| dc.description |
Given the declining price of solar panels and rising public awareness for environmental protection, solar photovoltaic distributed generation has become one of the fastest growing types of renewable sources being integrated onto distribution systems. The increasing penetration level of PV installations in the distribution system creates new opportunities but also raises several challenges for the power grid operators, therefore, there is an increasing need for both accurate models of the solar resource and advanced distribution system analysis methods to find the maximum number of PV installations that can be added at distribution system without affecting other distribution components. The primary challenge with high penetrations of solar power into power distribution feeders lies in the solar resource's inherent variability and uncertainty during partly cloudy days. The variability of solar power due to cloud shadows with high penetration levels can cause technical issues such as power fluctuation and voltage flicker in power distribution feeders. Besides this voltage fluctuation introduced by PV power injection, PV systems power injection may force the existing voltage regulating devices such as load tap changer (LTC) and line voltage regulators to operate continuously. The consequence is the reduction of the operating life of these control devices. To avoid overly conservative decisions on the amount of PV installed on a feeder and predict the likely impacts, distributed PV installations must go through an integration study before interconnecting with utility grid. The primary challenge to performing these studies is collecting high time-resolution PV data from multiple locations on the feeder where PV installation will be deployed in the near future. However, due to uniqueness of every study scenario and feeders under study, it is almost impossible or unreasonably expensive to take high-resolution solar measurements of all PV systems. To this end, this research presents a comprehensive simulation and analysis approach along with an easily implemented cloud shadow model to simulate realistic solar power variability of both distributed and large centralized PVs at various locations on a feeder at one second intervals. The practical usefulness of the proposed model is confirmed by simulating power of four actual PV plants using the cloud shadow model and comparing them to actual one second measured data. Simulations are made on an actual EPRI test feeder. EPRI’s OpenDSS distribution simulation program is driven through MATLAB. |
|