COMPARISON BETWEEN TWO CO-SIMULATION FRAMEWORKS, MOSAIK AND HELICS

The Energy systems combine physical domains related directly to the processes of generation, storage, distribution, and consumption, in addition to communication technologies and software infrastructure for data and control purposes. The complexity of these heterogeneous systems makes it difficult to study them since tools of different domains must interact. In recent years, emerge a new enabling technique where global simulation of a complex system can be achieved by composing the simulations of its parts called Co-simulation. This research has the purpose of making a comparison between two of these co-simulation frameworks, MOSAIK and HELICS. To this aim, the comparison is divided into two components theoretical and performance. The first component analyze the tools considering their conceptual architectures, giving particular importance to how each framework handles the time synchronization and the data exchange between all the co-simulation simulators. The second one center the study on the performance presented by the platforms HELICS and MOSAIK with two chosen case studies. To allow the configuration and set-up of each case study can be done in the same way, it uses a flexible platform that brings a plug-and-play integration of models, simulators, and scenarios, independently of the framework. In this platform, one or more models can be easily replaced without affecting the whole simulation engine, and it is possible to choose the framework you want to execute. Each study case is composed of different simulators that are combined in a shared simulation environment. Case study one represents a simple electrical network composed of four Python models used to simulate the grid, some photovoltaic panels, and buildings. Case study two models a greater electrical system where the performance of the building is simulated with Energyplus, the heat pump with its control strategy is modeled in Modelica, household occupancy, electrical loads, photovoltaic production, smart meters, weather, and grid employ Python simulators. Both frameworks simulate the case studies with a set of predefined scalability scenarios, i.e., each scenario run has more than one replica for one of its simulators. During these tests, data on time spent in the simulation and computational resources required for each case, each scenario, and each framework were obtained and stored. Finally, these results are presented as well as the analysis of the similarities and differences between MOSAIK and HELICS by performing the co-simulations.