C I R E D 22 nd International Conference on Electricity Distribution Stockholm, 10-13 June 2013 Paper 0596 CIRED2013 Session 3 Paper No 0596 A review of transient effects of different types of distributed generation units on overcurrent protection system Yousef FIROUZ Jacques LOBRY François VALLEE Olgan DURIEUX UMONS – Belgium UMONS – Belgium UMONS – Belgium ORES – Belgium yousef.firouz@umons.ac.be jacques.lobry@umons.ac.be francois.vallee@umons.ac.be olgan.durieux@ores.net ABSTRACT Adding distributed generation (DG) units to the passive electricity networks causes several significant changes in their characteristics like power flow direction, voltage profile and short circuit level. Therefore the currently used control and protection strategies can no longer work properly and they have to be revised and modified. One of the most important issues is protection problem. For a reliable and efficient protection system, both transient and steady state effects of DGs on fault current should be considered. In this paper the exact model of three different types of DGs with different grid interfaces (PSMG with full size converter, DFIG with partial size, commonly met on Belgian grids, and directly connected IG) including power electronics and control systems, are presented. After that, the fault currents generated by different size and location of DGs are compared with those one which are produced by ideal model of DGs in the same condition. The PSCAD software is used for simulation of transient contribution of DG in fault current, in an assumed medium voltage grid. INTRODUCTION Increasing the integration of DG, especially renewable energy based, in the distribution network and close to the consumption, provides several advantages like lower CO 2 emission, but also some severe difficulties such as voltage and frequency instability, voltage distortion and protection problems. Indeed, the added DGs to the grid could contribute to the fault current and increase the short circuit current in case of downstream fault. In addition to the fault current increase, DG reduces the grid fault current contribution and increases the risk of blinding of the related protection. Moreover, the presence of DG may change the direction of power flows and fault currents. Therefore, in the case of neighboring feeder faults, it can force the relay to make an inappropriate tripping command. The intensity of these problems depends on the penetration level of DG and its location. Therefore finding a comprehensive solution for protection issues needs accurate studies, analysis and simulation. Firstly, the effects of DGs should be identified. Lots of studies and researches have been done in this field and results are published in different papers [1-3]. In most of the publications, DG is modeled as a simple voltage source and its dynamics have not been considered ([4] and [5]). Using ideal model makes the simulation fast and simple but with a low accuracy in the results. Computations based on ideal source representations can, indeed, produce high short circuit current depending on the fault impedance while the short circuit power of DGs, especially renewable energy based, is limited and do not effectively contribute to fault current like an ideal source in the same conditions. Practically, the duration of fault current and its transient behavior depends on the type of generator, its interface to the grid and also depends on the control strategy which is used. The mentioned problems could be more tangible when power electronic is used as interface between DG and grid. The produced fault current is not greater than 2 or 2.5 times of converter rated current [2]. This issue will be investigated in more detail in the following sections. EFFECT OF IDEAL MODEL OF DG ON OVER CURRENT PROTECTION According to fig.1, for illustrating the effects of added DG on overcurrent protection, a test medium voltage network is considered and simulated in PSCAD. A 63kV sub transmission network is connected to the bus 1via a 100MVA 63kV/20kV transformer and through the breaker 1. Feeder 1 and feeder 2 are 3 and 6 km long respectively. In first Fig.1) Test MV grid including DG