This paper describes a new industrial case on automation, for large scale systems with high environmental impact: the \emph{mining ventilation} control systems. Ventilation control is essential for the operation of a mine in terms of safety ($CO$ and $NO_x$ regulation) and energy optimization. We first discuss a novel regulation architecture, highlighting the interest for a model-based control approach and the use of distributed sensing capabilities thanks to a wireless sensor network (WSN). We propose a new model for underground ventilation. The main components of the system dynamics are described with time-delays, transmission errors, energy losses and concentration profiles. Two different model-based control approaches, which can embody the complex dynamics of the system, are proposed. The first one resorts to a nonlinear model predictive control strategy (receding horizon) and aims to energy minimization thanks to a continuous operation of the fans. The second one, based on a hybrid description of the model and fans operation, provides automatic verification of the wireless control thanks to abstraction techniques. These control strategies are compared with simulations, in terms of regulation efficiency, energy consumption and need for computational capabilities. The industrial case description and control strategies open new vistas for the development of global system approaches that allow for the optimization of energy consumption of complex large-scale systems.