Time-Dependent Well Connectivity for Reservoir Model Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining well connectivity in hydrocarbon reservoirs are time-consuming, require physical intervention, and are prone to errors due to reliance on expert knowledge, especially in the early stages of field development.
Innovation Solution
A machine learning-based approach that utilizes production time-series data and metadata to train a model, predict future production trends, determine well connectivity scores, and detect changes in reservoir dynamics using change point methods, allowing for adaptive well development plans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional expert-based methods are used to determine well connectivity, then analysis may be performed with limited data, but the process is time-consuming and prone to errors
Solution Approach 1:
The patent replaces manual expert-based analysis with an automated machine learning system that processes production data to determine well connectivity. The ML model automatically identifies connectivity patterns and quantifies influence magnitudes without requiring expert intervention, thereby reducing analysis time while maintaining or improving accuracy through consistent algorithmic application.
Solution Approach 2:
The patent creates a virtual model of reservoir connectivity by training an ML system on historical production data. This digital twin or copy of the physical reservoir system allows for rapid analysis and prediction of well connectivity without requiring physical intervention or time-consuming field measurements, enabling quick assessment while preserving the essential connectivity relationships.
2Reliability
If physical intervention methods are used to assess reservoir connectivity, then direct measurements may be obtained, but the process requires physical intervention and is time-consuming
Solution Approach 1:
The patent substitutes physical intervention methods with a computational ML-based system that analyzes existing production data to infer connectivity relationships. This approach eliminates the need for physical well tests or direct measurements while providing reliable connectivity assessment through pattern recognition in production data, thereby simplifying operations while maintaining reliability.
Solution Approach 2:
The patent enables the reservoir system to reveal its own connectivity characteristics through production data that naturally records the influence of injectors on producers. The ML system simply needs to analyze this self-generated data without external physical intervention, allowing the reservoir to effectively assess its own connectivity through its production behavior.
3Productivity
If initial reservoir structure assumptions are used early in field development, then rapid initial analysis is possible, but assumptions may need modification as more data becomes available
Solution Approach 1:
The patent implements a dynamic connectivity assessment system using ML that can adapt as new production data becomes available. The system continuously learns from incoming data and updates connectivity predictions, allowing the model to evolve from initial assumptions to data-driven insights as field development progresses, thereby maintaining both efficiency and adaptability.
Solution Approach 2:
The patent incorporates feedback mechanisms where production data continuously informs and refines the ML model's connectivity predictions. As more production data is collected, the system learns from actual reservoir response and adjusts its connectivity assessments, enabling the model to adapt to new information while maintaining productive field development operations.
Data Source
AI summary
Systems and methods for optimal reservoir model adaptation based on spatiotemporal well connectivity analysis are disclosed. The methods include obtaining production time-series data and metadata from a plurality of wells in a subsurface; preprocessing the production time-series data and the metadata; training a ML model with the preprocessed production time-series data and metadata; predicting a future production times series data with the ML model; determining well connectivity scores of a subsurface with the ML model; detecting a change in reservoir dynamics with a change point method; and modifying a well development plan based on the change in reservoir dynamics.