Proxy Simulator for Real-Time Oil and Gas Production Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current oil and gas field production management is hindered by the time-consuming nature of physics-based simulators, which are complex and require extensive 'history matching' processes, making real-time decision-making difficult due to outdated data and the impracticality of running multiple simulations for large fields.
Innovation Solution
Implementing a proxy simulator that uses a decision management system to define control parameters, execute simulations, and generate a proxy model for real-time optimization, allowing for instantaneous predictions and proactive control decisions by eliminating non-essential parameters and using a neural network to calculate sensitivities and correlations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If physics-based simulators are used for production optimization, then production predictions can be made, but the simulation time is too long for real-time decision-making
Solution Approach 1:
The patent creates a proxy model that copies the essential behavior of the complex physics-based simulator. The proxy model is trained on historical simulation data and then used for real-time predictions, providing accurate production forecasts without requiring lengthy physics-based simulations. This allows real-time decision-making while maintaining prediction accuracy.
Solution Approach 2:
The patent performs preliminary actions by pre-training the proxy model using extensive physics-based simulations and historical data before real-time operation. The model learns patterns and relationships in advance, so that during actual production optimization, predictions can be made instantly without running full physics-based simulations again.
2Measurement precision
If multiple physics-based simulators are run for history matching, then model accuracy improves, but the complexity and time required increases significantly
Solution Approach 1:
The patent replaces multiple complex physics-based simulators with a single proxy model that captures the essential relationships. The proxy model is trained on data from the physics-based simulators but then stands alone, eliminating the need to run multiple complex simulations for history matching while maintaining model accuracy.
Solution Approach 2:
The patent transforms the complex physics-based simulation parameters into a simplified proxy model parameter set. By changing from detailed physics parameters to proxy model parameters that are trained on historical data, the system achieves comparable accuracy with much lower complexity and computational requirements.
3Loss of information
If continuous monitoring sensors are installed for real-time data collection, then data availability improves, but the ability to respond to detected issues in time deteriorates
Solution Approach 1:
The patent implements a feedback loop where real-time sensor data is continuously fed into the proxy model, which immediately predicts production outcomes and identifies issues. The system provides real-time feedback to operators, enabling them to respond to detected issues such as excess water production or equipment problems immediately rather than waiting for periodic analysis.
Solution Approach 2:
The patent replaces the manual mechanical process of analyzing periodic data with an automated computational system. The proxy model automatically processes continuous sensor data and provides instant predictions and alerts, eliminating the time delay associated with manual data analysis and enabling real-time response to production issues.
Data Source
AI summary
Methods, systems, and computer readable media are provided for real-time oil and gas field production optimization using a proxy simulator. A base model of a reservoir, well, pipeline network, or processing system is established in one or more physical simulators. A decision management system is used to define control parameters, such as valve settings, for matching with observed data. A proxy model is used to fit the control parameters to outputs of the physical simulators, determine sensitivities of the control parameters, and compute correlations between the control parameters and output data from the simulators. Control parameters for which the sensitivities are below a threshold are eliminated. The decision management system validates control parameters which are output from the proxy model in the simulators. The proxy model may be used for predicting future control settings for the control parameters.