Optimum Approximate-Factorization Schemes for Two-Dimensional Steady Potential Flows

A procedure is described for constructing optimum approximate-factorization algorithms for solution of the two-dimensional equations for steady transonic potential flows. The factorizations are optimum in the sense that they should lead to the fastest convergence rates. It is shown that it is important to split the effect of transformation derivatives, normally appearing in the flow equations, correctly between the two factors in the basic factorization. An optimum value for the relaxation parameter is determined, and a procedure is described for constructing an optimum sequence of values for the acceleration parameter.