Power Demand Estimation in Quasi-Real Time in a Medium Voltage Grid Area Under Conditions of Limited Observability

The paper presents the possibilities of practical use of available measurements to estimate in quasi-real time the demand for active (P) and reactive (Q) powers in a selected medium voltage (MV) grid area under conditions of limited observability. To estimate customers instantaneous power demand, the next day load forecasts, energy consumption profiles obtained off-line from AMI (Advanced Metering Infrastructure) balancing meters, P and Q power measurements in selected MV/ LV transformation points (reference substations), and meteorological measurements were used. To build the forecasting model, the artificial intelligence methods based on neural networks contained in MATLAB computing environment were used. The power demand and distribution grid operating status estimation method was used in the project “Development of a local balancing area as an element of increasing the distribution system operation security and energy efficiency”. The task was financed under the GEKON Program by NCBiR National Centre of Research and Development and NFOŚiGW National Fund for Environmental Protection and Water Management, and implemented with the participation of ENERGA-OPERATOR SA, the University of Zielona Gora and the Institute of Power Engineering, Gdansk Division.