The power grid infrastructure faces multiple challenges due to, not only the growing demands, but also the widespread deployment of renewable generation. The increasing level of renewable penetration in the energy mix requires to rethink the way the grid works, operates, and also how it is structured. This makes energy planning more critical as it will necessarily have to account for the effects of intermittence and variability of these sources, and the dynamic behaviour of the overall system. Power grid models can play an important role in performing that task. What is needed, is a new, faster computational model that can simulate large-scale grid operations, while capturing generating units' constraints, system flexibility and architecture. We present Spark!, a grid simulation model, for large scale future power grids over long term horizons. The model developed in Python, and built on a DEVS (Discrete Event System Specification) platform, captures the intermittent and stochastic nature of renewable energy resources and their associated forecast error, the thermal constraints of conventional generation resources, geographical and climate information, the transmission network, with a flexible time resolution.