Hydraulic Fracture Modeling for Well and Cluster Spacing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for hydraulic fracturing in unconventional reservoirs, such as tight reservoirs, are computationally expensive and do not adequately account for complex transport mechanisms, leading to inefficiencies in hydrocarbon production forecasting and well design.
Innovation Solution
Utilizing artificial intelligence techniques to generate multi-stage hydraulic fracture models integrated into reservoir models, reducing computational costs by training machine learning models with numerical simulation data to optimize well spacing and cluster spacing, and predicting hydrocarbon production performance before drilling or fracturing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If numerical simulations are used to model hydraulic fractures and forecast production, then prediction accuracy is improved, but computational cost and time increase
Solution Approach 1:
The patent applies preliminary action by pre-training machine learning models using numerical simulation data before actual production forecasting. The ML models are trained offline on comprehensive simulation datasets that capture complex fracture behaviors, so that during actual forecasting, predictions can be made rapidly without running full numerical simulations each time. This resolves the contradiction by performing the computationally intensive work in advance.
Solution Approach 2:
The patent uses copying by creating surrogate machine learning models that replicate the behavior of complex numerical simulations. Instead of running expensive numerical simulations for every forecasting scenario, the system copies the essential fracture propagation and production behavior into trained ML models that can be evaluated rapidly. This maintains prediction accuracy while dramatically reducing computational time.
2Productivity
If machine learning models are used to propagate fracture models, then computational cost is reduced, but complexity of transport mechanisms may not be adequately accounted for
Solution Approach 1:
The patent introduces an intermediary approach by using machine learning models as surrogate representations of complex physical systems. The ML models are trained on numerical simulation data that fully accounts for complex transport mechanisms, phase transitions, and nonlinear flow behaviors. During forecasting, the ML models serve as computationally efficient intermediaries that capture these complex mechanisms without requiring direct simulation of them. This resolves the contradiction by maintaining model fidelity through training data while achieving computational efficiency during deployment.
3Productivity
If multi-stage hydraulic fracturing is performed to increase reservoir conductivity, then hydrocarbon production is improved, but computational cost for modeling increases
Solution Approach 1:
The patent applies segmentation by dividing the multi-stage fracturing process into individual stage models that can be propagated independently using machine learning. Each fracture stage is modeled separately, and the ML-based propagation technique allows efficient prediction of fracture geometry and production for each stage without requiring full numerical simulation of the entire multi-stage process. This reduces computational cost while maintaining the ability to forecast improved hydrocarbon production from multi-stage fracturing.
Data Source
AI summary
Systems and methods for hydraulic fracturing a subsurface formation include obtaining a reservoir model representing a subsurface formation; obtaining a baseline hydraulic fracture model for one or more stages of a well in the subsurface formation based on the reservoir model. Additional hydraulic fracture models for one or more additional stages in one or more wells in the subsurface formation are generated using a machine learning model trained based on the baseline hydraulic fracture model. The additional hydraulic fracture models are integrated into the reservoir model. Hydrocarbon production from the subsurface formation is simulated using the reservoir model with the integrated additional hydraulic fracture models; and a well spacing for the one or more wells or a cluster spacing for hydraulic fractures in the one or more wells is determined based on the simulated hydrocarbon production.