Spline Interpolations for Water Hammer Analysis

The method of characteristics (MOC) with spline polynomials for interpolations was investigated for numerical water hammer analysis for a frictionless horizontal pipe. A new analysis quantifies numerical errors in a systematic and dimensionless way and is used to evaluate the performance of three spline polynomials and to compare them with other numerical schemes for water hammer analysis, as well as to assess the effect of required additional interpolation boundary conditions on the overall accuracy of spline and two-point fourth-order Hermite schemes. For the sudden valve closure test case investigated, the overall accuracy with splines is significantly improved compared to the MOC with linear interpolations or to second-order explicit finite difference techniques. Compared to the Hermite method, the spline scheme has about the same overall accuracy and has the advantage of offering a choice between several spline polynomials with different interpolating characteristics and of being unconditionally stable.