Probabilistic estimation of rated demand at bulk supply point

Abstract

This paper presents a Monte-Carlo based approach for probabilistic estimation of rated demand at bulk supply point at different times during the day. The required data are substation measured RMS real power, reactive power and voltage. Along with component-based load modeling approach, Monte-Carlo simulation is used to generate all possible combinations of percentages of different load categories participating under the total demand and calculate the corresponding per-unit total demand for every possible voltage. After the calculation, a probabilistic density function which shows the probabilistic distribution of the per-unit total demand can be generated. Rated demand is obtained by dividing the actual demand by the corresponding per-unit demand, and its probabilistic distribution can be obtained via either inverse cumulative distribution function (ICDF) regeneration in MATLAB or Gaussian distribution fitting. The result will help developing more accurate load models especially for load modeling with long timeframe or self-disconnections, and it will also enable validation of load disaggregation at given time in real system.