Share Thermal Insight, Not IP : Surrogate Thermal Models using Ansys Icepak

Overview

As semiconductor products transition from legacy chips to advanced 3D‑IC assemblies, thermal models must be accurate, scalable, and encrypted to share externally. Ansys Icepak’s Surrogate Thermal Models (STMs) enable this by converting a high‐fidelity 3D CFD simulation model into an embeddable, BC independent network model for system‑level use.

STMs capture detailed heat‑flow behavior from full 3D chip and package geometries while remaining sensitive to a wide range of HTC boundary conditions, ensuring reliable predictions across diverse cooling environments. Applicable to both traditional semiconductor designs and modern heterogeneous integrations, these models deliver the speed and accuracy required for integration into system level CFD models and encrypted distribution to the customers.

Icepak introduces an automated, easy to use toolkit to streamline training data generation and an advanced optimizer for accurate thermal network creation, with minimal user effort. In summary, STMs toolkit gives an efficient, encrypted, BC independent package thermal model with minimal effort that can be shared with internal/external customers without exposing IP. Attend this webinar to learn how STMs help semiconductor companies scale thermal insight securely from silicon to system.

What you will learn

• How to create BC independent, embeddable network models from 3D CFD model
• Where STMs fit for legacy chips and 3D‑IC designs
• How Boundary Condition Independent STMs predict performance across cooling conditions
• How Icepak automates STM training and generation

Who should attend

• System Integrators
• Semiconductor Thermal Experts
• Physical Design Teams & Managers

Speaker

• Tejas Jeurkar