Forecasting of Volumes of Power Consumption in the Electrical Distribution Networks Based on Hybrid Models

VALERII KRYSANOV, VIKTOR BURKOVSKII, ALEKSANDER DANILOV, KONSTANTIN GUSEV

Abstract


The article deals with increasing the prediction accuracy for electricity consumption in electrical grids under conditions of the modern wholesale electricity and capacity market and actual performance of grid operators. It contains the analysis of currently applied electricity consumption forecast methods. The article shows the necessity of accounting unpredictable and hardly definable factors, such as technical and economic factors (production schedule, end-consumer’s equipment load characteristics, electric power sector support and development costs), as well as climate factors (environment temperature, luminance, light day length, natural disasters). The most accurate accounting of these factors provides an opportunity to increase supervisory control quality and reduce electricity losses by reducing balancing market share, as well as output costs of generating company (end consumer). It is proposed to establish the use of Mamdani algorithm-based fuzzy neural networks as a core principle. The results show that electricity consumption forecast, based on using fuzzy neural networks is more accurate than the method currently specified in power supplier regulations. The article also contains an expert assessment of the economic viability of using said method for power supply to large regional industrial plants


DOI
10.12783/dteees/tpcase2018/30397

Full Text:

PDF

Refbacks

  • There are currently no refbacks.