We discuss the application of the FMI Co-Simulation technology to building energy performance simulation, where detailed physical building models are coupled to Modelica-based HVAC component and plant models. First, we describe the generation process of the building FMU from our stand-alone building simulation program NANDRAD and sketch out internal algorithms for FMI version 2 capabilities. Then, coupling scenarios are described and physical interface conventions are presented. Usability is addressed by automatic generation of building-model specific adapters and wrappers. The building FMU and plant FMUs are then simulated together using different Co-Simulation master algorithms. Finally, based on simulation results and performance analysis we conclude with recommendations on suitable master algorithm options and specific features of suitable building FMUs.
Error control in system simulation using co-simulation techniques is a task for the employed simulation master. With the availability of the FMI standard version 2.0 and rollback capabilities of simulation slaves, master algorithms can be implemented with support of error controlled integration. Particularly, for automated integration tools, the problem-specific dynamic adjustment of communication interval lengths becomes a necessity to obtain reliable co-simulation results while maintaining calculation efficiency. The article discusses various master algorithms and time step adjustment strategies using a test case with discontinuous input/output signals. As expected, fixed-step Gauss-Jacobi and Gauss-Seidel algorithms are found to be generally unsuited for the task. Iterationbased time step adjustment rules are an improvement, yet cannot recognize discontinuities resulting from timeevent. Since the traditional Richardson/step-doubling error estimate also fails to recognize discontinuous signal changes, a slope-based modified Richardson-test is introduced and successfully applied. Finally, it is concluded that a suitable master algorithm for such problems is the non-iterating Gauss-Seidel with modified Richardson communication interval adjustment.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.