Sensor-Less Digital Twin Comparison for Maintenance Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital twin technologies rely heavily on sensors, cloud computing, and machine learning, which are costly, computationally intensive, and require extensive data collection, often leading to inefficiencies and inaccuracies, especially in environments with limited connectivity and diverse systems from multiple OEMs.
Innovation Solution
A sensor-less digital twin system that uses a database subsystem, virtual sensor subsystem, and digital twin comparison subsystem to emulate operational data without physical sensors, incorporating environmental and physical profiles to estimate degradation and predict maintenance needs, utilizing a weighted comparison of multiple digital twins for decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor networks and cloud computing are used for digital twins, then measurement precision and reliability are improved, but device complexity and cost increase
Solution Approach 1:
The patent creates virtual copies (digital twins) of physical assets that replicate their operational characteristics and degradation patterns without requiring physical sensors on each asset. The virtual sensor subsystem generates synthetic operational data that mirrors what physical sensors would measure, eliminating the need for extensive sensor networks while maintaining measurement precision through mathematical modeling and environmental profile correlations
Solution Approach 2:
The patent replaces physical sensor systems with computational models. Instead of using mechanical/electronic sensors to directly measure asset conditions, the system uses virtual sensor algorithms that compute operational data from environmental profiles, asset specifications, and degradation models, thereby substituting a complex physical measurement system with a computational approach
2Measurement precision
If machine learning and extensive data collection are used, then prediction accuracy is improved, but loss of time and computational resources increase
Solution Approach 1:
The patent performs preliminary actions by pre-establishing environmental profiles, asset specifications, and degradation models before operational data collection begins. These pre-configured virtual representations enable the system to generate predictive maintenance insights immediately upon deployment without requiring extensive historical data collection and training periods, thus reducing the loss of time while maintaining prediction accuracy
Solution Approach 2:
The patent extracts only the essential parameters needed for accurate degradation modeling from complex asset systems, rather than collecting extensive comprehensive data. By identifying and utilizing key environmental factors, operational conditions, and asset-specific parameters, the system achieves maintenance prediction accuracy with minimal data collection requirements, thereby reducing computational resources and time loss
3Adaptability or versatility
If multiple digital twins with different models are used, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent implements a universal digital twin framework where a single virtual sensor subsystem can handle multiple asset types and OEMs through standardized environmental profiles and degradation models. The system is designed to be multi-functional, accommodating diverse physical assets from different manufacturers by using configurable asset specifications and environmental factors that can be adapted to various contexts without requiring fundamentally different system architectures, thus improving adaptability while controlling complexity through standardization
Data Source
AI summary
An operations and maintenance system for unique objects that includes a database subsystem that stores first and second distinct digital twins for each of the unique objects; at least one of the digital twins has an identifier that uniquely associates it with one of the unique objects and defines a unique virtual representation thereof. A virtual sensor subsystem emulates operational data for ones of the objects; the virtual sensor subsystem is not dependent on sensors physically associated with ones of the unique objects. A digital twin comparison subsystem compares outputs of the first and second distinct digital twins for each of the unique objects; the output of at least one of the first and second digital twins is a function of the emulated operational data for its associated object, and the system makes an operational or maintenance decision with respect to an object as a function of the comparison.