本文へジャンプ

成果報告書詳細
管理番号20190000000211
タイトル*平成30年度中間年報 超臨界地熱発電技術研究開発 超臨界地熱資源への調査井掘削に資する革新的技術開発 AIによる超臨界地熱資源評価・掘削技術
公開日2019/5/22
報告書年度2018 - 2018
委託先名国立研究開発法人産業技術総合研究所 国立大学法人京都大学 国立大学法人室蘭工業大学 国立大学法人東北大学 国立大学法人九州大学 地熱エンジニアリング株式会社 地熱技術開発株式会社 国立大学法人秋田大学 帝石削井工業株式会社 株式会社物理計測コンサルタント
プロジェクト番号P18008
部署名新エネルギー部
和文要約
英文要約Title: Technology for estimation of deep geothermal structures and drilling bit life using artificial intelligence

(1) Estimation of deep geothermal structures using artificial intelligence
For explorations of supercritical geothermal resources, it is important to predict the depth and temperature of potential reservoir targets before drilling. In this project, we aim to develop a novel artificial intelligence (AI)-based approach for deep geothermal structure estimation using various kinds of datasets available at geothermal sites, e.g., data from shallow wellbores, geophysical, geological, and geochemical data.
In FY2018, we have conducted the following studies:
(1-1) Review of existing studies in relevant topics
- Survey of existing studies applying AI techniques in the field of earth sciences and petroleum exploration has been completed.
- AI methods suited for the deep geothermal structure estimation has been suggested.
(1-2) Collection of training dataset for the AI models:
- Survey of high-temperature geothermal wells in Japan and also in other countries has been completed.
- Survey of available datasets of Kakkonda-Matsukawa geothermal area has been completed. Some of the data has already been collected as digital data.
- A literature review of granitic rock properties such as resistivity, Vp/Vs under various temperature and pressure conditions have been completed.

(2) Estimation of deep geothermal drilling bit life using artificial intelligence
The goal is to develop AI system for drilling operation assist that predicts and evaluates drilling bit life that can be realized at low operating costs by using existing mud logging and open software.
In FY2018, we have conducted the following studies:
(2-1) Overall system design
-The general coordination among each research items
-Conceptual design of the system to be the goal.
(2-2) Systematization of drilling phenomena
-A literature review on drilling models and drilling bit wear diagnostic techniques
-Selection of teacher data
-Selection of bit wear evaluation parameters applicable to machine learning with neural networks based on examination of drilling models.
(2-3) Sensor technology
-Preparation of the data set by extracting the necessary data items from the mud logging data of the actual wells.
(2-4) Machine learning technology
-A literature review on AI technology (mainly neural networks) used in the evaluation of drilling bits implemented in overseas oil and gas fields
-Extraction of effective parameters for drilling bit lifetime evaluation for the basic design.
ダウンロード成果報告書データベース(ユーザ登録必須)から、ダウンロードしてください。

▲トップに戻る