Publication: Data-driven modeling of aerodynamics in cycling motions
Loading...
Date
2024-07-01
Authors
Lee Jia Min
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Aerodynamic cycling refers to the study of how air resistance affects cycling performance. Past studies have reported significant correlations between body posture and aerodynamic drag in cycling. Despite these successful studies, there is a lack of the integration of biomechanical and aerodynamic factors to predict cycling performance using data-driven approach. Therefore, this study aims to uncover patterns and relationships between biomechanics and aerodynamics in cycling, focusing on body posture and its effect on drag area for enhanced cycling performance. Two static cycling case studies were analysed: Case Study 1 using experimental data and Case Study 2 using public domain data. The methodology included data collection, data extraction, statistical analysis using SPSS, and data preprocessing and classification using WEKA tools. The relationship between different body postures with specific torso angles and drag area was analysed in Case Study 1. Meanwhile, Case Study 2 focused on the relationship between completion time for a 40 km cycle and different body postures with specific torso angles. Significant correlations were found between drag area and attributes such as weight, height, BMI, and wind speeds in Case Study 1. Regression analysis identified torso angle, BMI, and wind speeds as significant predictors, with the model explaining 99% of the variance in drag area (R² = 0.990). In Case Study 2, positive correlations indicated that higher power output was associated with shorter cycling completion times, while larger frontal areas and higher drag areas led to longer times. The regression model for this case also showed high accuracy (R² = 0.994). Classification algorithms, including J48, LMT, and Simple Logistic, achieved the highest accuracy among the eight algorithms tested to predict outcomes based on the data. Results from the J48 tree diagram and equations from LMT and Simple Logistic indicated that larger torso angles and greater weight led to larger frontal areas, resulting in slower wind speed flow over the body surface due to higher drag area. The findings highlight the crucial impact of body posture on cycling aerodynamics. This study significantly contributes to sports biomechanics and aerodynamics by providing empirical data and predictive models that optimize body posture for enhanced cycling performance.