A stochastic analytic-probabilistic approach to distributed generation hosting capacity evaluation of active feeders

Abstract

The unregulated penetration of distributed generation (DG) on distribution networks can cause wide-ranging technical issues, such as voltage variation, thermal loading, and unbalance. To regulate penetration, a planner needs to determine the maximum penetration or hosting capacity (HC) of a network, which requires extensive studies assessing the technical impacts of DG on the most sensitive quality of supply variables. Several uncertainties complicate the HC computation, such as the stochasticity of customer loads, variable DG outputs, and unknown location and capacity of future DG installations. This uncertainty demands approaches with the capability of stochastic simulation of DG allocation and probabilistic assessment of the corresponding range of feeder performance. Existing uncertainty-based HC computation methods have inadequate probabilistic representation, high computational burden, and restricted scope of technical variables, reducing any confidence in the assessed HC. We use a stochastic analytic-probabilistic approach that embeds an analytic probabilistic load flow (PLF) transform in a Monte-Carlo simulation (MCS) of DG allocation, and applies a polynomial smoothing technique to further enhance computational efficiency. The acceptable variation of radial distribution feeder performance is identified by voltage limits, thermal capacity, and unbalance. The PLF has significant benefits in reducing the overall computational burden. The HC is quantified statistically using the beta probability density function, reflecting the composite risk from the multiple uncertainties. Compared with other formulations, this stochastic-analytic approach to HC improves feeder performance evaluation under a wide range of DG penetration scenarios and is appropriate for design and operational analysis.