A data-driven time-step determination approach for dynamic simulation of heat-electric coupled system

Dynamic simulation of heat-electric coupled system (HECS) can accurately describe the operating condition of networks and equipment of the system, where the space-time step plays a crucial role in the simulation accuracy and speed. To optimally determine the simulation space-time step, this paper identifies the main influencing parameters for simulation elapsed time by distance correlation (DC) analysis and establishes a simulation time prediction model with stepwise regression (SR) method. Meanwhile, the quadratic curve fitting (QCF) method is adopted to obtain the simulation error prediction model. A simulation space-time step optimal determination model considering the specific simulation requirements is further introduced to improve the simulation efficiency based on the aforementioned models. To validate the proposed methodologies, comprehensive analysis of various study cases on a 20-node testbed is presented. The results demonstrate that the average error in elapsed time and error prediction by the proposed model is merely 0.2426 s and 4.84% compared with the real data.