Reservoir Connectivity Emulation for Real-Time Well Production Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting well production profiles in hydrocarbon reservoirs, such as 3D spatial reservoir connectivity mapping, are limited by static property mapping that neglects dynamic connectivity and are computationally expensive.
Innovation Solution
An AI-based emulation of reservoir connectivity mapping using bi-directional long short-term memory models, sequential models, and bifurcated graph models integrates spatial and temporal domain data to estimate well production profiles, facilitating real-time predictions and continuous training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If static property mapping techniques are used for sweet spot mapping and ranking, then the method is simple and easy to implement, but the prediction accuracy of well production profiles is limited
Solution Approach 1:
The patent creates a virtual copy of the reservoir system through digital twin technology, replicating the physical reservoir's geometry, rock properties, and fluid characteristics in a computational model. This virtual replica enables accurate production predictions without requiring complex physical experiments or simplified static mappings, thus improving prediction accuracy while maintaining computational efficiency.
Solution Approach 2:
The patent transforms static reservoir property parameters into dynamic parameters that evolve over time within the digital twin model. By incorporating time-dependent variations in pressure, temperature, and fluid saturation, the model achieves higher prediction accuracy while remaining computationally tractable through efficient numerical algorithms.
2Measurement precision
If simulation-based techniques are used for sweet spot mapping and ranking, then the prediction accuracy of well production profiles is improved, but the computational cost increases significantly
Solution Approach 1:
The patent performs preliminary calibration and validation of the digital twin model using historical production data and reservoir test results before deploying it for predictive simulations. This pre-training phase establishes accurate parameter relationships and boundary conditions, enabling the model to generate high-accuracy predictions with reduced computational iterations and lower energy consumption during operational forecasting.
Solution Approach 2:
The patent replaces computationally intensive physics-based simulation mechanisms with a hybrid approach that combines simplified flow equations with machine learning algorithms trained on simulation data. This substitution maintains prediction accuracy while dramatically reducing computational cost and energy requirements for routine production forecasting and sweet spot identification.
3Measurement precision
If dynamic connectivity is integrated into reservoir modeling, then the prediction accuracy of well production profiles is improved, but the device complexity and data processing requirements increase
Solution Approach 1:
The patent divides the reservoir into discrete grid cells or geological units, each with its own dynamic properties and connectivity relationships. This segmentation allows the digital twin model to capture complex dynamic connectivity patterns through modular computational units, improving prediction accuracy while managing system complexity through hierarchical organization and efficient data structures.
Data Source
AI summary
Systems, devices, and methods for prediction of well production profiles in hydrocarbon reservoirs. Spatial domain data and temporal domain data are received, from one or more probes. The spatial domain data characterizes rock properties within a subterranean volume including a hydrocarbon reservoir and the temporal domain data characterizes temporal variations of well productions. The spatial domain data is formatted as shell-formatted data including a shell format of concentric shells processable by prediction models. The shell-formatted data is provided as input for the prediction models to generate a plurality of well production profiles and hydrocarbon reservoir depletion profiles. The prediction models are trained using the temporal domain data. The hydrocarbon production from the hydrocarbon reservoir is managed based on the production profiles and the hydrocarbon reservoir depletion profiles.