The rapid adoption of electric vehicles (EVs) creates new challenges for low-voltage distribution networks. Home EV chargers can consume as much power as an entire household; when many vehicles charge simultaneously, they risk overloading power lines and transformers. Time-of-use (ToU) and dynamic tariffs incentivise charging away from conventional peak periods. However, existing tariffs, which are based primarily on wholesale market prices, can concentrate demand within the same low-price periods and create new network peaks.

This research develops a day-ahead pricing scheme for a constrained low-voltage feeder in London that responds to available feeder capacity, mitigates rebound peaks, and reduces mean consumer charging costs. First, probability distributions for daily mileage, arrival time, and initial state of charge are derived to develop a bottom-up stochastic model of private EV charging demand. Monte Carlo sampling is used to generate individual charging sessions, which are aggregated to produce feeder-level load profiles across different levels of EV penetration.

These profiles are then used to construct a day-ahead, 30-minute tariff based on available feeder headroom. The tariff incorporates critical-peak pricing to discourage charging during periods of high network stress and is iteratively adjusted using a fixed-point method to account for the resulting changes in charging demand.

At 60% EV penetration on the studied London feeder, the proposed tariff reduces the probability of exceeding the 100.9~kW firm limit from 30.2% to 7.5%, while reducing charging costs by 33% relative to Octopus Go. It also outperforms Octopus Agile in the cost--risk analysis, reducing overload risk by 45.6% while achieving a lower peak charging price.