The ongoing energy transition towards large shares of renewable generation poses challenges for hydro power producers. We revisit the problem of optimising the operation of hydro power plants using mathematical modelling, but utilising computer science concepts in the design of the models and configuration of these models. We use a modular design allowing us to activate features, such as what markets or which technical aspects to consider, by activating or deactivating a specific module. In this paper we give an example of how our method can be used to configure which markets a model should operate on. Furthermore, we use a configuration process based on the SPEA2 evolutionary algorithm to explore the relationship between the scale of the model and the time required to solve it. Such methods assist in identifying configurations that are the best fit in terms of runtime, realism and accuracy.