Digital Twin Modeling for Multiphysical Systems With Real-Time ROM Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for constructing digital twins are limited in their ability to accurately and efficiently monitor and predict the operating states of multiphysical engineering systems, particularly in networks of multiple element facilities, due to difficulties in expanding 3D distribution combinations and real-time response requirements.
Innovation Solution
A method that defines a multiphysical engineering system as a network, establishes 0-D models for element facilities, constructs a system ROM by combining 3-D CAE data and machine learning, and uses gappy-POD and artificial intelligence to correct errors and optimize operating conditions in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If 3-D CAE ROM and measurement data are combined using coupled-POD, then the digital twin can accurately represent 3-D distribution, but it is difficult to expand the application to multiphysical systems configured as networks of multiple element facilities
Solution Approach 1:
The multiphysical system is segmented into multiple element facilities, each represented by a 0-D model. This segmentation allows the complex multiphysical system to be broken down into manageable components that can be individually modeled and then integrated through network topology, enabling scalability to large-scale systems while maintaining accuracy.
Solution Approach 2:
The patent transitions from 3-D spatial distribution models to 0-D point-based models for element facilities. This dimensional reduction simplifies the representation while maintaining essential system behavior, and the network topology provides the spatial relationships, making the system scalable and adaptable to multiphysical configurations.
2Measurement precision
If detailed 3-D CAE analysis is used for each element facility, then the model accuracy is high, but the calculation time increases and real-time response becomes difficult
Solution Approach 1:
The system separates the modeling approach by facility type: 0-D models for routine element facilities enable fast calculation, while 3-D CAE models are reserved for key element facilities requiring high accuracy. This segmentation allows real-time response for most operations while maintaining accuracy where needed.
Solution Approach 2:
Different levels of modeling detail are applied locally based on the importance and complexity of each element facility. Key facilities with significant impact on system performance use detailed 3-D models, while less critical facilities use simplified 0-D models, optimizing the balance between accuracy and computational efficiency.
3Productivity
If 0-D models are used for all element facilities, then the calculation speed is fast, but the ability to capture complex 3-D physical distributions is lost
Solution Approach 1:
The patent segments the system into 0-D element facility models connected through network topology. This segmentation enables fast calculation for system-level analysis while preserving the ability to incorporate 3-D physical distributions within key facilities when needed, achieving both speed and reliability.
Solution Approach 2:
The network topology acts as an intermediary that connects 0-D element models and can incorporate 3-D CAE results for key facilities. This intermediary structure allows the system to operate efficiently with 0-D models while selectively integrating detailed 3-D physical distributions where accuracy is critical.
4Measurement precision
If machine learning techniques are applied to correct ROM errors, then the prediction accuracy improves, but the complexity of the construction process increases
Solution Approach 1:
Machine learning techniques are applied to correct errors between ROM predictions and actual measurements by learning from historical data. This feedback mechanism continuously improves prediction accuracy by adjusting model parameters based on observed discrepancies, maintaining high accuracy while automating the correction process.
Solution Approach 2:
The patent modifies ROM parameters using machine learning corrections based on historical measurement data. By changing and optimizing model parameters through learned patterns rather than restructuring the entire model, the system improves accuracy while keeping the construction process manageable.
Data Source
AI summary
The purpose of the present invention is to provide a method for constructing a digital twin, enabling real-time monitoring, operation improvement, and coping with the occurrence of an accident in an industrial site by combining reduced order models of a multiphysical system, measurement data and artificial intelligence techniques, and the method for constructing a digital twin, according to the present invention, comprises: a network-defining step of defining a multiphysical engineering system as a network constituted by a combination of element facilities; an element model establishing step of establishing a relation-based 0-dimensional (0-D) model for each of the element facilities; a system model establishing step of closing all relations for a system by reflecting an additional relation by machine learning from a 3-dimensional computer aided engineering reduced order model (3-D CAE ROM) or data for key element facilities in the 0-D models established in the element model establishing step; a system ROM constructing step of constructing a system ROM for the system model established in the system model establishing step from calculation results for conditions sampled in an operating variable parameter space; a system ROM correcting step of minimizing an error between a model predicted value and measured data for the element facility and the system; and a real-time algorithm constructing step of constructing an algorithm for identifying an expected system state or an optimal operating condition in a virtual operating condition based on the real-time monitoring result.