Modelling of Demand Response in Distribution Systems

Doctoral thesis, 2014

Large scale integration of plug-in electric vehicles (PEVs) and intermittent renewable energy sources, such as solar PV and wind power, could affect the way in which the power system is planned, operated and controlled. In such systems, it could be desirable to utilize the flexibility of the demand, i.e. demand response (DR), to help maintain the balance between the supply and the demand, and to avoid or defer grid reinforcements. One approach to promote DR could be through the use of hourly electricity tariffs based on electricity prices on the day-ahead market. In this context, a PEV could be seen as a flexible load, due to the possibility to schedule its charging. The main purposes of this thesis are to investigate: the impact of PEVs on distribution systems; strategies to schedule the charging and other flexible loads; customers' benefits of scheduling their loads according to these strategies; and the influence of different network tariffs in combination with hourly electricity pricing. Furthermore, the possible synergy between DR and wind power is investigated. The results show a large difference in possible PEV penetration levels between the investigated areas. The distribution system in the commercial area investigated could handle a full PEV penetration without any overloading while the residential distribution system was experiencing transformer overloading even without PEVs. By scheduling the charging in order to minimize the losses, the residential distribution system all cars could be replaced by PEVs without increased peak demand. If instead the day-ahead market prices were used as scheduling signals, 24\% of the cars could be replaced by PEVs without increased peak demand. By considering other flexible loads, such as electric space-heating, the residential distribution system would experience a decreased peak demand, if up to 25\% of the customers were responsive to day-ahead market prices and if an energy based network tariff was used. For higher shares of responsive customers, the peak demand was increased, due to an increased coincident factor. By using a power based network tariff instead, the peak demand could be reduced even if all customers were responsive. From a customer perspective, a cost reduction of up to approximately 10\% could be achieved by actively managing the flexible loads. With PEVs available, a cost reduction of up to 13\% could be reached. For customers with large variations in their demand, a power based network tariff would be preferable, while for other customers, the difference in benefits between the network tariffs was found to be small. The possible economic benefit for wind power producers in power systems with DR, was found to be slightly decreased for several price-areas, although the total benefit of all wind power producers was increased. From a customer perspective, the interaction between DR and wind power could lead to a cost reduction for responsive customers in some areas whereas the cost was increased for responsive customers in other areas.