PHIL Simulation of an Island Grid: A Comparative Study with Classical Grid Simulation Software

Real-time simulator performance follows the evolution in computer technologies and can provide designers with an efficient approach to system prototyping and testing using “Power-Hardware-In-the-Loop” (PHIL) digital simulation. PHIL enables the test of advanced, real devices, such as Distributed Energy Storage Systems (DESS) or Distributed Generation (DG), by connecting them to a grid model that is simulated in real time. During simulation, the studied equipment behaves as if it were connected to the real network. As a result, it can be tested with a wide range of grid configurations, including fault conditions, without any risk.

In recent years, PHIL has delivered promising results within the framework of research projects dealing, for example with wind energy conversion systems, gas micro-turbine integration with distribution networks and small-scale distributed generation (DG) systems. While existing work typically deals with micro-grids, the potential of hardware-in-the-loop (HIL) simulations for large scale isolated grids should now be studied, since such grids are expected to accommodate more DG and other advanced grid equipment in the years to come.
 
Isolated power systems have specific features compared with large interconnected grids, such as significant dependence on oil generation plants, leading to higher production costs and emissions, and reduced sturdiness against electrical disturbances, resulting in frequent voltage/frequency fluctuations and power quality concerns. In addition, the variability brought by renewable-energy based DG and its current lack of contribution to ancillary services can impact the operation and stability of weak grids. This is the reason why island grids are said to have strong potential for the development of DESS for various applications. Non real-time modeling and simulation have traditionally been used for carrying out studies on this topic.
The purpose of the presentation is to illustrate, through the example of an innovative application of DESS on an island grid, how it is possible to move from a classical design approach to a “real-time” process, and to discuss the benefits gained from this switch. The main difficulty of the real-time simulation of a large scale isolated grid is related to the complexity of the grid’s structure and of the synchronous machine model. Hence, by simulating the complete network on a single processor of the real-time simulator, it was found that the required computation time was far beyond the maximum time-step to achieve real-time simulation. Work was therefore conducted to propose and validate solutions permitting either the distribution of calculations across several processor cores, or to simplify the network by reducing the number of simulated generators.
 
Brief Description of the Case Study
In cooperation with the integrated system operator of French isolated energy systems (EDF SEI), the use of distributed storage for frequency control was studied in the case of the electrically isolated archipelago of Guadeloupe. Located in the eastern Caribbean Sea, this overseas department of France comprises five islands: Basse-Terre (848 km2), Grande-Terre (588 km2, separated from Basse-Terre by a narrow sea channel) and the adjacent islands of La Désirade, Les Saintes and Marie-Galante. Guadeloupe covers about 1,600 square kilometers with a population of 407,000, as of 2008.
The peak power consumption of the network reached 242 MW in 2007. Fig. 1 shows the major production centers of the island and the structure of its HV transmission grid.
 

Fig. 1.  HV grid, major generation plants and DG location in Guadeloupe
 
The transmission system is operated at 63 kV and includes 13 substations. Each substation consists of 2 step-down 63/21kV transformers with on-load tap changers to adjust the voltage on downstream distribution grids.
The installed generation at the end of 2008 consisted of eight 21 MW diesel power plants at “Jarry-Nord” substation and two 32 MW coal-fired power stations at “Gardel” substation participating in primary frequency control. Three smaller diesel groups (5 MW) at “Jarry-Sud” substation, distributed hydro-electric plants (9,6MW) as well as geothermal power plants (5+10 MW) at “Bouillante” substation are often operated at their rated power. Four combustion turbines at “Jarry-Sud” substation are solicited only during peak demand (3x20+40 MW). Wind- (26.4 MW) and solar- (5 MWp) based power sources have been connected over the past few years but their shares in the annual energy mix remain rather marginal, 2.59 % and 0.16 % respectively, in 2006.
Several configurations of the network were taken into account to assess the potential of ultracapacitor-based DESS for frequency support. The various simulated incidents included the tripping of a coal-fired generator producing around 25 MW during off-peak situation (140 MW), with minimal power system inertia. Due to the relatively slow response of diesel plants while delivering their primary reserve, the grid frequency drops quickly below the load-shedding threshold in such a situation. An outage is therefore experienced by customers even if the primary reserve on remaining power plants is sufficient from a static point of view. DESS modeled as power or current injectors were connected at distribution levels in the simulated Guadeloupe power system. A specific control algorithm was developed so that DESS may dynamically assist the real power injection following generator tripping.
 
Objectives
This presentation focuses on the methodology to implement real-time simulation of the power grid and on the contribution of PHIL simulation in the design process of innovative grid equipment. The case study presented in Figure 1 will be used as an illustrative example.

See more about Electrical and Power Systems

Author(s): Ye Wang, Xavier Guillaud, Gauthier Delille, Frederic Colas, Bruno François,
Narrator: Ye Wang, Xavier Guillaud, Gauthier Delille, Frederic Colas, Bruno François,

Download Video : Universite-de-Lille.mp4

MPEG4 Mobile Feed :

Related Products