Eindhoven University of Technology & KU Leuven

Eindhoven University of Technology

Flow field within a cycling peloton, including contours of the pressure coefficient on the cyclist’s bodies and bicycles, and drag percentages showing drag reductions at different positions in the peloton.

Problem:

A cycling peloton is the main group of cyclists riding closely together to reduce aerodynamic drag and energy expenditure. While it is well known that the aerodynamic drag inside the peloton is significantly less than that at the front, it is not known how much this drag inside the peloton actually decreases and which positions are most beneficial in terms of drag reduction.

By using CFD, the aerodynamic drag of each cyclist can be assessed. To fully resolve the thin viscous/laminar sublayer of the boundary layer and the buffer layer, which is important to correctly reproduce boundary layer separation and laminar-to-turbulent transition, a high-resolution grid at the surface of every cyclist and bicycle is required. Grid-sensitivity analysis indicated the need for 40 prism layers with a wall-adjacent cell height of 20 μm and a growth ratio below 1.1. These stringent grid requirements would result in grids containing nearly 3 billion cells.

Solution:

The simulations were made possible by supercomputing with ANSYS Fluent CFD software on Cray supercomputers. The total grid consisted of 15 parts of approximately 200 million cells each, created in ANSYS Fluent Meshing. During preprocessing, all parts were connected in ANSYS Fluent via interfaces. The required supercomputing cycles were performed on a Cray XC-40 supercomputer. A total memory of 49,152 GB per job was required, and each job was run with 13,824 MPI ranks mapped one rank to one core.

The results were validated by wind tunnel tests, including a peloton of 121 quarter-scale models. Analyzing the drag at all 121 positions showed that all cyclists in the peloton experience a drag reduction compared to an isolated cyclist riding at the same speed. The leading rider has the largest drag (94 percent), followed by the cyclists at the outer front edges of the peloton who have a drag reduction in the range of 59–67 percent. For riders sufficiently embedded inside the peloton, the aerodynamic drag reduces strongly. Overall, the cyclists at the mid rear of the peloton have the largest drag reductions. Forty-eight of these riders have drag reductions down to 5–10 percent that of the isolated rider. This means that almost 40 percent of this peloton travels at very low cost in terms of energy. Hence, the peloton is a very energy-efficient transport mechanism.

These results can be used to improve cycling strategies, including breakaways and energy saving for the final sprint. They can also be used to improve the reliability of mathematical models of cycling that are also sometimes employed to develop breakaway strategies.

Software used: