Drilling Planning Knowledge Retrieval for Wellsite Risk Mitigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current drilling planning systems rely on manual searches and individual experience to identify mitigation and prevention measures for wellsite actions, lacking an automated system to quickly and accurately provide relevant measures based on well characteristics.
Innovation Solution
A method that automatically determines mitigation and prevention measures by generating query vectors from wellsite action characteristics, performing semantic searches in a knowledge bank, and displaying relevant measures on a graphical interface using an Approximate Nearest Neighbor AI approach.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual search through knowledge management system is used to identify mitigation and prevention measures, then the system complexity is low, but the time required and productivity are reduced
Solution Approach 1:
The patent replaces manual mechanical search processes with an automated AI-based semantic search system. The system uses natural language processing and vector embeddings to automatically query the knowledge management system, eliminating the need for manual browsing and significantly improving identification speed while accepting the necessary system complexity for automation.
Solution Approach 2:
The system enables self-service by automatically performing the knowledge retrieval task without human intervention. The automated workflow includes extracting well characteristics, generating queries, searching the knowledge base, and presenting results, allowing the system to serve itself in identifying mitigation measures without requiring manual operation.
2Reliability
If individual experience is relied upon to identify mitigation and prevention measures, then no additional system is needed, but the reliability and consistency of identified measures are reduced
Solution Approach 1:
The system incorporates feedback mechanisms by continuously learning from the knowledge management system's content and improving its semantic search capabilities. The AI model processes structured well characteristics and compares them against historical data and established procedures in the knowledge base, ensuring consistent and reliable identification of appropriate mitigation measures based on proven methodologies.
Solution Approach 2:
The system transforms unstructured individual experience into structured parameters by extracting key well characteristics (depth, formation, diameter, tools, field conditions) and using these as standardized inputs for query generation. This parameterization ensures consistent and reliable identification of mitigation measures across different users and situations.
3Measurement precision
If comprehensive well characteristics are analyzed to generate accurate risks, then the measurement precision is improved, but the quantity of data processing and time required increases
Solution Approach 1:
The system performs preliminary action by pre-processing well characteristics into structured formats and pre-generating query vectors before the actual search. The AI model prepares the data in advance by extracting relevant features and transforming them into semantic representations, which accelerates the subsequent risk identification process while maintaining high accuracy.
Solution Approach 2:
The system segments the comprehensive well characteristics into distinct, manageable parameters (depth, formation, diameter, tools, field conditions) that can be processed independently. This segmentation allows the AI to analyze each parameter's contribution to risk separately, improving measurement precision while reducing overall processing complexity and time.
Data Source
AI summary
A method for automatically determining mitigation and prevention measures that are related to a wellsite action. The method includes obtaining a plurality of characteristics of the wellsite action, inputting the plurality of characteristics into a graphical interface, and generating a plurality of risks that correspond to the plurality of characteristics and then displaying risks on the graphical interface. Next, the associated risks are converted into a query vector within a knowledge bank. The knowledge bank is then queried to provide a mitigation or prevention measure relevant to the query vector. Specifically, an approximate nearest neighbor search may be used which finds a vector representing a mitigation or prevention measure that is closest to the query vector. The method also includes displaying the mitigation or prevention measure within the graphical interface so that a user may perform a wellsite action in response to the displayed mitigation or prevention measure.