Well Assignment in Reservoir Simulation Using Bin-Packing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Reservoir simulations face increased computational runtime as the number of wells increases, leading to inefficiencies in parallel processing due to uneven distribution of well activities across computer processors.

Innovation Solution

Employing a bin-packing problem algorithm to dynamically assign wells to parallel processing stages, optimizing the distribution of well activities and reducing computational imbalance by using coarsened simulations to estimate well activity data before fine-grid simulations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the number of wells in a reservoir simulation is increased to improve production management accuracy, then the reliability of production predictions is improved, but the computational runtime increases

Engineering Contradiction:

Improveproduction prediction accuracyVSAvoidcomputational runtime

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The reservoir simulation is divided into multiple parallel processing stages, with each stage handling a subset of wells. This segmentation allows the simulation to process multiple wells simultaneously across different processor cores, reducing the overall computational runtime while maintaining the ability to model numerous wells for accurate production predictions

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic well assignment that adapts to varying well activity levels during different time periods. The bin-packing algorithm dynamically reassigns wells to processing stages based on current activity patterns, ensuring optimal utilization of computational resources throughout the simulation period, thereby reducing total runtime without sacrificing prediction accuracy

Inventive Principle:

Principle #15Dynamics

2Productivity

If more wells are assigned to parallel processing stages to improve productivity, then the productivity of simulation processing is improved, but computational imbalance increases due to uneven well activity distribution

Engineering Contradiction:

Improvesimulation processing throughputVSAvoidcomputational balance

Core Design Contradiction:

ProductivityVSStability of the object's composition

Solution Approach 1:

The system dynamically adjusts well assignments across parallel processing stages based on real-time well activity data. The bin-packing algorithm continuously reoptimizes the distribution of wells to processing stages, adapting to changing activity patterns and ensuring balanced computational loads across all processors throughout the simulation

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent incorporates feedback mechanisms where well activity data is collected and analyzed to inform subsequent well assignment decisions. The bin-packing algorithm uses this feedback to continuously optimize the distribution of wells across processing stages, maintaining computational balance while maximizing processing throughput

Inventive Principle:

Principle #23Feedback

3Ease of operation

If a simple well assignment method is used to reduce device complexity, then the ease of operation is improved, but computational imbalance occurs due to uneven distribution of well activities

Engineering Contradiction:

Improveassignment implementation simplicityVSAvoidcomputational balance

Core Design Contradiction:

Ease of operationVSStability of the object's composition

Solution Approach 1:

The bin-packing algorithm automatically performs the complex task of optimally assigning wells to parallel processing stages without requiring manual intervention. The system self-adjusts to varying well activity patterns and dynamically reoptimizes assignments, providing both simplicity of operation and computational balance through automated intelligent decision-making

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS11261707B2Method and system for well assignment in a reservoir simulation based on well activity

Publication Date: 2022.03.01 SAUDI ARABIAN OIL CO

AI summary

A method may include obtaining well activity data regarding various wells in a reservoir region of interest. The method may further include determining a predetermined well assignment for various parallel processing stages using a bin-packing problem algorithm and the well activity data. The predetermined well assignment may assign the wells to the parallel processing stages. A respective parallel processing stage among the parallel processing stages may perform a portion of a reservoir simulation using a respective computer processor and a well among the wells. The method may further include simulating the reservoir region of interest based on the wells being simulated according the predetermined well assignment for the parallel processing stages.