Geostatistical Characteristics of the Structure of Spatial Variation of Electrical Power in the National 110 KV Network Including Results of Variogram Model Components Filtering

The paper provides results of analysing the superficial variability of electrical power using two geostatistical methods – lognormal kriging and ordinary kriging. The research work was to provide detailed characterization and identification of the electrical load variability structure at nodes of a 110 kV network over the whole territory of Poland having been analyzed on the basis of results from kriging techniques applied. The paper proposes the methodology using two techniques of modelling and estimating average values Z* of electrical powers, i.e. lognormal kriging and ordinary kriging. The input data for calculations were electrical powers at nodes of 110 kV network related to the same time moment, i.e. 11:00 a.m. in summer and winter seasons. Kriging calculations were made for various variants of examinations. Filtering was carried out for assumed complex theoretical models of semivariograms of electrical powers, which means their dividing into several models components of the covariance (nugget effect, 1 spherical model, 2 spherical model), which were filtered out successively. Then, estimations were made for average values Z* of powers while particular components are passed over. The results of analyses made with considering particular components of semivariograms models were shown in raster maps providing distributions of estimated average values Z* of electrical powers. This allowed the orientation of variations in values of this parameter, both over the territory of the whole country and in time domain, for two seasons – summer and winter, and also when various models components were assumed of semivariograms of the loads. Detailed analysis of spatial-time variability of average values Z* of electrical loads over the country allowed to identify their range and nature of variability.