Home | english  | Impressum | Datenschutz | KIT

Integration of machine learning methods to predict energy production and consumption in demand response system

Integration of machine learning methods to predict energy production and consumption in demand response system
Forschungsthema:Wissensmanagement, Big Data Analytics, Smart Grid
Typ:Bachelorarbeit
Datum:27.02.2016
Betreuer:

Bearbeiter:Anggie Legiando Pratama

Due to the scarcity of fossil energy sources, buildings, factories, and cities are required to use nearly 100% decentral and renewable energy sources. This necessitates the introduction of an energy storage that is equipped with intelligent functionalities, such as prediction and control of energy production and consumption. These functionalities should allow minimum consump-tion of energy and maximum usage of renewable source. It aims to utilize electricity network has to be as low as possible.

Objective:

The main objective of this bachelor thesis is to develop a concept to predict the energy produc-tion and consumption in a small scale electricity grid equipped with energy storages, in order to generate recommendations of energy efficiency measures. The concept should base on data analytics and machine learning methods.

Work content:
This thesis consists of the following work:

  • Analysis of related fundamental theories, general requirements of the system and ob-jectives of the predictions
  • Identification of influencing parameters on energy consumption and prediction
  • Evaluation of existing data analytics/machine learning method and tools
  • Design the prediction system using the optimal methods and tools based on evaluation result
  • Verification using the real data
  • Presentation