Ski friction on snow and ice

Abstract

This thesis has tree main focus points; ski friction theory, materials and finish and the use of machine learning algorithms. The main results are in the machine learning segment. Olympiatoppen has recorded 16 000 ski pairs with up to 59 parameters which has been the basis for the machine learning algorithms. This thesis provides a working machine learning algorithm specialized for Olympiatoppens way of collecting data, with higher than 80\% accuracy. The algorithm takes in ten environmental parameters, and gives out the ten best products and their probability of success for the situation. The level of importance for the environmental parameters, according to the recorded data, is looked at and discussed. In the materials and finish section, zirconia and nonuniform grinding are found to be theoretically interesting results to apply to cross country skiing in the future. The thesis also explains the main influences to glide for sliders on snow and ice.