Virtual Sensor Readings for Failed Sensor Continuity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Sensors in inaccessible or hard-to-reach locations, such as subsea wells or unmanned spacecraft, fail frequently and are difficult or costly to replace, necessitating system shutdowns or conservative operation to ensure safety.
Innovation Solution
Generate simulated sensor readings using virtual sensors trained with historical data from operational sensors, allowing continued operation and extending the life of existing sensors by providing accurate readings within 10% of actual values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If physical sensors are replaced in inaccessible locations, then measurement reliability is improved, but system downtime and replacement costs increase
Solution Approach 1:
The patent creates virtual sensors that are digital copies of physical sensors, using machine learning models to replicate sensor behavior and generate simulated readings. This copying approach allows the system to maintain measurement capabilities even when physical sensors fail or need replacement, eliminating the need for system shutdowns during sensor maintenance.
Solution Approach 2:
The virtual sensor acts as an intermediary between the physical sensor and the control system. When physical sensors are inaccessible or failed, the virtual sensor provides intermediate simulated readings that allow the monitored system to continue operating without interruption, bridging the gap until physical sensor replacement can occur.
2Measurement precision
If physical sensors are replaced in hazardous environments, then measurement accuracy is improved, but safety risks and operational costs increase
Solution Approach 1:
Virtual sensors create digital replicas of physical sensors operating in hazardous environments, eliminating the need for personnel to physically access dangerous locations for sensor replacement or calibration. The virtual sensors maintain measurement precision through machine learning models trained on historical sensor data while keeping operators safe from environmental hazards.
Solution Approach 2:
The virtual sensor system provides self-service by automatically generating simulated readings based on data from other operational sensors, eliminating the need for manual intervention in hazardous environments. The system continuously updates and maintains sensor accuracy through automated machine learning processes without requiring physical access to dangerous locations.
3Productivity
If multiple virtual sensors are deployed, then system continuity is improved, but computational complexity and data processing requirements increase
Solution Approach 1:
The patent segments the virtual sensor system into multiple independent regression models, each handling specific sensor types or functions. This segmentation allows the system to maintain continuity by distributing computational load across multiple specialized models rather than using a single complex system, making the overall system more manageable and efficient.
Solution Approach 2:
The system dynamically adjusts parameters such as the number of virtual sensors deployed, the complexity of regression models, and the frequency of simulations based on system needs and available computational resources. This allows optimization of the balance between system continuity and computational complexity, deploying only the necessary number of virtual sensors to maintain productivity without excessive resource consumption.
Data Source
AI summary
A system includes an interface configured to receive first sensor data from multiple first sensors associated with a monitored system and also includes one or more processors. The one or more processors are configured to provide first input data to a plurality of regression models of a first virtual sensor corresponding to a particular sensor of the monitored system. The first input data is based on the first sensor data. The one or more processors are configured to obtain, from the plurality of regression models, output data based on the input data. The output data from each regression model of the plurality of regression models represents an estimated sensor reading of the particular sensor. The one or more processors are also configured to determine a simulated sensor reading for the particular sensor based on the output data, determine a confidence estimate associated with the simulated sensor reading, or both.