Computational studies on variable distributed energy systems

Abstract

In this work the large-scale integration of distributed wind turbine and photovoltaic power generation to medium voltage power distribution grid have been explored. The effects of the integration were mainly evaluated by steady-state power flow analysis. Compensation of unwanted overvoltage events was tested with the use of energy storage, utilization of grid topology, and redirecting of photovoltaic panels. In addition, this work examines the use of energy storage to compensate short and long term fluctuations in the power output of variable distributed generation. For this purpose three different storage control strategies were applied. For simulation purposes a detailed bottom-up consumer load model and a distribution grid power flow simulation model were developed. The stochastic approach in the load model is based on the classic bottom-up load model by Capasso et al. However, the emulation of the activities of individual people has been replaced by statistical use of household appliances where reference data is more readily available. The grid simulation model is based on established power flow computation methods. The model is especially designed for easy implementation of grid details as well as various types of distributed generation and storage and their control strategies. The distribution grid simulation results indicate that in a hypothetical crowded urban distribution grid an approximate limit for not causing grid disturbances for private households is around 0.5 kW of grid-connected photovoltaic power generation per apartment building household in both southern (Lisbon) and northern (Helsinki) climates. Using storage schemes the amount can be increased up to over 1 kW per household. However, grid details concerning storage size and siting topology greatly influence the exact limit. Further, the simulation results show that the integrated photovoltaic power generation can induce up to 34% reduction in the transmission losses of medium voltage distribution network. Corresponding analysis with wind power shows that if a MW-scale wind turbine is integrated to distribution grid the grid topology as well as the siting of distributed generation and storage are very important in avoiding unwanted overvoltage events. Overall, compensating power fluctuations of an individual MW-scale wind turbine needs to be viewed both on short and long time scales. Rapid minute-scale fluctuations can be most of the time smoothened by 50% with a storage unit in the capacity range of 25 kWh/MW. However, the long-term changes are difficult to compensate without large MWh-scale storage schemes. As a whole, the grid voltage issues proved to be much more crucial for the integration of variable distributed generation than the grid losses.