Optimal Control is a big opportunity for energy efficiency since it involves much smaller investments than those usually applied to technological building elements integration, by providing new ways for sustainable energy saving solutions. Nevertheless optimal management requires the development of a new class of predictive control logics, behaving consistently in changing environments. This class of control systems embed advanced predictive models, directly coupled with an environment monitoring sensor network, that are capable of interpreting sensed data (both indoor and environmental) and of forecasting future states. The design of the monitoring system is a critical passage as it is subjected to a number of operational and cost constraints, but also has to guarantee a reliable interpretation of the environmental behaviour of the building. This paper discusses this issue and presents the engineering framework and a methodology developed for supporting the design of monitoring networks integrated in optimal control system of large and complex environments. The application of the methodology to the case of a subway station is briefly presented too. This research is part of the European Research Project SEAM4US (Sustainable Energy Management for Underground Stations). The proposed solution is based on the definition and development of the set of models needed for both supporting the definition of the monitoring sensor network and being embedded in the final control system implementation. The development of this class of environmental models for large underground environments such as subway stations involves the elaboration and the integration of different simulation models. The simulation results constitute the knowledge that can support an efficient design of sensor networks even in large buildings.
