Models For The Real-time Control Of Subway Stations
Roberta Ansuini, Massimo Vaccarini, Alberto Giretti, Sara Ruffini · Building Simulation Conference proceedings · 2013
Th particular application domain raises a number of This paper discusses the role of models and modelling issues that make the development of the simulations in the perspective of the real-time control integrated station model a challenging engineering of complex buildings. It describes the approach used task (Ansuini et al., 2012). In fact, the modelling in the model engineering of the SEAM4US EU process involves multi-physics models with multiple research project, concerned with the optimal control of the Passeig De Graçia metro station systems in Barcelona. The paper introduces the model predictive control (MPC) architecture and the main features of time scales, different levels of detail and large spatial dimensions. The real-time management of the metro station underground space requires that models are able to support the human controller with scenario the large Modelica station model, made of over 80000 variables. It then details the statistical model reduction methodology adopted either for mapping the large station model into a deployable sensor network, and for embedding the model into the main analysis, which means they must be robust, invertible and able to propagate uncertainty throughout the co putational chain (Giretti et al., 2012). Finally, the run-time deployment of the models into the control loop imposes that models are computationally control loop. efficient, that they have a manageable size, and, most important, that they could be mapped into a