Artificial Neural Network for Water Injection Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for optimizing water injection in petroleum well fields are computationally expensive and slow, relying on numerical reservoir simulators that struggle with nonlinearly correlated parameters and require extensive iterations, making it difficult to predict optimal water injection strategies effectively.
Innovation Solution
A computer-implemented system using a machine learning engine, specifically an artificial neural network model, to optimize well field injection by processing water injection rate and voidage replacement data, providing a cumulative oil production forecast, and comparing results with numeric simulator outputs for accuracy confirmation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If numerical reservoir simulators are used to optimize water injection, then prediction accuracy is improved, but computational time and cost increase significantly
Solution Approach 1:
The patent applies preliminary action by pre-training a machine learning model using historical reservoir simulation data and expert knowledge before actual optimization tasks. The trained model captures complex nonlinear relationships between injection parameters and production outcomes, enabling rapid predictions without re-running expensive numerical simulations for each scenario.
Solution Approach 2:
The patent uses copying by creating a simplified surrogate model (machine learning model) that replicates the behavior of the complex numerical reservoir simulator. This surrogate model is trained on simulation data and can predict outcomes much faster than the original simulator, effectively copying its predictive capability while eliminating its computational bottleneck.
2Reliability
If multiple reservoir simulation iterations are performed to confirm parameter impacts, then prediction reliability is improved, but computational cost increases
Solution Approach 1:
The patent implements feedback by using historical simulation results and production data to continuously train and refine the machine learning model. This feedback loop allows the model to learn from past performance and improve its predictions, achieving reliable results without requiring repeated expensive simulation iterations for each parameter test.
Solution Approach 2:
The patent applies preliminary action by pre-training the machine learning model on extensive historical data before actual optimization tasks. This preliminary training incorporates the learnings from numerous simulation iterations, so that during actual use, the model can provide reliable predictions without re-running those expensive simulations for each new scenario.
3Ease of operation
If manual judgment and hybrid approaches are used for optimization, then ease of operation is improved, but computational efficiency deteriorates
Solution Approach 1:
The patent applies self-service by enabling the machine learning model to autonomously perform optimization analyses without requiring manual intervention for each scenario evaluation. The model automatically processes input parameters, predicts outcomes, and identifies optimal injection strategies, freeing engineers from repetitive computational tasks while maintaining ease of operation through simple model interaction.
Data Source
AI summary
Systems and methods for well field optimization are described. A system includes a well injection planner coupled to a machine learning (ML) engine. A computer-readable memory stores a trained model, input data, and predictive results data. The well injection planner is implemented on at least one processor and is configured to provide the trained model and the input data to an inference stage of the ML engine and to receive the predictive results data output from the inference stage. The input data includes at least water injection rate and voidage replacement data for a well field and the output predictive results data includes a cumulative oil production forecast result for the well field.