首页|New Machine Learning Study Findings Reported from Karlsruhe Institute of Technol ogy (KIT) (Machine Learning for Robust Structural Uncertainty Quantification In Fractured Reservoirs)
New Machine Learning Study Findings Reported from Karlsruhe Institute of Technol ogy (KIT) (Machine Learning for Robust Structural Uncertainty Quantification In Fractured Reservoirs)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News – Researchers detail new data in Machine Learning. According to news reporting out of Karlsruhe, Germany, by NewsRx edit ors, research stated, “Including uncertainty is essential for accurate decision- making in underground applications. We propose a novel approach to consider stru ctural uncertainty in two enhanced geothermal systems (EGSs) using machine learn ing (ML) models.” Financial support for this research came from Deutscher Akademischer Austausch D ienst (DAAD). Our news journalists obtained a quote from the research from the Karlsruhe Insti tute of Technology (KIT), “The results of numerical simulations show that a smal l change in the structural model can cause a significant variation in the tracer breakthrough curves (BTCs). To develop a more robust method for including struc tural uncertainty, we train three different ML models: decision tree regression (DTR), random forest regression (RFR), and gradient boosting regression (GBR). D TR and RFR predict the entire BTC at once, but they are susceptible to overfitti ng and underfitting. In contrast, GBR predicts each time step of the BTC as a se parate target variable, considering the possible correlation between consecutive time steps. This approach is implemented using a chain of regression models. Th e chain model achieves an acceptable increase in RMSE from train to test data, c onfirming its ability to capture both the general trend and small-scale heteroge neities of the BTCs. Additionally, using the ML model instead of the numerical s olver reduces the computational time by six orders of magnitude. This time effic iency allows us to calculate BTCs for 2 ‘ 000 different reservoir models, enabli ng a more comprehensive structural uncertainty quantification for EGS cases.”
KarlsruheGermanyEuropeCyborgsEme rging TechnologiesMachine LearningKarlsruhe Institute of Technology (KIT)
NETL
NSTL
