Subsurface Uncertainty Quantification using Machine Learning – Improved Reservoir Management through Data Driven Predictions

• Communicate the uncertainty range associated with subsurface predictions and therefore, convey the risks to peers & management.
• Explain what Conformal Prediction is and how it can be used to quantify subsurface uncertainty.
• Understand that Conformal Prediction can be applied to regression, classification and time series data.
• Apply Conformal Prediction regression techniques to estimate prediction intervals as opposed to just generating single predictions points.
• Classify Conformal Prediction classification to quantify binary and multi-class (facies) uncertainty conveyed by classification being in sets of potential classes rather than a single prediction.
• Develop Conformal Prediction to quantify oil and gas production forecast uncertainty applied to time series (production history) data to provide complementary non-Decline Curve Analysis (DCA) perspectives.
• Learn to use classification and regression metrics to assess the accuracy and effectiveness of predictions.
• Create various plots to assist in the interpretation of Conformal Prediction techniques.

• Petrophysicists involve in analyse and interpret well log data to determine the properties of subsurface formations.
• Reservoir Engineers participate in estimating and optimising the extraction of hydrocarbons from reservoirs.
• Geoscientists such as Geologists and Geophysicists that study the physical aspects of the Earth, particularly the subsurface structures that contain resources like oil and gas.
• Any technical professionals, regardless of their specific roles, that are interest in gaining a deeper understanding of subsurface uncertainty quantification.
• Data Scientists and Machine Learning Specialists in the Energy Sector whose role specialise in data science and machine learning and are looking to apply their skills to the subsurface domain.

• Intermediate

Pre-requisite: Participants are recommended to have a basic understanding of a programming language (preferably Python), particularly of machine learning concepts of classification and regression.

Learning Tools: Participants need to have access to a computer with internet access and a Google account (minimally consisting of Gmail and Google Drive) as most of the course entails the use of the Google Colab platform. Instructions will be provided as Google Colab notebooks and working data will be made available online. Participants can work in pairs or teams if individual participants are not proficient with programming languages.

Unlock the potential of your workforce with customized in-house training programs designed specifically for the energy sector. Our tailored, in-house courses not only enhance employee skills and engagement but also offer significant cost savings by eliminating travel expenses. Invest in your team’s success and achieve specific outcomes aligned with your organization’s goals through our expert training solutions. Request for further information regarding our on-site or in-house training opportunities.

In our ongoing commitment to sustainability and environmental responsibility, we will no longer providing hard copy training materials. Instead, all training content and resources will be delivered in digital format. Inspired by the oil and energy industry’s best practices, we are leveraging on digital technologies to reduce waste, lower our carbon emissions, ensuring our training content is always up-to-date and accessible. Click here to learn more.