Virtual Sensor Neural Network for Downhole Reservoir Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for characterizing and modeling petrophysical properties of reservoirs in natural resource exploration are limited by the need for ex-situ laboratory experiments, which prevent real-time forecasting and maintenance planning of downhole sensors, relying on a priori scheduling and assumptions.
Innovation Solution
The use of virtual sensors that leverage existing in-situ physical sensors to collect data, which is then input into a neural network model to predict formation and fluid properties, allowing for real-time forecasting and classification of responses to indicate maintenance needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ex-situ laboratory experiments are used to characterize reservoir petrophysical properties, then measurement precision is improved, but loss of time occurs due to inability to perform real-time in-situ measurements
Solution Approach 1:
The patent creates virtual sensors that are computational copies of physical sensors, using machine learning models to replicate sensor responses and predict formation properties. This allows real-time in-situ measurements without requiring actual physical sensor deployments, thereby eliminating the time loss associated with ex-situ laboratory experiments while maintaining measurement precision through trained virtual sensor models.
Solution Approach 2:
The patent replaces the mechanical system of physical sensor deployment and laboratory analysis with a computational system using virtual sensors and machine learning models. This substitution enables real-time prediction of formation properties directly in the wellbore environment, eliminating the time-consuming physical experiment process while preserving measurement accuracy through trained computational models.
2Productivity
If dedicated physical sensors are deployed for real-time monitoring, then productivity is improved, but device complexity and cost increase
Solution Approach 1:
The patent creates virtual copies of physical sensors through machine learning models that can predict sensor responses and formation properties in real-time. These virtual sensors provide the same monitoring capabilities as physical sensors without requiring actual hardware deployment, thereby maintaining productivity while reducing device complexity and costs associated with physical sensor installations.
Solution Approach 2:
The patent develops a universal machine learning framework that can predict multiple formation properties (permeability, porosity, saturation) using a single virtual sensor system. This multi-functional approach replaces the need for multiple dedicated physical sensors, reducing device complexity while maintaining real-time monitoring productivity across various reservoir parameters.
3Measurement precision
If physical sensors are installed in wellbore for in-situ measurements, then measurement precision is improved, but ease of operation deteriorates due to maintenance scheduling limitations
Solution Approach 1:
The patent creates virtual sensor copies that can predict formation properties and detect sensor performance degradation without requiring physical sensor installation. This allows operators to monitor formation properties continuously and predict when physical sensors will fail, enabling proactive maintenance scheduling and improving ease of operation while maintaining measurement precision through the virtual sensor predictions.
Solution Approach 2:
The patent implements a feedback mechanism where virtual sensors continuously predict formation properties and compare them against actual measurements from physical sensors. This feedback loop enables real-time detection of sensor degradation or failure, providing early warnings that allow operators to schedule maintenance before complete sensor failure occurs, thereby improving ease of operation while maintaining measurement precision.
4Loss of time
If a priori scheduling is used for sensor maintenance, then loss of time is reduced, but reliability deteriorates due to inability to predict actual sensor performance
Solution Approach 1:
The patent implements a feedback mechanism where virtual sensors continuously predict formation properties and compare them against actual physical sensor measurements. This real-time comparison provides feedback on sensor performance degradation, allowing operators to predict actual sensor failure points and schedule maintenance based on actual sensor condition rather than fixed schedules. This improves reliability by predicting actual sensor performance while optimizing maintenance timing to reduce downtime.
Solution Approach 2:
The patent uses virtual sensors to perform preliminary predictions of sensor performance degradation and formation properties before actual sensor failure occurs. This preliminary action allows operators to proactively schedule maintenance at optimal times based on predicted sensor performance trends, rather than relying on fixed a priori schedules. This approach maintains reliability by predicting actual sensor conditions while reducing unnecessary maintenance interruptions.
Data Source
AI summary
The present disclosure is related to improvements in methods for evaluating and predicting responses of virtual sensors to determine formation and fluid properties as well as classifying the predicted as plausible or outlier responses that can indicate the need for maintenance of downhole physical sensors. In one aspect, a method includes detecting a change to a system of operating a wellbore to yield a determination, the system including a virtual sensor, the virtual sensor including a physical sensor placed in the wellbore for collecting one or more physical properties inside the wellbore; and based on the determination, performing one of retraining a machine learning model for predicting an output of the virtual sensor or predicting an output of the virtual sensor using the machine learning mode, the predicted output being indicative of at least one of sub-surface formation or fluid properties inside the wellbore.