No abstract

Smart grids are complex electrical systems made up of power devices and communication systems that interact with each other, both locally through measures and generally through the exchange of information on communication networks. In this article, the simulation of a smart grid including the emulation of the communication network behavior together with the physical hardware devices is proposed using hardware-in-the-loop methodology. The original aspect of the technique is related to the fact that the models are generated in a tool that is independent of the real-time simulator using a model-based design approach, and then the code is generated and deployed in a real-time simulator coupled with hardware devices such as low voltage protection units. The proposed framework is used to test the protection coordination that employs the IEC-61850 communication protocol. A test case is shown demonstrating the coordination of switches and the emulation of the PV system in hardware-in-the-loop, focusing on the implementation of the interface algorithm over Ethernet using the UDP communication protocol. INDEX TERMS Co-simulation, hardware-in-the-loop, MATLAB-Simulink, model based design, PV system, protection devices, real-time simulation, smart grid.

Deterministic load flow analyses of power grids do not include the uncertain factors that affect the network elements; hence, their predictions can be very unreliable for distribution system operators and for the decision makers who deal with the expansion planning of the power network. Adding uncertain probability parameters in the deterministic load flow is vital to capture the wide variability of the currents and voltages. This is achieved by probabilistic load flow studies. Photovoltaic systems represent a remarkable source of uncertainty in the distribution network. In this study, we used a Gaussian copula to model the uncertainty in correlated photovoltaic generators. Correlations among photovoltaic generators were also included by exploiting the Gaussian copula technique. The large sets of samples generated with a statistical method (Gaussian copula) were used as the inputs for Monte Carlo simulations. The proposed methodologies were tested on two different networks, i.e., the 13 node IEEE test feeder and the non-synthetic European low voltage test network. Node voltage uncertainty and network health, measured by the percentage voltage unbalance factor, were investigated. The importance of including correlations among photovoltaic generators is discussed.

No abstract

No abstract