A computational strategy for determining the optimal scaled wind speed in icing wind tunnel experiments

Abstract

The icing wind tunnel experiment is one of the most important methods to investigate icing problems. Due to the blockage or capability limitations of icing wind tunnels, the geometric size of the testing model is usually needed to be sub-scaled and the corresponding icing conditions are required to be converted. However, the icing phenomenon is affected by various parameters, and how to determine the optimal subscale wind speed remains inconclusive. To solve this problem, a new computational strategy, named Improved Ruff Icing Scaling Method (IRISM), is proposed. In IRISM, the Ruff icing scaling theory is firstly applied to calculate the basic sub-scale temperature and the related cloudy parameters, etc., and then the icing numerical computation is utilized to evaluate the influence of sub-scale icing conditions on the similarity of the reference and subscale ice shapes. Finally, the optimal scaled wind speed is calculated by Differential Evolution Genetic Algorithm based on a high-efficient surrogate model, in which the ice shape consistency especially the ice horn structure is set as the optimization objective. The IRISM is comprehensively analyzed by icing numerical computations as well as icing wind tunnel experiments. The results show that the IRISM is capable of automatically providing the optimal scaled wind speed with the acceptable icing scaling error. As an important supplement to the current icing scaling methods, the IRISM can find the reasonable experimental parameters of scaled icing wind tunnel tests, which provides a new way for icing scaling analysis.