Data Analytics and Machine Learning for Subsurface Engineers
تحليل البيانات وتعلم الآلة لمهندسي ما تحت السطح
- Next date
- Available in-house
- Fees from
- —per delegate
- Formats
- In-house training0 sessions
- Duration
- 5 daysCertificate
Overview
Subsurface teams work with some of the largest and most complex datasets in any industry: seismic volumes, well logs, cores, pressure tests, production histories and drilling data collected over decades. Much of this data is underused because traditional workflows are slow and manual. Data analytics and machine learning give geoscientists and reservoir engineers new ways to clean and combine data, find patterns, predict rock and fluid properties, and speed up interpretation. Used well, these methods complement physical understanding rather than replace it, and they help teams make faster and better informed subsurface decisions.
This course introduces practical data analytics and machine learning methods for geoscientists and subsurface engineers. Participants learn the Python data science ecosystem, data cleaning and visualisation, and statistical analysis of well and production data. The course covers supervised learning for log prediction, facies classification and permeability estimation, unsupervised learning for clustering and rock typing, and time-series methods for production forecasting. It also explains neural networks for seismic interpretation, model validation and uncertainty, explainability, and how to deploy machine learning workflows within subsurface teams. Exercises use real datasets in prepared notebooks, so only basic programming knowledge is needed.
Objectives
- Use Python tools to load, clean and visualise subsurface datasets
- Apply statistical analysis to well log, core and production data
- Build supervised models for log prediction and facies classification
- Apply clustering methods for rock typing and data exploration
- Forecast production using decline and machine learning time-series methods
- Validate models, quantify uncertainty and explain model results
- Plan practical machine learning projects within subsurface teams
Who should attend
Geologists, geophysicists, petrophysicists, reservoir and production engineers, and data specialists working with subsurface data.
Training methodology
The course is strongly hands-on, with most sessions run in prepared Python notebooks using real well log, core, seismic and production datasets. Each method is introduced with a short explanation and a subsurface example, then participants apply it step by step and discuss results with the trainer. Exercises include predicting missing logs, classifying facies, clustering rock types and forecasting production. Group discussions focus on when machine learning adds value and when physics-based methods are better. The final day is dedicated to designing a realistic project for participants' own data.
Programme outline
Day1Data Foundations for Subsurface Work
- Why data analytics matters for exploration, development and production decisions today
- Types of subsurface data and their formats, including LAS, SEG-Y and production tables
- Introduction to Python, Jupyter notebooks and the main data science libraries
- Loading, cleaning and merging well log, core and production data from several sources
- Handling missing values, outliers, depth shifts and units in subsurface datasets
- Visualisation techniques such as crossplots, histograms and maps for data exploration
- Practical exercise: cleaning and visualising a multi-well log dataset in a notebook
Day2Statistics and Unsupervised Learning
- Descriptive statistics, distributions and correlation analysis applied to subsurface data
- Regression basics and the prediction of permeability from porosity and other logs
- Dimensionality reduction with principal component analysis for large multi-well log datasets
- Clustering methods such as k-means and Gaussian mixtures for rock typing
- Electrofacies identification and its comparison with facies descriptions made from core
- Geostatistics basics and the spatial nature of rock and fluid properties underground
- Practical exercise: clustering well logs to define rock types and comparing them with core
Day3Supervised Learning Applications
- Supervised learning concepts: training data, features, labels and model types
- Decision trees, random forests and gradient boosting for subsurface prediction tasks
- Predicting missing or poor quality logs such as sonic and density from other curves
- Facies classification from logs using labelled wells and blind well testing
- Permeability and production prediction from geological, petrophysical and completion parameters
- Avoiding data leakage and overfitting when working with correlated well data
- Practical exercise: training and testing a facies classification model on multi-well data
Day4Time Series, Production and Seismic Applications
- Production data analysis and data cleaning for decline curve studies
- Automated decline curve analysis and comparison with machine learning forecasts
- Time-series models for forecasting production and detecting well performance changes
- Neural networks and deep learning concepts explained for subsurface professionals
- Machine learning applications in seismic interpretation such as fault and salt detection
- Physics-informed and hybrid approaches that combine reservoir engineering principles with data
- Practical exercise: forecasting well production with decline analysis and a machine learning model
Day5Validation, Explainability and Deployment
- Model validation methods, performance metrics and the quantification of uncertainty
- Explainable machine learning techniques that help experts trust model results
- Data management, version control and reproducible workflows for subsurface teams
- Integrating machine learning into the existing interpretation and modelling software used by teams
- Building skills, collaboration between domain experts and data scientists, and governance
- Group exercise: designing a machine learning project for a real subsurface problem
- Final assessment and discussion of data analytics opportunities in participants' own teams
Dates and fees
There are no scheduled dates for this course at the moment. Request it as in-house training, or ask us about the next date.
Certificate
Every delegate who attends the full programme receives a Harfard Technology certificate of completion stating the course title, duration, dates and venue.
Frequently asked questions
Will I receive a certificate?
What is the difference between classroom and online?
Can you run this course for our company?
How do I register?
Related courses
Digital Transformation, AI & SAPArtificial Intelligence and Digital Drilling
Available in-houseDigital Transformation, AI & SAPDigital Twins and AI for Refineries and Gas Plants
Available in-houseDigital Transformation, AI & SAP