CIRED Workshop - Ljubljana, 7-8 June 2018 Paper 0454 Paper No 0454 Page 1 / 4 EVALUATION OF DIFFERENT LOCAL VAR CONTROL STRATEGIES IN LOW VOLTAGE GRIDS Daniel-Leon SCHULTIS Albana ILO Christian SCHIRMER TU Wien – Austria TU Wien – Austria TU Wien – Austria daniel-leon.schultis@tuwien.ac.at albana.ilo@tuwien.ac.at christian.schirmer@tuwien.ac.at ABSTRACT This paper compares for the first time the impact of different var-control strategies on the behaviour of low voltage grids. Besides existing control strategies like Q(U)- and cosφ(P)-control of PV-inverters also the new ones, namely L(U)-control with or without Q-Autarkic customers, are investigated. The assessment of different control strategies is made by means of social and technical criteria. Investigations show that involving the prosumer-owned inverters in voltage control entails in principle social issues like discrimination and threat to data privacy. Local cosφ(P)- and Q(U)-control cause relatively high grid losses, extensive Q-exchanges between medium and low voltage grids and thus also considerable distribution transformer loadings. The application of L(U)-control mitigates the social issues and fulfils best the technical criteria. In this case the network operator is able to perform an effective voltage control by using his own devices. This control strategy enables the prosumers to internally compensate their reactive power needs; thus acting Q-Autarkic. INTRODUCTION The increasing penetration of photovoltaic (PV) facilities in low voltage (LV) grids is challenging the traditional power system operation; the simultaneous PV-injections cause reverse active power flows which provoke violations of the upper voltage limit and increased equipment loadings and electric losses [1]. However, European distribution system operators (DSOs) have to ensure the compliance of their grid voltages with the EN 50160 limits of ±10% around rated voltage. An option for DSOs to mitigate the rise in voltage is to manipulate the reactive power flows within their grids, for instance by controlling the Q-provision of PV-inverters [2, 3] which are owned by prosumers or by installing and operating own Q-devices for voltage control. Such control concepts strongly impact the Q-balance of distribution grids and lead to uncontrolled Q-flows between different voltage levels [4]. Several control strategies for PV-inverters evolved over the past decade; their capabilities to produce reactive power is used for voltage control in LV- grids. Two established approaches are the Q(U)- or cosφ(P)-control [2]. Another control strategy for smart inverters is proposed in [5], where they are controlled to supply the reactive power which is needed by the loads in customer plant level at all times. To control grid voltages in case of such Q-Autarkic prosumers, variable shunt-coils with local L(U)-control are located at the ends of the violated feeders [5]. This paper evaluates for the first time different var- control strategies used in LV grids by means of social and technical criteria. Firstly, the theoretical test system is described. Secondly, relevant simulation scenarios are defined. In the following the evaluation criteria are defined. Finally, the assessment results are presented. TEST SYSTEMS DESCRIPTION This section gives a short description of the test LV-grid, the thereto connected prosumers and the considered var- controls. Low voltage grid Fig. 1 shows the theoretical test-grid which is used for the simulations. Figure 1: Theoretical test-grid It consists of two identical feeders:   with a cable structure and  ℎ with an overhead-line structure, which are connected to the MV-grid through a 20 kV / 0.4 kV, 160 kVA distribution transformer (DTR). Each feeder supplies 20 identical residential prosumers. Prosumers Fig. 2 shows the prosumer structure. It is characterized by the active   and reactive power consumption   of his internal loads and the active   and reactive power injection   of his PV-system. Voltage dependency of loads is modelled with an inherent ZIP model from [6]; an initial power factor of 0.95 is set for all loads. Each PV-system includes a PV-module with a rating of P r PV = 4  and an inverter with a rating of S r inv = P r PV 0.9 ⁄ . Power losses within PV-systems are neglected. Their reactive power injection is determined by the applied control strategy. If no-control is exercised, inverters inject by power factor one. Figure 2: Structure of a prosumer