Friction on ice:a thermographic analysis of surface temperature

Project Details


This was a 6 month project to span the period from the 2006 Olympics in February to the end of the 2005-6 curling season in June. We combined infra red thermography and pin-on-disk tribology to characterise the thermal change in ice due to rubbing against nylon and determine its effect on the coefficient of friction for ice.

Layman's description

Ice is a very slippery material because unlike other solids it can produce its own lubricant. When two objects slide pass each other the friction between them produces heat - this is why we warm our hands by rubbing them together on a cold day. In the case of ice, rubbing produces enough heat to melt its surface and provide a thin layer of lubricating water. Just as rubbing your hands faster produces more heat, the faster something slides on ice the more heat and lubricant is produced and the more slippery ice becomes.
Another key factor when considering the slipperiness of ice is temperature. Captain Scott noted during his Antarctic travels that once the temperature fell below -35°C it become incredibly hard to pull sleds through the snow. This was because the heat produced by the friction between sled and snow was not enough to warm the ice to its melting point (0°C) so no lubricating melt water was produced. Scott and his companions may as well have been pulling their sleds through sand. The closer the temperature of the ice is to its melting point the more melt lubricant is produced by sliding and the more slippery ice becomes.
Curlers use the dependence of ice friction on temperature to change the direction and length that a curling stone slides. Their vigorous sweeping in front of curling stone raises the temperature of the ice and reduces it coefficient of friction so that the stones slide further and curve less. When the British Women's team won the gold medal at the 2002 Salt Lake Games the victory came with the very last stone of the Olympic final. The trajectory of the gold medal winning stone was subtly corrected by having its path swept as it slid down the ice. The sweeping technique and fitness of the British athletes that corrected the passage of that final stone provided the difference between winning and losing.
There is a complex feed-back relationship between friction on ice and frictional heating, which affects both laboratory-based ice friction experiments and sweeping in curling. We will study the surface temperature of ice in friction experiments with a thermally sensitive infrared (IR) camera. This will help us determine how the interplay between friction and heating affects ice friction measurements. We will also use the IR camera to study the heat generated by sweeping in the sport of curling, which will help British curlers modify their sweeping style and may provide a useful training tool for the Scottish Institute of Sport.

Key findings

Data from the laboratory based sweeping rig were used to calibrate the thermo-mechanical model. The velocity dependence of μ for a non-melting body sliding on ice has the form:
μ = a + bv–½
where a is a dimensionless parameter and b is a parameter with the dimensions m½s–½. The numerical model was calibrated by modifying these parameters and comparing the results to the change in ice temperature that was measured experimentally using the laboratory-based sweeping rig (Fig. 3). The model used a 10 s ramping period over which time the frequency of the brush head was increased from zero its eventual running frequency. The ramping period was incorporated so that the model most accurately replicated the experimental methods used to measure the change in temperature. It was found that the model reproduced the experimental results best when a friction law with a = 0.13 and b = 0.5 m½s–½ was used. The comparison of experimental and model changes in temperature 2 mm below the ice surface for different frequencies and applied pressures are shown in Fig. 3b.
The parameters used in equation 3 are consistent with tribological results found for a nylon pin on an ice disk. The model was also successfully tested by comparing its temperature change data obtained using IR thermography for nylon rubbing on ice. The model was then used to determine the effectiveness of two common sweeping styles in curling; a conventional style where players sweep across the path of the approaching curling stone with a low attack angle with respect to the direction of the stone’s motion and a high-attack angle style with respect to the stone’s motion. It was found that while the high-attack angle style produces greater temperature rises, the conventional style is more effective because it produces a thermal pattern where the temperature peaks are closer to the stone. The calibrated model was also used to determine the additional distance that stones travel when swept. It was found that sweeping in front of a stone using a high-attack angle for the last 6 m of its journey results in the stone sliding an addition 0.24 m compared an unswept stone, while using a conventional style resulted in an additional 0.56 m.
Brush head design improvements may be made based on our IR thermography and tribology data and results from the numerical model. Increased temperature rises in the ice can be generated by using materials with low thermal diffusivity. The geometry of the brush heads can optimised to produce thermal patterns on the ice that will lead to stones sliding further.
Effective start/end date1/04/0630/06/08


  • EPSRC: £55,602.00