Forecasting future electricity demand on low voltage (LV) distribution networks is particularly challenging in the face of new technologies such as Electric Vehicles (EVs) where the uptake and charging behaviours are highly uncertain. To support their planning processes, Orion has undertaken a staged assessment of LV network performance in the face of increasing levels of EV charging. The first stage of EV modelling involved a high-level deterministic simulation of all Orion’s LV networks to identify areas that may be more vulnerable to constraints. From the results of the first stage, a smaller subset of vulnerable LV networks was selected for further modelling. This paper focuses on the second stage of the assessment process where 236 selected distribution transformers and their downstream LV networks were modelled using a probabilistic methodology developed by the EPECentre. This model captures a wide range of scenarios and encompasses data from New Zealand-centric inputs, including national travel surveys, household smart meter demand profiles and the current NZ EV fleet composition. In order to meet the large simulation burden required of a probabilistic methodology, a fast power flow solver was developed. This paper reports the forecast congestion levels and probabilities that these selected circuits may face as EV charging increases on Orion’s network. The learnings and challenges of working with imperfect distribution asset data sets to develop these models are also shared, such as approximating LV phase imbalance and electricity demand profiles with only the LV distribution transformer’s Maximum Demand Indicator (MDIs) measurements.