Controlling Ev Charging and Pv Generation in a Low Voltage Grid

Increasing amounts of local power generation and loads are challenging the management of low voltage (LV) grids. This paper proposes a low voltage grid monitoring and control solution enabling a grid operator to: 1) assess available capacity in an LV grid for planning purposes, 2) provide grid constrained electrical vehicle (EV) charging control, and 3) optimize the injection of photovoltaic (PV) energy in the grid. The ability to increase LV grid utilization and to minimize critical grid events are shown in Park & Ride and Shopping Mall EV charging scenarios. Relying on the prediction of domestic demand and PV generation, the ability of the EV charging strategy to handle prediction errors is analyzed. INTRODUCTION Two factors, that will threaten the power quality in the low voltage grid in the future, are the massive deployment of distributed generation (DG) using renewable resources (photovoltaic, wind) and of charging stations for electric vehicles. In order to tackle these and other future problems, it is necessary to gradually replace the current, tedious and inaccurate offline grid planning process, which is usually, based on standard profiles, with an automated, online process. In this paper we propose a grid component called low voltage grid controller (LVGC) that enables: ● Online monitoring of the LV grid using existing smart meter data to provide a grid state estimate ● Computation of a feasible region in which the grid can be loaded within grid constraints (voltages, currents, transformer limits, etc.) considering fairness among users ● Planning the Electrical Vehicle (EV) demand under uncertain forecasts of photovoltaic (PV) generation and domestic power. ● Controlled EV charging and limitation of PV active power in case of over-generation. Recent work shows that new LV grid load scenarios from e.g. electrical vehicles require a degree of control [1]-[4]. Analysis on macroscopic level shows how smart-charging allows an increase of EV penetration to more than 50% without overload of current grids (Portugal [1]), while the authors of [3] show how grid investments can be reduced by up to 25% for transformers with 75% of household owners owning a car (year 2040 scenarios, Holland). For household charging scenarios (aiming for 100% charge) different control techniques have been proposed. Key examples area) basic house-hold peak avoidance approaches [3] b) dynamic charging intensity control cooptimized with grid constraints for energy loss optimization [4],and c) sophisticated charging schedule planning on a day basis based on historical trip forecasting, desired load curves (price controlled) and grid constraints [2]. Our work recognizes the recent approach of [2] to derive power limitations on busses in LV grids for EV planning purposes, but extends the view by: ● Proposing a new formulation of how available power resources are estimated to ensure fair allocation of free bus resources, with applications for both planning, charging control and general control purposes. ● Proposing an explicit smart meter based architecture for LV grid management enabling short-term (15 min) adaptation of charging schedules based on: plug-in events and changes in the available power resources (caused by errors in the load prediction and PV output). ● Defining a smart charging control approach aiming to eliminate grid overload events. ● Studying new EV public charging scenarios (Park & Ride and Shopping Center). We clarify their different requirements and the achieved charging service quality when minimizing grid events and accommodating to photovoltaic production. The rest of the paper is structured as follows: in Section 2 we describe the system high-level architecture. Section 3 presents details on the developed control and support mechanisms. In Section 4, we present results from selected evaluation scenarios, and finally in section 5 conclusions and further research directions are provided. 2. SYSTEM ARCHITECTURE The architecture studied in this work is based on a LV Grid Controller (LVGC) located at a secondary substation level, as depicted in Figure 1. A main motivation for this approach is to utilize the availability of smart meter measurements and provide grid operators new options for LV grid management. In practice the LV grid controller enables the following features: A) Online LV grid monitoring that enables operators to identify situations, where real consumer patterns frequently lead to grid events. A grid event is here defined as an overcurrent (OC), over-voltage (OV) or under-voltage (UV). B) What-if analysis to enable simulated deployment of new EV charging stations or PV generation which may be power curtailed in certain cases. C) Provide online load and generation management here emphasized by electrical vehicle charging control as well as power limitation control of photovoltaic systems. This paper only focuses on the charging control. This proposed architecture is compatible with other hierarchically controlled smart grid architectures. Thus, future versions the LVGC will comply with the objectives of virtual power plant controllers [6]. In its current version, C I R E D 22nd International Conference on Electricity Distribution Stockholm, 10-13 June 2013