Browsing by Author "Waswa, Lewis"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    A probabilistic estimation of the capacity of solar PV SSEGs installed on a LV feeder network
    (Stellenbosch : Stellenbosch University, 2020-03) Waswa, Lewis; Bekker, Bernard; Chihota, Justice; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.
    ENGLISH ABSTRACT: Increased solar photovoltaic (PV) installation on to the grid has led to increased technical challenges in electricity network operations. These challenges mainly stem from the design structure of the grid, which only allows unidirectional power flow. This results in several challenges including violation of voltage limits, tripping of network protection systems and distribution line overloads among other issues. These challenges are mainly restricted to the distribution networks, as most solar PV small-scale embedded generators (PV SSEGs) are connected to the distribution networks, whose conditions are, in most cases, not remotely monitored. This results in increased challenges experienced by the networks in terms of network planning, distribution network operations, maintenance, regulation and grid control. To manage these challenges, the distribution operator needs to estimate the total capacity of solar PV installed on the distribution network, in addition to how much of that capacity is embedded in the network’s net demand, which is important in determining the condition of the network at any particular time. Several methods have been used to estimate the capacity of solar PV SSEGs installed in an area. Most studies apply remote sensing and computer vision algorithms to count the number of solar PV panels found in an area. Analysis of these studies indicate that the results obtained cannot be used in determining the condition of the network as they only determine the capacity of solar PV in an area. Secondly, disaggregation studies have largely been used to quantify the installed solar PV capacity embedded in the net demand of a feeder or network. These methods assume a multi-variable approach which requires multiple inputs that are not readily available. This study introduces a novel probabilistic method that applies Monte Carlo methods to quantify the solar PV SSEGs embedded in the net demand of a low voltage feeder. Historical demand, net demand and the solar PV output is used to determine the solar PV capacity embedded in the net demand of a feeder. The accuracy of the method is tested using simulated net demand and actual measured net demand metered from households connected on carefully selected feeders. Results demonstrate that the method performs well where the historical demand and the net metered demand are obtained from similar customer classes. Therefore, it is concluded that it is possible to estimate the capacity of solar PV SSEGs embedded in the net demand obtained from a feeder by analysing and comparing the net demand of that feeder and the historical demand of a similar customer class feeder.