Hybrid Modeling for Two-Phase Cooling System Simulation Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Two-phase liquid cooling systems for electronics are complex and computationally intensive to model, limiting their optimization and large-scale use due to the lack of accurate thermal modeling tools for predicting temperatures and two-phase flow in electronic device packages.
Innovation Solution
A hybrid modeling technique that performs detailed simulations based on three-dimensional models for a subset of simulation domains and reduced-order simulations for other domains, generating a hybrid model that improves accuracy and reduces computational steps, allowing for efficient optimization of two-phase cooling systems and design trade-off analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detailed simulations are performed for all simulation domains, then modeling accuracy is improved, but computational time and complexity increase significantly
Solution Approach 1:
The device under test is divided into multiple simulation domains, with different modeling approaches applied to each domain. Critical regions receive detailed simulations while non-critical regions use reduced-order models, enabling selective allocation of computational resources to maintain accuracy where needed while reducing overall computational burden.
Solution Approach 2:
Different levels of modeling detail are applied to different spatial regions based on their importance to the overall thermal performance. High-accuracy detailed models are concentrated in regions with significant thermal gradients or design criticality, while simpler models are used in regions with more uniform conditions, optimizing the balance between accuracy and computational efficiency.
2Measurement precision
If detailed simulations are performed for all simulation domains, then two-phase flow prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The computational domain is segmented into regions where two-phase flow phenomena are critical versus regions where single-phase approximation suffices. Detailed two-phase flow simulations are performed only in regions where phase change significantly impacts thermal performance, while reduced-order models handle regions with minimal phase change effects.
Solution Approach 2:
The modeling approach dynamically adjusts based on local thermal conditions, transitioning between detailed two-phase flow models and simplified single-phase models according to parameters such as vapor quality, heat flux magnitude, and temperature gradients. This parameter-driven selection reduces computational complexity while maintaining accuracy in critical regions.
3Productivity
If reduced-order simulations are used for all domains, then computational time is reduced, but modeling accuracy deteriorates
Solution Approach 1:
The device is partitioned into critical and non-critical simulation domains based on thermal importance. Reduced-order simulations are applied to non-critical domains to maximize computational efficiency, while detailed simulations are reserved for critical domains where accurate prediction of thermal properties is essential for design validation.
Solution Approach 2:
Different simulation fidelities are assigned to different spatial locations based on local thermal conditions and design requirements. Regions with high heat flux, steep temperature gradients, or design sensitivity receive detailed modeling treatment, while regions with more benign conditions use reduced-order models, ensuring adequate accuracy throughout the device without uniform over-modeling.
Data Source
AI summary
Techniques that facilitate hybrid modeling for a device under test associated with a cooling system (e.g., a two-phase cooling system) are provided. In one example, information indicative of a first model of a device under test associated with a cooling system is determined. Second information indicative of constraints that define values for an operational quantity related to the cooling system is also determined. Information indicative of a second model for the device under test is generated based on the information indicative of the first model and the second information indicative of the one or more constraints. In an aspect, a first simulation process is performed to determine first thermal properties for a first simulation domain associated with the device under test. In another aspect, a second simulation process is performed to determine second thermal properties for a second simulation domain associated with the device under test.