Abstract
The integration of electric vehicles (EVs) plays an integral part in reducing GHG emissions from the transport sector. In recent years, the number of EVs has increased rapidly. Due to government policies and technological advancement, the growth is expected to continue. The EVs have a limited range, but it will often be sufficient for daily routines. However, to enable long-distance travel, a network of fast charging stations (FCSs) is needed. Fast charging of EVs is characterized by its stochastic nature, high power, and short charging time. This can potentially result in bottlenecks in the grid. To face the challenges that come with the integration of FCSs into the distribution grid, an optimal planning scheme is needed. The main objective of this master thesis was to develop a model to decide the optimal planning of an FCS network. In this thesis, an EV mobility model, FCS load model and distribution grid model are combined in an optimization model to decide the optimal planning of FCSs. The FCS load model is developed to determine the load profile at different FCSs. The FCS load model includes the EV traffic flow, EV charging curves and temperature-dependent driving consumption. The available traffic data was inadequate. Hence, a mobility model was developed to create a more detailed traffic flow of EVs in the system. The EV mobility model determines the route of each EV. The FCS load model determines the charging need of the different EVs and which FCS they will charge at. Then, by aggregating the charging needs of the EVs, the charging demand at each FCSs is determined. The data about the Norwegian distribution grid is not open to the public. Thus, a novel distribution grid model was developed, which creates and dimension distribution grids. The proposed distribution grid model is based on power system planning principles, taking into consideration both economic and power system aspects. The aforementioned models are combined in the optimization model. The optimal planning of FCSs is a nonlinear problem and a particle swarm optimization (PSO) algorithm is implemented to solve the problem. The proposed optimization model is a two-step model, the first step determines the location of the FCSs, and the second step determines the number of charging points. The performance of the developed optimization model was tested on a 74 km stretch of highway between Gardermoen and Hamar. There are many aspects to consider when planning an FCS network. Thus, different objective functions were used in the optimization model. The first case study minimized the additional energy losses in the distribution grid due to the integration of FCS. For the second case study, the cost of FCSs was added to the objective function. For the final case, the perspective of EV owners was taken into consideration, by assigning a cost to EV detours. Thus, for the last objective function, the perspective of the DSO, FCS operator and EV owners were included. The results illuminate how the optimal number of FCSs and their location is highly dependent on the objective function. For the three case studies performed, all got a different optimal number of FCSs. The proposed optimization model was able to find the optimum solution with all the three objective functions. To compare the different objective functions, the social cost was computed for all three cases. The results showed that the social cost was highest for case 2, which only considers the DSO and FCS operator perspective. This resulted in a 25.2% higher social cost for case 2 than case 3, with most of the increase due to a 5400% increase in the detour cost. Thus, emphasizing the importance of considering the perspective of all the interested parties when planning an FCS network. The effects on the serviceability of an FCS when reducing its peak power were investigated. This showed promising results as the peak power of the FCS could be reduced significantly, with little impact on the serviceability of the FCS.