日本地球惑星科学連合2019年大会

講演情報

[J] ポスター発表

セッション記号 M (領域外・複数領域) » M-GI 地球科学一般・情報地球科学

[M-GI37] 情報地球惑星科学と大量データ処理

2019年5月26日(日) 17:15 〜 18:30 ポスター会場 (幕張メッセ国際展示場 8ホール)

コンビーナ:村田 健史(情報通信研究機構)、本田 理恵(高知大学自然科学系理工学部門)、野々垣 進(国立研究開発法人 産業技術総合研究所 地質情報研究部門 情報地質研究グループ)、堀之内 武(北海道大学地球環境科学研究院)

[MGI37-P06] Analysis of vessel status information of past research cruises and prediction of downtime for future cruise

*山岸 保子1谷津 健2柏瀬 憲彦2牧 哲司2山室 悠太2,3阪口 秀1 (1.国立研究開発法人海洋研究開発機構 数理科学・先端技術研究分野、2.国立研究開発法人海洋研究開発機構 海洋工学センター、3.日本海洋事業株式会社)

Japan Agency for Marine-Earth Science and Technology (JAMSTEC) owns several research vessels and is responsible for the operation of the research cruise by these vessels. The ship time of research vessels belonging to JAMSTEC has declined over the last 10 years due to decrease in the number of vessels and the budget for research cruises. On the other hand, the occurrence of trench type large earthquake, global warming, and seabed resource exploration have increased the importance of the oceanic observation. Under these circumstances, it is indispensable to plan the most optimal research cruises to carry out the necessary oceanic observations. At present, JAMSTEC has to plan all research cruises for next fiscal year one year before. In order to perform the most effective research cruises, our project aims to provide useful information when planning future research cruise. Our project started in FY 2016. We first started to investigate the efficiency of the past research cruises. We have compiled time-sequence data of vessel status and performed observations in the past cruise, and estimated downtime for each cruise, which is a period when the observation or cruise is interrupted due to rough weather, machine trouble, etc. The downtime is one of the indicators for the efficiency of the research cruise. We also have developed database system to store the time-sequence data and information of each research cruises including downtime. The time-sequence data has been compiled based on E-mail sent from vessel to ground management office in the cruise. In E-mail, the information is written in text, ie in natural language. Therefore, the time-sequence data should be compiled by manually, it takes a lot of time to digitize it. Currently, we are also developing an application that automatically create the time-sequence data of vessel status and performed observation. Next, we have analyzed the downtime and the information of the past research cruises and searched for important factors that affect the downtime. And then, we have developed an application to predict the downtime of future cruises by using machine learning. In this study, we will introduce the analysis results of the downtime and the progress of the application development. Acknowledgments: we are grateful to Mr. Morisaki and Ms. Sada for their supports to make data stored into the database system.