THE SOAR METHOD FOR AUTOMATICALLY OPTIMIZING SIMPLE RELAXATION FACTORS

Abstract

The SOAR method (SIMPLE Optimized and Automated Relaxation Factors) is proposed to automatically determine the optimum values of relaxation factors for the SIMPLE algorithm while improving SIMPLE's convergence characteristics. The factors are calculated from a term that represents the difference between the approximated equations for the iterative SIMPLE method and the exact equations for the Newton-Raphson direct method. Verification calculations for a cavity driven flow problem demonstrate that SOAR can dramatically reduce the number of iterations for convergence and the calculation time without requiring excessive calculation resources. Further, automatic determination of the relaxation factors greatly facilitates use of SIMPLE.