Predictive Wellbore Model Using Reference Fiber Optic Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high cost and time-consuming process of deploying fiber optic cables in each wellbore for measuring parameters during wellbore operations, particularly in target wellbores, hinder efficient calibration and optimization of completion operations like hydraulic fracturing.
Innovation Solution
A pressure and rate-based model is generated using data from a reference wellbore with a fiber optic cable, which is then applied to target wellbores without fiber optics, estimating a uniformity index to output commands for adjusting completion operations, such as pump-rate and proppant control, thereby optimizing wellbore operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fiber optic cables are deployed in each target wellbore to measure parameters during wellbore operations, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent creates a virtual copy of the fiber optic measurement system by training a machine learning model on data from a reference wellbore equipped with fiber optics. This trained model is then applied to target wellbores to generate estimated measurements, effectively copying the measurement capability without physically deploying fiber optic cables in each target wellbore.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between the actual physical measurements taken in the reference wellbore and the estimated measurements needed for target wellbores. This intermediary processes the reference data and transforms it into predictions for target wellbores, eliminating the need for direct fiber optic deployment in each target wellbore.
2Measurement precision
If fiber optic cables are deployed in each target wellbore to measure parameters during wellbore operations, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent performs preliminary action by collecting and processing fiber optic data from a reference wellbore in advance to train a machine learning model. This pre-trained model can then be rapidly applied to multiple target wellbores without requiring time-consuming fiber optic deployment in each wellbore, significantly reducing the overall time required for measurements.
3Measurement precision
If fiber optic cables are deployed in each target wellbore to measure parameters during wellbore operations, then measurement precision is improved, but loss of substance increases
Solution Approach 1:
The patent uses the trained machine learning model to copy the measurement capabilities of fiber optic systems, allowing accurate parameter estimation in target wellbores without the need for physical fiber optic cable deployment. This virtual copying eliminates the material consumption associated with installing and removing fiber optics in each wellbore.
Data Source
AI summary
A system can receive data that can indicate a flow rate with respect to perforations in a reference wellbore. The data can be received from a fiber optic cable in the reference wellbore in a geographic area of interest. The system can determine a uniformity index, which can indicate a uniformity of flow with respect to the perforations, based on the data. The system can generate a pressure-based model by training a model using the uniformity index for applying the pressure-based model to a target wellbore in the geographic area of interest to determine controls to rate and proppant with respect to the target wellbore.