Hybrid AI Multiphase Flow Prediction for Well Parameter Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional multiphase flow meters for determining oil, gas, and water flow rates in wells are expensive and require maintenance, necessitating a more cost-effective and reliable method for accurate phase flow rate measurement.
Innovation Solution
A hybrid AI model integrated with physics-based simulations uses temperature and pressure data to predict flow rates, optimizing well operation parameters through an optimizer, and automatically adjusts these parameters for improved accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional multiphase flow meters are installed to measure flow rates, then measurement accuracy is improved, but cost and maintenance requirements increase
Solution Approach 1:
The patent creates a virtual copy of the multiphase flow metering system using AI models and physics-based simulations. Instead of relying on physical flow meters, the system uses sensor data (temperature, pressure) combined with machine learning models to predict flow rates, thereby eliminating the need for expensive physical measurement devices while maintaining measurement capability through computational simulation
Solution Approach 2:
The patent replaces the mechanical/physical flow metering system with an information-processing system. The physical flow meters that directly measure fluid properties are substituted with AI models that process temperature and pressure sensor data to infer flow rates, transforming a mechanical measurement problem into an computational prediction problem
2Measurement precision
If multiphase flow meters are used to determine phase flow rates, then flow rate accuracy is improved, but device cost and maintenance needs worsen
Solution Approach 1:
The AI model serves multiple functions: it acts as a flow meter, a diagnostic tool, and an optimization system. The same computational model that predicts flow rates can also identify meter malfunctions and optimize well operations, eliminating the need for separate dedicated devices for each function and reducing overall system cost
Solution Approach 2:
The system creates a virtual replica of the flow measurement and control system using AI models. This virtual copy processes sensor data to predict flow rates and optimize operations, replacing expensive physical flow meters and control systems with computational models that have no moving parts and require minimal maintenance
3Measurement precision
If real-time flow prediction is implemented using AI models, then measurement accuracy is improved, but computational complexity increases
Solution Approach 1:
The system performs preliminary training of AI models using extensive simulation data generated from physics-based multiphase flow models. This pre-training phase creates a knowledge base that the model uses for rapid real-time predictions, separating the computationally intensive model development phase from the fast execution phase, thereby reducing real-time computational complexity
Solution Approach 2:
The patent introduces physics-based simulation models as an intermediary layer between raw sensor data and AI predictions. These physics models generate training data and provide physical constraints that guide the AI model, acting as a mediator that translates complex physical phenomena into a form that machine learning algorithms can process efficiently
4Productivity
If automated optimization of well operation parameters is implemented, then productivity is improved, but automation extent increases
Solution Approach 1:
The system implements a closed-loop feedback mechanism where AI-predicted flow rates continuously inform optimization decisions, and the results of parameter adjustments are fed back into the model for continuous learning and refinement. This feedback loop enables the system to automatically adapt to changing well conditions and optimize production without human intervention
Solution Approach 2:
The well optimization system performs self-service by automatically adjusting its own operation parameters based on real-time sensor data and AI predictions. The system monitors its own performance, identifies optimization opportunities, and executes parameter adjustments autonomously, eliminating the need for external human operators to manually tune well parameters
Data Source
AI summary
A method for determining multiphase flow rates from a well in real-time, the method includes obtaining temperature and pressure data of flow of a multiphase fluid that includes a first phase at one or more locations in a wellbore conveying the flow. The flow is controlled, at least in part, by a set of well operation parameters. The method further includes determining, with an artificial intelligence (AI) model, a first predicted flow rate of the first phase at a first location given an input that includes the temperature and pressure data. The method further includes determining, with an optimizer applied to the AI model, an optimal set of well operation parameters based on, at least, the first predicted flow rate. The method further includes adjusting, automatically, the set of well operation parameters to the optimal set of well operation parameters.