Management of Voltage on LV Distribution Networks with Pre-Existing High Levels of Uncontrolled PV Systems & Inverters

Abstract

Low voltage urban distributors servicing single-phase residential or small commercial loads and hosting large numbers of photovoltaic (PV) systems can become hosting capacity limited for new PV installations owing to the phenomenon of voltage-rise at times of peak PV generation. We investigate the efficacy of adding smart PV systems, capable of both reactive power output and power curtailment via battery storage, to existing networks as a means of managing voltage within acceptable or regulated limits and thus enabling higher penetrations of PV systems with existing network infrastructure. We model numerous complete 3-phase 4-wire LV distributors with random allocation of loads and random deployment of PV systems, containing varying fractions of smart, controlled PV systems, and using as an example network, load and generation data from the Canberra region. We confirm that indeed voltage rise is likely to be a major problem for most networks without controlled PV systems, that voltage-dependent reactive power control is an effective means for managing voltage but that this approach has limited capability for many networks if pre-existing uncontrolled PV penetration already exceeds in the order of 25%. We show that reactive power control when combined with active power curtailment is a very effective method for controlling voltage, with as little as 25 – 50% of deployed systems enabled with this control regime being able to effectively manage voltage on the majority of networks up to very high penetration levels, albeit with increasing storage capacity demands if fewer systems are equipped – typically between 5 and 15 kWh being required at the extreme, but for example up to 25 kWh required at worst for a PV systems at the very end of some networks if loaded with 75% pre-existing uncontrolled systems. Finally, the modelling results point to the importance for network operators to ensure the deployment of PV systems balanced across phases, so as to avoid both large current imbalance and high voltage rise. We propose, via the approach presented, a methodology which could be readily applied to any generic network configuration or indeed to specific known real networks, in order to determine the best strategy for managing very high penetration of photovoltaic systems. Show more