Robust Price-based EV Load Management Considering Human-choice Uncertainty

This paper presents a robust bi-level optimization model with near-optimal solution in lower-level to manage the electric vehicle (EV) load. The diversity of human preferences gives rise to uncertain human-choices when multiple options share the same charging cost. This uncertainty contributes to markedly distinct load profiles, resulting in severe ramping events, congestion, and voltage issues. To address it, this paper proposes a Location-Time-of-Use (LToU) charging price-based EV demand management approach that aims to flatten the net load curve in the face of uncertainties related to human choices. It is important to note that EV owners may have multiple near-optimal solutions to choose from, resulting in different demand response profiles. Therefore, it is essential to model EV owner's choice uncertainties, so that LToU rate can be correctly set. Addressing lower-level solution uncertainty in optimization models remains an open question in the literature. This work aims to bridge the gap with the proposed model. By exploring the model structure, we propose a group of optimality conditions and an algorithm to solve the problem. We perform comprehensive simulations with a power-transportation system. The results show the proposed approach can help flatten the load curve under human-choice uncertainty.