Using equation-based languages for generating embedded code for smart building applications

While significant research has been done on applying equation-oriented object languages, such as Modelica, to the simulation of complex systems, much research remains to use such languages for generating application-specific embedded code. We describe a method for using a hybrid system language (as a reference model), from which we generate reduced-order models suitable for creating embedded code for tasks such as control and diagnostics. We apply our approach to the generation of embedded diagnostics code for the operation of heating, ventilation and air-conditioning (HVAC) for complex buildings.