM에 비해 계절에 따른 청명한 날과 흐린 날 모두 효과적으로 예측할 수 있음을 확인하였다. The supply of solar power is increasing owing to global energy market trends and the improvement of new and renewable energy-related systems. solar power is a promising alternative energy source in that it can make electricity without generating carbon. However, since the power generation output is unstable due to the nature of using solar energy, a power generation prediction, maintenance, and management system is required. To predict the amount of solar power generation, it is necessary to forecast solar radiation that has the most influence on the solar power generation. Therefore, research related to this has been continuously being conducted. In this paper, we propose a method for forecasting solar radiation based on k-Nearest Neighbor and Long Short-Term Memory (kNN-LSTM). The proposed method uses kNN to search training data similar to the previous day"s solar radiation pattern and then uses it as an input variable to LSTM to forecasting the hourly solar radiation of the next day. The experimental results show that the proposed method can effectively predict both seasonal sunny and cloudy days compared to the conventional method by comparing it with Support Vector Regression, Artificial Neural Network and LSTM.
3D Radar Objects Tracking and Reflectivity Profiling
Kim, Yong Hyun,Lee, Hansoo,Kim, Sungshin Korean Institute of Intelligent Systems 2012 INTERNATIONAL JOURNAL of FUZZY LOGIC and INTELLIGE Vol.12 No.4
The ability to characterize feature objects from radar readings is often limited by simply looking at their still frame reflectivity, differential reflectivity and differential phase data. In many cases, time-series study of these objects' reflectivity profile is required to properly characterize features objects of interest. This paper introduces a novel technique to automatically track multiple 3D radar structures in C,S-band in real-time using Doppler radar and profile their characteristic reflectivity distribution in time series. The extraction of reflectivity profile from different radar cluster structures is done in three stages: 1. static frame (zone-linkage) clustering, 2. dynamic frame (evolution-linkage) clustering and 3. characterization of clusters through time series profile of reflectivity distribution. The two clustering schemes proposed here are applied on composite multi-layers CAPPI (Constant Altitude Plan Position Indicator) radar data which covers altitude range of 0.25 to 10 km and an area spanning over hundreds of thousands $km^2$. Discrete numerical simulations show the validity of the proposed technique and that fast and accurate profiling of time series reflectivity distribution for deformable 3D radar structures is achievable.
Positioning and Driving Control of Fork-type Automatic Guided Vehicle With Laser Navigation
Kim, Jaeyong,Cho, Hyunhak,Kim, Sungshin Korean Institute of Intelligent Systems 2013 INTERNATIONAL JOURNAL of FUZZY LOGIC and INTELLIGE Vol.13 No.4
We designed and implemented a fork-type automatic guided vehicle (AGV) with a laser guidance system. Most previous AGVs have used two types of guidance systems: magnetgyro and wire guidance. However, these guidance systems have high costs, are difficult to maintain with changes in the operating environment, and can drive only a pre-determined path with installed sensors. A laser guidance system was developed for addressing these issues, but limitations including slow response time and low accuracy remain. We present a laser guidance system and control system for AGVs with laser navigation. For analyzing the performance of the proposed system, we designed and built a fork-type AGV, and performed repetitions of our experiments under the same working conditions. The results show an average positioning error of 51.76 mm between the simulated driving path and the driving path of the actual fork-type AGV. Consequently, we verified that the proposed method is effective and suitable for use in actual AGVs.