Voltage Control with PV Inverters in Low Voltage Networks—In Depth Analysis of Different Concepts and Parameterization Criteria

Introduction:

Solar photovoltaic (PV) power is considered in most scenarios as an important energy resource for meeting the medium and long term renewable energy targets. A very specific property of PV generation is its highly decentralized nature: according to more than 70% of the installed PV capacity (more than 38gwin 2014) is embedded in the low voltage (LV) network. In some areas, the local PV penetration reaches very high values (200 kwp/km2 ) and the hosting capacity (HC) of LV networks is locally exhausted triggering expensive network reinforcement measures.

One of the main limitations towards the HC is the voltage rise caused by the PV infeed. Currently, the voltage at remote nodes of feeders with a high penetration of PV generation, might exceed the planning limits of distribution system operators (DSO) which are set to allow them to comply with the +10% limit specified in the voltage quality standard.

Existing system:

The impact on the effectiveness of the voltage control Q(U), due to asymmetrical power infeed is investigated. In, the impact on the control due to neutral point shifting is analysed. Several methods have been proposed in order to reduce or actively compensate the possible asymmetry introduced by single-phase PV installations.

Ina concept for determining the optimal phase connection with smart meters has been proposed. One of the most important conclusions is that distributing the PV power over the three phases should be the first measure to counteract voltage unbalance. A step model with increasing complexity is proposed to control the voltage in LV networks. The proposed concepts include coordinated control in which settings are sent to an on-load-tap-changer and to the PV inverters, depending on the voltage measurements provided by smart meters.

Dis-advantages:

  • Increasing complexity.
  • More COST

Proposed system:

The proposed VVI proved to be a suitable indicator for the comparison of different potential settings of the Q(U)-control (i.e., dead-band, droop value). These settings have a large impact on the consumed reactive energy (factor six between the least effective and the most effective Q(U) control) and on the effectiveness of the control (ability to compensate the voltage rise caused by the infeed).

The proposed network is a continuous feeder (OH line or cable)with an infinite number of PV generators. A simple computationshowed that using 100 nodes (or generators) is sufficientto approximate the theoretical case. The HC is considered to belimited by the maximum admissible voltage. It was confirmedby verifying that the loading of the cable or the OH line always stays below 100% at full reactive power consumption.

The PVgeneration is considered to be symmetrical, homogenously andcontinuously distributed along the feeder.An optimal power flow (OPF) has been implemented for benchmarking purposes. Although the implementation of an OPF is not realistic in LV networks due to high requirements in terms of control and communication, it allows for the comparison of the results to a reference.

Advantages:

  • Potential financial implications depending on the charging system between transmission and distribution networks.
  • Maximal effectiveness at minimal reactive energy consumption.

Applications:

  • Photovoltaic power systems.
  • Power distribution.

Block Diagram: