Auto-Tuning Under-Relaxation Factor for Diverged Numerical Simulations

Solution Overview

Problem

Numerical simulations, particularly in Computational Fluid Dynamics (CFD), often face divergence issues and require extensive computational resources, with manual intervention needed to select suitable control parameters for convergence, leading to inefficiencies and resource wastage.

Innovation Solution

A processor-implemented method and system that stabilize diverged simulations and accelerate converged ones by receiving past residues, using a classifier to determine simulation status and predicting an under-relaxation factor through control logic, integrating the predicted output to stabilize and accelerate the simulation process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual intervention is used to select control parameters for convergence, then convergence can be achieved, but time consumption and computational resource waste increase significantly

Engineering Contradiction:

ImproveconvergenceVSAvoidtime consumption

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system enables self-service by implementing an automatic control mechanism where the simulation system monitors its own convergence status and autonomously adjusts control parameters without external manual intervention. The controller continuously receives residue information, determines convergence status, and modifies parameters to maintain convergence, making the system self-regulating and eliminating the need for human operators to manually tune parameters throughout the simulation process.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system implements feedback by establishing a closed-loop control mechanism where the controller continuously monitors simulation residue and convergence status, then uses this feedback information to dynamically adjust control parameters. The feedback loop ensures that parameter adjustments are based on real-time simulation performance, allowing the system to respond to convergence changes and maintain optimal operation throughout the simulation process.

Inventive Principle:

Principle #23Feedback

2Reliability

If manual intervention is used to select control parameters for convergence, then convergence can be achieved, but computational resource waste increases

Engineering Contradiction:

ImproveconvergenceVSAvoidcomputational resource waste

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The system enables self-service by implementing an automatic control mechanism where the simulation system monitors its own convergence status and autonomously adjusts control parameters without external manual intervention. The controller continuously receives residue information, determines convergence status, and modifies parameters to maintain convergence, making the system self-regulating and eliminating the need for human operators to manually tune parameters throughout the simulation process.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system implements feedback by establishing a closed-loop control mechanism where the controller continuously monitors simulation residue and convergence status, then uses this feedback information to dynamically adjust control parameters. The feedback loop ensures that parameter adjustments are based on real-time simulation performance, allowing the system to respond to convergence changes and maintain optimal operation throughout the simulation process.

Inventive Principle:

Principle #23Feedback

3Device complexity

If iterative numerical simulation is performed without automatic control, then implementation is simple, but divergence occurs and simulation stability is poor

Engineering Contradiction:

Improveimplementation simplicityVSAvoidsimulation stability

Core Design Contradiction:

Device complexityVSStability of the object's composition

Solution Approach 1:

The system implements feedback by establishing a closed-loop control mechanism where the controller continuously monitors simulation residue and convergence status, then uses this feedback information to dynamically adjust control parameters. The feedback loop ensures that parameter adjustments are based on real-time simulation performance, allowing the system to respond to convergence changes and maintain optimal operation throughout the simulation process.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system applies parameter changes by dynamically modifying control parameters based on real-time simulation status. The controller adjusts parameters such as under-relaxation factors according to the determined convergence status, enabling the simulation to adapt to changing conditions and maintain stability without requiring complex manual intervention or pre-configured parameter sets.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20240126943A1System and method for stabilizing and accelerating iterative numerical simulation

Publication Date: 2024.04.18 TATA CONSULTANCY SERVICES LTD

AI summary

Simulation of dynamic physical systems is done using iterative solvers. However, this iterative process is a time consuming and compute intensive process and, for a given set of simulation parameters, the solution does not always converge to a physically meaningful solution, resulting in huge waste of man hours and computation resource. Embodiments herein provide a method and system for stabilizing a diverged numerical simulation and accelerating a converged numerical simulation by changing one or more control parameters. An automatic monitoring mechanism of residue history (to interpret convergence or divergence) and a subsequent control logic to auto-tune the under-relaxation factor would help in stabilizing a diverging simulation and reaching faster convergence by accelerating converging simulation.