Modeling Of Power System Short-term Load Forecasting For Special Day In All-weather

In recent years,with the rapid increase of scale and difference between peak and Valley load and power network short-term load forecasting accuracy,stability and rapidly put forward higher requirements.Special daily load because of outside weather,characteristics and the impact of national leave policy,and many other factors,presents the inherent uncertainty and volatility.Special daily load forecasting accuracy than a lower,unable to meet the requirements of power load forecasting accuracy.Engineering applications based on the reality of the subject through research on special days(holidays,weekends,typhoons)load variations considering special meteorological information,date type,the State paid leave policies and other factors influence.For holidays,weekends,typhoons are proposed a new method of load forecast.The main results are as follows:1)Holiday short-term load forecasting based on the fractal characteristics modified meteorological similar day.Based on similar principles,establishment of curve analysis functions to calculate date difference function looks similar to weather.And fractal self-similarity applies to holiday short-term load forecasting for the first time in the field,through accurate weather similar to find results,increased holiday load forecasting accuracy.2)Based on weather-load interaction of semi-parametric regression analysis method in weekend load forecasting.From the perspective of multi-source data interaction,the weekend load forecast as a result is the coupling of multi-source data.Using Semiparametric theory,dig deep coupling between the two weather-load relationship.And taking into account the effect of temperature accumulation,and amendments to the forecast model.3)Based on weather information for particle reduction of Typhoon-period load forecasting method.According to tracking load and the relevance of urban power load change affected by the typhoon,Typhoon daily load based on load losses and load divided into two parts.Focus on load losses in Typhoon-affected cities,and with three-stage changes in weather will load trends during the Typhoon as a piecewise function;with weather information for particle size reduction,seeking similar Typhoon-free properties environments to establish the basis prediction model.In this paper,simulation of the actual load data in Guangxi power grid,the results show that the prediction model for special days can improve the precision of short term load forecasting.