Co-simulation Extrapolation Using Mathematical Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing co-simulation methods face challenges in accurately extrapolating coupling variables over long time intervals, leading to coupling errors and disruptions in real-time simulations, especially due to dead times caused by communication and measurement processes.
Innovation Solution
A method and simulation device that uses a mathematical model to extrapolate coupling variables, allowing for reliable prediction and compensation of dead times, enabling real-time co-simulation by determining the time behavior of subsystems and handling errors through model-based extrapolation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If signal-based extrapolation methods are used for coupling variables, then the co-simulation can be performed with simple implementation, but the extrapolation accuracy deteriorates over long time intervals
Solution Approach 1:
The patent changes the fundamental parameter of extrapolation from signal-based to model-based approaches. By using mathematical models that incorporate system dynamics and physics principles, the extrapolation accuracy is significantly improved over long time intervals while maintaining computational efficiency through model reduction techniques.
Solution Approach 2:
The patent introduces mathematical models as intermediaries between the subsystems in co-simulation. These models serve as mediators that predict coupling variables more accurately by incorporating system knowledge, thereby improving extrapolation accuracy without requiring frequent data exchange between subsystems.
2Measurement precision
If the sampling steps or exchange intervals are kept small to reduce coupling error, then the extrapolation accuracy improves, but the computing time increases
Solution Approach 1:
The patent applies preliminary action by using mathematical models to predict future states of coupling variables before the actual exchange time. This allows the system to prepare accurate coupling variable values in advance, reducing the need for frequent exchanges and thereby decreasing computing time while maintaining accuracy.
Solution Approach 2:
The patent employs dynamic model-based extrapolation that adapts to changing system conditions. The mathematical models dynamically adjust predictions based on current system states, allowing for accurate coupling variable estimation over variable time intervals without requiring uniformly small sampling steps.
3Loss of time
If model-based extrapolation is used to extrapolate over several coupling time steps, then real-time co-simulation becomes possible, but the device complexity increases
Solution Approach 1:
The patent segments the overall system into independent subsystems, each with its own mathematical model. This segmentation allows each subsystem to be simulated independently with model-based extrapolation, enabling real-time co-simulation while managing complexity through modular architecture and localized modeling efforts.
Data Source
Figure 1~4
Figure 3
Figure 5
AI summary
In order to achieve real-time co-simulation of subsystems of a complete system (1), said subsystems being reciprocally coupled by coupling variables (y1, y2), a mathematical model (M) of the subsystems (TS1, TS2) which is valid at the actual point of operation of the complete system (1) is determined from input variables (x1, x2) and/or measurement variables (w1, w2) of said subsystems (TS1, TS2) using a data-based model identification method, and from this model (M), the coupling variables (y1, y2) are extrapolated for a subsequent coupling time step and supplied to the subsystems (TS1, TS2).