Objectives  This study focuses on the feasibility of a ship resistance model test in an ice field of small ice floes made of substitute material in order to reveal the resistance components and thereby provide technical support for the design of ice-going ships.

  Methods  Ship resistance test in ice floes made of polypropylene (PP) instead of natural refrigerated ice is conducted. By adjusting the sizes, shapes, numbers of ice floes, the random ice field with a given concentration is generated. The geometric phase transition theory predicts that there exists a critical concentration which divides the random ice field into discrete phase (concentration is less than critical value) and connected phase (concentration is greater than critical value).

  Results  The main components of ice resistance in the discrete phase are open water resistance and ship-ice collision resistance, while ice resistance in the connected phase includes ice friction resistance, open water friction resistance and collision resistance. If the fractal dimension of the random ice field is used to redefine the ice resistance coefficient, it is nearly constant in the trial range (speed 0.3–0.9 m/s) when the concentration is smaller than the critical value. When the concentration is greater than the critical value, the ice friction resistance is inversely proportional to speed.

  Conclusions  Polypropylene can replace frozen ice in the prediction of ice resistance. The pure ice resistance of an ice field is divided into two components: ice resistance arising from collision and ice friction resistance arising from accumulation.