Digital Twin Virtual Sensors for Unmeasurable Real-Time Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital twin technologies require data from real sensors, limiting their application to locations where sensors cannot be installed, and they cannot provide virtual sensor data when real sensors are not operating or are cyber-attacked.
Innovation Solution
A computing system using a physics-based model and machine learning to implement virtual sensors by reproducing state information and generating virtual sensor data based on real sensor data, even in locations where real sensors are absent, and inferring data for unmeasurable ranges or dimensions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If virtual sensor data is generated using digital twin technology, then the data range of digital twin is expanded, but virtual sensor data cannot be obtained at locations where real sensors are not installed
Solution Approach 1:
The patent creates virtual sensors as digital copies of real sensors within the digital twin model. These virtual sensors replicate the measurement capabilities and characteristics of physical sensors, allowing the system to generate sensor data at locations where physical sensors cannot be installed. The virtual sensor model is trained using data from corresponding real sensors to ensure accurate replication of measurement behavior.
Solution Approach 2:
The patent introduces virtual sensors as intermediary elements between the physical system and the digital twin model. When real sensors are unavailable, damaged, or cannot be installed at certain locations, the virtual sensors serve as mediators to provide the necessary measurement data. These virtual sensors are positioned at strategic locations within the digital twin model to capture system states that would otherwise be unobservable.
2Measurement precision
If real sensors are used for data collection, then accurate sensor data can be obtained, but the system cannot provide data when real sensors stop working or are cyber-attacked
Solution Approach 1:
The patent prepares virtual sensor models in advance that can serve as backup data sources. The virtual sensor models are trained and validated using historical data from real sensors, creating a pre-established alternative data generation capability. When real sensors fail or are compromised, the system can immediately switch to using virtual sensor data, preventing data gaps without requiring real-time sensor functionality.
Solution Approach 2:
The patent changes the source and nature of sensor data parameters by transitioning from direct physical measurements to model-generated data. The virtual sensor model learns the relationship between system states and sensor readings, enabling it to generate data that mimics real sensor behavior. This parameter transformation allows the system to maintain data availability even when physical sensing capabilities are compromised.
3Loss of information
If physical sensors are installed at all required locations, then complete data coverage is achieved, but physical and economic constraints prevent mounting all necessary sensors
Solution Approach 1:
The patent makes the digital twin model multi-functional by enabling it to both simulate system behavior and generate sensor data. Instead of requiring separate physical sensors for each measurement point, the digital twin model serves multiple functions: it models the entire system dynamics and simultaneously acts as a virtual sensor array. This universal approach eliminates the need for extensive physical sensor deployment while maintaining complete data coverage.
Solution Approach 2:
The patent creates virtual replicas of sensor measurement capabilities within the digital twin model. Rather than installing physical sensors at every possible location, the system creates virtual sensor copies that can be positioned anywhere in the digital model. These virtual sensors replicate the measurement functions of physical sensors, providing complete spatial coverage without the physical constraints of sensor installation.
Data Source
AI summary
Disclosed herein is a computing system for implementing a physics-based model, including a physics-based model configured to describe a physical complex system and perform deductive inference, and at least one processor. The at least one processor is configured to receive first data obtained such that the state information of first nodes is included therein in association with at least one of the passage of time and spatial distribution, to identify the state information for each of the first nodes in the physics-based model based on the first data in association with at least one of the passage of time and spatial distribution, and to determine parameters between variables in the physics-based model so that the state information can be reproduced for each of the first nodes in association with at least one of the passage of time and spatial distribution.