Since more than 50 years, the critical power model provides a mathematical and physiological framework to study human exercise performance and fatigue development. However, only recently, the model has been applied to intermittent exercise, making it a hot topic again for sport and exercise scientists. This modern application bears the potential to predict changes in the energy balance of an individual at any time point during exercise. Insight into this balance is of paramount interest to athletes who want to pace their race in an optimal way, but just as much for physiotherapists who need to prescribe rehabilitation programs tailored to the patient. Nevertheless, in spite of the broad spectrum of applications, especially given the emergence of sports wearables which can provide real-time feedback, current modelling of the energy balance is insufficient and lacks fundamental knowledge of the underlying physiology. Although the consumption of energy can be easily determined, quantification of its recharge is much more complex. During the past years, we have thoroughly investigated this recovery and acquired important insights into its physiological determinants. However, we still miss essential information before a true predictive model with strong practical feasibility can be developed. The present project therefore builds on this research with the goal of developing and validating an innovative model that enables the real-time prediction of energy balance during exercise.