SmartFarm – Data based optimization for optimal energy management

Abstract

In the project SmartFarm we develop a method that automatically decides which renewable energies are profitable to use at the farm for the next hours. To this means optimization methods are used on various levels. Initially, all necessary data is measured on a demonstration object. This data forms the basis for data based learning methods for the modelling of the consumers (e.g. milk cooling), the producers (e.g. solar plant) and the energy storage systems (e.g. battery storage device). In this paper we will focus on the data based models of the producers and energy storage systems. Optimization methods are used to identify optimal parameters for a general ansatz function, which is derived from a Taylor approximation. We develop all models with real data of the demonstration object and verify them with this data as well. Numerical results and the further course of action will be discussed in the paper.