Neural Network Water Cut Estimation via Pressure Drop Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current water cut measurement systems in oil and hydrocarbon flowmeters are intermittent, prone to drifting, and costly, leading to inaccurate results and a need for alternative methods to estimate water cut in real-time.
Innovation Solution
An intelligent water cut estimation system utilizing two pressure sensors, temperature sensors, and a neural network model, coupled with a data processor, to generate real-time estimates by calculating pressure drops and potential energy losses, iteratively refining the water cut estimation until convergence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a multi-phase flowmeter is placed at each well to measure water cut continuously, then measurement precision and reliability are improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent uses pressure sensors to create indirect copies of water cut information through pressure drop measurements, rather than directly measuring water cut with expensive flowmeters at each well. The pressure-based estimation system replicates the functionality of direct water cut measurement using simpler, cheaper sensors.
Solution Approach 2:
The patent introduces pressure sensors as intermediary devices that indirectly measure water cut through the relationship between pressure drop and water cut in multiphase flow. This intermediary approach allows water cut estimation without requiring direct water cut measurement instruments at each well.
2Device complexity
If a single multi-phase flowmeter is shared among multiple wells, then device complexity and cost are reduced, but measurement precision and reliability deteriorate due to intermittent measurements and flowmeter drifting
Solution Approach 1:
The patent creates continuous water cut estimates by processing pressure data through neural network models, generating virtual water cut measurements that fill the gaps between intermittent flowmeter readings. This copying approach reconstructs the continuous water cut information that would otherwise be lost.
Solution Approach 2:
The system uses neural network models that continuously process pressure sensor data to generate water cut estimates, providing real-time feedback on water cut conditions. This feedback mechanism maintains measurement precision without requiring continuous direct measurement.
3Measurement precision
If flowmeters are calibrated frequently to maintain accuracy, then measurement precision is improved, but loss of time and productivity increase due to calibration interruptions
Solution Approach 1:
The patent implements a feedback mechanism using neural network models that continuously monitor pressure data and generate water cut estimates in real-time. This continuous feedback system maintains measurement accuracy without requiring periodic calibration interruptions, as the model adapts to changing conditions through continuous data processing.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables continuous, accurate, and cost-effective real-time water cut measurement, instantly identifies malfunctioning flowmeters, optimizes calibration frequencies, and determines production allocation per well, reducing inefficiencies and enhancing measurement quality.
Implementation Method 1
determine a pressure drop between the two points based on the received pressure data from the at least two pressure sensors
Implementation Method 2
estimate a potential energy loss based on the dynamic pressure loss and the pressure drop
Data Source
AI summary
Systems and methods for generating pressure data from at least two pressure sensors, storing one or more parameters indicative of water cut in a neural network model, receiving pressure data from at least the two pressure sensors respectively indicative of the pressure at two points of a well bore, determining a pressure drop between the two points, generating an input water cut estimate, estimating a dynamic pressure loss to initiate an iterative process, estimating a potential energy loss, inverse modeling a water cut estimate, comparing the water cut estimate to the input water cut estimate to generate a water cut Δ, utilizing the water cut estimate as the input water cut estimate for the iterative process when the water cut Δ exceeds a threshold, and continuing the iterative process until the water cut Δ is below the threshold.