History Matching Simulation Model Using Self Organizing Maps

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current history matching methods for numerical simulation models, particularly in reservoir fields, rely on static geological information and rectangular boxes that do not align with natural reservoir characteristics, leading to inefficient parameter adjustment and numerous simulation runs.

Innovation Solution

The method employs Self Organizing Maps (SOM) to dynamically group grid blocks into regions based on similar behavior, considering various parameters like permeability and porosity, using weight factors and rules to identify regions automatically, thereby simplifying the history matching process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If rectangular boxes are used to define regions in the reservoir field, then the region selection process is simplified and standardized, but the regions do not align with natural reservoir characteristics leading to inaccurate history matching

Engineering Contradiction:
Improveease of region definitionVSAvoidaccuracy of region alignment with reservoir characteristics
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The patent transforms the static, fixed rectangular box region definition into a dynamic process where regions are automatically generated based on reservoir behavior patterns. The system dynamically identifies and defines regions that adapt to the actual reservoir characteristics and flow patterns, rather than forcing reservoir data into predetermined geometric shapes.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The invention changes the parameters used to define regions from simple geometric coordinates (rectangular boxes) to complex behavioral parameters including production history, pressure data, and flow patterns. This parameter transformation allows regions to be defined by actual reservoir performance characteristics rather than arbitrary geometric boundaries.

Inventive Principle:
Principle #35Parameter changes

2Ease of manufacture

If static geological information is used to define regions, then the region definition process is straightforward and based on available data, but the regions do not reflect hydraulic parameters and production-related changes over time

Engineering Contradiction:
Improveease of region definitionVSAvoidadaptability to hydraulic parameters and production changes
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary analysis of production history and hydraulic behavior patterns before finalizing region definitions. By pre-processing the dynamic production data and identifying behavioral patterns in advance, the system can then automatically define regions that reflect both geological structure and hydraulic performance characteristics.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The invention implements a feedback mechanism where production history and hydraulic parameter changes are continuously monitored and used to refine and update region definitions. The system uses actual reservoir performance feedback to adjust and optimize region boundaries, ensuring they remain aligned with current reservoir behavior rather than becoming outdated.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If traditional history matching methods are used with manual region adjustment, then the process allows for detailed control and understanding of parameter changes, but numerous simulation runs are required making the process time-consuming

Engineering Contradiction:
Improvequality of history matchVSAvoidtime for simulation runs
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system implements self-service automation where the software automatically performs region identification, parameter calibration, and history matching optimization without requiring numerous manual simulation runs. The automated system serves itself by using algorithms to identify patterns and adjust parameters efficiently, reducing the need for repetitive manual intervention and extensive simulation testing.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The invention replaces the mechanical trial-and-error process of manual history matching with automated computational algorithms. Instead of physically running numerous simulations with manual parameter adjustments, the system uses computer-based optimization algorithms and pattern recognition to automatically determine optimal parameters and region definitions, substituting computational intelligence for manual mechanical processes.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP1994481B1Method for history matching a simulation model using self organizing maps to generate regions in the simulation model
Publication Date: 2018.10.24 LOGINED
  • EP1994481B1 patent drawingFigure 1
  • EP1994481B1 patent drawingFigure 2~3
  • EP1994481B1 patent drawingFigure 3A

AI summary

A method of history matching a simulation model is disclosed comprising: (a) defining regions exhibiting similar behavior in the model thereby generating the model having a plurality of regions, each of the plurality of regions exhibiting a similar behavior; (b) introducing historically known input data to the model; (c) generating output data from the model in response to the historically known input data; (d) comparing the output data from the model with a set of historically known output data; (e) adjusting the model when the output data from the model does not correspond to the set of historically known output data, the adjusting step including the step of arithmetically changing each of the regions of the model; and (f) repeating steps (b), (c), (d), and (e) until the output data from the model does correspond to the set of historically known output data.