http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 저자 : Radhi Muammar (검색결과 1 건)
- 무료
- 기관 내 무료
- 유료
-
Reconstruction of old well log data using deep learning imputation
Gian Antariksa,Radhi Muammar,Agung Nugraha,Jihwan Lee 대한산업공학회 2022 대한산업공학회 추계학술대회논문집 Vol.2022 No.11
Well logs are important datasets for interpreting subsurface geology because they represent the physical properties of the logged formations. However, at certain intervals, such information may be missing and/or incorrect due to drilling issues (e.g., significantly larger mud weight than formation integrity, resulting in formation damage), an inefficient logging process, or tool operational concerns. In the case of logging tool faults, such flaws are often identified by the existence of abnormal data spikes or erroneously low log readings in the damaged formation. To solve this issue, a new deep learning-based system was tested to imputate such missing values for sonic log (DT) data. The missing well log values were predicted using data-driven machine learning methods, specifically the GRU and LSTM, given time step sequence inputs from a window. To establish the ideal parameters for achieving the best validation score, an empirical research technique was used. Throughout the training phase, the relative value of various input parameters was analyzed in order to eliminate insensitive measurements and prioritize data with a high connection to the target variables. This paper highlights the study’s important findings, problems, and potential improvements for future studies.
연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
