Applied Geostatistics

Establish why stochastic methods are useful in reservoir modelling. Use this to plan which statistical ideas and methods we need to study.

Goals - At the end of the module you should:

• appreciate why this course is important
• be able place the remainder of the course in a context of what is useful in geology.

The data

• preparation
• exploratory analysis

Statistical basics

• use of basic tools for summarising and understanding data
• use of basic tools for preparing data for further statistical analysis or for geological analysis

Goals - At the end of the module you should:

• be able to choose suitable techniques for analysis of data
• use every day tools such as Microsoft Excel® to check data quality and undertake the basic statistical analysis
• prepare data, using such common tools, for input to Petrel

Why do we need statistics?

The data requirements of a simulator

• collecting data together from different disciplines, on different support, for modelling.

Goals - At the end of the module you should:

• understand what to do when inferring from incomplete data: e.g. known values at wells, what happens between?
• know what to do when we cannot measure a parameter directly but have to predict from one we can measure

Sources of error

• confidence
• variance
• reliability

Goals - At the end of the module you should:

• be able to quantify the confidence in a prediction.
• know limitations of Petrel, when to transform and when to re-analyse.

How does a spatial dependence in the data affect statistics?

How to analyse data for spatial dependence.

Goals - At the end of these modules you should:

• be able to choose suitable techniques for analysis of data
• use everyday tools such as Microsoft Excel® to undertake the basic statistical analysis
• Take a spatial data set and plot it using Excel®

Geostatistical basics

• use of basic tools

Goals - At the end of the module you should:

• be able to establish trends
• use tools such as directional semivariograms as tools in their own right

Support and upscaling

• averaging and simulation techniques
• static and dynamic data

Goals - At the end of this module you should:

• be able to correctly choose techniques for upscaling different data types
• be able to distinguish correctly upscaled data from badly upscaled data

Estimation, including Kriging:

• understand the basics
• simple exercises to illustrate how kriging incorporates spatial variation and allows estimation of variance
• Using co-located co-kriging for better prediction of important parameters betwee

More advanced techniques

• Making use of new data:
  • History matching
  • Bayesian statistics
• Objects:
  
  • Sequential / Indicator simulation
• Object modelling in Petrel:
  • defining objects
  • incorporating objects in models
• Multi-point statistics:
  • limitations of semivariograms
  • combining with object modelling

Goals - At the end of this module you should:

• be able to correctly and appropriately use techniques presented in Petrel.

Recap and reinforcement:

• cover any techniques shown by the exercises to be shaky

Goals - At the end of this module you should:

• leave the course with a firm understanding of the basics and the ability to use all of the techniques covered in the course, in everyday work
• know how to re-familiarise (get back up to speed) after not using the techniques for any length of time