Goran Banjac

Load forecast of a university building for application in microgrid power flow optimization

Microgrid is defined as a cluster of distributed generation sources, storages and loads that cooperate together in order to improve power supply reliability and overall power system stability. Short-term power production and load profile prediction is very important for power flow optimization in a microgrid, thus enhancing the management of distributed generation sources and storages in order to improve the microgrid reliability, as well as the economics of energy trade with electricity markets. However, short-term load prediction is a complex procedure, mainly because of the highly nonsmooth and nonlinear behaviour of the load time series. In this paper we develop and verify a neural-network-based short-term load profile prediction model. Neural network inputs are lagged load data, as well as meteorological and time data, while neural network output is load at the particular moment. Neural network training and validation is performed on load data recorded at University of Zagreb Faculty of Electrical Engineering and Computing, and on meteorological data obtained from Meteorological and Hydrological Service of Croatia, in period 2011-2013.

Global linear convergence in operator splitting methods

We establish necessary and sufficient conditions for linear convergence of operator splitting methods for a general class of convex optimization problems where the associated fixed-point operator is averaged. Most existing results establishing linear convergence in such methods require restrictive assumptions regarding strong convexity and smoothness of the constituent functions in the optimization problem. However, there are several examples in the literature showing that linear convergence is possible even when these properties do not hold. We provide a unifying analysis method for establishing linear convergence based on linear regularity and show that many existing results are special cases of our approach. Moreover, we propose a novel linearly convergent splitting method for linear programming.