Delft simulation model for optimum performance during team time trials
During his Mechanical Engineering degree programme, TU Delft student Mats Overtoom came up with a mathematical model that could provide answers to such questions as ‘What is the optimum changeover schedule for a cyclist in a team time trial?’ and ‘How fast does he need to cycle up a mountain to reach optimum performance?’. The input for the simulation model was data from cyclists from Team Sunweb and specific track information. It gives the Dumoulin team strategic tips for optimum performance, the best order of the cyclists and the length of turns on the front during a team time trial.
TU Delft and Team Sunweb have been working together for years to use scientific research to improve sporting performances. The team time trial project is a recent example of this. Insights gained during the project will be evident in the 2018 Tour de France. Where in the past the Sunweb team made a predictive simulation for individual time trials as a handy tool for choosing between different racing strategies, now the team has been looking with TU Delft to predict finishing time for the team time trial.
Positioning and speed
‘Optimising the team time trial is far more complex than the individual time trial because the performance depends on several racers,’ explains Teun van Erp, scientific expert of Team Sunweb. ‘You need to use positioning and speed to manage the physiological condition of several racers. So there are more components in the strategy for a team time trial. If the team drops one or more racers during the race, for example, when would be the best time to do that? We are working together with TU Delft to gain more insight into that.’
‘The project began with reading theories on physiological models, resistance models and models describing the aerodynamic interaction between cyclists cycling in a group. After a month with no results, Team Sunweb became world champions in the team time trial, and the pressure to perform in the research increased immediately,’ says Overtoom. The student took the models found in the literature and combined them in a computer simulation in order to calculate the physiological states that corresponded to a specific race plan. Genuine race data were used to check whether the model was able to make realistic predictions. ‘Analysing various races and scenarios enabled us to ascertain which factors slowed the performance of the cyclists. This was followed by the optimisation phase in which we looked, for example, at the order of the racers and the ideal length of the turn on the front.’
Team time trial strategy
‘This gave us answers to various questions concerning the strategy for the team time trial. For example what is the right time to drop a cyclist? How can the load be shared across the cyclists? And what is the optimum starting order for a specific group of cyclists? What group composition gives the best finishing time if a racer can’t hold on? And is it faster to drop him or do we need to keep him to the end?,’ says Overtoom. ‘These are all concrete questions from our team coaches. We’ve used the answers to adjust the race protocols and we are expecting improved performances in future races,’ says Van Erp.
The above project is part of the Tour de France innovation video of Team Sunweb:
Team Sunweb's new Tour de France Innovations
About the TU Delft bicycle lab of Arend Schwab
About the collaboration between TU Delft and Team Sunweb
Press release A 3D printed mannequin of Tom Dumoulin in the TU Delft wind tunnel helps gain a competitive advantage
About the TU Delft Sports Engineering Institute