Optimal sitting and sizing of hydrogen refilling stations in distribution networks under locational marginal prices

The decarbonization of the mobility sector motivates to increase the penetration of battery and fuel-cell electric vehicles. The proliferation of these mobility modes will be accompanied by the massive installation of charging and refilling infrastructures into existing networks. This work focuses on fuel-cell vehicles, for which refilling points are needed. Difficulties in hydrogen transportation can be circumvented by deploying onsite hydrogen generation assets (electrolysers) and storage, which can be partially or fully supplied through local renewable generators. Nevertheless, such assets require a considerable initial investment, being necessary the use of planning tools in order to maximize revenues for private investors. This paper focuses on this issue. In particular, an optimal sitting and sizing tool for hydrogen refilling stations with onsite storage and electrolysers is developed. The developed methodology considers the influence of locational marginal prices, which are cleared by the distribution system operator in order to translate the real electricity cost per node. This pricing strategy helps to best allocate assets through the network and thus resulting valuable for planners in order to site refilling infrastructures properly. An original multi-year iterative algorithm based on the multi-cut Benders' decomposition is proposed in order to alleviate the intrinsic high computational cost of the planning tool while accommodate long-term inflation and degradation rates of parameters. A number of simulations are performed on the well-known IEEE 33-bus system. Results verify that locational marginal pricing effectively translates the nodal electricity cost to end-users. Remark, the total electrolysis capacity turns out to be the most significant parameter, reducing further the cost of the project, while storage capacity has a limited influence. Results highlight the importance of the infrastructure capacity when determining the placement and sizing of electrolysers, thus supporting decisions when upgrading existing infrastructure. The impact of the number of stations to be installed and the budget cap is also analysed, showing that both parameters have similar influence and may reduce the total project cost by 70% approximately. The typical scheduling behaviour of the electrolysis-storage facilities is discussed, showing how storage is capable to provide energy arbitrage exploiting locational marginal prices. Finally, the computational performance of the developed algorithm is assessed, verifying that the new tool is efficient and portable.