Optimal Charging of Electric Buses in an Existing Schedule Utilizing the Day-Ahead Electricity Market

Abstract

In recent years the push to renewable energy has increased substantially. Households are encouraged to disconnect gas lines and install solar panels to create their own renewable energy. Governments are trying to stimulate both households as well as companies to rethink their energy usage and reduce or transition their energy usage. In the case of public transit in The Netherlands, agreements have been made to solely have tank-to-wheel emission free fleets by the year 2030, with the introduction of new vehicles to be tank-to-wheel emission free from the year 2025 onwards. This thesis report is the product of my research performed at Qbuzz which tries to help them reduce the charging costs of their electric buses. We have compared two different ways of scheduling the charging of the electric bus fleet located at bus depot Remiseweg, Nieuwegein, The Netherlands. The first scheduling is a naive charging schedule. A bus will charge whenever it is at a charging location for as much as possible. The second charging schedule is a smart charging scheduling. This means a bus will only charge if the schedule tells it to. The schedule is optimised using a fortyone hours look ahead algorithm which will schedule the charging to be as cheap as possible, whilst adhering to a set of rules. Using the smart charging algorithm the electricity costs can be reduced by 28.8% to 31.4%. This is the result when the buses drive exactly on schedule. However, in the real world, the buses do not always drive exactly on schedule. This will result in buses arriving earlier or later than planned. To see how much of the expected savings could be realised a simulation of the operations is created. This simulation showed that a savings between 22.0% and 23.9% is to be expected.