Horizontal Well Placement Optimization Using Neural Network Forecasts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Intuitive engineering judgment is ineffective for optimizing well placement, and existing computational methods are time-consuming and computationally expensive due to the complexity of geological and petrophysical parameters, especially when nonlinear correlations evolve over time.
Innovation Solution
Utilizing an artificial neural network (ANN) system to predict and optimize well placement by integrating reservoir properties and well locations, which is trained using numerical reservoir simulations to provide rapid well placement forecasts with high accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If numerical reservoir simulation is used to predict well placement and production performance, then prediction accuracy is improved, but computational time and cost increase significantly
Solution Approach 1:
The patent creates a simplified copy of the complex reservoir simulation system in the form of an artificial neural network. The ANN learns from training data generated by numerical simulations and provides rapid predictions without requiring extensive computational resources, thus achieving high prediction accuracy with significantly reduced computational time.
Solution Approach 2:
The patent transforms the problem from direct numerical simulation to a parameter-based neural network model. By changing the approach from solving complex differential equations to using trained neural network parameters, the system achieves fast predictions while maintaining accuracy, effectively resolving the time-accuracy tradeoff.
2Reliability
If numerical reservoir simulation is used to evaluate well placement options, then production performance prediction is improved, but device complexity and computational cost increase
Solution Approach 1:
The patent creates a simplified copy of the complex reservoir simulation system in the form of an artificial neural network. The ANN learns from training data generated by numerical simulations and provides rapid predictions without requiring extensive computational resources, thus achieving high prediction accuracy with significantly reduced computational time.
Solution Approach 2:
The patent replaces the complex mechanical computational simulation system with an artificial neural network system. This substitution eliminates the need for complex numerical solvers and extensive computational resources while maintaining reliable production performance predictions, thereby reducing device complexity.
3Ease of operation
If intuitive engineering judgment is used for well placement, then ease of operation is improved, but prediction accuracy deteriorates
Solution Approach 1:
The patent introduces an artificial neural network as an intermediary between engineering judgment and well placement decisions. The ANN processes complex geological and reservoir data to provide objective, accurate predictions, serving as a mediator that enhances both the ease of operation and prediction accuracy by automating the analysis of multiple parameters.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An artificial neural network (ANN) system for predicting and optimizing well placement, productivities, and development includes: an ANN; computer hardware for building, training, using, and storing the ANN; and computer software for programming and processing the ANN. The ANN includes an input layer of input nodes representing at least one input parameter, an output layer of output nodes representing at least one output parameter, and at least one hidden layer operatively coupling the input layer to the output layer.