Ship Hull Resistance Prediction from Model Tests with a Resistance Similitude Simulator

Abstract

Depending on model sizes and test speeds, model Reynolds numbers are typically one to two orders of magnitude less than those occurring at full scale. The region of laminar flow is significantly larger on a model surface than at full scale. Turbulence stimulators (TSs) are extensively applied in model resistance tests to simulate ship boundary layer flow. Because of lack of theoretical guidance, the size, shape, and location of turbulent stimulators used in each test facility are typically selected based on experience gained in each laboratory, and parasitic drag induced by TSs is subtracted from measured model resistance in some test facilities and not subtracted in other facilities. If turbulence stimulation can be specified in such a way as to achieve momentum thickness similarity between the model and full scale, the form factor will be constant, so a simple scaling formula can be applied to predict ship resistance. Ship resistance can be predicted from measured model resistance with the application of this simple scaling formula. The article describes how for a given body geometry, ship Reynolds number and model Reynolds number, a unique TS (size, shape, and location) can be designed based on three equations derived in this article. This uniquely designed TS is termed the "resistance similitude simulator" in this article. The parasitic drag induced by the simulator is a part of the similarity solution and does not need to be subtracted from measured model resistance. 1. Introduction The ability to predict resistance plays an important role in the design of a ship or submarine. An accurate prediction of the drag of the vessel is required for the propulsor design and to determine the required power. The resistance of the vessel also strongly influences its lifetime operation costs. Although computational methods are now commonly used to estimate ship resistance, model tests are still heavily relied on, and scaling formulae are used to relate model data to full-scale performance.