Optimized solutions for smart microgrids
The share of distributed energy generation is growing at a rapid pace. The dropping cost of photovoltaic panels and Governments' incentives are making more and more convenient the installation of photovoltaic panels for privates all around the world. In this thesis, data from 18 houses in the Netherlands is collected and analyzed to verify the effect of a large concentration of photovoltaic energy generation on the distribution grid. The study reveals that during Spring and Summer problems for the grid may arise due to the large amount of current injected into the grid. Distributed storage, through the installation of batteries, and load shifting are simulated to test their effectiveness in the reduction of the over-injection problem. The results of the physical model are then studied from the economic perspective to verify which option is the most profitable. Finally, different machine learning algorithms are implemented to predict the load consumption and photovoltaic energy generation one-day ahead.