메뉴 건너뛰기
.. 내서재 .. 알림
소속 기관/학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
로그인 회원가입 고객센터 ENG
주제분류

추천
검색
질문

논문 기본 정보

자료유형
학술저널
저자정보
정민우 (안양대학교) 이주용 (안양대학교) 왕경희 (안양대학교) 이채연 (안양대학교) 한승희 (안양대학교) 김희진 (안양대학교) 손승민 (안양대학교) 정필수 (안양대학교) 최대련 (안양대학교) 윤희영 (안양대학교)
저널정보
한국대기환경학회 한국대기환경학회지(국문) 한국대기환경학회지 제41권 제1호
발행연도
2025.2
수록면
67 - 82 (16page)
DOI
10.5572/KOSAE.2025.41.1.067

이용수

표지
📌
연구주제
📖
연구배경
🔬
연구방법
🏆
연구결과
AI에게 요청하기
추천
검색
질문

초록· 키워드

오류제보하기
The purpose of this study is to develop an AI model for forecasting the spatial distribution of PM<sub>2.5</sub> concentrations over the Korean Peninsula for the next 24 hours. To achieve this, four Conv-LSTM (Convolutional Long Short-Term Memory) based AI models were proposed, and their forecasting performance was compared with that of the CMAQ (Com- munity Multiscale Air Quality) model to identify the most suitable AI model for this task. The proposed AI models share an identical network structure but differ based on the application of the ReLU activation function: Case 1 (no ReLU function applied), Case 2 (ReLU applied to encoder and decoder layers), Case 3 (ReLU applied to the output layer), and Case 4 (ReLU applied to the encoder, decoder, and output layers). The results showed that although the proposed AI models did not outperform the CMAQ model in terms of data characteristic simulation, they exhibited improved prediction accuracy and spatial pattern simulation performance. Specifically, the Case 3 model exhibited the smallest range of RMSE values, with hourly RMSE ranging from 8.16 to 8.76 μg/m³ and spatial RMSE from 6.60 to 7.26 μg/m³. On average, the Case 3 model showed an improvement of 1.81 μg/m³ in RMSE and 1.31 μg/m³ in spatial RMSE over the CMAQ model, demonstrating the best prediction accuracy and spatial pattern simulation performance among the proposed AI models. Therefore, Case 3 was selected as the most suitable AI model for forecasting the 24-hour PM<sub>2.5</sub> spatial concentration distribution over the Korean Peninsula. To evaluate the spatial distribution simulation performance of the selected Case 3, the PM<sub>2.5</sub> concentrations during the selected high-concentration period were forecasted. The results showed that the model effectively simulated the movement trajectory and distribution, closely resembling the actual data, and accurately forecasted the high-concentration areas. The results of this study suggest that the predictive performance of AI models can vary depending on the use and placement of the ReLU activation function, and demonstrate the potential of the proposed Conv-LSTM-based AI models to overcome the limitations of chemical transport models like the CMAQ model. However, the AI model exhibited limitations in predicting peak concentration values at high-concentration points. Therefore, future research will focus on adjusting the proportion of high-concentration cases to enhance the AI model’s prediction performance.

목차

Abstract
1. 서론
2. 연구 방법
3. 결과 및 고찰
4. 결론
References

참고문헌 (0)

참고문헌 신청

함께 읽어보면 좋을 논문

논문 유사도에 따라 DBpia 가 추천하는 논문입니다. 함께 보면 좋을 연관 논문을 확인해보세요!

이 논문의 저자 정보

이 논문과 함께 이용한 논문

최근 본 자료

전체보기

댓글(0)

0

UCI(KEPA) : I410-151-25-02-092303537