Search documents

Browse topics

Document details

Simulation of Distribution Grids with Photovoltaics by Means of Stochastic Load Profiles and Irradiance Data
C. Bucher, J. Betcke, G. Andersson, B. Bletterie, L. Küng
Simulation, Grid Stability, Active and Reactive Power Control, Monte-Carlo Method
PV Systems
Subtopic: Large PV Power Plants and Distributed PV: System Aspects and Grid Connection
Event: 27th European Photovoltaic Solar Energy Conference and Exhibition
Session: 5CO.7.5
3795 - 3800
ISBN: 3-936338-28-0
Paper DOI: 10.4229/27thEUPVSEC2012-5CO.7.5
0,00 EUR
Document(s): paper


A method to perform distribution grid power flow calculations using stochastic load profiles and meteorological data is presented in this paper. The method is based on two previously presented algorithms: The first algorithm generates stochastic household load profiles and the second generates stochastic irradiance data, both using a sampling rate of one minute. Applying this data to an electrical power distribution grid containing a significant number of photovoltaic systems, realistic power flow and system voltage scenarios are obtained. The results are used to precisely answer a broad range of questions concerning grid integration of photovoltaic systems. The main purpose is to find the maximum photovoltaic power which can be installed in a given distribution grid and to evaluate different measures to raise this limit, such as respecting the correlation between irradiance and loads, active power curtailment, reactive power control or power management algorithms. In contrast to previous works, this paper not only addresses certain situations such as best or worst case scenarios, but also analyses the entire range of possible outcomes in Monte Carlo style simulations. Unacceptable situations in the power grid (e.g. overvoltages) can be quantified in probability of occurrence, time of day and their durations. The presented method is demonstrated with a five-node radial test feeder and artificial domestic loads with a resolution of one minute, based on a reference load profile from ewz Zurich.