Methods for Balancing Electricity Generation from Renewables on the Electricity Markets

Abstract

Power generation mix of each country depends mainly on the availability of usable resources in its territory or the possibility of importing them. In connection with the fulfilment of ambitious objectives to be reached by 2020 power generation mix is changing rapidly and is heading toward low carbon electricity generation footprint. Power generation mix in individual countries is more and more affected by their geographical location and its potential for utilization of renewables. In the beginning of year 2016 OTC electricity prices for 2017 base-load delivery in Germany were attacking 20 EUR level while half a year ago it was traded for 30 EUR. [37] Subsidized renewables caused price curve distortion resulting in conventional hard coal and lignite power plants that accounted in 2015 for 42% of total electricity generation, to operate close to their variable costs. Share of Germany’s electricity generation from renewables reached in 2015 30% and still new projects are planned or already under construction. [13] Low prices are not only affecting electricity markets where renewables are installed but also adjacent electricity markets in the neighbouring countries that are well-interconnected. With increasing renewable power capacity serious problems are experienced in the whole meshed central European network that may eventually lead to blackout situations. For electricity market participants with renewables in their generation portfolio is crucial to have all possible means to stay balanced. Interconnected electricity markets with high liquidity covering every day 24 hours and offering short-term products that can be traded as close as possible to delivery have to be on place. In case of missing trading platform, low liquidity or long time lag between delivery and product trading deadline, market participants are unable to balance themselves and are consequently exposed to balancing costs that can account for big part of total costs endangering profitability. In any case balance responsible parties have to do everything to be balanced. In my thesis I am testing possibility of using neural networks to improve generation forecast for wind farm from three independent meteorological data providers. Results from test scenarios which differ in various combinations of inputs were analysed and compared with regard to the most important indicators from the perspective of the wind farm owner. Balancing cost saving criterion is the most important measure to evaluate whether obtained results are better compared to the results of the most accurate meteorological data provider. Test scenarios for both day ahead and intraday time frame have been tested. Test period of six months proved that cost savings for balancing deviations can be achieved.