About This Classroom Training
Quantifying uncertainty in subsurface environment is a complex challenge that machine learning techniques can effectively address and assist. The heterogeneous nature of subsurface formations, characterised by varying properties like porosity and permeability, often defies traditional modelling approaches. Machine learning algorithms excel at identifying patterns in this complexity, offering insights that conventional models might have missed.
The energy sector's data acquisition limitations often result in sparse, unevenly distributed data points. Machine learning mitigates this through techniques like data augmentation and advanced interpolation, enabling more reliable predictions from limited data. Furthermore, these models can better account for measurement errors in well logs and core data by learning to recognise and adjust for data inaccuracies.
This 4-day course is tailored for energy sector professionals aiming to enhance their expertise in managing subsurface uncertainties. It covers essential topics including regression, classification, and time series analysis, with a focus on applying Conformal Prediction for estimating prediction intervals. By bridging traditional geoscience with modern data science, the course equips participants to effectively communicate uncertainty and make informed decisions in exploration and production activities by utilising the presence of data.
Participants will gain a comprehensive understanding of leveraging machine learning to quantify subsurface uncertainty, enhancing their capabilities in reservoir management and production forecasting. This training not only elevates technical skills but also prepares professionals for the challenges of the evolving energy landscape.
This course will be delivered face-to-face over 4-day sessions, comprising of 8 hours per day, 1 hour lunch and 2 breaks of 15 minutes per day. Course Duration: 32 hours in total, 32 CPD points. This course can also be delivered through Virtual Inspector Led Training.