Spatial Load Forecasting of Distribution Network Based on Artificial Intelligent Algorithm

For the influence of the popularity of rooftop photovoltaic (PV) and the use of large-scale electric vehicles (EVs) on the power grid load, a space load forecasting method of urban distribution power grid was proposed, which took into accounts the spatial and temporal distribution of PV and EVs. Through analytic hierarchy process and fuzzy comprehensive evaluation method, various factors affecting the rooftop distributed PV output power were fully considered, and the rooftop PV output power of each planning area was forecasted by combining the least squares support vector machine and particle swarm optimization. Based on the time-space transfer probability matrix of EVs in different planning areas, Monte Carlo algorithm was used to simulate the time-space distribution of charging load with high probability. Taking an urban area as an example, the forecasted rooftop PV output power, EV charging load and traditional power load are superimposed on different planning areas to obtain the forecasted spatial load values.