Virtual Flow Meter for Well Production Rate Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for testing well production rates are limited by the high costs of equipment and facilities, leading to infrequent testing, which hinders the optimization of production behavior on finer time scales and lacks incentive for frequent monitoring.
Innovation Solution
A machine-learning based system utilizing deep learning and neural networks to estimate production rates in producing wells by applying input parameters such as injection rates, pressure changes, lift rates, and operational history, effectively creating a virtual flow meter for more accurate and frequent monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional test separators are used for periodic well tests, then production rate measurement is achieved, but equipment cost and facility cost increase significantly
Solution Approach 1:
The patent creates a virtual flow meter that replicates the measurement function of a physical test separator using machine learning models. The system processes well test data through trained neural networks to generate flow rate measurements without requiring actual test separator equipment, thereby achieving measurement precision while eliminating the need for expensive physical infrastructure
Solution Approach 2:
The patent replaces the mechanical test separator system with a computational machine learning system. Instead of using physical equipment to separate and measure fluid flow, the system uses trained models that process input parameters (pressure, temperature, well depth, production history) to predict flow rates, substituting mechanical measurement with intelligent algorithms
2Productivity
If test separators are deployed for frequent testing, then production behavior optimization on finer time scales is enabled, but equipment cost and operational cost increase
Solution Approach 1:
The virtual flow meter creates a digital replica of the measurement capability, allowing frequent assessments of production behavior without the capital expenditure required for physical test separators. The system can evaluate production at any desired time scale by processing available well data through the trained machine learning model
Solution Approach 2:
The system utilizes existing well operational data and parameters that are already being collected during normal production operations. By processing this existing data through the machine learning model, the system provides continuous production assessment without requiring additional dedicated testing equipment or interrupting normal well operations
3Measurement precision
If periodic well testing is performed with test separators, then production rate determination is achieved, but time loss between tests increases
Solution Approach 1:
The virtual flow meter enables continuous production rate determination by processing well data through the machine learning model without interruption. Instead of periodic measurements separated by time intervals, the system can continuously assess production behavior as new data becomes available, eliminating the time loss between discrete testing events
Solution Approach 2:
The system creates a continuous virtual measurement record by repeatedly applying the trained model to incoming well data, generating an ongoing series of production rate estimates that replace the discrete, time-separated measurements from physical test separators
Data Source
AI summary
Various aspects described herein relate to a system that utilized deep learning and neural networks to estimate/predict an amount of natural resource production in a well given a set of parameters indicative of physical changes to the well. In one aspect, a virtual flow meter includes memory having computer-readable instructions stored therein and one or more processors configured to execute the computer-readable instructions to receive one or more input parameters indicative of physical changes to at least one well; apply the one or more input parameters to a trained neural network architecture; and determine one or more outputs of the trained neural network architecture, the one or more outputs corresponding to predicted fluid output of the at least one well.