Well Control Expert System Using Bayesian Decision Networks

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current drilling and extraction methods for oil and natural gas are time-consuming, costly, and produce inconsistent results, failing to incorporate recent practices and opinions effectively.

Innovation Solution

A well control expert system utilizing a Bayesian decision network (BDN) model that provides recommendations for well control operations based on inputs, including kick indicators, verifications, details, and circulations, to optimize drilling processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional field experience and laboratory work techniques are used to develop and select drilling components and operational practices, then expertise and practical knowledge are utilized, but the process becomes time-consuming and expensive

Engineering Contradiction:

Improvequality of well control recommendationsVSAvoidtime for developing and selecting drilling components

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent creates a virtual copy of expert knowledge by training an artificial neural network on historical well control data and expert opinions. This digital model replicates the decision-making process of experienced drilling engineers, allowing rapid analysis of drilling scenarios without requiring physical presence or manual review of all historical cases. The neural network model serves as a reproducible copy of expert judgment that can be applied consistently across different drilling situations.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The system performs preliminary analysis by pre-training the neural network model with extensive historical data and expert knowledge before actual drilling operations. This preliminary preparation allows the system to rapidly provide well control recommendations during critical situations without requiring time-consuming manual analysis. The model has already processed and learned from numerous historical cases in advance, enabling fast decision support when needed.

Inventive Principle:

Principle #10Preliminary action

2Reliability

If traditional field experience and laboratory work techniques are used to develop drilling practices, then practical expertise is applied, but the results are inconsistent and fail to incorporate recent changes in practices and opinions

Engineering Contradiction:

Improveconsistency of drilling resultsVSAvoidincorporation of recent changes in practices

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The system incorporates feedback mechanisms by continuously updating the neural network model with new well control data and expert opinions. The model learns from historical outcomes and adjusts its recommendations based on what has worked or failed in previous drilling operations. This feedback loop ensures that the system consistently applies lessons learned while adapting to new practices and changing conditions in the drilling industry.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent implements a dynamic system where the neural network model can be retrained and updated with new data and expert knowledge over time. Unlike static traditional methods, this system evolves to incorporate recent changes in drilling practices, technologies, and expert opinions. The model's weights and parameters are adjusted based on new information, allowing it to adapt while maintaining consistent decision-making frameworks.

Inventive Principle:

Principle #15Dynamics

3Reliability

If extensive field experience and laboratory work are used to select drilling components and operational practices, then thorough evaluation is achieved, but costs increase significantly

Engineering Contradiction:

Improvequality of component and practice selectionVSAvoidcost of developing and selecting drilling components

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The patent replaces the mechanical and human-intensive process of expert review and laboratory testing with an computational neural network system. Instead of physically testing components and manually reviewing operational practices, the system uses automated machine learning algorithms to evaluate drilling components and practices based on historical data patterns. This substitution dramatically reduces the time and cost while maintaining or improving the quality of selections.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The neural network model performs self-evaluation by automatically analyzing drilling scenarios and providing well control recommendations without requiring continuous human intervention. The system serves itself by learning from historical data and autonomously making predictions about optimal drilling components and practices. This self-service capability eliminates the need for expensive and time-consuming manual expert analysis for each new drilling situation.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS9202175B2Systems and methods for an expert system for well control using Bayesian intelligence

Publication Date: 2015.12.01 SAUDI ARABIAN OIL CO

AI summary

Systems and methods are provided for a well control expert system that provides well control recommendations for a drilling system. The well control expert system includes a well control Bayesian decision network (BDN) model that receives inputs and outputs recommendations based on Bayesian probability determinations. The well control BDN model includes a circulation section, a well control practices section, and a troubleshooting section.