Summary
The distribution networks are in a transition stage from being “passive” (consuming energy and typically with unidirectional power flows) into “active” (consuming/producing energy with bidirectional power flows). This transition has exposed the networks to stochastically behaving risks such as violation of the prescribed limits of power quality or overloading the network elements. The current paper presents a new risk analysis to quantify the risks of violating operational constraints due to a large number of small‐size Photovoltaics (PVs). Risk's likelihood and severity are estimated based on the relative frequency of the number and the relative frequency of the accumulative depth of a violation, respectively. This article has two objectives. The first is to determine the hosting capacity of the targeted network. The second is to address the effect of spatial correlation on risk quantification, specifically the effect of perfectly positive correlation (PEC) to the effect of partially positive correlation (PAC). Modelling PAC through spatial rank correlation is established utilizing the Nataf transformation due to the non‐Gaussianity of the probability densities of PV uncertainties. The approach is implemented on two distribution networks: IEEE 13 bus test distribution feeder and a large distribution network with multiple zones situated in South Australia. The approach is location‐specific and time‐varying. Hosting capacity is determined with the results of PEC showed an overestimation of the problem in comparison with the more realistic simulation under PAC. The analysis outcomes can help distribution network operators in managing and regulating the growing risks of high PV penetration.